]> git.ipfire.org Git - thirdparty/gcc.git/blame - gcc/ChangeLog
Add --enable-host-shared configuration option
[thirdparty/gcc.git] / gcc / ChangeLog
CommitLineData
459260ec
DM
12013-10-15 David Malcolm <dmalcolm@redhat.com>
2
3 * Makefile.in (PICFLAG): New.
4 (enable_host_shared): New.
5 (INTERNAL_CFLAGS): Use PICFLAG.
6 (LIBIBERTY): Use pic build of libiberty.a if configured with
7 --enable-host-shared.
8 * configure.ac: Add --enable-host-shared, setting up new
9 PICFLAG variable.
10 * configure: Regenerate.
11 * doc/install.texi (--enable-shared): Add note contrasting it
12 with...
13 (--enable-host-shared): New option.
14
16022041
RS
152013-10-15 Richard Biener <rguenther@suse.de>
16
17 * tree-tailcall.c (find_tail_calls): Don't use tail-call recursion
18 for built-in functions.
19
b114bfb4
ZC
202013-10-15 Zhenqiang Chen <zhenqiang.chen@arm.com>
21
22 * tree-ssa-reassoc.c: Include rtl.h and tm_p.h.
23 (optimize_range_tests_1): New function,
24 extracted from optimize_range_tests.
25 (optimize_range_tests_xor): Similarly.
26 (optimize_range_tests_diff): New function.
27 (optimize_range_tests): Use optimize_range_tests_1.
28
69b8f2f9
CH
292013-10-15 Cong Hou <congh@google.com>
30
31 * tree-vect-loop.c (vect_is_simple_reduction_1): Relax the
32 requirement of the reduction pattern so that one operand of the
33 reduction operation can come from outside of the loop.
34
8ff5a4e4
JG
352013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
36
37 * config/arm/neon-schedgen.ml: Remove.
38 * config/arm/cortex-a9-neon.md: Remove comment regarding
39 neon-schedgen.ml.
40
52b69ad2
JG
412013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
42
43 * config/arm/types: Remove old neon types.
44
e7c6a556
JG
452013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
46
47 * config/arm/cortex-a7.md
48 (cortex_a7_neon_type): New.
49 (cortex_a7_neon_mul): Update for new types.
50 (cortex_a7_neon_mla): Likewise.
51 (cortex_a7_neon): Likewise.
52
795dd8d5
JG
532013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
54
55 * config/arm/cortex-a15-neon.md
56 (cortex_a15_neon_type): New,
57
58 (cortex_a15_neon_int_1): Remove.
59 (cortex_a15_neon_int_2): Likewise.
60 (cortex_a15_neon_int_3): Likewise.
61 (cortex_a15_neon_int_4): Likewise.
62 (cortex_a15_neon_int_5): Likewise.
63 (cortex_a15_neon_vqneg_vqabs): Likewise.
64 (cortex_a15_neon_vmov): Likewise.
65 (cortex_a15_neon_vaba): Likewise.
66 (cortex_a15_neon_vaba_qqq): Likewise.
67 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
68 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
69 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
70 Likewise.
71 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
72 (cortex_a15_neon_mla_qqq_8_16): Likewise.
73 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
74 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
75 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
76 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
77 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
78 (cortex_a15_neon_shift_1): Likewise.
79 (cortex_a15_neon_shift_2): Likewise.
80 (cortex_a15_neon_shift_3): Likewise.
81 (cortex_a15_neon_vshl_ddd): Likewise.
82 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
83 (cortex_a15_neon_vsra_vrsra): Likewise.
84 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
85 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
86 (cortex_a15_neon_bp_3cycle): Likewise.
87 (cortex_a15_neon_ldm_2): Likewise.
88 (cortex_a15_neon_stm_2): Likewise.
89 (cortex_a15_neon_mcr): Likewise.
90 (cortex_a15_neon_mrc): Likewise.
91 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
92 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
93 (cortex_a15_neon_fp_vmul_ddd): Likewise.
94 (cortex_a15_neon_fp_vmul_qqd): Likewise.
95 (cortex_a15_neon_fp_vmla_ddd): Likewise.
96 (cortex_a15_neon_fp_vmla_qqq): Likewise.
97 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
98 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
99 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
100 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
101 (cortex_a15_neon_bp_simple): Likewise.
102 (cortex_a15_neon_bp_2cycle): Likewise.
103 (cortex_a15_neon_bp_3cycle): Likewise.
104 (cortex_a15_neon_vld1_1_2_regs): Likewise.
105 (cortex_a15_neon_vld1_3_4_regs): Likewise.
106 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
107 (cortex_a15_neon_vld2_4_regs): Likewise.
108 (cortex_a15_neon_vld3_vld4): Likewise.
109 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
110 (cortex_a15_neon_vst1_3_4_regs): Likewise.
111 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
112 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
113 (cortex_a15_neon_vst3_vst4): Rename to...
114 (cortex_a15_neon_vst4): This, update for new attributes.
115 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
116 (cortex_a15_neon_vld3_vld4_lane): Likewise.
117 (cortex_a15_neon_vst1_vst2_lane): Likewise.
118 (cortex_a15_neon_vst3_vst4_lane): Likewise.
119 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
120 (cortex_a15_neon_ldm_2): Likewise.
121 (cortex_a15_neon_stm_2): Likewise.
122 (cortex_a15_neon_mcr): Likewise.
123 (cortex_a15_neon_mcr_2_mcrr): Likewise.
124 (cortex_a15_neon_mrc): Likewise.
125 (cortex_a15_neon_mrrc): Likewise.
126
127 (cortex_a15_neon_abd): New.
128 (cortex_a15_neon_abd_q): Likewise.
129 (cortex_a15_neon_aba): Likewise.
130 (cortex_a15_neon_aba_q): Likewise.
131 (cortex_a15_neon_acc): Likewise.
132 (cortex_a15_neon_acc_q): Likewise.
133 (cortex_a15_neon_arith_basic): Likewise.
134 (cortex_a15_neon_arith_complex): Likewise.
135 (cortex_a15_neon_multiply): Likewise.
136 (cortex_a15_neon_multiply_q): Likewise.
137 (cortex_a15_neon_mla): Likewise.
138 (cortex_a15_neon_mla_q): Likewise.
139 (cortex_a15_neon_sat_mla_long): Likewise.
140 (cortex_a15_neon_shift_acc): Likewise.
141 (cortex_a15_neon_shift_imm_basic): Likewise.
142 (cortex_a15_neon_shift_imm_complex): Likewise.
143 (cortex_a15_neon_shift_reg_basic): Likewise.
144 (cortex_a15_neon_shift_reg_basic_q): Likewise.
145 (cortex_a15_neon_shift_reg_complex): Likewise.
146 (cortex_a15_neon_shift_reg_complex_q): Likewise.
147 (cortex_a15_neon_fp_negabs): Likewise
148 (cortex_a15_neon_fp_arith): Likewise
149 (cortex_a15_neon_fp_arith_q): Likewise
150 (cortex_a15_neon_fp_cvt_int): Likewise
151 (cortex_a15_neon_fp_cvt_int_q): Likewise
152 (cortex_a15_neon_fp_cvt_16): Likewise
153 (cortex_a15_neon_fp_mul): Likewise
154 (cortex_a15_neon_fp_mul_q): Likewise
155 (cortex_a15_neon_fp_mla): Likewise
156 (cortex_a15_neon_fp_mla_q): Likewise
157 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
158 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
159 (cortex_a15_neon_bitops): Likewise.
160 (cortex_a15_neon_bitops_q): Likewise.
161 (cortex_a15_neon_from_gp): Likewise.
162 (cortex_a15_neon_from_gp_q): Likewise.
163 (cortex_a15_neon_tbl3_tbl4): Likewise.
164 (cortex_a15_neon_zip_q): Likewise.
165 (cortex_a15_neon_to_gp): Likewise.
166 (cortex_a15_neon_load_a): Likewise.
167 (cortex_a15_neon_load_b): Likewise.
168 (cortex_a15_neon_load_c): Likewise.
169 (cortex_a15_neon_load_d): Likewise.
170 (cortex_a15_neon_load_e): Likewise.
171 (cortex_a15_neon_load_f): Likewise.
172 (cortex_a15_neon_store_a): Likewise.
173 (cortex_a15_neon_store_b): Likewise.
174 (cortex_a15_neon_store_c): Likewise.
175 (cortex_a15_neon_store_d): Likewise.
176 (cortex_a15_neon_store_e): Likewise.
177 (cortex_a15_neon_store_f): Likewise.
178 (cortex_a15_neon_store_g): Likewise.
179 (cortex_a15_neon_store_h): Likewise.
180 (cortex_a15_vfp_to_from_gp): Likewise.
181
bc5e57e3
JG
1822013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
183
184 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
185
186 (cortex_a9_neon_vshl_ddd): Remove.
187 (cortex_a9_neon_vst3_vst4): Likewise.
188 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
189
190 (cortex_a9_neon_bit_ops_q): New.
191
192 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
193 (cortex_a9_neon_int_2): Likewise.
194 (cortex_a9_neon_int_3): Likewise.
195 (cortex_a9_neon_int_4): Likewise.
196 (cortex_a9_neon_int_5): Likewise.
197 (cortex_a9_neon_vqneg_vqabs): Likewise.
198 (cortex_a9_neon_vmov): Likewise.
199 (cortex_a9_neon_vaba): Likewise.
200 (cortex_a9_neon_vaba_qqq): Likewise.
201 (cortex_a9_neon_shift_1): Likewise.
202 (cortex_a9_neon_shift_2): Likewise.
203 (cortex_a9_neon_shift_3): Likewise.
204 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
205 (cortex_a9_neon_vsra_vrsra): Likewise.
206 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
207 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
208 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
209 Likewise.
210 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
211 (cortex_a9_neon_mla_qqq_8_16): Likewise.
212 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
213 Likewise.
214 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
215 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
216 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
217 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
218 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
219 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
220 (cortex_a9_neon_fp_vsum): Likewise.
221 (cortex_a9_neon_fp_vmul_ddd): Likewise.
222 (cortex_a9_neon_fp_vmul_qqd): Likewise.
223 (cortex_a9_neon_fp_vmla_ddd): Likewise.
224 (cortex_a9_neon_fp_vmla_qqq): Likewise.
225 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
226 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
227 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
228 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
229 (cortex_a9_neon_bp_simple): Likewise.
230 (cortex_a9_neon_bp_2cycle): Likewise.
231 (cortex_a9_neon_bp_3cycle): Likewise.
232 (cortex_a9_neon_ldr): Likewise.
233 (cortex_a9_neon_str): Likewise.
234 (cortex_a9_neon_vld1_1_2_regs): Likewise.
235 (cortex_a9_neon_vld1_3_4_regs): Likewise.
236 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
237 (cortex_a9_neon_vld2_4_regs): Likewise.
238 (cortex_a9_neon_vld3_vld4): Likewise.
239 (cortex_a9_neon_vld1_vld2_lane): Likewise.
240 (cortex_a9_neon_vld3_vld4_lane): Likewise.
241 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
242 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
243 (cortex_a9_neon_vst1_3_4_regs): Likewise.
244 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
245 (cortex_a9_neon_vst1_vst2_lane): Likewise.
246 (cortex_a9_neon_vst3_vst4_lane): Likewise.
247 (cortex_a9_neon_mcr): Likewise.
248 (cortex_a9_neon_mcr_2_mcrr): Likewise.
249 (cortex_a9_neon_mrc): Likewise.
250 (cortex_a9_neon_mrrc): Likewise.
251
9deda8cb
JG
2522013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
253
254 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
255
256 (cortex_a8_neon_vshl_ddd): Remove.
257 (cortex_a8_neon_vst3_vst4): Likewise.
258 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
259
260 (cortex_a8_neon_bit_ops_q): New.
261
262 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
263 (cortex_a8_neon_int_2): Likewise..
264 (cortex_a8_neon_int_3): Likewise.
265 (cortex_a8_neon_int_5): Likewise.
266 (cortex_a8_neon_vqneg_vqabs): Likewise.
267 (cortex_a8_neon_int_4): Likewise.
268 (cortex_a8_neon_vaba): Likewise.
269 (cortex_a8_neon_vaba_qqq): Likewise.
270 (cortex_a8_neon_shift_1): Likewise.
271 (cortex_a8_neon_shift_2): Likewise.
272 (cortex_a8_neon_shift_3): Likewise.
273 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
274 (cortex_a8_neon_vsra_vrsra): Likewise.
275 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
276 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
277 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
278 Likewise.
279 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
280 (cortex_a8_neon_mla_qqq_8_16): Likewise.
281 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
282 Likewise.
283 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
284 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
285 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
286 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
287 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
288 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
289 (cortex_a8_neon_fp_vsum): Likewise.
290 (cortex_a8_neon_fp_vmul_ddd): Likewise.
291 (cortex_a8_neon_fp_vmul_qqd): Likewise.
292 (cortex_a8_neon_fp_vmla_ddd): Likewise.
293 (cortex_a8_neon_fp_vmla_qqq): Likewise.
294 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
295 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
296 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
297 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
298 (cortex_a8_neon_bp_simple): Likewise.
299 (cortex_a8_neon_bp_2cycle): Likewise.
300 (cortex_a8_neon_bp_3cycle): Likewise.
301 (cortex_a8_neon_ldr): Likewise.
302 (cortex_a8_neon_str): Likewise.
303 (cortex_a8_neon_vld1_1_2_regs): Likewise.
304 (cortex_a8_neon_vld1_3_4_regs): Likewise.
305 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
306 (cortex_a8_neon_vld2_4_regs): Likewise.
307 (cortex_a8_neon_vld3_vld4): Likewise.
308 (cortex_a8_neon_vld1_vld2_lane): Likewise.
309 (cortex_a8_neon_vld3_vld4_lane): Likewise.
310 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
311 (cortex_a8_neon_vst1_3_4_regs): Likewise.
312 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
313 (cortex_a8_neon_vst1_vst2_lane): Likewise.
314 (cortex_a8_neon_vst3_vst4_lane): Likewise.
315 (cortex_a8_neon_mcr): Likewise.
316 (cortex_a8_neon_mcr_2_mcrr): Likewise.
317 (cortex_a8_neon_mrc): Likewise.
318 (cortex_a8_neon_mrrc): Likewise.
319
0f686aa9
JG
3202013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
321
322 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
323 (fp): New.
324 * config/aarch64/aarch64-simd.md (neon_type): Remove.
325 (aarch64_simd_dup<mode>): Add "type" attribute.
326 (aarch64_dup_lane<mode>): Likewise.
327 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
328 (*aarch64_simd_mov<mode>): Likewise.
329 (aarch64_simd_mov_from_<mode>low): Likewise.
330 (aarch64_simd_mov_from_<mode>high): Likewise.
331 (orn<mode>3): Likewise.
332 (bic<mode>3): Likewise.
333 (add<mode>3): Likewise.
334 (sub<mode>3): Likewise.
335 (mul<mode>3): Likewise.
336 (*aarch64_mul3_elt<mode>): Likewise.
337 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
338 (*aarch64_mul3_elt_to_128df): Likewise.
339 (*aarch64_mul3_elt_to_64v2df): Likewise.
340 (neg<mode>2): Likewise.
341 (abs<mode>2): Likewise.
342 (abd<mode>_3): Likewise.
343 (aba<mode>_3): Likewise.
344 (fabd<mode>_3): Likewise.
345 (*fabd_scalar<mode>3): Likewise.
346 (and<mode>3): Likewise.
347 (ior<mode>3): Likewise.
348 (xor<mode>3): Likewise.
349 (one_cmpl<mode>2): Likewise.
350 (aarch64_simd_vec_set<mode>): Likewise.
351 (aarch64_simd_lshr<mode>): Likewise.
352 (aarch64_simd_ashr<mode>): Likewise.
353 (aarch64_simd_imm_shl<mode>): Likewise.
354 (aarch64_simd_reg_sshl<mode): Likewise.
355 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
356 (aarch64_simd_reg_shl<mode>_signed): Likewise.
357 (aarch64_simd_vec_setv2di): Likewise.
358 (aarch64_simd_vec_set<mode>): Likewise.
359 (aarch64_mla<mode>): Likewise.
360 (*aarch64_mla_elt<mode>): Likewise.
361 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
362 (aarch64_mls<mode>): Likewise.
363 (*aarch64_mls_elt<mode>): Likewise.
364 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
365 (<su><maxmin><mode>3): Likewise.
366 (move_lo_quad_<mode>): Likewise.
367 (aarch64_simd_move_hi_quad_<mode>): Likewise.
368 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
369 (vec_pack_trunc_<mode>): Likewise.
370 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
371 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
372 (*aarch64_<su>mlal_lo<mode>): Likewise.
373 (*aarch64_<su>mlal_hi<mode>): Likewise.
374 (*aarch64_<su>mlsl_lo<mode>): Likewise.
375 (*aarch64_<su>mlsl_hi<mode>): Likewise.
376 (*aarch64_<su>mlal<mode>): Likewise.
377 (*aarch64_<su>mlsl<mode>): Likewise.
378 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
379 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
380 (add<mode>3): Likewise.
381 (sub<mode>3): Likewise.
382 (mul<mode>3): Likewise.
383 (div<mode>3): Likewise.
384 (neg<mode>2): Likewise.
385 (abs<mode>2): Likewise.
386 (fma<mode>4): Likewise.
387 (*aarch64_fma4_elt<mode>): Likewise.
388 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
389 (*aarch64_fma4_elt_to_128df): Likewise.
390 (*aarch64_fma4_elt_to_64v2df): Likewise.
391 (fnma<mode>4): Likewise.
392 (*aarch64_fnma4_elt<mode>): Likewise.
393 (*aarch64_fnma4_elt_<vswap_width_name><mode>
394 (*aarch64_fnma4_elt_to_128df): Likewise.
395 (*aarch64_fnma4_elt_to_64v2df): Likewise.
396 (<frint_pattern><mode>2): Likewise.
397 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
398 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
399 (vec_unpacks_lo_v4sf): Likewise.
400 (aarch64_float_extend_lo_v2df): Likewise.
401 (vec_unpacks_hi_v4sf): Likewise.
402 (aarch64_float_truncate_lo_v2sf): Likewise.
403 (aarch64_float_truncate_hi_v4sf): Likewise.
404 (aarch64_vmls<mode>): Likewise.
405 (<su><maxmin><mode>3): Likewise.
406 (<maxmin_uns><mode>3): Likewise.
407 (reduc_<sur>plus_<mode>): Likewise.
408 (reduc_<sur>plus_v2di): Likewise.
409 (reduc_<sur>plus_v2si): Likewise.
410 (reduc_<sur>plus_<mode>): Likewise.
411 (aarch64_addpv4sf): Likewise.
412 (clz<mode>2): Likewise.
413 (reduc_<maxmin_uns>_<mode>): Likewise.
414 (reduc_<maxmin_uns>_v2di): Likewise.
415 (reduc_<maxmin_uns>_v2si): Likewise.
416 (reduc_<maxmin_uns>_<mode>): Likewise.
417 (reduc_<maxmin_uns>_v4sf): Likewise.
418 (aarch64_simd_bsl<mode>_internal): Likewise.
419 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
420 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
421 (aarch64_get_lane<mode>): Likewise.
422 (*aarch64_combinez<mode>): Likewise.
423 (aarch64_combine<mode>): Likewise.
424 (aarch64_simd_combine<mode>): Likewise.
425 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
426 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
427 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
428 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
429 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
430 (aarch64_<sur>h<addsub><mode>): Likewise.
431 (aarch64_<sur><addsub>hn<mode>): Likewise.
432 (aarch64_<sur><addsub>hn2<mode>): Likewise.
433 (aarch64_pmul<mode>): Likewise.
434 (aarch64_<su_optab><optab><mode>): Likewise.
435 (aarch64_<sur>qadd<mode>): Likewise.
436 (aarch64_sqmovun<mode>): Likewise.
437 (aarch64_<sur>qmovn<mode>): Likewise.
438 (aarch64_s<optab><mode>): Likewise.
439 (aarch64_sq<r>dmulh<mode>): Likewise.
440 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
441 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
442 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
443 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
444 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
445 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
446 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
447 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
448 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
449 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
450 (aarch64_sqdmull<mode>): Likewise.
451 (aarch64_sqdmull_lane<mode>_internal): Likewise.
452 (aarch64_sqdmull_n<mode>): Likewise.
453 (aarch64_sqdmull2<mode>_internal): Likewise.
454 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
455 (aarch64_sqdmull2_n<mode>_internal): Likewise.
456 (aarch64_<sur>shl<mode>): Likewise.
457 (aarch64_<sur>q<r>shl<mode>
458 (aarch64_<sur>shll_n<mode>): Likewise.
459 (aarch64_<sur>shll2_n<mode>): Likewise.
460 (aarch64_<sur>shr_n<mode>): Likewise.
461 (aarch64_<sur>sra_n<mode>): Likewise.
462 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
463 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
464 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
465 (aarch64_cm<optab><mode>): Likewise.
466 (aarch64_cm<optab>di): Likewise.
467 (aarch64_cm<optab><mode>): Likewise.
468 (aarch64_cm<optab>di): Likewise.
469 (aarch64_cmtst<mode>): Likewise.
470 (aarch64_cmtstdi): Likewise.
471 (aarch64_cm<optab><mode>): Likewise.
472 (*aarch64_fac<optab><mode>): Likewise.
473 (aarch64_addp<mode>): Likewise.
474 (aarch64_addpdi): Likewise.
475 (sqrt<mode>2): Likewise.
476 (vec_load_lanesoi<mode>): Likewise.
477 (vec_store_lanesoi<mode>): Likewise.
478 (vec_load_lanesci<mode>): Likewise.
479 (vec_store_lanesci<mode>): Likewise.
480 (vec_load_lanesxi<mode>): Likewise.
481 (vec_store_lanesxi<mode>): Likewise.
482 (*aarch64_mov<mode>): Likewise.
483 (aarch64_ld2<mode>_dreg): Likewise.
484 (aarch64_ld2<mode>_dreg): Likewise.
485 (aarch64_ld3<mode>_dreg): Likewise.
486 (aarch64_ld3<mode>_dreg): Likewise.
487 (aarch64_ld4<mode>_dreg): Likewise.
488 (aarch64_ld4<mode>_dreg): Likewise.
489 (aarch64_tbl1<mode>): Likewise.
490 (aarch64_tbl2v16qi): Likewise.
491 (aarch64_combinev16qi): Likewise.
492 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
493 (aarch64_st2<mode>_dreg): Likewise.
494 (aarch64_st2<mode>_dreg): Likewise.
495 (aarch64_st3<mode>_dreg): Likewise.
496 (aarch64_st3<mode>_dreg): Likewise.
497 (aarch64_st4<mode>_dreg): Likewise.
498 (aarch64_st4<mode>_dreg): Likewise.
499 (*aarch64_simd_ld1r<mode>): Likewise.
500 (aarch64_frecpe<mode>): Likewise.
501 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
502 (aarch64_frecps<mode>): Likewise.
503
f7379e5e
JG
5042013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
505
506 * config/arm/iterators.md (V_elem_ch): New.
507 (q): Likewise.
508 (VQH_type): Likewise.
509 * config/arm/arm.md (is_neon_type): New.
510 (conds): Use is_neon_type.
511 (anddi3_insn): Update type attribute.
512 (xordi3_insn): Likewise.
513 (one_cmpldi2): Likewise.
514 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
515 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
516 (*movmisalign<mode>_neon_store): Likewise.
517 (*movmisalign<mode>_neon_load): Likewise.
518 (vec_set<mode>_internal): Likewise.
519 (vec_set<mode>_internal): Likewise.
520 (vec_setv2di_internal): Likewise.
521 (vec_extract<mode>): Likewise.
522 (vec_extract<mode>): Likewise.
523 (vec_extractv2di): Likewise.
524 (*add<mode>3_neon): Likewise.
525 (adddi3_neon): Likewise.
526 (*sub<mode>3_neon): Likewise.
527 (subdi3_neon): Likewise.
528 (fma<VCVTF:mode>4): Likewise.
529 (fma<VCVTF:mode>4_intrinsic): Likewise.
530 (*fmsub<VCVTF:mode>4): Likewise.
531 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
532 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
533 (ior<mode>3): Likewise.
534 (and<mode>3): Likewise.
535 (orn<mode>3_neon): Likewise.
536 (orndi3_neon): Likewise.
537 (bic<mode>3_neon): Likewise.
538 (bicdi3_neon): Likewise.
539 (xor<mode>3): Likewise.
540 (one_cmpl<mode>2): Likewise.
541 (abs<mode>2): Likewise.
542 (neg<mode>2): Likewise.
543 (negdi2_neon): Likewise.
544 (*umin<mode>3_neon): Likewise.
545 (*umax<mode>3_neon): Likewise.
546 (*smin<mode>3_neon): Likewise.
547 (*smax<mode>3_neon): Likewise.
548 (vashl<mode>3): Likewise.
549 (vashr<mode>3_imm): Likewise.
550 (vlshr<mode>3_imm): Likewise.
551 (ashl<mode>3_signed): Likewise.
552 (ashl<mode>3_unsigned): Likewise.
553 (neon_load_count): Likewise.
554 (ashldi3_neon_noclobber): Likewise.
555 (ashldi3_neon): Likewise.
556 (signed_shift_di3_neon): Likewise.
557 (unsigned_shift_di3_neon): Likewise.
558 (ashrdi3_neon_imm_noclobber): Likewise.
559 (lshrdi3_neon_imm_noclobber): Likewise.
560 (<shift>di3_neon): Likewise.
561 (widen_ssum<mode>3): Likewise.
562 (widen_usum<mode>3): Likewise.
563 (quad_halves_<code>v4si): Likewise.
564 (quad_halves_<code>v4sf): Likewise.
565 (quad_halves_<code>v8hi): Likewise.
566 (quad_halves_<code>v16qi): Likewise.
567 (reduc_splus_v2di): Likewise.
568 (neon_vpadd_internal<mode>): Likewise.
569 (neon_vpsmin<mode>): Likewise.
570 (neon_vpsmax<mode>): Likewise.
571 (neon_vpumin<mode>): Likewise.
572 (neon_vpumax<mode>): Likewise.
573 (*ss_add<mode>_neon): Likewise.
574 (*us_add<mode>_neon): Likewise.
575 (*ss_sub<mode>_neon): Likewise.
576 (*us_sub<mode>_neon): Likewise.
577 (neon_vadd<mode>_unspec): Likewise.
578 (neon_vaddl<mode>): Likewise.
579 (neon_vaddw<mode>): Likewise.
580 (neon_vhadd<mode>): Likewise.
581 (neon_vqadd<mode>): Likewise.
582 (neon_vaddhn<mode>): Likewise.
583 (neon_vmul<mode>): Likewise.
584 (neon_vfms<VCVTF:mode>): Likewise.
585 (neon_vmlal<mode>): Likewise.
586 (neon_vmls<mode>): Likewise.
587 (neon_vmlsl<mode>): Likewise.
588 (neon_vqdmulh<mode>): Likewise.
589 (neon_vqdmlal<mode>): Likewise.
590 (neon_vqdmlsl<mode>): Likewise.
591 (neon_vmull<mode>): Likewise.
592 (neon_vqdmull<mode>): Likewise.
593 (neon_vsub<mode>_unspec): Likewise.
594 (neon_vsubl<mode>): Likewise.
595 (neon_vsubw<mode>): Likewise.
596 (neon_vqsub<mode>): Likewise.
597 (neon_vhsub<mode>): Likewise.
598 (neon_vsubhn<mode>): Likewise.
599 (neon_vceq<mode>): Likewise.
600 (neon_vcge<mode>): Likewise.
601 (neon_vcgeu<mode>): Likewise.
602 (neon_vcgt<mode>): Likewise.
603 (neon_vcgtu<mode>): Likewise.
604 (neon_vcle<mode>): Likewise.
605 (neon_vclt<mode>): Likewise.
606 (neon_vcage<mode>): Likewise.
607 (neon_vcagt<mode>): Likewise.
608 (neon_vtst<mode>): Likewise.
609 (neon_vabd<mode>): Likewise.
610 (neon_vabdl<mode>): Likewise.
611 (neon_vaba<mode>): Likewise.
612 (neon_vabal<mode>): Likewise.
613 (neon_vmax<mode>): Likewise.
614 (neon_vmin<mode>): Likewise.
615 (neon_vpaddl<mode>): Likewise.
616 (neon_vpadal<mode>): Likewise.
617 (neon_vpmax<mode>): Likewise.
618 (neon_vpmin<mode>): Likewise.
619 (neon_vrecps<mode>): Likewise.
620 (neon_vrsqrts<mode>): Likewise.
621 (neon_vqabs<mode>): Likewise.
622 (neon_vqneg<mode>): Likewise.
623 (neon_vcls<mode>): Likewise.
624 (clz<mode>2): Likewise.
625 (popcount<mode>2): Likewise.
626 (neon_vrecpe<mode>): Likewise.
627 (neon_vrsqrte<mode>): Likewise.
628 (neon_vget_lane<mode>_sext_internal): Likewise.
629 (neon_vget_lane<mode>_zext_internal): Likewise.
630 (neon_vdup_n<mode>): Likewise.
631 (neon_vdup_n<mode>): Likewise.
632 (neon_vdup_nv2di): Likewise.
633 (neon_vdup_lane<mode>_interal): Likewise.
634 (*neon_vswp<mode>): Likewise.
635 (neon_vcombine<mode>): Likewise.
636 (float<mode><V_cvtto>2): Likewise.
637 (floatuns<mode><V_cvtto>2): Likewise.
638 (fix_trunc<mode><V_cvtto>2): Likewise.
639 (fixuns_trunc<mode><V_cvtto>2
640 (neon_vcvt<mode>): Likewise.
641 (neon_vcvt<mode>): Likewise.
642 (neon_vcvtv4sfv4hf): Likewise.
643 (neon_vcvtv4hfv4sf): Likewise.
644 (neon_vcvt_n<mode>): Likewise.
645 (neon_vcvt_n<mode>): Likewise.
646 (neon_vmovn<mode>): Likewise.
647 (neon_vqmovn<mode>): Likewise.
648 (neon_vqmovun<mode>): Likewise.
649 (neon_vmovl<mode>): Likewise.
650 (neon_vmul_lane<mode>): Likewise.
651 (neon_vmul_lane<mode>): Likewise.
652 (neon_vmull_lane<mode>): Likewise.
653 (neon_vqdmull_lane<mode>): Likewise.
654 (neon_vqdmulh_lane<mode>): Likewise.
655 (neon_vqdmulh_lane<mode>): Likewise.
656 (neon_vmla_lane<mode>): Likewise.
657 (neon_vmla_lane<mode>): Likewise.
658 (neon_vmlal_lane<mode>): Likewise.
659 (neon_vqdmlal_lane<mode>): Likewise.
660 (neon_vmls_lane<mode>): Likewise.
661 (neon_vmls_lane<mode>): Likewise.
662 (neon_vmlsl_lane<mode>): Likewise.
663 (neon_vqdmlsl_lane<mode>): Likewise.
664 (neon_vext<mode>): Likewise.
665 (neon_vrev64<mode>): Likewise.
666 (neon_vrev32<mode>): Likewise.
667 (neon_vrev16<mode>): Likewise.
668 (neon_vbsl<mode>_internal): Likewise.
669 (neon_vshl<mode>): Likewise.
670 (neon_vqshl<mode>): Likewise.
671 (neon_vshr_n<mode>): Likewise.
672 (neon_vshrn_n<mode>): Likewise.
673 (neon_vqshrn_n<mode>): Likewise.
674 (neon_vqshrun_n<mode>): Likewise.
675 (neon_vshl_n<mode>): Likewise.
676 (neon_vqshl_n<mode>): Likewise.
677 (neon_vqshlu_n<mode>): Likewise.
678 (neon_vshll_n<mode>): Likewise.
679 (neon_vsra_n<mode>): Likewise.
680 (neon_vsri_n<mode>): Likewise.
681 (neon_vsli_n<mode>): Likewise.
682 (neon_vtbl1v8qi): Likewise.
683 (neon_vtbl2v8qi): Likewise.
684 (neon_vtbl3v8qi): Likewise.
685 (neon_vtbl4v8qi): Likewise.
686 (neon_vtbl1v16qi): Likewise.
687 (neon_vtbl2v16qi): Likewise.
688 (neon_vcombinev16qi): Likewise.
689 (neon_vtbx1v8qi): Likewise.
690 (neon_vtbx2v8qi): Likewise.
691 (neon_vtbx3v8qi): Likewise.
692 (neon_vtbx4v8qi): Likewise.
693 (*neon_vtrn<mode>_insn): Likewise.
694 (*neon_vzip<mode>_insn): Likewise.
695 (*neon_vuzp<mode>_insn): Likewise.
696 (neon_vld1<mode>): Likewise.
697 (neon_vld1_lane<mode>): Likewise.
698 (neon_vld1_lane<mode>): Likewise.
699 (neon_vld1_dup<mode>): Likewise.
700 (neon_vld1_dup<mode>): Likewise.
701 (neon_vld1_dupv2di): Likewise.
702 (neon_vst1<mode>): Likewise.
703 (neon_vst1_lane<mode>): Likewise.
704 (neon_vst1_lane<mode>): Likewise.
705 (neon_vld2<mode>): Likewise.
706 (neon_vld2<mode>): Likewise.
707 (neon_vld2_lane<mode>): Likewise.
708 (neon_vld2_lane<mode>): Likewise.
709 (neon_vld2_dup<mode>): Likewise.
710 (neon_vst2<mode>): Likewise.
711 (neon_vst2<mode>): Likewise.
712 (neon_vst2_lane<mode>): Likewise.
713 (neon_vst2_lane<mode>): Likewise.
714 (neon_vld3<mode>): Likewise.
715 (neon_vld3qa<mode>): Likewise.
716 (neon_vld3qb<mode>): Likewise.
717 (neon_vld3_lane<mode>): Likewise.
718 (neon_vld3_lane<mode>): Likewise.
719 (neon_vld3_dup<mode>): Likewise.
720 (neon_vst3<mode>): Likewise.
721 (neon_vst3qa<mode>): Likewise.
722 (neon_vst3qb<mode>): Likewise.
723 (neon_vst3_lane<mode>): Likewise.
724 (neon_vst3_lane<mode>): Likewise.
725 (neon_vld4<mode>): Likewise.
726 (neon_vld4qa<mode>): Likewise.
727 (neon_vld4qb<mode>): Likewise.
728 (neon_vld4_lane<mode>): Likewise.
729 (neon_vld4_lane<mode>): Likewise.
730 (neon_vld4_dup<mode>): Likewise.
731 (neon_vst4<mode>): Likewise.
732 (neon_vst4qa<mode>): Likewise.
733 (neon_vst4qb<mode>): Likewise.
734 (neon_vst4_lane<mode>): Likewise.
735 (neon_vst4_lane<mode>): Likewise.
736 (neon_vec_unpack<US>_lo_<mode>): Likewise.
737 (neon_vec_unpack<US>_hi_<mode>): Likewise.
738 (neon_vec_<US>mult_lo_<mode>): Likewise.
739 (neon_vec_<US>mult_hi_<mode>): Likewise.
740 (neon_vec_<US>shiftl_<mode>): Likewise.
741 (neon_unpack<US>_<mode>): Likewise.
742 (neon_vec_<US>mult_<mode>): Likewise.
743 (vec_pack_trunc_<mode>): Likewise.
744 (neon_vec_pack_trunc_<mode>): Likewise.
745 (neon_vabd<mode>_2): Likewise.
746 (neon_vabd<mode>_3): Likewise.
747
a9e66678
JG
7482013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
749
750 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
751 (load_pair): Update type attribute.
752 (store_pair): Update type attribute.
753 * config/aarch64/iterators.md (q): New.
754
20445ed6
JG
7552013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
756
757 * config/arm/types.md: Add new types for Neon insns.
758
afb4ac68
AI
7592013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
760 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
761 Sergey Lega <sergey.s.lega@intel.com>
762 Anna Tikhonova <anna.tikhonova@intel.com>
763 Ilya Tocar <ilya.tocar@intel.com>
764 Andrey Turetskiy <andrey.turetskiy@intel.com>
765 Ilya Verbin <ilya.verbin@intel.com>
766 Kirill Yukhin <kirill.yukhin@intel.com>
767 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
768
769 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
770 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
771 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
772 (rcp14<mode>): New.
773 (srcp14<mode>): Ditto.
774 (rsqrt14<mode>): Ditto.
775 (rsqrt14<mode>): Ditto.
776 (avx512f_vmscalef<mode>): Ditto.
777 (avx512f_scalef<mode>): Ditto.
778 (avx512f_getexp<mode>): Ditto.
779 (avx512f_sgetexp<mode>): Ditto.
780 (avx512f_fixupimm<mode>): Ditto.
781 (avx512f_sfixupimm<mode>): Ditto.
782 (avx512f_rndscale<mode>): Ditto.
783 (*avx512er_exp2<mode>): Ditto.
784 (*avx512er_rcp28<mode>): Ditto.
785 (avx512er_rsqrt28<mode>): Ditto.
786 (avx512f_getmant<mode>): Ditto.
787 (avx512f_getmant<mode>): Ditto.
788 (avx512f_rndscale<mode>): Fix formatting.
789
a1b499a3
JL
7902013-10-15 Martin Jambor <mjambor@suse.cz>
791
792 * ipa-utils.h (ipa_edge_within_scc): Declare.
793 * ipa-cp.c (edge_within_scc): Moved...
794 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
afb4ac68 795
2e2206fa
AI
7962013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
797 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
798 Sergey Lega <sergey.s.lega@intel.com>
799 Anna Tikhonova <anna.tikhonova@intel.com>
800 Ilya Tocar <ilya.tocar@intel.com>
801 Andrey Turetskiy <andrey.turetskiy@intel.com>
802 Ilya Verbin <ilya.verbin@intel.com>
803 Kirill Yukhin <kirill.yukhin@intel.com>
804 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
805
806 * config/i386/predicates.md (const_8_to_15_operand): New.
807 (const_16_to_31_operand): Ditto.
808 * config/i386/sse.md (V8FI): New.
809 (V16FI): Ditto.
810 (reduc_splus_v8df): Ditto.
811 (reduc_splus_v16sf): Ditto.
812 (avx512f_vextract<shuffletype>32x4_1): Ditto.
813 (vec_extract_hi_<mode>): Ditto.
814 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
815 (vec_set_lo_<mode>): Ditto.
816 (vec_set_hi_<mode>): Ditto.
817 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
818 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
819 (avx512f_pshufd_1): Ditto.
820 (avx512f_broadcast<mode>): Ditto.
821 (avx512f_broadcast<mode>): Ditto.
822 (define_split): Split vec_extract_lo into move.
823 (ssequartermode): Ditto.
824 (ssedoublemode): Extened with wider modes.
825 (vec_extract_lo_<mode>): Ditto.
826
0fe65b75
AI
8272013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
828 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
829 Sergey Lega <sergey.s.lega@intel.com>
830 Anna Tikhonova <anna.tikhonova@intel.com>
831 Ilya Tocar <ilya.tocar@intel.com>
832 Andrey Turetskiy <andrey.turetskiy@intel.com>
833 Ilya Verbin <ilya.verbin@intel.com>
834 Kirill Yukhin <kirill.yukhin@intel.com>
835 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
836
837 * config/i386/predicates.md (register_or_constm1_operand): New.
838 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
839 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
840 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
841 UNSPEC_SCATTER_PREFETCH
842 (VI48_512): New.
843 (avx512f_ucmp<mode>3): Ditto.
844 (avx512f_vternlog<mode>): Ditto.
845 (avx512f_align<mode>): Ditto.
846 (<shift_insn><mode>3): Ditto.
847 (avx512f_<rotate>v<mode>): Ditto.
848 (avx512f_<rotate><mode>): Ditto.
849 (avx512f_eq<mode>3): Ditto.
850 (avx512f_eq<mode>3_1): Ditto.
851 (avx512f_gt<mode>3): Ditto.
852 (avx512f_testm<mode>3): Ditto.
853 (avx512f_testnm<mode>3): Ditto.
854 (avx512pf_gatherpf<mode>): Ditto.
855 (*avx512pf_gatherpf<mode>_mask): Ditto.
856 (*avx512pf_gatherpf<mode>): Ditto.
857 (avx512pf_scatterpf<mode>): Ditto.
858 (*avx512pf_scatterpf<mode>_mask): Ditto.
859 (*avx512pf_scatterpf<mode>): Ditto.
860 (avx512f_vec_dup_gpr<mode>): Ditto.
861 (clz<mode>2): Ditto.
862 (conflict<mode>): Ditto.
863 (REDUC_SMINMAX_MODE): Extened with wider modes.
864 (reduc_<code>_<mode>): Ditto.
865 (vlshr<mode>3): Ditto.
866 (vashl<mode>3): Ditto.
867
ab931c71
AI
8682013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
869 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
870 Sergey Lega <sergey.s.lega@intel.com>
871 Anna Tikhonova <anna.tikhonova@intel.com>
872 Ilya Tocar <ilya.tocar@intel.com>
873 Andrey Turetskiy <andrey.turetskiy@intel.com>
874 Ilya Verbin <ilya.verbin@intel.com>
875 Kirill Yukhin <kirill.yukhin@intel.com>
876 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
877
878 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
879 UNSPEC_SCATTER.
880 (VI48F_512): New.
881 (avx512fmaskmode): Ditto.
882 (bcstscalarsuff): Ditto.
883 (avx512f_blendm<mode>): Ditto.
884 (cmp_imm_predicate): Ditto.
885 (avx512f_cmp<mode>3): Ditto.
886 (avx512f_vec_dup<mode>): Ditto.
887 (avx512f_vec_dup_mem<mode>): Ditto.
888 (avx512f_vpermi2var<mode>3): Ditto.
889 (avx512f_vpermt2var<mode>3): Ditto.
890 (vec_init<mode>): Ditto.
891 (avx512f_gathersi<mode>): Ditto.
892 (*avx512f_gathersi<mode>): Ditto.
893 (*avx512f_gathersi<mode>_2): Ditto.
894 (avx512f_gatherdi<mode>): Ditto.
895 (*avx512f_gatherdi<mode>): Ditto.
896 (*avx512f_gatherdi<mode>_2): Ditto.
897 (avx512f_scattersi<mode>): Ditto.
898 (*avx512f_scattersi<mode>): Ditto.
899 (avx512f_scatterdi<mode>): Ditto.
900 (*avx512f_scatterdi<mode>): Ditto.
901 (sseintprefix): Extened with wider modes.
902 (VEC_GATHER_IDXSI): Ditto.
903 (VEC_GATHER_IDXDI): Ditto.
904 (VEC_GATHER_SRCDI): Ditto.
905
c9acb877
MGD
9062013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
907 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
908
909 * config/arm/t-aprofile: New file.
910 * config.gcc: Handle --with-multilib-list option.
911
ecbda484
BS
9122013-10-15 Bernd Schmidt <bernds@codesourcery.com>
913
914 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
915 r2.
916
24f161fd
RB
9172013-10-15 Richard Biener <rguenther@suse.de>
918
919 * tree-loop-distribution.c (build_empty_rdg): Inline into
920 single user.
921 (rdg_flag_vertex): Inline into single user.
922 (rdg_flag_vertex_and_dependent): Likewise.
923 (build_rdg_partition_for_vertex): Remove processed bitmap.
924 (rdg_build_partitions): Simplify.
925
5de989ed
RB
9262013-10-15 Richard Biener <rguenther@suse.de>
927
928 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
929 Restructure forwarding through conversions and copies to
930 avoid performing copy-propagation the wrong way. Adjust
931 recursion invocations.
932 (forward_propagate_addr_expr): Add argument stating if we
933 are recursing from a single-use.
934 (ssa_forward_propagate_and_combine): Adjust.
935
47e0da37
DM
9362013-10-14 David Malcolm <dmalcolm@redhat.com>
937
938 * dumpfile.h (gcc::dump_manager): New class, to hold state
939 relating to dumpfile management.
940 (get_dump_file_name): Remove in favor of method of dump_manager.
941 (dump_initialized_p): Likewise.
942 (dump_start): Likewise.
943 (dump_finish): Likewise.
944 (dump_switch_p): Likewise.
945 (dump_register): Likewise.
946 (get_dump_file_info): Likewise.
947 * context.c (gcc::context::context): Construct the dump_manager
948 instance.
949 * context.h (gcc::context::get_dumps): New.
950 (gcc::context::m_dumps): New.
951 * coverage.c (coverage_init): Port to dump_manager API.
952 * dumpfile.c (extra_dump_files): Convert to field of
953 gcc::dump_manager.
954 (extra_dump_files_in_use): Likewise.
955 (extra_dump_files_alloced): Likewise.
956 (gcc::dump_manager::dump_manager): New.
957 (dump_register): Convert to...
958 (gcc::dump_manager::dump_register): ...method, replacing
959 function-static next_dump with m_next_dump field.
960 (get_dump_file_info): Convert to...
961 (gcc::dump_manager::get_dump_file_info): ...method.
962 (get_dump_file_name): Convert to...
963 (gcc::dump_manager::get_dump_file_name): ...method.
964 (dump_start): Convert to...
965 (gcc::dump_manager::dump_start): ...method.
966 (dump_finish): Convert to...
967 (gcc::dump_manager::dump_finish): ...method.
968 (dump_begin): Replace body with...
969 (gcc::dump_manager::dump_begin): ...new method.
970 (dump_phase_enabled_p): Convert to...
971 (gcc::dump_manager::dump_phase_enabled_p): ...method.
972 (dump_phase_enabled_p): Convert to...
973 (gcc::dump_manager::dump_phase_enabled_p): ...method.
974 (dump_initialized_p): Convert to...
975 (gcc::dump_manager::dump_initialized_p): ...method.
976 (dump_flag_name): Replace body with...
977 (gcc::dump_manager::dump_flag_name): ...new method.
978 (dump_enable_all): Convert to...
979 (gcc::dump_manager::dump_enable_all): ...new method.
980 (opt_info_enable_passes): Convert to...
981 (gcc::dump_manager::opt_info_enable_passes): ...new method.
982 (dump_switch_p_1): Convert to...
983 (gcc::dump_manager::dump_switch_p_1): ...new method.
984 (dump_switch_p): Convert to...
985 (gcc::dump_manager::dump_switch_p): ...new method.
986 (opt_info_switch_p): Port to dump_manager API.
987 (enable_rtl_dump_file): Likewise.
988 * opts-global.c (handle_common_deferred_options): Port to new
989 dump_manager API.
990 * passes.c (pass_manager::finish_optimization_passes): Likewise.
991 (pass_manager::register_one_dump_file): Likewise.
992 (pass_manager::register_pass): Likewise.
993 (pass_init_dump_file): Likewise.
994 (pass_fini_dump_file): Likewise.
995 * statistics.c (statistics_early_init): Likewise.
996
83fd5d11
RB
9972013-10-14 Richard Biener <rguenther@suse.de>
998
999 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
1000 iterative_hash_canonical_type, gimple_canonical_type_hash,
1001 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
1002 gimple_register_canonical_type, print_gimple_types_stats,
1003 free_gimple_type_tables): Move to lto/lto.c
1004 (gt-gimple.h): Do not include.
1005 * gimple.h (gimple_register_canonical_type,
1006 print_gimple_types_stats, free_gimple_type_tables): Remove.
1007 * Makefile.in (GTFILES): Remove gimple.c.
1008
a6dbd0f7
TS
10092013-10-14 Travis Snoozy <quandary@remstate.com>
1010
1011 PR target/58716
1012 * config/msp430/msp430.c (msp430_option_override): Correct thinko
1013 scanning for msp430x targets.
1014
b5ccb9ed
EB
10152013-10-14 Eric Botcazou <ebotcazou@adacore.com>
1016
1017 PR bootstrap/58509
1018 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
1019 (registers_ok_for_ldd_peep): Move around.
1020 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
1021 * config/sparc/sparc.md (widening peepholes): Use it.
1022
fd7e5084
RB
10232013-10-14 Richard Biener <rguenther@suse.de>
1024
1025 PR middle-end/58712
1026 PR middle-end/55358
1027 * gimple.c (iterative_hash_canonical_type): Make sure to
1028 record the hash into the correct hashtable slot.
1029
3246afaa
EB
10302013-10-13 Eric Botcazou <ebotcazou@adacore.com>
1031
1032 PR rtl-optimization/58662
1033 * combine.c (try_combine): Take into account death nodes on I2 when
1034 splitting a PARALLEL of two independent SETs. Fix dump message.
1035
5d30dc5b
OE
10362013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1037
1038 PR target/51244
1039 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1040 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1041 * config/sh/t-sh (sh_treg_combine.o): New entry.
1042 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1043 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1044 (register_sh_passes): New function. Register sh_treg_combine pass.
1045 (sh_option_override): Invoke it.
1046 (sh_canonicalize_comparison): Handle op0_preserve_value.
1047 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1048 opportunities. Canonicalize branch condition.
1049 (nott): Allow only if pseudos can be created for non-SH2A.
1050
37d1c5db
L
10512013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1052
1053 PR target/58690
1054 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1055 (ix86_expand_movmem): Replace copy_addr_to_reg with
1056 ix86_copy_addr_to_reg.
1057 (ix86_expand_setmem): Likewise.
1058
b86a70ff
AM
10592013-10-12 Alexander Monakov <amonakov@ispras.ru>
1060
1061 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1062 provided by ix86_fp_compare_mode instead of CCFPUmode.
1063
6e1d5961
JG
10642013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1065
1066 * config/aarch64/arm_neon.h
1067 (vtbx<1,3>_<psu>8): Fix register constriants.
1068
ef3cfba2
JL
10692013-10-11 Jeff Law <law@redhat.com>
1070
1071 PR tree-optimization/58640
adcfd489
UB
1072 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1073 threading paths that cross over two loop entry points.
ef3cfba2 1074
6e8b7d9c
BS
10752013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1076
1077 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1078 handle vector float as well.
1079 (*vsx_le_perm_load_v4si): Likewise.
1080 (*vsx_le_perm_store_v2di): Likewise.
1081 (*vsx_le_perm_store_v4si): Likewise.
1082
f200869a
BS
10832013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1084
1085 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1086 directly to circumvent subtract from splat{31} workaround.
1087 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1088 prototype.
1089 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1090 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1091 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1092 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1093 separate big and little endian logic.
1094 (*altivec_vperm_<mode>_internal): New define_insn.
1095 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1096 separate big and little endian logic.
1097 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1098 (vec_permv16qi): Add little endian logic.
1099
8d28e3fc
MG
11002013-10-11 Marc Glisse <marc.glisse@inria.fr>
1101
1102 * doc/extend.texi (returns_nonnull): Remove arguments.
1103
3c87b77b
AI
11042013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1105 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1106 Sergey Lega <sergey.s.lega@intel.com>
1107 Anna Tikhonova <anna.tikhonova@intel.com>
1108 Ilya Tocar <ilya.tocar@intel.com>
1109 Andrey Turetskiy <andrey.turetskiy@intel.com>
1110 Ilya Verbin <ilya.verbin@intel.com>
1111 Kirill Yukhin <kirill.yukhin@intel.com>
1112 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1113
1114 * config/i386/sse.md (VI48F_256_512): New.
1115 (avx2_permvar<mode>): Change to ...
1116 (<avx2_avx512f>_permvar<mode>): This.
1117
50e60d7d
AI
11182013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1119 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1120 Sergey Lega <sergey.s.lega@intel.com>
1121 Anna Tikhonova <anna.tikhonova@intel.com>
1122 Ilya Tocar <ilya.tocar@intel.com>
1123 Andrey Turetskiy <andrey.turetskiy@intel.com>
1124 Ilya Verbin <ilya.verbin@intel.com>
1125 Kirill Yukhin <kirill.yukhin@intel.com>
1126 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1127
1128 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1129 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1130 * config/i386/sse.md (VI4_AVX): New.
1131 (sf2simodelower): Ditto.
1132 (sse2_cvtps2dq): Change to ...
1133 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1134
f62ce24f
AI
11352013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1136 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1137 Sergey Lega <sergey.s.lega@intel.com>
1138 Anna Tikhonova <anna.tikhonova@intel.com>
1139 Ilya Tocar <ilya.tocar@intel.com>
1140 Andrey Turetskiy <andrey.turetskiy@intel.com>
1141 Ilya Verbin <ilya.verbin@intel.com>
1142 Kirill Yukhin <kirill.yukhin@intel.com>
1143 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1144
1145 * config/i386/sse.md (V_512): New.
1146 (VI_512): Ditto.
1147 (vcond<V_512:mode><VF_512:mode>): Ditto.
1148 (vcond<V_512:mode><VI_512:mode>): Ditto.
1149 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1150
558d9f79
AI
11512013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1152 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1153 Sergey Lega <sergey.s.lega@intel.com>
1154 Anna Tikhonova <anna.tikhonova@intel.com>
1155 Ilya Tocar <ilya.tocar@intel.com>
1156 Andrey Turetskiy <andrey.turetskiy@intel.com>
1157 Ilya Verbin <ilya.verbin@intel.com>
1158 Kirill Yukhin <kirill.yukhin@intel.com>
1159 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1160
1161 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1162 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1163 (FMAMODE): Ditto.
1164 (fma<mode>4): Removed condition.
1165 (fms<mode>4): Ditto.
1166 (fnma<mode>4): Ditto.
1167 (fnms<mode>4): Ditto.
1168 (fma4i_fmadd_<mode>): Ditto.
1169 (*fma_fmadd_<mode>): Ditto.
1170 (*fma_fmsub_<mode>): Ditto.
1171 (*fma_fnmadd_<mode>): Ditto.
1172 (*fma_fnmsub_<mode>): Ditto.
1173 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1174 (*fma_fmaddsub_<mode>): Ditto.
1175 (*fma_fmsubadd_<mode>): Ditto.
1176 (*fmai_fmadd_<mode>): Ditto.
1177 (*fmai_fmsub_<mode>): Ditto.
1178 (*fmai_fnmadd_<mode>): Ditto.
1179 (*fmai_fnmsub_<mode>): Ditto.
1180
e8d08206
AI
11812013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1182 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1183 Sergey Lega <sergey.s.lega@intel.com>
1184 Anna Tikhonova <anna.tikhonova@intel.com>
1185 Ilya Tocar <ilya.tocar@intel.com>
1186 Andrey Turetskiy <andrey.turetskiy@intel.com>
1187 Ilya Verbin <ilya.verbin@intel.com>
1188 Kirill Yukhin <kirill.yukhin@intel.com>
1189 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1190
1191 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1192 (VI124_256): Changed to ...
1193 (VI124_256_48_512): This.
1194 (ssepackmode): Extended with wider modes.
1195 (<code><mode>3): Changed iterator.
1196 (*avx2_<code><mode>3): Ditto.
1197 (vec_pack_trunc_<mode>): Ditto.
1198
16821545
AI
11992013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1200 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1201 Sergey Lega <sergey.s.lega@intel.com>
1202 Anna Tikhonova <anna.tikhonova@intel.com>
1203 Ilya Tocar <ilya.tocar@intel.com>
1204 Andrey Turetskiy <andrey.turetskiy@intel.com>
1205 Ilya Verbin <ilya.verbin@intel.com>
1206 Kirill Yukhin <kirill.yukhin@intel.com>
1207 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1208
1209 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1210 (VI8F_256_512): Ditto.
1211 (abs<mode>2): Changed iterator.
1212 (avx2_perm<mode>): Changed to ...
1213 (<avx2_avx512f>_perm<mode>): This.
1214 (avx2_perm<mode>_1): Changed to ...
1215 (<avx2_avx512f>_perm<mode>_1): This.
1216
5348cff8
AI
12172013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1218 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1219 Sergey Lega <sergey.s.lega@intel.com>
1220 Anna Tikhonova <anna.tikhonova@intel.com>
1221 Ilya Tocar <ilya.tocar@intel.com>
1222 Andrey Turetskiy <andrey.turetskiy@intel.com>
1223 Ilya Verbin <ilya.verbin@intel.com>
1224 Kirill Yukhin <kirill.yukhin@intel.com>
1225 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1226
1227 * config/i386/sse.md (VI48_AVX512F): New.
1228 (VI48_AVX2): Changed to ...
1229 (VI48_AVX2_48_AVX512F): This.
1230 (avx2_ashrv<mode>): Changed to ...
1231 (<avx2_avx512f>_ashrv<mode>): This.
1232 (avx2_<shift_insn>v<mode>): Changed to ...
1233 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1234
f5f41d88
AI
12352013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1236 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1237 Sergey Lega <sergey.s.lega@intel.com>
1238 Anna Tikhonova <anna.tikhonova@intel.com>
1239 Ilya Tocar <ilya.tocar@intel.com>
1240 Andrey Turetskiy <andrey.turetskiy@intel.com>
1241 Ilya Verbin <ilya.verbin@intel.com>
1242 Kirill Yukhin <kirill.yukhin@intel.com>
1243 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1244
1245 * config/i386/sse.md (VI4_AVX512F): New.
1246 (VI8_AVX2_AVX512F): Ditto.
1247 (mul<mode>3): Extended with wider modes.
1248 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1249 (mul<mode>3): Ditto.
1250 (vec_widen_<s>mult_odd_<mode>): Ditto.
1251
3bdf6340
AI
12522013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1253 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1254 Sergey Lega <sergey.s.lega@intel.com>
1255 Anna Tikhonova <anna.tikhonova@intel.com>
1256 Ilya Tocar <ilya.tocar@intel.com>
1257 Andrey Turetskiy <andrey.turetskiy@intel.com>
1258 Ilya Verbin <ilya.verbin@intel.com>
1259 Kirill Yukhin <kirill.yukhin@intel.com>
1260 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1261
1262 * config/i386/sse.md (VI2_AVX512F): New.
1263 (VI124_AVX512F): Ditto.
1264 (sseunpackmode): Extended with wider modes.
1265 (sseunpackfltmode): Ditto.
1266 (vec_unpacks_float_hi_<mode>): Ditto.
1267 (vec_unpacks_float_lo_<mode>): Ditto.
1268 (vec_unpacku_float_hi_<mode>): Ditto.
1269 (vec_unpacku_float_lo_<mode>): Ditto.
1270 (vec_unpacks_lo_<mode>): Ditto.
1271 (vec_unpacks_hi_<mode>): Ditto.
1272 (vec_unpacku_lo_<mode>): Ditto.
1273 (vec_unpacku_hi_<mode>): Ditto.
1274
67f783cb
AI
12752013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1276 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1277 Sergey Lega <sergey.s.lega@intel.com>
1278 Anna Tikhonova <anna.tikhonova@intel.com>
1279 Ilya Tocar <ilya.tocar@intel.com>
1280 Andrey Turetskiy <andrey.turetskiy@intel.com>
1281 Ilya Verbin <ilya.verbin@intel.com>
1282 Kirill Yukhin <kirill.yukhin@intel.com>
1283 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1284
1285 * config/i386/i386.md (multdiv): New.
1286 (multdiv_mnemonic): Ditto.
1287 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1288 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1289 (<sse>_vmdiv<mode>3): Removed.
1290
ec5e777c
AI
12912013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1292 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1293 Sergey Lega <sergey.s.lega@intel.com>
1294 Anna Tikhonova <anna.tikhonova@intel.com>
1295 Ilya Tocar <ilya.tocar@intel.com>
1296 Andrey Turetskiy <andrey.turetskiy@intel.com>
1297 Ilya Verbin <ilya.verbin@intel.com>
1298 Kirill Yukhin <kirill.yukhin@intel.com>
1299 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1300
1301 * config/i386/sse.md (V): Extended with wider modes.
1302 (VF2): Ditto.
1303 (ssehalfvecmode): Ditto.
1304 (i128): Ditto.
1305 (ssepackfltmode): Ditto.
1306 (avx_vec_concat<mode>): Ditto.
1307 (V_256_512): New iterator.
1308 (VF2_512_256): Ditto.
1309 (si2dfmode): New attribute.
1310 (si2dfmodelower): Ditto.
1311 (sf2dfmode): Ditto.
1312 (concat_tg_mode): Ditto.
1313 (floatv4siv4df2): Changed to ...
1314 (float<si2dfmodelower><mode>2): This.
1315 (avx_cvtps2pd256): Changed to ...
1316 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1317 (vec_pack_trunc_v4df): Changed to ...
1318 (vec_pack_trunc_<mode>): This.
1319 (avx_vpermil<mode>): Changed to ...
1320 (<sse2_avx_avx512f>_vpermil<mode>): This.
1321 (<fixsuffix>fix_truncv8dfv8si2): New.
1322 (vec_pack_sfix_trunc_v8df): Ditto.
1323 (avx512f_rndscale<mode>): Ditto.
1324 (avx512f_roundpd512): Ditto.
1325 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1326
a9ccbba2
AI
13272013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1328 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1329 Sergey Lega <sergey.s.lega@intel.com>
1330 Anna Tikhonova <anna.tikhonova@intel.com>
1331 Ilya Tocar <ilya.tocar@intel.com>
1332 Andrey Turetskiy <andrey.turetskiy@intel.com>
1333 Ilya Verbin <ilya.verbin@intel.com>
1334 Kirill Yukhin <kirill.yukhin@intel.com>
1335 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1336
1337 * config/i386/i386.md (any_fix): New iterator.
1338 (fixsuffix): New attribute.
1339 * config/i386/sse.md (VF1): Extened with wider modes.
1340 (VI): Ditto.
1341 (VI_AVX2): Ditto.
1342 (VI8): Ditto.
1343 (sseintvecmodelower): Ditto.
1344 (ssescalarmode): Ditto.
1345 (ssescalarnum): Ditto.
1346 (VF1_128_256): New.
1347 (ssexmmmode): Ditto.
1348 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1349 (<sse>_rcp<mode>2): Change iterator.
1350 (rsqrt<mode>2): Ditto.
1351 (<sse>_rsqrt<mode>2): Ditto.
1352 (avx2_vec_dup<mode>): Ditto.
1353 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1354 (round<mode>2_sfix): Ditto.
1355 (avx2_pbroadcast<mode>): Ditto.
1356 (*andnot<mode>3): Handle XI mode.
1357 (*<code><mode>3): Ditto.
1358 (AVXTOSSEMODE): Removed.
1359 (avx_vpermil<mode>): Changed to ...
1360 (<sse2_avx_avx512f>_vpermil<mode>): This.
1361
2b1ebb0c
AI
13622013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1363 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1364 Sergey Lega <sergey.s.lega@intel.com>
1365 Anna Tikhonova <anna.tikhonova@intel.com>
1366 Ilya Tocar <ilya.tocar@intel.com>
1367 Andrey Turetskiy <andrey.turetskiy@intel.com>
1368 Ilya Verbin <ilya.verbin@intel.com>
1369 Kirill Yukhin <kirill.yukhin@intel.com>
1370 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1371
1372 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1373 (<sse>_comi): Ditto.
1374 (<sse>_ucomi): Ditto.
1375 (sse_cvtss2siq_2): Ditto.
1376 (sse2_cvtsd2si): Ditto.
1377 (sse2_cvtsd2siq): Ditto.
1378 (sse2_cvttsd2si): Ditto.
1379 (sse2_cvttsd2siq): Ditto.
1380 (<shift_insn><mode>3): Ditto.
1381 (sse2_cvtsi2sdq): Update constraint and prefix.
1382 (sse_cvtsi2ss): Update prefix.
1383 (sse_cvtsi2ssq): Ditto.
1384
d89124ed
JJ
13852013-10-11 Jakub Jelinek <jakub@redhat.com>
1386
1387 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1388 ignore internal calls.
1389
e6ad28c3
RB
13902013-10-11 Richard Biener <rguenther@suse.de>
1391
1392 * tree-pretty-print.c (dump_generic_node): Allow to dump
1393 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1394 (OVF) if TREE_OVERFLOW is set.
1395
d1227447
TS
13962013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1397
1ac4b8e4
TS
1398 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1399
7d05cebb
TS
1400 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1401 * gimple.h: Likewise.
1402
adcfd489 1403 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
6545f0b3 1404
1a79fb8a
TS
1405 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1406 macros.
1407
3afd2873
TS
1408 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1409
adcfd489 1410 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
d1227447
TS
1411 * configure: Regenerate.
1412
acf0174b
JJ
14132013-10-11 Jakub Jelinek <jakub@redhat.com>
1414
1415 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1416 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1417 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1418 similarly pp_right_paren.
1419 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1420 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1421 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1422 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1423 clauses.
1424 (omp_declare_simd_clauses_equal,
1425 omp_remove_redundant_declare_simd_attrs): New functions.
1426 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1427 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1428 * tree.h (OMP_LOOP_CHECK): Define.
1429 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1430 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1431 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1432 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1433 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1434 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1435 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1436 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1437 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1438 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1439 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1440 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1441 (omp_remove_redundant_declare_simd_attrs): New prototype.
1442 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1443 GIMPLE_OMP_TEAMS): New codes.
1444 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1445 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1446 fields.
1447 (target_nesting_level): New variable.
1448 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1449 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1450 for collapse > 1 static schedule unless ordered.
1451 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1452 (determine_parallel_type): Adjust get_ws_args_for caller.
1453 (install_var_field): Handle mask & 4 for double indirection.
1454 (scan_sharing_clauses): Ignore shared clause on teams construct.
1455 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1456 (create_omp_child_function): If inside target or declare target
adcfd489 1457 constructs, set "omp declare target" attribute on the child function.
acf0174b
JJ
1458 (find_combined_for): New function.
1459 (scan_omp_parallel): Handle combined loops.
1460 (scan_omp_target, scan_omp_teams): New functions.
1461 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1462 restrictions and set ctx->cancellable for cancellable constructs.
1463 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1464 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1465 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
adcfd489 1466 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
acf0174b
JJ
1467 (omp_clause_aligned_alignment): New function.
1468 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1469 on decls.
1470 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1471 on teams constructs. Handle user defined reductions and new
1472 OpenMP 4.0 clauses.
1473 (lower_reduction_clauses): Don't set placeholder to address of ref
1474 if it has already the right type.
1475 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1476 (expand_parallel_call): Use the new non-_start suffixed builtins,
1477 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1478 and GOMP_parallel_end after the call.
1479 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1480 (expand_omp_for_init_counts): Handle combined loops.
1481 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1482 loops.
1483 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1484 end of cancellable loops.
1485 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1486 Likewise. Handle collapse > 1 loops.
1487 (expand_omp_simd): Handle combined loops.
1488 (expand_omp_for): Add inner_stmt argument, adjust callers of
1489 expand_omp_for* functions, use expand_omp_for_static*chunk even
1490 for collapse > 1 unless ordered.
1491 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1492 of cancellable sections.
1493 (expand_omp_single): Remove need_barrier variable, just rely on
1494 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1495 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1496 (expand_omp_atomic_load, expand_omp_atomic_store,
1497 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1498 (expand_omp_target): New function.
1499 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1500 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1501 (build_omp_regions_1): Immediately close region for
1502 GF_OMP_TARGET_KIND_UPDATE.
1503 (maybe_add_implicit_barrier_cancel): New function.
1504 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1505 cancellation.
1506 (lower_omp_single): Likewise. Add clobber after the barrier.
1507 (lower_omp_taskgroup): New function.
1508 (lower_omp_for): Handle combined loops. Adjust
1509 lower_rec_input_clauses caller. Handle cancellation.
1510 (lower_depend_clauses): New function.
1511 (lower_omp_taskreg): Lower depend clauses. Adjust
1512 lower_rec_input_clauses caller. Add clobber after the call. Handle
1513 cancellation.
1514 (lower_omp_target, lower_omp_teams): New functions.
1515 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1516 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1517 and GOMP_cancellation_point calls.
1518 (lower_omp): Fold stmts inside of target region.
1519 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1520 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1521 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1522 (BT_FN_VOID_OMPFN_PTR_UINT,
1523 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1524 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1525 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1526 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1527 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1528 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1529 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1530 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1531 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1532 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1533 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1534 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1535 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1536 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1537 BUILT_IN_GOMP_PARALLEL_END.
1538 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1539 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1540 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1541 GF_OMP_FOR_KIND_DISTRIBUTE.
1542 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1543 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1544 (dump_gimple_omp_return): Print lhs if it has any.
1545 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1546 gimple_omp_atomic_seq_cst_p.
1547 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1548 and GIMPLE_OMP_TEAMS.
1549 * langhooks.c (lhd_omp_mappable_type): New function.
1550 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1551 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1552 hook.
1553 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1554 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1555 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1556 ORT_TARGET.
1557 (struct gimplify_omp_ctx): Add combined_loop field.
1558 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1559 on stmts inside of target region.
1560 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1561 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1562 ORT_TARGET and ORT_TARGET_DATA.
1563 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1564 Handle GOVD_MAP.
1565 (omp_notice_threadprivate_variable): Complain about threadprivate
1566 variables in target region.
1567 (omp_notice_variable): Complain about vars with non-mappable type
1568 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1569 (omp_check_private): Ignore ORT_TARGET* regions.
1570 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1571 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1572 (find_combined_omp_for): New function.
1573 (gimplify_omp_for): Handle gimplification of combined loops.
1574 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1575 OMP_TEAMS.
1576 (gimplify_omp_target_update): New function.
1577 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1578 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1579 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1580 (gimplify_body): If fndecl has "omp declare target" attribute, add
1581 implicit ORT_TARGET context around it.
1582 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1583 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1584 * tree-nested.c (convert_nonlocal_reference_stmt,
1585 convert_local_reference_stmt, convert_gimple_call): Handle
1586 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1587 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1588 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1589 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1590 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1591 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1592 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1593 BUILT_IN_GOMP_LOOP_END_CANCEL,
1594 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1595 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1596 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1597 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1598 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1599 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1600 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1601 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1602 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1603 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1604 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1605 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1606 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1607 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1608 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1609 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1610 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1611 gimple_build_omp_teams): New functions.
1612 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1613 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1614 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1615 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1616 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1617 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1618 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1619 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1620 GF_OMP_ATOMIC_SEQ_CST): New.
1621 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1622 gimple_build_omp_teams): New prototypes.
1623 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1624 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1625 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1626 GIMPLE_OMP_SINGLE as end of range.
1627 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1628 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1629 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1630 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1631 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1632 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1633 gimple_omp_target_kind, gimple_omp_target_set_kind,
1634 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1635 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1636 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1637 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1638 gimple_omp_teams_set_clauses): New inlines.
1639 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1640 and GIMPLE_OMP_TASKGROUP.
adcfd489 1641 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
acf0174b
JJ
1642 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1643 enum omp_clause_proc_bind_kind): New.
1644 (union omp_clause_subcode): Add depend_kind, map_kind and
1645 proc_bind_kind fields.
1646 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1647 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1648 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1649 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1650 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1651
67fa7880
TJ
16522013-10-10 Teresa Johnson <tejohnson@google.com>
1653
1654 * predict.c (tree_estimate_probability): Add new parameter
1655 for estimate_bb_frequencies.
1656 (estimate_bb_frequencies): Add new parameter to force estimation.
1657 (rebuild_frequencies): When max frequency in function is small,
1658 recompute counts from frequencies.
1659 * predict.h (estimate_bb_frequencies): New parameter.
1660
b591a8b7
DM
16612013-10-10 David Malcolm <dmalcolm@redhat.com>
1662
1663 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1664 optimizations are disabled.
1665
782f0db2
DM
16662013-10-10 David Malcolm <dmalcolm@redhat.com>
1667
1668 * coverage.c (coverage_finish): Fix leak of da_file_name.
1669
13ef00fa
JH
16702013-10-10 Jan Hubicka <jh@suse.cz>
1671
1672 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1673 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1674 Bobcat and generic.
1675
a1d50386
JJ
16762013-10-10 Jakub Jelinek <jakub@redhat.com>
1677
1678 PR middle-end/58670
1679 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1680 if any labels are in FALLTHRU_BB, use a special label emitted
1681 immediately after the asm goto insn rather than label_rtx
1682 of the LABEL_DECL.
1683 (expand_asm_stmt): Adjust caller.
1684 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1685 edge if the last insn in predecessor is a jump with single successor,
1686 but it isn't simplejump_p.
1687
e2c2fde2
RB
16882013-10-10 Richard Biener <rguenther@suse.de>
1689
1690 PR tree-optimization/58656
1691 * tree-ssa-pre.c (phi_translate): Do not cache failed
1692 translations.
1693
de6a5608
AM
16942013-10-10 Andrew MacLeod <amacleod@redhat.com>
1695
adcfd489
UB
1696 * gimplify.c: Include expr.h and tm_p.h for targets with special
1697 va-arg padding requirements.
de6a5608 1698
b184c8f1
AM
16992013-10-10 Andrew MacLeod <amacleod@redhat.com>
1700
1701 * tree-flow.h: Move some prototypes to gimple.h.
1702 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1703 * gimple.h: Relocate some prototypes from tree-flow.h
1704 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1705 Move to gimplify.c.
1706 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1707 (build_va_arg_indirect_ref): Relocate and make static.
1708 (std_gimplify_va_arg_expr): Relocate here.
1709 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1710 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1711
13b72c22
AK
17122013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1713
1714 * doc/md.texi: Document the mnemonic attribute.
1715
5db40447
AK
17162013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1717
1718 PR target/57377
1719 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1720 (set_attr_alternative x ...) when searching for user defined
1721 mnemonic attribute.
1722
ed018d4b
AM
17232013-10-10 Andrew MacLeod <amacleod@redhat.com>
1724
1725 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1726
4d5b5e9f
ER
17272013-10-09 Easwaran Raman <eraman@google.com>
1728
1729 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1730 * cfgexpand.c (defer_stack_allocation): ...use here
1731 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1732
c9ef86a1
ZC
17332013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1734
1735 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1736 (operand_equal_for_value_replacement): New function, extracted from
1737 value_replacement and enhanced to catch more cases.
1738 (value_replacement): Use operand_equal_for_value_replacement.
1739
2431114f
AM
17402013-10-09 Andrew MacLeod <amacleod@redhat.com>
1741
adcfd489 1742 * loop-doloop.c (doloop_modify, doloop_optimize): Use
2431114f
AM
1743 get_max_loop_iterations.
1744
1dd2a9a0
KT
17452013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1746
1747 * config/arm/aarch-common.c (arm_early_load_addr_dep):
1748 Place comment above function.
1749
c1bf2a39
AM
17502013-10-09 Andrew MacLeod <amacleod@redhat.com>
1751
1752 * tree-flow.h: Remove all remaining prototypes, enums and structs that
1753 are not related to tree-cfg.c.
1754 * tree-ssa-address.h: New file. Relocate prototypes.
1755 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
1756 (addr_for_mem_ref): New. Combine call to get_address_description and
1757 return addr_for_mem_ref.
1758 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
1759 * tree-ssa-live.h: Adjust prototypes.
1760 * passes.c: Include tree-ssa-live.h.
1761 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
1762 * graphite.c (graphite_transform_loops): Make static.
1763 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
adcfd489 1764 make_pass_graphite, pass_data_graphite_transforms,
c1bf2a39
AM
1765 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
1766 * ipa-pure-const.c (warn_function_noreturn): Make static.
1767 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1768 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1769 Relocate from tree-cfg.c
1770 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
1771 static.
1772 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1773 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1774 Move to ipa-pure-const.c.
adcfd489
UB
1775 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
1776 Relocate from tree-optimize.c.
c1bf2a39
AM
1777 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
1778 make_pass_fixup_cfg): Move to tree-cfg.c.
1779 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
1780 Relocate some prototypes.
1781 * tree-data-ref.h (tree_check_data_deps) Add prototype.
1782 * tree-dump.c (dump_function_to_file): Remove prototype.
1783 Add tree-flow.h to the include file.
1784 * tree-dump.h: Remove prototype.
1785 * tree-parloops.h: New File. Add prototypes.
1786 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
1787 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
1788 from tree-ssa-loop.c.
1789 * tree-predcom.c (run_tree_predictive_commoning,
1790 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
1791 Relocate here from tree-ssa-loop.c.
adcfd489 1792 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
c1bf2a39
AM
1793 ssa_name_values.release ().
1794 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
1795 (ssa_name_values): Relocate from tree-flow.h.
1796 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
1797 * tree-ssa-loop.c (run_tree_predictive_commoning,
1798 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
1799 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1800 make_pass_graphite, pass_data_graphite_transforms,
1801 make_pass_graphite_transforms, gate_tree_parallelize_loops,
1802 tree_parallelize_loops, pass_data_parallelize_loops,
1803 make_pass_parallelize_loops): Move to other files.
1804 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
1805 moved here.
1806 * tree.h: Remove prototypes from tree-address.c.
1807
cc524fc7
AM
18082013-10-09 Andrew MacLeod <amacleod@redhat.com>
1809
1810 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
1811 (struct int_tree_map): Move to tree-hasher.h
1812 (SCALE, LABEL, PERCENT): Move to gimple.h
1813 * tree-flow-inline.h: Delete. Move functions to other files.
1814 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
1815 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
1816 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
1817 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
1818 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
1819 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
1820 * tree-hasher.h: Don't include tree-flow.h.
1821 (struct int_tree_map): Relocate from tree-flow.h.
1822 * tree-sra.c (contains_view_convert_expr_p): Relocate from
1823 tree-flow-inline.h and make static.
adcfd489
UB
1824 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
1825 tree-flow-inline.h.
cc524fc7
AM
1826 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
1827 tree-flow-inline.h and make static.
1828 * tree.h (is_global_var, may_be_aliased): Relocate from
1829 tree-flow-inline.h.
1830 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
1831 * value-prof.c: No longer include tree-flow-inline.h.
1832 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
1833
71343877
AM
18342013-10-09 Andrew MacLeod <amacleod@redhat.com>
1835
1836 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
1837 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
1838 (enum move_pos): Move to tree-ssa-loop-im.h
1839 * cfgloop.h: Move some prototypes.
1840 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
1841 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
1842 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
1843 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
1844 (loop_containing_stmt): Relocate from tree-flow-inline.h.
1845 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
1846 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
1847 (enum move_pos): Relocate here.
1848 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
1849 tree-ssa-loop.c.
1850 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
adcfd489
UB
1851 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
1852 make_pass_lim): Relocate here from tree-ssa-loop.c.
1853 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
71343877
AM
1854 tree-ssa-loop.c.
1855 (loop_edge_to_cancel, unloop_loops): Make static.
1856 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
1857 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
1858 (tree_complete_unroll, gate_tree_complete_unroll,
1859 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
1860 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
1861 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
1862 * tree-ssa-loop-ivopts.c: Remove local prototypes.
1863 (stmt_invariant_in_loop_p): Remove unused function.
1864 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
1865 * tree-ssa-loop-manip.h: New file. Add prototypes.
1866 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
1867 (gcov_type_to_double_int): Move to cfgloop.h.
1868 (double_int_cmp, bound_index,
1869 estimate_numbers_of_iterations_loop): Make static.
1870 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
1871 (max_loop_iterations): Factor out get_max_loop_iterations.
1872 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
1873 cfgloop.c.
1874 * tree-ssa-loop-niter.h: New file. Add prototypes.
1875 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
1876 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
1877 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
1878 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
1879 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
1880 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
1881 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
1882 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
1883 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
1884 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
1885 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
1886 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
1887 pass_data_complete_unroll, make_pass_complete_unroll,
1888 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
1889 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
1890 tree-ssa-loop-ivcanon.c.
1891 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
1892 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
1893 tree-ssa-loop-prefetch.c.
1894 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
1895 tree-ssa-loop-im.c.
1896 (get_lsm_tmp_name): Relocate and add suffix parameter.
1897 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
1898 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
1899 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
1900 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
adcfd489 1901 (get_estimated_loop_iterations): Factor out accessor from
71343877 1902 estimated_loop_iterations in tree-ssa-loop-niter.c.
adcfd489
UB
1903 (get_max_loop_iterations): Factor out accessor from
1904 _max_loop_iterations in tree-ssa-niter.c.
71343877
AM
1905 * loop-unroll.c (decide_unroll_constant_iterations,
1906 decide_unroll_runtime_iterations, decide_peel_simple,
1907 decide_unroll_stupid): Use new get_* accessors.
1908
826cacfe
MG
19092013-10-09 Marc Glisse <marc.glisse@inria.fr>
1910
1911 PR tree-optimization/20318
1912 * doc/extend.texi (returns_nonnull): New function attribute.
1913 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
1914 attribute.
1915 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
1916 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
1917
378f8976
EB
19182013-10-09 Eric Botcazou <ebotcazou@adacore.com>
1919
1920 PR middle-end/58570
1921 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
1922 false if both components are bitfields.
1923
0fe04f5c
AV
19242013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1925
1926 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
1927 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
1928 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
1929 * config/aarch64/aarch64.h
1930 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
1931 * config/aarch64/aarch64-simd-builtins.def
1932 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
1933
bed9bae4
AV
19342013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1935
1936 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
1937
7f3d8b19
AV
19382013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1939
1940 * config/aarch64/arm_neon.h (vdiv_f64): Added.
1941
7df625a6
AV
19422013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1943
1944 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
1945 (vneg_f64): New intrinsic.
1946 (vneg_s8): Asm replaced with C.
1947 (vneg_s16): Likewise.
1948 (vneg_s32): Likewise.
1949 (vneg_s64): New intrinsic.
1950 (vnegq_f32): Asm replaced with C.
1951 (vnegq_f64): Likewise.
1952 (vnegq_s8): Likewise.
1953 (vnegq_s16): Likewise.
1954 (vnegq_s32): Likewise.
1955 (vnegq_s64): Likewise.
1956
2b86fca7
RL
19572013-10-09 Renlin Li <Renlin.Li@arm.com>
1958
1959 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
1960
e179df83
AK
19612013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1962
1963 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
1964 packed stack special handling.
1965 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
1966 back to fixed stack slots for FPRs saved due to stdarg.
1967
74129172
AK
19682013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1969
1970 * config/s390/s390.c (s390_frame_info): Restructure function.
1971
6455a49e
AK
19722013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1973
1974 * config/s390/s390.c (struct s390_frame_layout): New field
1975 gpr_save_slots.
1976 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
1977 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
1978 regs_ever_clobbered to char*.
adcfd489
UB
1979 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
1980 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
6455a49e
AK
1981 (s390_register_info): Enable FPR save slots. Move/Copy some
1982 functionality into ...
1983 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
1984 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
1985 function.
1986 (s390_frame_info): Do gpr slot allocation here now. stdarg does
1987 not imply a stack frame.
1988 (s390_init_frame_layout): Remove variable clobbered_regs.
1989 (s390_update_register_info): Remove function.
1990 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
1991 cannot be used for register renaming.
1992 (s390_hard_regno_scratch_ok): New function.
1993 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
1994 (s390_initial_elimination_offset): Change offset calculation of
1995 the return address pointer.
adcfd489 1996 (save_gprs): Deal with only r6 being saved from the call-saved regs.
6455a49e 1997 (restore_gprs): Set frame related flag.
adcfd489 1998 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
6455a49e
AK
1999 (s390_emit_prologue): Call s390_register_info instead of
2000 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
2001 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
2002 (s390_optimize_prologue): Call s390_optimize_register_info.
2003 Try to remove also FPR slot save/restore INSNs. Remove frame
2004 related flags from restore INSNs.
2005
59d96342
DD
20062013-10-08 DJ Delorie <dj@redhat.com>
2007
0dc7adad
DD
2008 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
2009 (movhi): Likewise.
2010
59d96342
DD
2011 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
2012 avoid conflict with the MI use of %c.
2013 * config/rl78/rl78-real.md: change %c to %C throughout.
2014 * config/rl78/rl78-virt.md: Likewise.
2015
a81169d8
JH
20162013-10-08 Jan Hubicka <jh@suse.cz>
2017
2018 * config/i386/i386.c (ix86_option_override_internal): Switch
2019 to SSE math for -ffast-math when target ISA supports SSE2.
2020
481d1b81
AM
20212013-10-08 Andrew MacLeod <amacleod@redhat.com>
2022
2023 * tree-flow.h: Remove some prototypes.
2024 * tree.h: Remove some protypes, add a couple.
2025 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
2026 using_eh_for_cleanups_p): Add interface routines for front ends.
2027 * tree-eh.h: New file. Add protoptyes.
2028 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
2029 (add_stmt_to_eh_lp_fn): Make static.
2030 (lower_try_finally): Use new using_eh_for_cleanups_p.
2031 * emit-rtl.c: Include tree-eh.h.
2032 * gimple.h: Include tree-eh.h.
2033
0e6a0e48
MG
20342013-10-08 Marc Glisse <marc.glisse@inria.fr>
2035
2036 PR tree-optimization/58480
2037 * tree-vrp.c (infer_nonnull_range): New function.
2038 (infer_value_range): Call infer_nonnull_range.
2039
ef6179d1
DC
20402013-10-08 Dehao Chen <dehao@google.com>
2041
2042 PR tree-optimization/58619
2043 * tree-inline.c (copy_phis_for_bb): Combine location data
2044 only if non-null.
2045
07d964d5
ZC
20462013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2047
2048 PR target/58423
2049 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2050 RTX_FRAME_RELATED_P on INSN.
2051
19e9b2a3
BS
20522013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2053
2054 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2055 (altivec_expand_vec_perm_const): Call it.
2056
0cf68694
BS
20572013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2058
2059 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2060 sequences for LE VSX loads and stores at expand time.
2061 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2062 prototype.
2063 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2064 (rs6000_gen_le_vsx_permute): New.
2065 (rs6000_gen_le_vsx_load): New.
2066 (rs6000_gen_le_vsx_store): New.
2067 (rs6000_gen_le_vsx_move): New.
2068 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2069 (*vsx_le_perm_load_v4si): New.
2070 (*vsx_le_perm_load_v8hi): New.
2071 (*vsx_le_perm_load_v16qi): New.
2072 (*vsx_le_perm_store_v2di): New.
2073 (*vsx_le_perm_store_v4si): New.
2074 (*vsx_le_perm_store_v8hi): New.
2075 (*vsx_le_perm_store_v16qi): New.
2076 (*vsx_xxpermdi2_le_<mode>): New.
2077 (*vsx_xxpermdi4_le_<mode>): New.
2078 (*vsx_xxpermdi8_le_V8HI): New.
2079 (*vsx_xxpermdi16_le_V16QI): New.
2080 (*vsx_lxvd2x2_le_<mode>): New.
2081 (*vsx_lxvd2x4_le_<mode>): New.
2082 (*vsx_lxvd2x8_le_V8HI): New.
2083 (*vsx_lxvd2x16_le_V16QI): New.
2084 (*vsx_stxvd2x2_le_<mode>): New.
2085 (*vsx_stxvd2x4_le_<mode>): New.
2086 (*vsx_stxvd2x8_le_V8HI): New.
2087 (*vsx_stxvd2x16_le_V16QI): New.
2088
9520e1eb
RL
20892013-10-07 Renlin Li <Renlin.Li@arm.com>
2090
2091 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2092
36c0bd4f
AK
20932013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2094
2095 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2096 loop to work also for 31bit ABI.
2097 Save the stack pointer for frame_size > 0.
2098
ee163e72
AK
20992013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2100
2101 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2102 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2103 constraint letters from expanders.
2104 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2105 retry count to general_operand.
2106 ("tabort"): Give operand 0 a mode.
2107 ("tabort_1"): Add mode and constraint letter for operand 0.
2108 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2109
45d99234
JL
21102013-10-04 Jeff Law <law@redhat.com>
2111
2112 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2113
adcfd489
UB
2114 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2115 out of ...
45d99234
JL
2116 (thread_across_edge): Here. Call it.
2117
7c327f7b
CC
21182013-10-04 Cary Coutant <ccoutant@google.com>
2119
2120 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
adcfd489 2121 discarding a location list expression (or a piece of one).
7c327f7b 2122
6867e128
JH
21232013-10-03 Jan Hubicka <jh@suse.cz>
2124
65efa7e7
UB
2125 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2126 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
3700bce3 2127 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
6867e128 2128
c7f36d55
JH
21292013-10-03 Jan Hubicka <jh@suse.cz>
2130
65efa7e7 2131 * config/i386/i386.c (ix86_option_override_internal): Do not enable
c7f36d55
JH
2132 accumulate-outgoing-args when producing unwind info.
2133
4f0bee4c
WM
21342013-10-03 Wei Mi <wmi@google.com>
2135
2136 * lra-constraints.c (insert_move_for_subreg): New function
2137 extracted from simplify_operand_subreg.
2138 (simplify_operand_subreg): Add reload for paradoxical subreg.
2139
111c3f39
RX
21402013-10-03 Rong Xu <xur@google.com>
2141
65efa7e7
UB
2142 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2143 the candidate of builtin_expect such that we should fix the
2144 size/time estimation.
2145 (estimate_function_body_sizes): Do the acutally size/time fix-up
2146 for builtin_expect.
111c3f39 2147
942df739
RX
21482013-10-03 Rong Xu <xur@google.com>
2149
65efa7e7
UB
2150 * predict.c (tree_predict_by_opcode): Get the probability
2151 for builtin_expect from param builtin_expect_probability.
2152 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2153 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2154 * doc/invoke.texi: Add documentation for builtin-expect-probability.
942df739 2155
2284b034
MG
21562013-10-03 Marc Glisse <marc.glisse@inria.fr>
2157
2158 PR c++/19476
2159 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2160 * calls.c (alloca_call_p): Use get_callee_fndecl.
2161 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2162 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2163 Likewise.
2164 (vrp_visit_stmt): Remove duplicated code.
2165
0609bdf2
MM
21662013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2167
2168 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2169 ceildf2, btruncdf2, instead of vsx_* name.
2170
2171 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2172 iterators to only do V2DF and V4SF here. Move the DF code to
2173 rs6000.md where it is combined with SF mode. Replace <VSv> with
2174 just 'v' since only vector operations are handled with these insns
2175 after moving the DF support to rs6000.md.
2176 (vsx_sub<mode>3): Likewise.
2177 (vsx_mul<mode>3): Likewise.
2178 (vsx_div<mode>3): Likewise.
2179 (vsx_fre<mode>2): Likewise.
2180 (vsx_neg<mode>2): Likewise.
2181 (vsx_abs<mode>2): Likewise.
2182 (vsx_nabs<mode>2): Likewise.
2183 (vsx_smax<mode>3): Likewise.
2184 (vsx_smin<mode>3): Likewise.
2185 (vsx_sqrt<mode>2): Likewise.
2186 (vsx_rsqrte<mode>2): Likewise.
2187 (vsx_fms<mode>4): Likewise.
2188 (vsx_nfma<mode>4): Likewise.
2189 (vsx_copysign<mode>3): Likewise.
2190 (vsx_btrunc<mode>2): Likewise.
2191 (vsx_floor<mode>2): Likewise.
2192 (vsx_ceil<mode>2): Likewise.
2193 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2194 (vsx_sminsf3): Likewise.
2195 (vsx_fmadf4): Likewise.
2196 (vsx_fmsdf4): Likewise.
2197 (vsx_nfmadf4): Likewise.
2198 (vsx_nfmsdf4): Likewise.
2199 (vsx_cmpdf_internal1): Likewise.
2200
2201 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2202 simpler to select whether a target has SPE or traditional floating
2203 point support in iterators.
2204 (TARGET_DF_SPE): Likewise.
2205 (TARGET_SF_FPR): Likewise.
2206 (TARGET_DF_FPR): Likewise.
2207 (TARGET_SF_INSN): Macros to say whether floating point support
2208 exists for a given operation for expanders.
2209 (TARGET_DF_INSN): Likewise.
2210
2211 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2212 combining of SF/DF mode operations, using both traditional and VSX
2213 registers.
2214 (Fvsx): Likewise.
2215 (Ff): Likewise.
2216 (Fv): Likewise.
2217 (Fs): Likewise.
2218 (Ffre): Likewise.
2219 (FFRE): Likewise.
2220 (abs<mode>2): Combine SF/DF modes using traditional floating point
2221 instructions. Add support for using the upper DF registers with
2222 VSX support, and SF registers with power8-vector support. Update
2223 expanders for operations supported by both the SPE and traditional
2224 floating point units.
2225 (abs<mode>2_fpr): Likewise.
2226 (nabs<mode>2): Likewise.
2227 (nabs<mode>2_fpr): Likewise.
2228 (neg<mode>2): Likewise.
2229 (neg<mode>2_fpr): Likewise.
2230 (add<mode>3): Likewise.
2231 (add<mode>3_fpr): Likewise.
2232 (sub<mode>3): Likewise.
2233 (sub<mode>3_fpr): Likewise.
2234 (mul<mode>3): Likewise.
2235 (mul<mode>3_fpr): Likewise.
2236 (div<mode>3): Likewise.
2237 (div<mode>3_fpr): Likewise.
2238 (sqrt<mode>3): Likewise.
2239 (sqrt<mode>3_fpr): Likewise.
2240 (fre<Fs>): Likewise.
2241 (rsqrt<mode>2): Likewise.
2242 (cmp<mode>_fpr): Likewise.
2243 (smax<mode>3): Likewise.
2244 (smin<mode>3): Likewise.
2245 (smax<mode>3_vsx): Likewise.
2246 (smin<mode>3_vsx): Likewise.
2247 (negsf2): Delete SF operations that are merged with DF.
2248 (abssf2): Likewise.
2249 (addsf3): Likewise.
2250 (subsf3): Likewise.
2251 (mulsf3): Likewise.
2252 (divsf3): Likewise.
2253 (fres): Likewise.
2254 (fmasf4_fpr): Likewise.
2255 (fmssf4_fpr): Likewise.
2256 (nfmasf4_fpr): Likewise.
2257 (nfmssf4_fpr): Likewise.
2258 (sqrtsf2): Likewise.
2259 (rsqrtsf_internal1): Likewise.
2260 (smaxsf3): Likewise.
2261 (sminsf3): Likewise.
2262 (cmpsf_internal1): Likewise.
2263 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2264 (negdf2): Delete DF operations that are merged with SF.
2265 (absdf2): Likewise.
2266 (nabsdf2): Likewise.
2267 (adddf3): Likewise.
2268 (subdf3): Likewise.
2269 (muldf3): Likewise.
2270 (divdf3): Likewise.
2271 (fred): Likewise.
2272 (rsqrtdf_internal1): Likewise.
2273 (fmadf4_fpr): Likewise.
2274 (fmsdf4_fpr): Likewise.
2275 (nfmadf4_fpr): Likewise.
2276 (nfmsdf4_fpr): Likewise.
2277 (sqrtdf2): Likewise.
2278 (smaxdf3): Likewise.
2279 (smindf3): Likewise.
2280 (cmpdf_internal1): Likewise.
2281 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2282 (btrunc<mode>2): Delete separate expander, and combine with the
2283 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2284 (btrunc<mode>2_fpr): Likewise.
2285 (ceil<mode>2): Likewise.
2286 (ceil<mode>2_fpr): Likewise.
2287 (floor<mode>2): Likewise.
2288 (floor<mode>2_fpr): Likewise.
2289 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2290 Add support for using the upper registers with VSX and
2291 power8-vector. Move insns to be closer to the define_expands. On
2292 VSX systems, prefer the traditional form of FMA over the VSX
2293 version, since the traditional form allows the target not to
2294 overlap with the inputs.
2295 (fms<mode>4_fpr): Likewise.
2296 (nfma<mode>4_fpr): Likewise.
2297 (nfms<mode>4_fpr): Likewise.
2298
5bea0c6c 22992013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
65efa7e7 2300 Richard Earnshaw <richard.earnshaw@arm.com>
5bea0c6c
KT
2301
2302 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2303 (struct mult_cost_table): Likewise.
2304 (struct mem_cost_table): Likewise.
2305 (struct fp_cost_table): Likewise.
2306 (struct vector_cost_table): Likewise.
2307 (cpu_cost_table): Likewise.
2308 * config/arm/arm.opt (mold-rts-costs): New option.
2309 (mnew-generic-costs): Likewise.
2310 * config/arm/arm.c (generic_extra_costs): New table.
2311 (cortexa15_extra_costs): Likewise.
2312 (arm_slowmul_tune): Use NULL as new costs.
2313 (arm_fastmul_tune): Likewise.
2314 (arm_strongarm_tune): Likewise.
2315 (arm_xscale_tune): Likewise.
2316 (arm_9e_tune): Likewise.
2317 (arm_v6t2_tune): Likewise.
2318 (arm_cortex_a5_tune): Likewise.
2319 (arm_cortex_a9_tune): Likewise.
2320 (arm_v6m_tune): Likewise.
2321 (arm_fa726te_tune): Likewise.
2322 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2323 (arm_cortex_tune): Use generict_extra_costs.
2324 (shifter_op_p): New function.
2325 (arm_unspec_cost): Likewise.
2326 (LIBCALL_COST): Define.
2327 (arm_new_rtx_costs): New function.
2328 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2329 table is available. Use old costs otherwise unless mnew-generic-costs
2330 is specified.
2331 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2332 (cpu_cost_table): Declare.
2333
24c56925
MS
23342013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2335
2336 PR target/58460
2337 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2338 (*subs_mul_imm_<mode>)
2339 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2340 (*sub_<shift>_<mode>)
2341 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2342 Remove k constraint.
2343
2ab8f063
IB
23442013-10-03 Ian Bolton <ian.bolton@arm.com>
2345
2346 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2347 code.
2348 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2349
79221839
TJ
23502013-10-02 Teresa Johnson <tejohnson@google.com>
2351
2352 * predict.c (probably_never_executed): New function.
2353 (probably_never_executed_bb_p): Invoke probably_never_executed.
2354 (probably_never_executed_edge_p): Ditto.
2355 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2356 Treat profile insanities conservatively.
2357
d441afe0
JDA
23582013-10-02 John David Anglin <danglin@gcc.gnu.org>
2359
2360 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2361
80f466c4
VM
23622013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2363
2364 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2365 use it. Use smaller increase for scratch. Don't increase reject
2366 for early clobber scratch.
2367 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2368 setting eliminated regs except setting fp from hfp.
2369 (lra_eliminate): Check lra_insn_recog_data on NULL.
2370
6e228b4b
MM
23712013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2372
2373 PR target/58587
2374 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
65efa7e7 2375 setting -mvsx-timode by default until the underlying problem is fixed.
6e228b4b
MM
2376 (RS6000_CPU, power7 defaults): Likewise.
2377
11b54a5a
UB
23782013-10-02 Uros Bizjak <ubizjak@gmail.com>
2379
2380 * config/x-linux (host-linux.o): Remove header dependencies.
2381 Use $(COMPILE) and $(POSTCOMPILE).
2382 * config/t-linux-android (linux-android.o): Ditto.
2383
bbc02b69
UB
23842013-10-02 Uros Bizjak <ubizjak@gmail.com>
2385
2386 * Makefile.in (expmed.o-warn): Remove.
2387
440917de
AM
23882013-10-02 Andrew MacLeod <amacleod@redhat.com>
2389
2390 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2391 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2392
99206ca9
TJ
23932013-10-02 Teresa Johnson <tejohnson@google.com>
2394
2395 * dojump.c (do_jump_1): Divide probability between
2396 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2397
56e82b14
TT
23982013-10-02 Tom Tromey <tromey@redhat.com>
2399
2400 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2401
3d9c733e
AM
24022013-10-02 Andrew MacLeod <amacleod@redhat.com>
2403
2404 * tree-flow.h: Remove some prototypes.
2405 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2406 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2407 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2408 mark_virtual_phi_result_for_renaming): Relocate here.
2409 * tree-into-ssa.h: Add prototypes.
bbc02b69 2410 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
3d9c733e
AM
2411 single_pred_before_succ_order.
2412 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2413 (nonfreeing_call_p) Move to gimple.c.
2414 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2415 tree-ssa-phiopt.c.
2416 * basic-block.h (single_pred_before_succ_order): Add prototype.
2417 * gimple.c (nonfreeing_call_p): Relocate here.
2418 * gimple.h: Add prototype.
2419 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2420 * tree-ssa-dom.h: New file. Relocate prototypes here.
2421 * tree-ssa.h: Include tree-ssa-dom.h.
2422
78cedfb1
UB
24232013-10-02 Uros Bizjak <ubizjak@gmail.com>
2424
2425 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2426 Use $(COMPILE) and $(POSTCOMPILE).
2427
2428 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2429
744730a4
AM
24302013-10-02 Andrew MacLeod <amacleod@redhat.com>
2431
2432 * tree-flow.h: Remove some prototypes.
2433 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2434 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2435 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2436 propagate_tree_value*): Move from here to...
2437 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2438 propagate_tree_value*): Relocate here.
2439 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2440 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2441 * gimple-fold.c: Remove gimple-fold.h from include list.
2442 * tree-vrp.c: Remove gimple-fold.h from include list.
2443 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2444 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2445 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2446 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2447 * sese.c: Add tree-ssa-propagate.h to include list.
2448
826a536d
RB
24492013-10-02 Richard Biener <rguenther@suse.de>
2450
2451 * tree-loop-distribution.c: Include tree-vectorizer.h for
2452 find_loop_location.
2453 (enum partition_kind): Remove PKIND_REDUCTION.
78cedfb1 2454 (struct partition_s): Remove has_writes member, add reduction_p member.
826a536d
RB
2455 (partition_alloc): Adjust.
2456 (partition_builtin_p): Likewise.
2457 (partition_has_writes): Remove.
2458 (partition_reduction_p): New function.
2459 (partition_merge_into): Likewise.
2460 (generate_code_for_partition): Commonize builtin partition
2461 handling tail.
2462 (rdg_cannot_recompute_vertex_p): Remove.
2463 (already_processed_vertex_p): Likewise.
2464 (rdg_flag_vertex): Do not set has_writes.
2465 (classify_partition): Adjust.
2466 (rdg_build_partitions): Do not set has_writes, treat all
2467 partitions as useful.
78cedfb1 2468 (distribute_loop): Record number of library calls generated. Adjust.
826a536d
RB
2469 (tree_loop_distribution): Report number of loops and library
2470 calls generated as opt-info.
2471
4b403ece
AM
24722013-10-02 Andrew MacLeod <amacleod@redhat.com>
2473
2474 * tree-flow.h: Include new .h files. Move prototypes.
2475 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2476 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2477 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2478 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2479 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2480 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2481 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2482 * tree.h (get_ref_base_and_extent): Move prototype out.
78cedfb1 2483 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
4b403ece
AM
2484 tree-dfa.h.
2485 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2486 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2487 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2488 * tree-scalar-evolution.c: Include tree.h.
2489 * sese.c: Include tree.h.
2490 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2491 * dwarf2out.c: Include tree-dfa.h.
2492 * tree-chrec.c: Include tree.h.
2493 * tree-data-ref.c: Include tree.h.
2494
1d2151c6
YZ
24952013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2496
2497 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2498 Fix whitespace.
2499
441ad147
RO
25002013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2501
2502 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2503 Use $(COMPILE) and $(POSTCOMPILE).
2504 (sol2-cxx.o): Likewise.
2505 (sol2-stubs.o): Likewise.
2506 (sol2.o): Likewise.
2507 * config/x-solaris (host-solaris.o): Likewise.
2508
2509 * config/sparc/t-sparc (sparc.o): Remove.
2510 (sparc-c.o): Remove header dependencies.
2511 Use $(COMPILE) and $(POSTCOMPILE).
2512 * config/sparc/x-sparc: Likewise.
2513
2deaf8b0
JR
25142013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2515
2516 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2517 * config/arc/arc700.md: Likewise.
2518 * config/arc/arc-modes.def: Likewise.
2519 * config/arc/arc-simd.h: Likewise.
2520 * config/arc/t-arc-uClibc: Likewise.
2521 * config/arc/t-arc-newlib: Likewise.
2522
f6fe771a
RL
25232013-10-02 Renlin Li <renlin.li@arm.com>
2524
78cedfb1
UB
2525 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2526 plus_constant.
f6fe771a
RL
2527 (aarch64_expand_epilogue): Likewise.
2528
0916f876
YZ
25292013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2530 Yufeng Zhang <yufeng.zhang@arm.com>
2531
2532 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2533 declaration.
2534 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2535 'base_in' represent a conversion and legal_cast_p_1 holds; set
2536 'base_in' with the returned value from get_unwidened.
2537
ebfcd719
KT
25382013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2539
2540 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2541 plus_constant is not used.
2542
157ca3e9
WM
25432013-10-01 Wei Mi <wmi@google.com>
2544
78cedfb1 2545 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
157ca3e9
WM
2546 * config/i386/i386.md: Add define_peephole2 to
2547 break partial reg stall for cvtss2sd/cvtsd2ss.
2548
cd4dd8f0
JR
25492013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2550
2551 * config/arc/arc.c (pass_arc_ifcvt::clone):
78cedfb1 2552 Update for ctxt_ -> m_ctxt change.
cd4dd8f0 2553
1465e5cf
JL
25542013-10-01 Jeff Law <law@redhat.com>
2555
2556 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2557 outgoing_edge and intermediate_edge fields. Instead store the path.
2558 (redirection_data::hash): Hash on the last edge's destination index.
2559 (redirection_data::equal): Check the entire thread path.
2560 (lookup_redirectio_data): Corresponding changes.
2561 (create_edge_and_update_destination_phis): Likewise.
2562 (thread_single_edge): Likewise.
2563
25c606cb 25642013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
0ccbc132 2565 Diego Novillo <dnovillo@google.com>
25c606cb
JR
2566
2567 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2568 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2569 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2570 versions of these patterns.
2571
0ccbc132
JR
2572 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2573 (__builtin_arc_aligned): Likewise.
2574
f55d4a20
JR
2575 * config/arc/arc.md: Expand adc_0 comment stating the intended
2576 purpose and why it isn't ready.
2577 Replace commented out call_value_via_label_mixed with a
2578 plain comment about bl_s.
2579
5719867d 2580 * config/arc/arc.c (stdio.h): Don't include directly.
5719867d
JR
2581 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2582 Remove if (1) condition.
2583 (arc_encode_section_info): Fix comment.
2584
6462fab0
JR
25852013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2586
2587 * config/arc/arc.c (arc_conditional_register_usage):
2588 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2589 Also set reg_alloc_order for DMA config regs.
2590
5d5f6720
JR
25912013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2592 Jeremy Bennett <jeremy.bennett@embecosm.com>
2593
2594 * doc/install.texi (--with-cpu): Mention ARC.
2595 (arc-*-elf32): New paragraph.
2596 (arc-linux-uclibc): Likewise.
2597 * doc/md.texi (Machine Constraints): Add ARC part.
2598 * doc/invoke.texi: (menu): Add ARC Options.
2599 (Machine Dependent Options) <ARC Options>: Add synopsis.
2600 (node ARC Options): Add.
2601 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2602 (ARC Built-in Functions): New section defining
2603 generic ARC built-in functions.
2604 (ARC SIMD Built-in Functions): New section defining SIMD specific
2605 built-in functions.
2606 (Declaring Attributes of Functions): Extended
2607 description of short_call and long_call attributes for ARC and
2608 added index entries.
2609
526b7aee
SV
26102013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2611 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2612 Joern Rennecke <joern.rennecke@embecosm.com>
2613 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2614 Brendan Kehoe <brendan@zen.org>
2615 Michael Eager <eager@eagercon.com>
2616 Simon Cook <simon.cook@embecosm.com>
2617 Jeremy Bennett <jeremy.bennett@embecosm.com>
2618
2619 * config/arc, common/config/arc: New directories.
2620
53426f6c
JR
26212013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2622 Brendan Kehoe <brendan@zen.org>
2623 Simon Cook <simon.cook@embecosm.com>
2624
2625 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2626
1df9f5a9
AM
26272013-10-01 Andrew MacLeod <amacleod@redhat.com>
2628
2629 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2630 * tree-ssa-coalesce.h: New. Move prototype to here.
2631 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2632 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2633 (gimple_can_coalesce_p): Move to...
2634 * gimple.c (gimple_can_coalesce_p): Here.
2635
80560f95
AM
26362013-10-01 Andrew MacLeod <amacleod@redhat.com>
2637
2638 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2639 (dump_decl_set): Move to gimple.c.
2640 * gimple.h: Don't include tree-ssa-operands.h.
2641 (dump_decl_set): Add prototype.
2642 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2643 Move to gimple-ssa.h.
2644 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2645 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2646 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2647 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2648 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2649 than PHI_ARG_DEF.
2650 (dump_decl_set): Relocate here.
2651 * gimple-ssa.h: New file.
2652 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2653 Relocate from gimple.h.
2654 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2655 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2656 swap_tree_operands and remove non-ssa path.
2657 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2658 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2659 swap_ssa_operands.
2660 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2661 vect_is_simple_reduction_1): Use swap_ssa_operands.
2662 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2663 (enum need_phi_state): Move to tree-into-ssa.c.
2664 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2665 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2666 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2667 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2668 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2669 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2670 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2671 num_imm_uses): Move to ssa-iterators.h.
2672 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2673 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2674 tree-phinodes.h.
2675 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2676 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2677 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2678 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2679 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2680 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2681 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2682 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2683 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2684 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2685 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2686 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2687 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2688 (set_phi_nodes): Move to tree-phinodes.h.
2689 * tree-ssa-operands.h (enum ssa_op_iter_type,
2690 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2691 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2692 (dump_decl_set): Remove prototype.
2693 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2694 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2695 (set_phi_nodes): Relocate from tree-flow-inline.h.
2696 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2697 tree-flow-inline.h
2698 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2699 include list. Temporarily add gimple.h to include list.
2700 * ssa-iterators.h: New file.
2701 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2702 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2703 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2704 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2705 Relocate from tree-ssa-operands.h.
2706 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2707 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2708 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2709 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2710 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2711 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2712 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2713 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2714 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2715 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2716 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2717 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2718 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2719 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2720 Relocate from tree-flow-inline.h.
2721 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2722
8da00d65
VP
27232013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2724
2725 * aarch64-simd.md
2726 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2727 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2728 Insert '\t' to output template.
2729 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2730 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2731 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2732 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2733
e2ebe1c2
UB
27342013-10-01 Uros Bizjak <ubizjak@gmail.com>
2735
2736 * doc/install.texi (Host/target specific installation notes for GCC):
2737 Put @anchor before @heading.
2738 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2739 Use @email for email addresses.
2740
aee2d611
JL
27412013-10-01 Jeff Law <law@redhat.com>
2742
2743 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
2744 a vec. Only delete the path if we create one without successfully
2745 registering a jump thread.
2746 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
2747 as a pointer.
2748 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
2749 (paths): New vector of jump threading paths.
2750 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
2751 (THREAD_PATH): New accessor macro for the entire thread path.
2752 (lookup_redirection_data): Get intermediate and final outgoing edge
2753 from the thread path.
2754 (create_edge_and_update_destination_phis): Copy the threading path.
2755 (ssa_fix_duplicate_block_edges): Get edges and block types from the
2756 jump threading path.
2757 (ssa_redirect_edges): Get edges and block types from the jump threading
2758 path. Free the path vector.
2759 (thread_block): Get edges from the jump threading path. Look at the
2760 entire path to see if we thread to a loop exit. If we cancel a jump
2761 thread request, then free the path vector.
2762 (thread_single_edge): Get edges and block types from the jump threading
2763 path. Free the path vector.
2764 (thread_through_loop_header): Get edges and block types from the jump
2765 threading path. Free the path vector.
2766 (mark_threaded_blocks): Iterate over the vector of paths and store
2767 the path on the appropriate edge. Get edges and block types from the
2768 jump threading path.
2769 (mark_threaded_blocks): Get edges and block types from the jump
2770 threading path. Free the path vector.
2771 (thread_through_all_blocks): Use the vector of paths rather than
2772 a vector of 3-edge sets.
2773 (register_jump_thread): Accept pointer to a path vector rather
2774 than the path vector itself. Store the path vector for later use.
2775 Simplify.
2776
966f97ac 27772013-10-01 Jakub Jelinek <jakub@redhat.com>
8da00d65 2778 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
966f97ac
JJ
2779
2780 PR target/58574
2781 * config/s390/s390.c (s390_split_branches): Modify check for table
2782 jump insns.
2783 (s390_chunkify_start): Rearrange table jump insn check in order to
2784 deal with compare and branch insns correctly.
2785
3a323a38
KV
27862013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
2787
2788 PR target/58578
2789 Revert
2790 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
2791 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
2792 define_insn_and_split.
2793 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
2794 (shiftsi3_compare): New pattern.
2795 (rrx): New pattern.
2796 * config/arm/unspecs.md (UNSPEC_RRX): New.
2797
90444831
AM
27982013-10-01 Alan Modra <amodra@gmail.com>
2799
2800 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
2801 casing inout operands.
2802
818625cf
RB
28032013-10-01 Richard Biener <rguenther@suse.de>
2804
2805 PR tree-optimization/58553
2806 * tree-loop-distribution.c (struct partition_s): Add niter member.
2807 (classify_partition): Populate niter member for the partition
2808 and properly identify whether the relevant store happens before
2809 or after the loop exit.
2810 (generate_memset_builtin): Use niter member from the partition.
2811 (generate_memcpy_builtin): Likewise.
2812
30f641cd
RS
28132013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
2814
2815 * vec.h (vec_prefix, vec): Prefix member names with "m_".
2816 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
2817
65d3284b
RS
28182013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
2819
2820 * basic-block.h (edge_list): Prefix member names with "m_".
2821 * context.h (context): Likewise.
2822 * domwalk.h (dom_walker): Likewise.
2823 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
2824 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2825 * hash-table.h (hash_table): Likewise.
2826 * machmode.h (bit_field_mode_iterator): Likewise.
2827 * pass_manager.h (pass_list): Likewise.
2828 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
2829 * tree-pass.h (pass_data): Likewise.
2830 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
2831 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
2832 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2833 * asan.c (pass_data_asan): Update accordingly.
2834 * cfganal.c (control_dependences::find_control_dependence): Likewise.
2835 (control_dependences::control_dependences): Likewise.
2836 (control_dependences::~control_dependences): Likewise.
2837 (control_dependences::~control_dependences): Likewise.
2838 (control_dependences::get_edges_dependent_on): Likewise.
2839 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
2840 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
2841 * context.c (gcc::context::context): Likewise.
2842 * cprop.c (pass_rtl_cprop::clone): Likewise.
2843 * domwalk.c (dom_walker::walk): Likewise.
2844 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
2845 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
2846 * mode-switching.c (pass_mode_switching::clone): Likewise.
2847 * passes.c (opt_pass::opt_pass): Likewise.
2848 (pass_manager::pass_manager): Likewise.
2849 * predict.c (pass_strip_predict_hints::clone): Likewise.
2850 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
2851 (pass_split_all_insns::clone): Likewise.
2852 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
2853 Likewise.
2854 (bit_field_mode_iterator::next_mode): Likewise.
2855 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
2856 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
2857 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
2858 * tree-complex.c (pass_lower_complex::clone): Likewise.
2859 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
2860 * tree-object-size.c (pass_object_sizes::clone): Likewise.
2861 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
2862 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
2863 (pass_fold_builtins::clone): Likewise.
2864 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
2865 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
2866 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
2867 (pass_cd_dce::clone): Likewise.
2868 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
2869 (pass_phi_only_cprop::clone): Likewise.
2870 * tree-ssa-dse.c (pass_dse::clone): Likewise.
2871 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
2872 * tree-ssa-loop.c (pass_lim::clone): Likewise.
2873 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
2874 * tree-ssa-pre.c (pass_fre::clone): Likewise.
2875 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
2876 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
2877 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
2878 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
2879 * tree-vrp.c (pass_vrp::clone): Likewise.
2880 * tsan.c (pass_tsan::clone): Likewise.
2881
f66d0891
JJ
28822013-09-30 Jakub Jelinek <jakub@redhat.com>
2883
d2a365a8
JJ
2884 PR middle-end/58564
2885 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
2886 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
2887
f66d0891
JJ
2888 PR middle-end/58564
2889 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
2890 optimization, punt if sign_bit_p looked through any zero extension.
2891
05357ac3
TJ
28922013-09-30 Teresa Johnson <tejohnson@google.com>
2893
2894 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
2895 Update redirected out edge count in joiner case.
2896 (ssa_redirect_edges): Common the joiner and non-joiner cases
2897 so that joiner case gets profile updates.
2898
ca406576
RB
28992013-09-30 Richard Biener <rguenther@suse.de>
2900
2901 PR tree-optimization/58554
e2ebe1c2
UB
2902 * tree-loop-distribution.c (classify_partition): Require
2903 unconditionally executed stores for memcpy and memset recognition.
ca406576
RB
2904 (tree_loop_distribution): Calculate dominance info.
2905
92d649c4
VK
29062013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2907
2908 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
2909 (NO_PROFILE_COUNTERS): Likewise.
2910 (PROFILE_HOOK): Likewise.
2911 (FUNCTION_PROFILER): Likewise.
2912 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
2913
bd9534e2
IS
29142013-09-30 Iain Sandoe <iain@codesourcery.com>
2915
2916 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
2917 calls and defines in TARGET_MACHO conditional.
2918 (load_macho_picbase_di): Likewise.
2919 (reload_macho_picbase): Likewise.
2920 (reload_macho_picbase_si): Likewise.
2921 (reload_macho_picbase_di): Likewise.
2922 (nonlocal_goto_receiver): Likewise.
2923
4494fbc9
NC
29242013-09-30 Nick Clifton <nickc@redhat.com>
2925
2926 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
2927 that use the MSP430X ISA.
2928 (msp430_option_override): Scan -mmcu command line option for any
2929 MCU name that supports the MSP430X ISA.
2930 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
2931 -mmcu options which enable the MSP430X ISA.
2932
791d9044
RB
29332013-09-30 Richard Biener <rguenther@suse.de>
2934
2935 PR middle-end/58532
2936 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
2937 before looking for setjmp-like calls.
2938
e1c5c877
IS
29392013-09-29 Iain Sandoe <iain@codesourcery.com>
2940
2941 PR target/10901
2942 * config/darwin-protos.h (machopic_get_function_picbase): New.
2943 * config/darwin.c (machopic_get_function_picbase): New.
2944 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
2945 label for a new func. (load_macho_picbase_di): Likewise.
2946 (reload_macho_picbase): New expand.
2947 (reload_macho_picbase_si): New insn.
2948 (reload_macho_picbase_di): New insn.
2949 (nonlocal_goto_receiver): New define and split.
2950 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
2951 (unspecv enum): Add UNSPECV_NLGR.
2952
749278c5
IS
29532013-09-29 Iain Sandoe <iain@codesourcery.com>
2954
2955 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
2956 that altivec registers are correctly sized on Darwin.
2957
2c43a51c
IS
29582013-09-29 Iain Sandoe <iain@codesourcery.com>
2959
2960 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
2961 darwin-driver.o): Use COMPILE and POSTCOMPILE.
2962 * config/x-darwin (host-darwin.o): Likewise.
2963 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
2964 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
2965 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
2966
7a5911d3
UB
29672013-09-29 Uros Bizjak <ubizjak@gmail.com>
2968
2969 * doc/invoke.texi: Fix usage of @tie{} command.
2970
3f67a73d
EB
29712013-09-29 Eric Botcazou <ebotcazou@adacore.com>
2972
2973 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
2974
9ac2f538
JH
29752013-09-28 Jan Hubicka <jh@suse.cz>
2976
2977 * config/i386/x86-tune.def: Add documentation for each of the options;
2978 add whitespace.
2979
b0ff06da
JH
29802013-09-28 Jan Hubicka <jh@suse.cz>
2981
7a5911d3
UB
2982 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
2983 generic.
b0ff06da
JH
2984 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
2985 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
2986 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
2987 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
2988 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
2989
c3284718
RS
29902013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
2991
2992 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
2993 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
2994 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
2995 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
2996 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
2997 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
2998 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
2999 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
3000 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
3001 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
3002 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
3003 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
3004 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
3005 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
3006 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
3007 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
3008 gimple.h, godump.c, graphite-clast-to-gimple.c,
3009 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
3010 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
3011 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
3012 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
3013 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
3014 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
3015 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
3016 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
3017 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
3018 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
3019 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
3020 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
3021 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
3022 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
3023 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
3024 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
3025 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
3026 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
3027 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
3028 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
3029 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
3030 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
3031 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
3032 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
3033 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3034 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3035 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3036 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3037 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3038 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3039 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3040 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3041 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3042 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3043 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3044 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3045 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3046 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3047 whitespace before "(".
3048
c6285bd7
SL
30492013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3050
3051 * expr.h (extract_bit_field): Remove packedp parameter.
3052 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3053 from forward declaration.
3054 (store_split_bit_field): Remove packedp arg from calls to
3055 extract_fixed_bit_field.
3056 (extract_bit_field_1): Remove packedp parameter and packedp
3057 argument from recursive calls and calls to extract_fixed_bit_field.
3058 (extract_bit_field): Remove packedp parameter and corresponding
3059 arg to extract_bit_field_1.
3060 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3061 to issue warnings.
3062 (extract_split_bit_field): Remove packedp arg from call to
3063 extract_fixed_bit_field.
3064 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3065 (copy_blkmode_from_reg): Likewise.
3066 (copy_blkmode_to_reg): Likewise.
3067 (read_complex_part): Likewise.
3068 (store_field): Likewise.
3069 (expand_expr_real_1): Likewise.
3070 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3071 to extract_bit_field.
3072 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3073 call to extract_bit_field.
3074 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3075 call to extract_bit_field.
3076 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3077 and special packedp behavior from -fstrict-volatile-bitfields
3078 documentation.
3079
522d4efc
JBG
30802013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3081
3082 * lra-eliminations.c (init_elim_table): Guard value_p.
3083
19cdb489
MM
30842013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3085
3086 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3087 DFmode, DImode, and SFmode in the upper VSX registers based on the
3088 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3089 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3090 of the VSX register handling.
3091
3092 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3093 (f32_sv): Likewise.
3094 (zero_extendsidi2_lfiwzx): Add support for loading into the
3095 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3096 only do VSX memory options on Altivec registers.
3097 (extendsidi2_lfiwax): Likewise.
3098 (extendsfdf2_fpr): Likewise.
3099 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3100 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3101 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3102 (movdi_internal64): Likewise.
3103
d6d11272
XDL
31042013-09-27 Xinliang David Li <davidxl@google.com>
3105
3106 * opts.c (finish_options): Adjust parameters
3107 according to vect cost model.
3108 (common_handle_option): Set dynamic vect cost
3109 model for FDO.
3110 targhooks.c (default_add_stmt_cost): Compute stmt cost
3111 unconditionally.
3112 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3113 Use helper function.
3114 * tree-vectorizer.h (unlimited_cost_model): New function.
3115 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3116 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3117 function.
3118 (vect_enhance_data_refs_alignment): Ditto.
3119 * flag-types.h: New enum.
3120 * common/config/i386/i386-common.c (ix86_option_init_struct):
3121 No need to initialize vect_cost_model flag.
3122 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3123 unconditionally.
3124
ac1857a3
DN
31252013-09-27 Diego Novillo <dnovillo@google.com>
3126
3127 * gimple.h (enum ssa_mode): Remove.
3128
3c468b4c
PM
31292013-09-27 Paulo Matos <pmatos@broadcom.com>
3130
3131 PR middle-end/58463
3132 * gcc.dg/pr58463.c: New test.
3133
0d0e2af6
PM
31342013-09-27 Paulo Matos <pmatos@broadcom.com>
3135
3136 * cfgloop.h (number_of_loops): Fix typo in check for null.
3137
09dc585e
JJ
31382013-09-27 Jakub Jelinek <jakub@redhat.com>
3139
3140 PR middle-end/58551
3141 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3142 are children of outermost saved_cfun's loop, and set it up to
3143 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3144 if loop != loop0 and SESE region contains bbs that belong to loop0.
3145
ec5a3504
RS
31462013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3147
3148 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3149 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3150 (set_address_segment, set_address_base, set_address_index)
3151 (set_address_disp): Accept the argument unconditionally.
3152 (baseness): Remove must_be_base_p and must_be_index_p checks.
3153 (decompose_normal_address): Classify as much as possible in the
3154 main loop.
3155
f91aec98
RS
31562013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3157
3158 * cse.c (count_reg_usage): Handle INT_LIST.
3159 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3160 * reginfo.c (reg_scan_mark_refs): Likewise.
3161 * reload1.c (eliminate_regs_1): Likewise.
3162
a9195970
IS
31632013-09-27 Iain Sandoe <iain@codesourcery.com>
3164
3165 PR middle-end/58547
3166 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3167 signed.
3168
e6f1c509
RB
31692013-09-27 Richard Biener <rguenther@suse.de>
3170
3171 PR tree-optimization/58459
3172 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3173 restriction not propagating into loops.
3174
84f48495
FW
31752013-09-26 Florian Weimer <fw@deneb.enyo.de>
3176
3177 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3178 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3179 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3180
74fc8b8a
RB
31812013-09-26 Richard Biener <rguenther@suse.de>
3182
7a5911d3 3183 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
74fc8b8a 3184
b4ada065
RB
31852013-09-26 Richard Biener <rguenther@suse.de>
3186
3187 * alias.h (component_uses_parent_alias_set): Rename to ...
3188 (component_uses_parent_alias_set_from): ... this.
3189 * alias.c (component_uses_parent_alias_set): Rename to ...
3190 (component_uses_parent_alias_set_from): ... this and return
3191 the desired parent.
3192 (reference_alias_ptr_type_1): Use the result from
3193 component_uses_parent_alias_set_from instead of stripping
3194 components one at a time.
3195 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3196
78bca40d
AM
31972013-09-26 Andrew MacLeod <amacleod@redhat.com>
3198
7a5911d3
UB
3199 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3200 Move prototypes to...
78bca40d
AM
3201 * tree-ssa-ter.h: New File. Move prototypes here.
3202 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3203 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3204 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
7a5911d3 3205 from is_replaceable_p.
78bca40d
AM
3206 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3207 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3208 (process_replaceable): Use ter_is_replaceable_p.
3209 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3210 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3211 newly refactored ssa_is_replaceable_p.
3212 * cfgexpand.c: Include tree-outof-ssa.h.
3213 * ssaexpand.h: Delete.
3214
ff2a63a7
AM
32152013-09-26 Andrew MacLeod <amacleod@redhat.com>
3216
3217 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3218 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3219 tree-ssa.c
3220 (create_gimple_tmp): Delete.
3221 (get_expr_type, build_assign, build_type_cast): Move to...
3222 * gimple-builder.c: New File.
3223 (get_expr_type): Relocate from gimple.c.
3224 (build_assign, build_type_cast): Change to only create ssanames.
3225 * gimple.h: Move prototypes to...
3226 * gimple-builder.h: New File. Here.
3227 * tree-ssa.h: And here.
3228 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3229 count_uses_and_derefs): Relocate from gimple.c.
3230 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3231 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3232 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3233 gimple_replace_ssa_lhs.
3234 * asan.c: Include gimple-builder.h.
3235 * Makefile.in: Add gimple-builder.o.
3236
a2544177
RB
32372013-09-26 Richard Biener <rguenther@suse.de>
3238
3239 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3240 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3241 (loe_visit_block): Use gcc_checking_assert.
3242 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3243 gimple_assign_ssa_name_copy_p.
3244 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3245 change.
3246
6d1ca9a0
DE
32472013-09-26 David Edelsohn <dje.gcc@gmail.com>
3248
3249 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3250 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3251
46e950db
RB
32522013-09-26 Richard Biener <rguenther@suse.de>
3253
3254 PR tree-optimization/58539
3255 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3256 the fact that debug statements are not taking part in loop-closed
3257 SSA construction.
3258
40ada30a
NC
32592013-09-26 Nick Clifton <nickc@redhat.com>
3260
3261 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3262 time warning message.
3263 (msp430_print_operand_raw): Delete unused letter parameter.
3264 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3265 (msp430_print_operand_address): New function.
3266 (msp430_print_operand): Move address printing code from here to
3267 new function.
3268 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3269 assembler.
3270 (zero_extendpsisi2): Likewise.
3271 (extendpsisi2): New pattern.
3272 (andneghi3): New pattern.
3273
38e8f663
YR
32742013-09-26 Yvan Roux <yvan.roux@linaro.org>
3275
3276 * config/aarch64/aarch64.opt (mlra): New option.
3277 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3278 (TARGET_LRA_P): Define.
3279
4167e8d4
EB
32802013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3281
3282 * expr.c (expand_assignment): Remove obsolete comment.
3283
5254eac4
JL
32842013-09-25 Jeff Law <law@redhat.com>
3285
3286 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3287 tree-ssa-threadupdate.h.
3288 (register_jump_thread): Similarly.
3289 * tree-ssa-threadupdate.h: New header file.
3290 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3291 * tree-vrp.c: Likewise.
3292 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3293 (thread_around_empty_blocks): Change type of path vector argument to
3294 an edge,type pair from just an edge. Initialize both elements when
3295 appending to a jump threading path. Tweak references to elements
3296 appropriately.
3297 (thread_across_edge): Similarly. Release memory for the elements
3298 as needed.
3299 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3300 (dump_jump_thread_path): New function broken out from
3301 register_jump_thread.
3302 (register_jump_thread): Use dump_jump_thread_path. Change type of
3303 path vector entries. Search the path for NULL edges and dump
3304 the path if one is found. Tweak the conversion of path to 3-edge
3305 form to use the block copy type information embedded in the path.
3306
20f114a3
YR
33072013-09-25 Yvan Roux <yvan.roux@linaro.org>
3308
3309 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3310
3936bafc
YR
33112013-09-25 Yvan Roux <yvan.roux@linaro.org>
3312 Vladimir Makarov <vmakarov@redhat.com>
3313
3314 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3315 from the least significant bit.
3316 (strip_address_mutations): Add bitfield operations handling.
3317 (must_be_index_p): Add shifting and rotate operations handling.
3318 (set_address_base): Use must_be_base_p predicate.
4167e8d4 3319 (set_address_index): Use must_be_index_p predicate.
3936bafc 3320
b86f6e9e
AI
33212013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3322 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3323 Sergey Lega <sergey.s.lega@intel.com>
3324 Anna Tikhonova <anna.tikhonova@intel.com>
3325 Ilya Tocar <ilya.tocar@intel.com>
3326 Andrey Turetskiy <andrey.turetskiy@intel.com>
3327 Ilya Verbin <ilya.verbin@intel.com>
3328 Kirill Yukhin <kirill.yukhin@intel.com>
3329 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3330
3331 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3332 Use new names.
3333 (ix86_expand_vector_move_misalign): Support new unaligned load and
3334 stores and use new names.
3335 (CODE_FOR_sse2_storedqu): Rename to ...
3336 (CODE_FOR_sse2_storedquv16qi): ... this.
3337 (CODE_FOR_sse2_loaddqu): Rename to ...
3338 (CODE_FOR_sse2_loaddquv16qi): ... this.
3339 (CODE_FOR_avx_loaddqu256): Rename to ...
3340 (CODE_FOR_avx_loaddquv32qi): ... this.
3341 (CODE_FOR_avx_storedqu256): Rename to ...
3342 (CODE_FOR_avx_storedquv32qi): ... this.
3343 * config/i386/i386.md (fpint_logic): New.
3344 * config/i386/sse.md (VMOVE): Extend for AVX512.
3345 (VF): Ditto.
3346 (VF_128_256): New.
3347 (VF_512): Ditto.
3348 (VI_UNALIGNED_LOADSTORE): Ditto.
3349 (sse2_avx_avx512f): Ditto.
3350 (sse2_avx2): Extend for AVX512.
3351 (sse4_1_avx2): Ditto.
3352 (avx2_avx512f): New.
3353 (sse): Extend for AVX512.
3354 (sse2): Ditto.
3355 (sse4_1): Ditto.
3356 (avxsizesuffix): Ditto.
3357 (sseintvecmode): Ditto.
3358 (ssePSmode): Ditto.
3359 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3360 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3361 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3362 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3363 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3364 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3365 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3366 (STORENT_MODE): Extend for AVX512.
3367 (*absneg<mode>2): Replace constraint "x" with "v".
3368 (*mul<mode>3): Ditto.
3369 (*ieee_smin<mode>3): Ditto.
3370 (*ieee_smax<mode>3): Ditto.
3371 (avx_cmp<mode>3): Replace VF with VF_128_256.
3372 (*<sse>_maskcmp<mode>3_comm): Ditto.
3373 (<sse>_maskcmp<mode>3): Ditto.
3374 (<sse>_andnot<mode>3): Extend for AVX512.
3375 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3376 (<code><mode>3, fpint_logic): New.
3377 (*<code><mode>3): Extend for AVX512.
3378 (avx512flogicsuff): New.
3379 (avx512f_<logic><mode>): Ditto.
3380 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3381 VF_128_256.
3382 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3383 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3384 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3385 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3386 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3387 (xop_vpermil2<mode>3): Ditto.
3388 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3389 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3390 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3391 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3392
4d44d03c
TT
33932013-09-25 Tom Tromey <tromey@redhat.com>
3394
3395 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3396 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3397 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3398 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3399 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3400 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3401 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3402 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3403 (GRAPHITE_HTAB_H): Remove.
3404
86aaf68d
TT
34052013-09-25 Tom Tromey <tromey@redhat.com>
3406
3407 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3408
c987deb8
TT
34092013-09-25 Tom Tromey <tromey@redhat.com>
3410
3411 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3412
9439747e
TT
34132013-09-25 Tom Tromey <tromey@redhat.com>
3414
3415 * config/i386/t-i386 (i386.o): Remove.
3416 (i386-c.o): Use COMPILE and POSTCOMPILE.
3417
aefc31a1
TT
34182013-09-25 Tom Tromey <tromey@redhat.com>
3419
3420 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3421
b6541edc
TT
34222013-09-25 Tom Tromey <tromey@redhat.com>
3423
3424 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3425 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3426 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3427 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3428 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3429 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3430 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3431 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3432 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3433 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3434 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3435 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3436 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3437 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3438 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3439 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3440 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3441 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3442 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3443 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3444 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3445 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3446 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3447 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3448 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3449 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3450 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3451 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3452 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3453 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3454 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3455 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3456 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3457 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3458 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3459 (graphite-clast-to-gimple.o, graphite-dependences.o)
3460 (graphite-interchange.o, graphite-poly.o)
3461 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3462 (graphite-optimize-isl.o, tree-vect-loop.o)
3463 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3464 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3465 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3466 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3467 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3468 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3469 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3470 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3471 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3472 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3473 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3474 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3475 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3476 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3477 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3478 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3479 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3480 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3481 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3482 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3483 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3484 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3485 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3486 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3487 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3488 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3489 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3490 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3491 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3492 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3493 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3494 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3495 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3496 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3497 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3498 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3499 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3500 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3501 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3502 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3503 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3504 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3505 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3506 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3507 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3508 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3509 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3510 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3511 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3512 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3513 (gcov-dump.o): Remove.
3514 (default-c.o): Use COMPILE and POSTCOMPILE.
3515 (CFLAGS-gcc.o): New variable.
3516 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3517
8c796d19
TT
35182013-09-25 Tom Tromey <tromey@redhat.com>
3519
3520 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3521 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3522 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3523 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3524 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3525 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3526 (c-family/array-notation-common.o, c-family/stub-objc.o)
3527 (c-family/c-ubsan.o): Remove.
3528
0a6c2227
TT
35292013-09-25 Tom Tromey <tromey@redhat.com>
3530
3531 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3532
f3bc55f0
TT
35332013-09-25 Tom Tromey <tromey@redhat.com>
3534
3535 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3536 to add -DENABLE_SHARED_LIBGCC.
3537 (gcc.o): Don't use subshell.
3538
c11c795e
TT
35392013-09-25 Tom Tromey <tromey@redhat.com>
3540
3541 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3542 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3543 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3544 * configure, config.in: Rebuild.
3545
043378c3
TT
35462013-09-25 Tom Tromey <tromey@redhat.com>
3547
3548 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3549 (COMPILE, POSTCOMPILE): New variables.
3550 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3551 (DEPFILES): New variable.
3552 Include ".Po" files.
3553 * configure.ac: Add checks for dependency checking.
3554 * configure, aclocal.m4: Regenerate.
3555
eab34643
TT
35562013-09-25 Tom Tromey <tromey@redhat.com>
3557
3558 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
7a5911d3 3559 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
eab34643 3560
d2db5e91
TT
35612013-09-25 Tom Tromey <tromey@redhat.com>
3562
3563 * Makefile.in (generated_files): Add options.h,
3564 target-hooks-def.h, insn-opinit.h,
3565 common/common-target-hooks-def.h, pass-instances.def,
3566 c-family/c-target-hooks-def.h.
3567
af33629e
JL
35682013-09-25 Jeff Law <law@redhat.com>
3569
3570 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3571 than foo[foo.length () - 1] to access last member in a vec.
3572 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3573
ed0d3051
RB
35742013-09-25 Richard Biener <rguenther@suse.de>
3575
3576 PR middle-end/58521
3577 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3578
123485ca
JH
35792013-09-25 Jan Hubicka <jh@suse.cz>
3580
3581 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3582 test.
3583
d70b0f1f
MP
35842013-09-25 Marek Polacek <polacek@redhat.com>
3585
3586 PR sanitizer/58420
3587 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3588 when determining the type name.
3589
e3f02798
OE
35902013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3591
3592 * config/sh/sh.md: Fix formatting.
3593
7a5911d3 35942013-09-24 Xinliang David Li <davidxl@google.com>
4f17aa0b 3595
7a5911d3
UB
3596 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3597 max peel iterations parameter.
4f17aa0b
XDL
3598 * param.def: New parameter.
3599 * doc/invoke.texi: Document New parameter.
3600
0498471b
CL
36012013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3602
3603 * gimple-pretty-print.c: Various whitespace tweaks.
3604 * tree-core.h: Likewise.
3605 * tree-pretty-print.c: Likewise.
3606 * tree-ssa-alias.c: Likewise.
3607 * tree-ssa-copy.c: Likewise.
3608 * tree-ssanames.c: Likewise.
3609 * tree-ssanames.h: Likewise.
3610 * tree-vrp.c: Likewise.
3611
5a8c2b57
AM
36122013-09-24 Alan Modra <amodra@gmail.com>
3613
3614 PR middle-end/57134
3615 PR middle-end/57586
7a5911d3
UB
3616 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3617 for output operands that disallow regs. Don't use EXPAND_WRITE on
3618 inout operands.
5a8c2b57 3619
f40333af
RB
36202013-09-24 Richard Biener <rguenther@suse.de>
3621
3622 PR middle-end/58513
3623 * tree.c (reference_alias_ptr_type): Move ...
3624 * alias.c (reference_alias_ptr_type): ... here and implement
3625 in terms of the new reference_alias_ptr_type_1.
3626 (ref_all_alias_ptr_type_p): New helper.
3627 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3628 use ref_all_alias_ptr_type_p.
3629 (get_deref_alias_set): Add flag_strict_aliasing check here.
3630 (reference_alias_ptr_type_1): New function, split out from ...
3631 (get_alias_set): ... here.
3632 (alias_ptr_types_compatible_p): New function.
3633 * alias.h (reference_alias_ptr_type): Declare.
3634 (alias_ptr_types_compatible_p): Likewise.
3635 * tree.h (reference_alias_ptr_type): Remove.
3636 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3637 to compare MEM_REF alias types.
3638
583e8bf5
RB
36392013-09-24 Richard Biener <rguenther@suse.de>
3640
3641 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3642
5e8586d7
MM
36432013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3644
19c3e797
MM
3645 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3646 reload helper function arrays into a single array reg_addr.
3647 (reload_fpr_gpr): Likewise.
3648 (reload_gpr_vsx): Likewise.
3649 (reload_vsx_gpr): Likewise.
3650 (struct rs6000_reg_addr): Likewise.
3651 (reg_addr): Likewise.
3652 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3653 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3654 (rs6000_init_hard_regno_mode_ok): Likewise.
3655 (rs6000_secondary_reload_direct_move): Likewise.
3656 (rs6000_secondary_reload): Likewise.
3657
5e8586d7
MM
3658 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3659 constraints: wu, ww, and wy. Repurpose wv constraint added during
3660 power8 changes. Put wg constraint in alphabetical order.
3661
3662 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3663 for future work to add ISA 2.07 VSX single precision support.
3664 (-mvsx-scalar-double): Change default from -1 to 1, update
3665 documentation comment.
3666 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3667 (-mupper-regs-df): New debug switch to control whether DF values
3668 can go in the traditional Altivec registers.
3669 (-mupper-regs-sf): New debug switch to control whether SF values
3670 can go in the traditional Altivec registers.
3671
3672 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3673 and wy constraints.
3674 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3675 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3676 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3677 (rs6000_debug_legitimate_address_p): Print if we are running
3678 before, during, or after reload.
3679 (rs6000_secondary_reload): Add a comment.
3680 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3681
3682 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3683 constraints. Update documentation string.
3684 (wd constraint): Likewise.
3685 (wf constraint): Likewise.
3686 (wg constraint): Likewise.
3687 (wn constraint): Likewise.
3688 (ws constraint): Likewise.
3689 (wt constraint): Likewise.
3690 (wx constraint): Likewise.
3691 (wz constraint): Likewise.
3692 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3693 instructions.
3694 (ww constraint): Likewise.
3695 (wy constraint): Likewise.
3696 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3697 the previous submissions.
3698 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3699
e5af9ddd
RS
37002013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3701
3702 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3703 (REG_BR_PROB): Say that the probability is stored in an int_list.
3704 * reg-notes.def: Update commentary to mention INT_LIST.
3705 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3706 (INT_LIST): New rtx.
3707 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3708 * rtlanal.c (int_reg_note_p): New function.
3709 (alloc_reg_note): Assert that the note does not have an int argument.
3710 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3711 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3712 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3713 rather than an INSN_LIST. Handle INT_LIST.
3714 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3715 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3716 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3717 Manipulate them as ints rather than rtxes.
3718 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3719 * regmove.c (copy_src_to_dest): Likewise.
3720 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3721
3722 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3723 into the cases that need it.
3724 * config/arm/arm.c (arm_unwind_emit): Likewise.
3725
3726 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3727 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3728 * loop-doloop.c (add_test, doloop_modify): Likewise.
3729 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3730 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3731 * predict.c (combine_predictions_for_insn): Likewise.
3732 * print-rtl.c (print_rtx): Handle INT_LIST.
3733 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3734 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3735 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3736 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3737 (ix86_split_fp_branch, predict_jump): Likewise.
3738 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3739 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3740 * config/spu/spu.c (ea_load_store_inline): Likewise.
3741
3742 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3743 value of a REG_BR_PROB note.
3744 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3745 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3746 * emit-rtl.c (try_split): Likewise.
3747 * predict.c (br_prob_note_reliable_p): Likewise.
3748 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
3749 * reorg.c (mostly_true_jump): Likewise.
3750 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
3751 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
3752 * config/i386/i386.c (ix86_print_operand): Likewise.
3753 * config/ia64/ia64.c (ia64_print_operand): Likewise.
3754 * config/mmix/mmix.c (mmix_print_operand): Likewise.
3755 * config/rs6000/rs6000.c (output_cbranch): Likewise.
3756 * config/s390/s390.c (s390_expand_tbegin): Likewise.
3757 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
3758 * config/sparc/sparc.c (output_cbranch): Likewise.
3759 * config/spu/spu.c (get_branch_target): Likewise.
3760 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
3761 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
3762
450ad0cd
JH
37632013-09-23 Jan Hubicka <jh@suse.cz>
3764
3765 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
3766 for ipa-devirt.
3767 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
3768 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
3769 of external calls
3770 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
3771 (gimple_fold_call): Dump inconsistent devirtualizations; add
3772 sanity check for type based devirtualizations.
3773 * ipa-prop.c: Include ipa-utils.h
3774 (ipa_intraprocedural_devirtualization): Add sanity check.
3775 (try_make_edge_direct_virtual_call): Likewise.
3776
cefce769
EB
37772013-09-23 Eric Botcazou <ebotcazou@adacore.com>
3778
3779 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
3780 assignment statements.
3781
a895a2b8
KV
37822013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
3783
3784 * gimple-pretty-print.c (dump_ssaname_info): New function.
3785 (dump_gimple_phi): Call it.
3786 (pp_gimple_stmt_1): Likewise.
3787 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
3788 (range_info_def): Declare.
3789 * tree-pretty-print.c (pp_double_int): New function.
3790 (dump_generic_node): Call it.
3791 * tree-pretty-print.h (pp_double_int): Declare.
3792 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
3793 * tree-ssanames.h (range_info_def): New structure.
3794 (value_range_type): Move definition here.
3795 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
3796 Declare.
3797 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
3798 initialization.
3799 (set_range_info): New function.
3800 (get_range_info): Likewise.
3801 (duplicate_ssa_name_range_info): Likewise.
3802 (duplicate_ssa_name_fn): Check pointer type and call
3803 duplicate_ssa_name_range_info.
3804 * tree-ssa-copy.c (fini_copy_prop): Likewise.
3805 * tree-vrp.c (value_range_type): Remove definition, now in
3806 tree-ssanames.h.
7a5911d3 3807 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
a895a2b8
KV
3808 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
3809 (SSA_NAME_RANGE_INFO): New macro.
3810
984af6ac
RB
38112013-09-23 Richard Biener <rguenther@suse.de>
3812
3813 PR tree-optimization/58464
3814 * tree-ssa-pre.c (phi_trans_lookup): Remove.
7a5911d3 3815 (phi_trans_add): Change to add conditionally on being not present.
984af6ac
RB
3816 (phi_translate_1): Remove recursion detection here.
3817 (phi_translate): Pre-seed the cache with NULL to catch
3818 recursion here in a more generic way.
3819 (bitmap_find_leader): Adjust comment.
3820 (get_representative_for): Dump value-numbers.
3821 (create_expression_by_pieces): Likewise.
3822 (insert_into_preds_of_block): Likewise.
3823
0e26cf79
CB
38242013-09-23 Christian Bruel <christian.bruel@st.com>
3825
3826 PR target/58475
3827 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
3828 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
3829
4d0e5b54
JG
38302013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
3831
3832 Revert r202780:
3833 2013-09-20 Renlin Li <renlin.li@arm.com>
3834
7a5911d3
UB
3835 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3836 plus_constant.
4d0e5b54
JG
3837 (aarch64_expand_epilogue): Likewise.
3838 (aarch64_legitimize_reload_address): Likewise.
3839
ca081cc8
EB
38402013-09-22 Eric Botcazou <ebotcazou@adacore.com>
3841
3842 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
3843 NULL_TREE before pushing them onto the vector. Likewise for labels.
3844
8fc53a5f
EB
38452013-09-21 Eric Botcazou <ebotcazou@adacore.com>
3846
3847 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
3848 comparison operators when -fno-trapping-math is in effect.
3849 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
3850 comparison operators in TFmode and assert that unsupported operators
3851 cannot reach here.
3852 (ia64_print_operand): Likewise.
3853
38542013-09-21 Jan Hubicka <jh@suse.cz>
0ca6c49f
JH
3855
3856 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
3857 (sse_typeless_stores): Enable for core
3858 (sse_load0_by_pxor): Likewise.
3859 (four_jump_limit): Disable for core.
3860 (pad_returns): Likewise.
3861 (avoid_vector_decode): Likewise.
3862 (fuse_cmp_and_branch): Enable for cores.
3863 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
3864
578e51e6
JDA
38652013-09-20 John David Anglin <danglin@gcc.gnu.org>
3866
96ce28a3
JDA
3867 PR middle-end/56791
3868 * config/pa/pa.c (pa_option_override): Disable auto increment and
3869 decrement instructions until reload is completed.
3870
578e51e6
JDA
3871 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
3872 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
3873 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
3874
6fcd3a13 38752013-09-20 DJ Delorie <dj@redhat.com>
7a5911d3 3876 Nick Clifton <nickc@redhat.com>
6fcd3a13
DD
3877
3878 * config/rl78/rl78.c: Various whitespace and comment tweaks.
3879 (need_to_save): Save bank 0 on interrupts.
3880 (characterize_address): Strip far address wrappers.
3881 (rl78_as_legitimate_address): Likewise.
3882 (transcode_memory_rtx): Likewise.
3883 (rl78_peep_movhi_p): Disable this peephole after devirt.
3884 (rl78_propogate_register_origins): Forget all origins when a
3885 CLOBBER is seen.
3886 * config/rl78/rl78-virt.md: Various whitespace tweaks.
3887 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
3888 peephole2's.
3889 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
3890 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
3891 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
3892 relocating.
3893 * config/rl78/constraints.md: Various whitespace and paren tweaks.
3894
4a08db33
JDA
38952013-09-20 John David Anglin <danglin@gcc.gnu.org>
3896
3897 * config/pa/pa.md: In "scc" insn patterns, change output template to
3898 handle const0_rtx in reg_or_0_operand operands.
3899
7691132c
JBG
39002013-09-20 Martin Husemann <martin@NetBSD.org>
3901
3902 PR target/56875
3903 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
3904 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
3905
a3cc13cc
RB
39062013-09-20 Richard Biener <rguenther@suse.de>
3907
3908 PR middle-end/58484
3909 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
3910 remembering SSA name version and block index.
3911 (new_scev_info_str): Adjust.
3912 (hash_scev_info): Likewise. Also hash the block index.
3913 (eq_scev_info): Adjust.
3914 (find_var_scev_info): Likewise.
3915 (struct instantiate_cache_entry): Remove.
3916 (struct instantiate_cache_type): Use a htab to map name, block
3917 to chrec.
3918 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3919 (get_instantiated_value_entry): Likewise.
3920 (hash_idx_scev_info, eq_idx_scev_info): New functions.
3921 (instantiate_scev_name): Adjust.
3922
925f3871
JL
39232013-09-20 Jeff Law <law@redhat.com>
3924
3925 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
3926
2888c331
YZ
39272013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
3928
3929 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
3930 Call aarch64_simd_expand_args to update op[argc].
3931
0a811e96
BS
39322013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3933
7a5911d3
UB
3934 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
3935 plugin argument.
0a811e96 3936
0078f462
BS
39372013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3938
7a5911d3
UB
3939 * gengtype.c (file_rules): Added rule for *.cc files.
3940 (get_output_file_with_visibility): Give fatal message when no
3941 rules found.
0078f462 3942
d20e48b0
RL
39432013-09-20 Renlin Li <renlin.li@arm.com>
3944
3945 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
3946 (aarch64_expand_epilogue): Likewise.
3947 (aarch64_legitimize_reload_address): Likewise.
3948
59bb1544
BE
39492013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
3950
3951 PR middle-end/57748
3952 * expr.c (expand_assignment): Remove misalignp code path.
3953
a5430019
MP
39542013-09-20 Marek Polacek <polacek@redhat.com>
3955
3956 PR sanitizer/58413
3957 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
3958 TYPE_PRECISION. Add asserts.
3959
9fed7f3a
RB
39602013-09-20 Richard Biener <rguenther@suse.de>
3961
3962 PR tree-optimization/58453
3963 * tree-loop-distribution.c (distribute_loop): Apply the cost
3964 model for -ftree-loop-distribute-patterns, too.
3965
a820c834
RB
39662013-09-20 Richard Biener <rguenther@suse.de>
3967
3968 PR middle-end/58473
3969 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
3970 make type comparison less strict.
3971
3cbe17f7
AM
39722013-09-20 Alan Modra <amodra@gmail.com>
3973
3974 * configure: Regenerate.
3975 * aclocal.m4: Regenerate.
3976
87fccdbb
MP
39772013-09-20 Marek Polacek <polacek@redhat.com>
3978
3979 PR other/58467
3980 * doc/extend.texi: Document that attribute used is meant to be used
3981 on variables with static storage duration.
3982
8928eff3
JJ
39832013-09-19 Jakub Jelinek <jakub@redhat.com>
3984
3985 PR tree-optimization/58472
3986 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
3987 simd_lane_access set inv_p = false.
3988 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
3989 the simduid magic VAR_DECL.
3990
b7b28c53
JH
39912013-09-19 Jan Hubicka <jh@suse.cz>
3992
3993 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
3994
83ae86f5
JL
39952013-09-17 Jeff Law <law@redhat.com>
3996
3997 * tree-ssa-dom.c (record_temporary_equivalences): New function
3998 split out of dom_opt_dom_walker::after_dom_children.
3999 (dom_opt_dom_walker::thread_across_edge): Move common code
4000 in here from dom_opt_dom_walker::after_dom_children.
4001 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
4002
9d532162
JH
40032013-09-19 Jan Hubicka <jh@suse.cz>
4004
4005 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
4006 (TARGET_GENERIC): Use PROCESOR_GENERIC
4007 (enum processor_type): Unify generic32 and 64.
4008 * i386.md (cpu): Likewise.
4009 * x86-tune.def (use_leave): Enable for generic32.
4010 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
4011 * athlon.md: Change generic64 to generic in all occurences.
4012 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
4013 (ix86_target_macros_internal): Likewise.
4014 * driver-i386.c (host_detect_local_cpu): Likewise.
7a5911d3
UB
4015 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
4016 to ..
9d532162
JH
4017 (generic_memcpy, generic_memset, generic_cost): This one.
4018 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
4019 (m_GENERIC32, m_GENERIC64): Remove.
4020 (m_GENERIC): Turn into one flag.
4021 (processor_target): Unify generic tunnings.
4022 (ix86_option_override_internal): Replace generic32/64 by generic.
4023 (ix86_issue_rate): Likewise.
4024 (ix86_adjust_cost): Likewise.
4025
188c7d00
JH
40262013-09-19 Jan Hubicka <jh@suse.cz>
4027
4028 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
4029 of speculative flag.
4030
4befd127
JJ
40312013-09-19 Jakub Jelinek <jakub@redhat.com>
4032
4033 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4034 GOMP_sections_start, even if !exit_reachable.
4035
8a26ad39
VM
40362013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4037
4038 * lra-constraints.c (need_for_all_save_p): Use macro
4039 HARD_REGNO_CALL_PART_CLOBBERED.
4040 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4041 set up pseudo conflict hard regs.
4042
09487185
MM
40432013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4044
4045 PR target/58452
4046 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4047 operands.
4048
080cbf9e
VM
40492013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4050
4051 PR rtl-optimization/58438
4052 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4053 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4054 from previous subpasses.
4055
a0f70fcb
RE
40562013-09-18 Richard Earnshaw <rearnsha@arm.com>
4057
4058 * arm.c (arm_get_frame_offsets): Validate architecture supports
4059 LDRD/STRD before accepting the tuning preference.
4060 (arm_expand_prologue): Likewise.
4061 (arm_expand_epilogue): Likewise.
4062
0547c9b6
RB
40632013-09-18 Richard Biener <rguenther@suse.de>
4064
4065 PR tree-optimization/58417
4066 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4067 have chrecs with symbols defined in the loop as operands.
4068 (chrec_fold_multiply): Likewise.
4069 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4070 parameters before folding binary operations.
4071 (struct instantiate_cache_entry_hasher): Remove.
4072 (struct instantiate_cache_type): Use a pointer-map.
4073 (instantiate_cache_type::instantiate_cache_type): New function.
4074 (instantiate_cache_type::get): Likewise.
4075 (instantiate_cache_type::set): Likewise.
4076 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4077 (get_instantiated_value_entry): Likewise.
4078 (global_cache): New global.
4079 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4080 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4081 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4082 (instantiate_scev_name): Adjust.
4083 (instantiate_scev): Construct global instead of local cache.
4084 (resolve_mixers): Likewise.
4085
309dc1aa
DM
40862013-09-18 Daniel Morris <danielm@ecoscentric.com>
4087 Paolo Carlini <paolo.carlini@oracle.com>
4088
4089 PR c++/58458
4090 * doc/implement-cxx.texi: Fix references to the C++ standards.
4091
995a1b4a
JJ
40922013-09-18 Jakub Jelinek <jakub@redhat.com>
4093
4094 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4095 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4096 simd_lane_access drs, update also DR_ALIGNED_TO.
4097
ce6923c5
MP
40982013-09-18 Marek Polacek <polacek@redhat.com>
4099
4100 PR sanitizer/58411
4101 * doc/extend.texi: Document no_sanitize_undefined attribute.
4102 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4103 no_sanitize_undefined attribute.
4104
d30d00a2
NC
41052013-09-18 Nick Clifton <nickc@redhat.com>
4106
4107 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4108 (ASM_DECLARE_FUNCTION_NAME): Define.
4109
4d9192b5
TS
41102013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4111
4112 * compare-elim.c (find_comparison_dom_walker): New class
4113 (find_comparisons_in_bb): Rename to
4114 find_comparison_dom_walker::before_dom_children
4115 (find_comparisons): Adjust
4116 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4117 adjust.
4118 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4119 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4120 (init_walk_dominator_tree): Remove declaration.
4121 (fini_walk_dominator_tree): Remove declaration.
4122 * fwprop.c (single_def_use_dom_walker): New class
4123 (single_def_use_enter_block): Convert to
4124 single_def_use_dom_walker::before_dom_children.
4125 (single_def_use_leave_block): Convert to
4126 single_def_use_dom_walker::after_dom_children.
4127 (build_single_def_use_links): Adjust.
4128 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4129 class.
4130 (find_candidates_in_block): Convert to
4131 find_candidates_dom_walker::before_dom_children.
4132 (execute_strength_reduction): Adjust.
4133 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4134 (sese_dom_walker): New class.
4135 (sese_dom_walker::sese_dom_walker): New constructor.
4136 (sese_dom_walker::~sese_dom_walker): New destructor.
4137 (build_sese_conditions_before): Convert to
4138 sese_dom_walker::before_dom_children.
4139 (build_sese_conditions_after): Convert to
4140 sese_dom_walker::after_dom_children.
4141 (build_poly_scop): Adjust
4142 * tree-into-ssa.c (rewrite_dom_walker): New class
4143 (rewrite_enter_block): Convert to
4144 rewrite_dom_walker::before_dom_children.
4145 (rewrite_leave_block): Convert to
4146 rewrite_dom_walker::after_dom_children.
4147 (rewrite_update_dom_walker): New class.
4148 (rewrite_update_enter_block): Convert to
4149 rewrite_update_dom_walker::before_dom_children.
4150 (rewrite_update_leave_block): Convert to
4151 rewrite_update_dom_walker::after_dom_children.
4152 (rewrite_blocks, rewrite_into_ssa): Adjust.
4153 (mark_def_dom_walker): New class.
4154 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4155 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4156 (mark_def_sites_blocks): Convert to
4157 mark_def_dom_walker::before_dom_children.
4158 (mark_def_site_blocks): Remove.
4159 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4160 (tree_ssa_dominator_optimize): Adjust.
4161 (dom_thread_across_edge): Convert to method
4162 dom_opt_dom_walker::thread_across_edge.
4163 (dom_opt_enter_block): Convert to member function
4164 dom_opt_dom_walker::before_dom_children.
4165 (dom_opt_leave_block): Convert to member function
4166 dom_opt_dom_walker::after_dom_children.
4167 * tree-ssa-dse.c (dse_dom_walker): New class.
4168 (dse_enter_block): Convert to member function
4169 dse_dom_walker::before_dom_children.
4170 (tree_ssa_dse): Adjust.
4171 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4172 (determine_invariantness_stmt): Convert to method
4173 invariantness_dom_walker::before_dom_children.
4174 (determine_invariantness): Remove
4175 (move_computations_dom_walker): New class.
4176 (move_computations_stmt): Convert to method
4177 move_computations_dom_walker::before_dom_children.
4178 (move_computations, tree_ssa_lim): Adjust.
4179 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4180 (nt_init_block): Make method
4181 notrappping_dom_walker::before_dom_children.
4182 (nt_fini_block): Make
4183 method nontrapping_dom_walker::after_dom_children.
4184 (get_non_trapping): Adjust.
4185 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4186 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4187 (eliminate_leave_block): Make method.
4188 eliminate_dom_walker::after_dom_children.
4189 (eliminate): Adjust
4190 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4191 (strlen_enter_block): Make method
4192 strlen_dom_walker::before_dom_children.
4193 (strlen_leave_block): Make
4194 method strlen_dom_walker::after_dom_children.
4195 (tree_ssa_strlen): Adjust.
4196 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4197 (tree_ssa_uncprop): Adjust.
4198 (uncprop_leave_block): Make method
4199 uncprop_dom_walker::after_dom_children.
4200 (uncprop_leave_block): Make method
4201 uncprop_dom_walker::before_dom_children.
4202
3b3cc26b
BC
42032013-09-18 Bin Cheng <bin.cheng@arm.com>
4204
7a5911d3
UB
4205 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4206 branch in same basic block. Check both src and dest of the move insn.
3b3cc26b 4207
b5ba3a9b
NC
42082013-09-17 Nick Clifton <nickc@redhat.com>
4209
4210 * config/rl78/rl78-real.md (bf): New pattern.
4211 (bt): New pattern.
4212 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4213 (rl78_print_operand): Do not put a # before a %B.
4214 * config/rl78/rl78.opt: Tweak doc strings.
4215
09d5094b
DD
42162013-09-17 DJ Delorie <dj@redhat.com>
4217
4218 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4219 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4220 (rl78_option_override): Likewise, if -mallregs.
4221 (is_virtual_register): Likewise.
4222 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4223 (REGNO_OK_FOR_BASE_P): Likewise.
4224 * config/rl78/rl78.opt (-mallregs): New.
4225
2ba1dca1
NC
42262013-09-17 Nick Clifton <nickc@redhat.com>
4227
4228 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4229 For interrupt functions: save all call clobbered registers if the
4230 interrupt handler is not a leaf function.
4231 (rl78_expand_prologue): Always recompute the frame information.
4232 For interrupt functions: only select bank 0 if one of the bank 0
4233 registers is going to be psuhed.
4234
2e7c3f21
DD
42352013-09-17 DJ Delorie <dj@redhat.com>
4236
4237 * config/rl78/constraints.md: For each W* constraint, rename to C*
4238 and create a W* constraint that checks for an optional ES: prefix
4239 pattern also.
4240 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4241 (es_addr): New. Used to wrap far addresses.
4242 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4243 (rl78_es_base): New.
4244 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4245 wrapped far addresses.
4246 (rl78_print_operand_1): Unwrap far addresses before processing.
4247 (rl78_lo16): Wrap far addresses in unspecs.
4248 (rl78_es_addr): New.
4249 (rl78_es_base): New.
4250 (insn_ok_now): Check for not-yet-wrapped far addresses.
4251 (transcode_memory_rtx): Properly re-wrap far addresses.
4252
5c26a69a
SH
42532013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4254
4255 * config/sparc/t-rtems: Add leon3 multilibs.
4256
181f5f3e
CH
42572013-09-17 Cong Hou <congh@google.com>
4258
4259 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4260 when checking the dot production pattern. The type of rhs operand
4261 of multiply is now checked correctly.
4262
8d34e421
JL
42632013-09-17 Jeff Law <law@redhat.com>
4264
4265 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4266 edge implied equivalences into successor phis.
4267 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4268 here from tree-ssa-threadedge.c.
4269 (mark_threaded_blocks): When threading through a joiner, if both
4270 successors of the joiner's clone reach the same block, verify the
4271 PHI arguments are equal. If not, cancel the jump threading request.
4272 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4273 tree-ssa-threadupdate.c
4274 (thread_across_edge): Don't check PHI argument equality when
4275 threading through joiner block here.
7a5911d3 4276
c152901f
AM
42772013-09-17 Andrew MacLeod <amacleod@redhat.com>
4278
4279 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4280 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
7a5911d3
UB
4281 (warn_uninit, warn_uninitialized_vars,
4282 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4283 Move to tree-ssa-uninit.c.
78cedfb1 4284 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
7a5911d3 4285 (has_undefined_value_p): New. Pass dependant parts of
c152901f
AM
4286 ssa_undefined_value_p.
4287 (uninit_undefined_value_p): Use has_undefined_value_p.
7a5911d3
UB
4288 (warn_uninit, warn_uninitialized_vars,
4289 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4290 Move from tree-ssa.c.
4291 * tree-ssa.h: Adjust prototypes.
c152901f 4292
b631d45a
JH
42932013-09-17 Jan Hubicka <jh@suse.cz>
4294
4295 PR middle-end/58332
4296 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4297 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4298 FUNCTION_NOT_OPTIMIZED.
4299 * ipa-inline-analysis.c (compute_inline_parameters): Function
4300 not optimized is not inlinable unless it is alwaysinline.
4301 (inline_analyze_function): Force calls in not optimized
4302 function not inlinable.
4303
5b79657a
JH
43042013-09-17 Jan Hubicka <jh@suse.cz>
4305
4306 PR middle-end/58329
4307 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4308 to return NULL.
4309 * ipa.c (function_and_variable_visibility): Likewise.
4310 * ipa-profile.c (ipa_profile): Likewise.
4311
27442c24
BE
43122013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4313
4314 PR ipa/58398
4315 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4316 attribute, and don't inline the resolver in this case.
4317
753b9bc3
TJ
43182013-09-17 Teresa Johnson <tejohnson@google.com>
4319
4320 * coverage.c (get_coverage_counts): Add missing newline.
4321
e3d3cfb4
KT
43222013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4323
4324 PR tree-optimization/58088
4325 * fold-const.c (mask_with_trailing_zeros): New function.
4326 (fold_binary_loc): Make sure we don't recurse infinitely
4327 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4328 Use mask_with_trailing_zeros where appropriate.
4329
9a0ac98f
YR
43302013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4331
4332 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4333 of 'prev' var to get better distance estimation.
4334
78bbd765
EB
43352013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4336
4337 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4338 * tree-inline.c (is_parameter_of): New predicate.
4339 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4340 a parameter has been remapped.
4341 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4342 (optimize_inline_calls): Initialize transform_parameter.
4343 (copy_gimple_seq_and_replace_locals): Likewise.
4344 (tree_function_versioning): Likewise.
4345 (maybe_inline_call_in_expr): Likewise.
4346
cad055a4
NC
43472013-09-17 Nick Clifton <nickc@redhat.com>
4348
4349 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4350 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4351 interrupt handlers.
4352 (is_attr_func): New function.
4353 (msp430_is_interrupt_func): New function.
4354 (is_naked_func): New function.
4355 (is_reentrant_func): New function.
4356 (is_critical_func): New function.
4357 (msp430_start_function): Add annotations for function attributes.
4358 (msp430_attr): New function.
4359 (msp430_attribute_table): New.
4360 (msp430_function_section): New function.
4361 (TARGET_ASM_FUNCTION_SECTION): Define.
4362 (msp430_builtin): New enum.
4363 (msp430_init_builtins): New function.
4364 (msp430_builtin_devl): New function.
4365 (msp430_expand_builtin): New function.
4366 (TARGET_INIT_BUILTINS): Define.
4367 (TARGET_EXPAND_BUILTINS): Define.
4368 (TARGET_BUILTIN_DECL): Define.
4369 (msp430_expand_prologue): Add support for naked, interrupt,
4370 critical and reentrant functions.
4371 (msp430_expand_epilogue): Likewise.
4372 (msp430_print_operand): Handle 'O' character.
4373 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4374 NO_TRAMPOLINES.
4375 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4376 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4377 (pushm): Use a 'n' rather than an 'i' constraint.
4378 (msp_return): Add generation of the interrupt return instruction.
4379 (disable_interrupts): New pattern.
4380 (enable_interrupts): New pattern.
4381 (push_intr_state): New pattern.
4382 (pop_intr_state): New pattern.
4383 (bic_SR): New pattern.
4384 (bis_SR): New pattern.
4385 * doc/extend.texi: Document MSP430 function attributes and builtin
4386 functions.
4387
deb6c11a
RB
43882013-09-17 Richard Biener <rguenther@suse.de>
4389
4390 PR tree-optimization/58432
4391 * tree-loop-distribution.c (tree_loop_distribution): Also
4392 scan PHIs for outside loop uses and seed a partition from them.
4393
c068654b
BC
43942013-09-17 Bin Cheng <bin.cheng@arm.com>
4395
4396 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4397 (restructure_reference): Call backtrace_base_for_ref.
4398
567969e4
AM
43992013-09-17 Alan Modra <amodra@gmail.com>
4400
4401 PR target/57589
7a5911d3
UB
4402 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4403 patch.
567969e4 4404
5c0029de
DD
44052013-09-16 DJ Delorie <dj@redhat.com>
4406
4407 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4408 vregs location for RL78/G10.
4409 (rl78_expand_prologue): Avoid SEL on G10.
4410 (rl78_expand_epilogue): Likewise.
4411 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4412 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4413 __RL78_G10__ when appropriate.
4414 (ASM_SPEC): Pass -mg10 along to the assembler.
4415 * config/rl78/rl78.md (sel_rb): Disable for G10.
4416 * config/rl78/rl78.opt: Add -mg10 option.
4417 * config/rl78/t-rl78: Add -mg10 multilib.
4418
ea0f3e87
XDL
44192013-09-16 Xinliang David Li <davidxl@google.com>
4420
4421 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4422 * omp-low.c (omp_max_vf): Ditto.
4423 (expand_omp_simd): Ditto.
4424 * tree-vectorizer.c (vectorize_loops): Ditto.
4425 (gate_vect_slp): Ditto.
4426 (gate_increase_alignment): Ditto.
4427 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4428 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4429 (gate_tree_vectorize): Name change.
4430 (tree_vectorize): Ditto.
4431 (pass_vectorize::gate): Call new function.
4432 (pass_vectorize::execute): Ditto.
7a5911d3 4433 * opts.c: O3 default setting change.
ea0f3e87
XDL
4434 (finish_options): Check new flag.
4435 * doc/invoke.texi: Document new flags.
4436 * common.opt: New flags.
4437
4f3da2e5
AS
44382013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4439
4440 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4441 (Condition Code): Don't reference it.
4442 * doc/tm.texi: Regenerate.
4443
aa940f7c
VM
44442013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4445
4446 PR middle-end/58418
4447 * lra-constraints.c (undo_optional_reloads): Consider all optional
4448 reload even if it did not get a hard reg.
4449
e645e942
TJ
44502013-09-16 Teresa Johnson <tejohnson@google.com>
4451
4452 * dumpfile.c (dump_loc): Remove newline emission.
4453 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4454 emission to dump_printf_loc calls where missing.
4455 (vect_mark_for_runtime_alias_test): Ditto.
4456 (vect_analyze_data_ref_dependence): Ditto.
4457 (vect_analyze_data_ref_dependences): Ditto.
4458 (vect_slp_analyze_data_ref_dependence): Ditto.
4459 (vect_slp_analyze_data_ref_dependences): Ditto.
4460 (vect_compute_data_ref_alignment): Ditto.
4461 (vect_update_misalignment_for_peel): Ditto.
4462 (vect_verify_datarefs_alignment): Ditto.
4463 (vector_alignment_reachable_p): Ditto.
4464 (vect_get_data_access_cost): Ditto.
4465 (vect_enhance_data_refs_alignment): Ditto.
4466 (vect_find_same_alignment_drs): Ditto.
4467 (vect_analyze_data_refs_alignment): Ditto.
4468 (vect_analyze_group_access): Ditto.
4469 (vect_analyze_data_ref_access): Ditto.
4470 (vect_analyze_data_ref_accesses): Ditto.
4471 (vect_prune_runtime_alias_test_list): Ditto.
4472 (vect_analyze_data_refs): Ditto.
4473 (vect_create_addr_base_for_vector_ref): Ditto.
4474 (vect_create_data_ref_ptr): Ditto.
4475 (vect_grouped_store_supported): Ditto.
4476 (vect_grouped_load_supported): Ditto.
4477 * value-prof.c (check_counter): Ditto.
4478 (check_ic_target): Ditto.
4479 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4480 (vect_recog_widen_mult_pattern): Ditto.
4481 (vect_recog_widen_sum_pattern): Ditto.
4482 (vect_recog_over_widening_pattern): Ditto.
4483 (vect_recog_widen_shift_pattern): Ditto.
4484 (vect_recog_rotate_pattern): Ditto.
4485 (vect_recog_vector_vector_shift_pattern): Ditto.
4486 (vect_recog_divmod_pattern): Ditto.
4487 (vect_recog_mixed_size_cond_pattern): Ditto.
4488 (vect_recog_bool_pattern): Ditto.
4489 (vect_pattern_recog_1): Ditto.
4490 (vect_pattern_recog): Ditto.
4491 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4492 (vect_is_simple_iv_evolution): Ditto.
4493 (vect_analyze_scalar_cycles_1): Ditto.
4494 (vect_get_loop_niters): Ditto.
4495 (vect_analyze_loop_1): Ditto.
4496 (vect_analyze_loop_form): Ditto.
4497 (vect_analyze_loop_operations): Ditto.
4498 (vect_analyze_loop_2): Ditto.
4499 (vect_analyze_loop): Ditto.
4500 (report_vect_op): Ditto.
4501 (vect_is_slp_reduction): Ditto.
4502 (vect_is_simple_reduction_1): Ditto.
4503 (vect_get_known_peeling_cost): Ditto.
4504 (vect_estimate_min_profitable_iters): Ditto.
4505 (vect_model_reduction_cost): Ditto.
4506 (vect_model_induction_cost): Ditto.
4507 (get_initial_def_for_induction): Ditto.
4508 (vect_create_epilog_for_reduction): Ditto.
4509 (vectorizable_reduction): Ditto.
4510 (vectorizable_induction): Ditto.
4511 (vectorizable_live_operation): Ditto.
4512 (vect_loop_kill_debug_uses): Ditto.
4513 (vect_transform_loop): Ditto.
4514 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4515 (vect_stmt_relevant_p): Ditto.
4516 (process_use): Ditto.
4517 (vect_mark_stmts_to_be_vectorized): Ditto.
4518 (vect_model_simple_cost): Ditto.
4519 (vect_model_promotion_demotion_cost): Ditto.
4520 (vect_model_store_cost): Ditto.
4521 (vect_get_store_cost): Ditto.
4522 (vect_model_load_cost): Ditto.
4523 (vect_get_load_cost): Ditto.
4524 (vect_init_vector_1): Ditto.
4525 (vect_get_vec_def_for_operand): Ditto.
4526 (vect_finish_stmt_generation): Ditto.
4527 (vectorizable_call): Ditto.
4528 (vectorizable_conversion): Ditto.
4529 (vectorizable_assignment): Ditto.
4530 (vectorizable_shift): Ditto.
4531 (vectorizable_operation): Ditto.
4532 (vectorizable_store): Ditto.
4533 (vectorizable_load): Ditto.
4534 (vectorizable_condition): Ditto.
4535 (vect_analyze_stmt): Ditto.
4536 (vect_transform_stmt): Ditto.
4537 (vect_is_simple_use): Ditto.
4538 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4539 (vect_can_advance_ivs_p): Ditto.
4540 (vect_update_ivs_after_vectorizer): Ditto.
4541 (vect_do_peeling_for_loop_bound): Ditto.
4542 (vect_gen_niters_for_prolog_loop): Ditto.
4543 (vect_update_inits_of_drs): Ditto.
4544 (vect_create_cond_for_alias_checks): Ditto.
4545 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4546 (vect_build_slp_tree_1): Ditto.
4547 (vect_supported_load_permutation_p): Ditto.
4548 (vect_analyze_slp_instance): Ditto.
4549 (vect_analyze_slp): Ditto.
4550 (vect_make_slp_decision): Ditto.
4551 (vect_detect_hybrid_slp): Ditto.
4552 (vect_bb_vectorization_profitable_p): Ditto.
4553 (vect_slp_analyze_bb_1): Ditto.
4554 (vect_update_slp_costs_according_to_vf): Ditto.
4555 (vect_get_mask_element): Ditto.
4556 (vect_transform_slp_perm_load): Ditto.
4557 (vect_schedule_slp_instance): Ditto.
4558 (vect_schedule_slp): Ditto.
4559 (vect_slp_transform_bb): Ditto.
4560 * profile.c (read_profile_edge_counts): Ditto.
4561 (compute_branch_probabilities): Ditto.
4562 * coverage.c (get_coverage_counts): Ditto.
4563
46f851f3
DN
45642013-09-16 Diego Novillo <dnovillo@google.com>
4565
7a5911d3 4566 * tree-core.h: Add missing comment lines from refactoring of tree.h.
46f851f3 4567
1632a686
JH
45682013-09-16 Jan Hubicka <jh@suse.cz>
4569
4570 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
7a5911d3 4571 abstract functions; for static functions check the presence of body.
1632a686 4572
828e70c1
JG
45732013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4574
4575 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4576 * config/aarch64/aarch64-simd.md
4577 (aarch64_mla_elt<mode>): New.
4578 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4579 (aarch64_mls_elt<mode>): Likewise.
4580 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4581 (aarch64_fma4_elt<mode>): Likewise.
4582 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4583 (aarch64_fma4_elt_to_128v2df): Likewise.
4584 (aarch64_fma4_elt_to_64df): Likewise.
4585 (fnma<mode>4): Likewise.
4586 (aarch64_fnma4_elt<mode>): Likewise.
4587 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4588 (aarch64_fnma4_elt_to_128v2df): Likewise.
4589 (aarch64_fnma4_elt_to_64df): Likewise.
4590 * config/aarch64/iterators.md (VDQSF): New.
4591 * config/aarch64/arm_neon.h
4592 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4593 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4594
779aea46
JG
45952013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4596
4597 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4598 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4599 (aarch64_mul3_elt_to_128df): Likewise.
4600 (aarch64_mul3_elt_to_64v2df): Likewise.
4601 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4602 (VMUL): New.
4603 (VMUL_CHANGE_NLANES) Likewise.
4604 (h_con): Likewise.
4605 (f): Likewise.
4606 * config/aarch64/arm_neon.h
4607 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4608
a407a750
JG
46092013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4610
4611 * config/aarch64/arm_neon.h
4612 (vcvtx_high_f32_f64): Fix parameters.
4613
f77e2d2d 46142013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
a5bd71fa 4615 Uros Bizjak <ubizjak@gmail.com>
f77e2d2d 4616
a5bd71fa 4617 * config/alpha.c: Include tree-ssanames.h.
f77e2d2d 4618
36875e8f
RB
46192013-09-16 Richard Biener <rguenther@suse.de>
4620
4621 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4622 (dot_rdg_1): Handle control_dd.
4623 (create_edge_for_control_dependence): New function.
4624 (create_rdg_edges): Add control dependences if asked for.
4625 (build_rdg): Likewise.
4626 (generate_loops_for_partition): If there are not necessary
4627 control stmts remove all their dependencies.
4628 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4629 (distribute_loop): Pass on control dependences.
4630 (tree_loop_distribution): Compute control dependences and remove
4631 restriction on number of loop nodes.
4632
96e24d49
JJ
46332013-09-16 Jakub Jelinek <jakub@redhat.com>
4634
4635 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4636 for internal calls.
4637
40dbb05c
RS
46382013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4639
4640 * cse.c (try_const_anchors): Punt on CC modes.
4641
97191ad0
JBG
46422013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4643
4644 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4645
403a3fb7
JDA
46462013-09-14 John David Anglin <danglin@gcc.gnu.org>
4647
4648 PR target/58382
4649 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4650 calls to word_mode.
4651
fb5610fb
IS
46522013-09-14 Iain Sandoe <iain@codesourcery.com>
4653
4654 PR target/48094
4655 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
f77e2d2d
JBG
4656 seen.
4657 (darwin_objc1_section): Likewise.
fb5610fb
IS
4658 (darwin_file_end): Emit Image Info section when required.
4659
afdec9bd
JH
46602013-09-14 Jan Hubicka <jh@suse.cz>
4661
4662 * tree-into-ssa.c (gate_into_ssa): New.
4663 (pass_data_build_ssa): Use it.
4664 * cgraph.h (expand_thunk): Update prototype.
4665 * cgraphunit.c (analyze_function): Expand thunks early.
4666 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4667 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4668 set lowered flag; do not add new function.
4669 (assemble_thunks_and_aliases): Update.
4670 * tree-ssa.c (gate_init_datastructures): New gate.
4671 (pass_data_init_datastructures): Use it.
4672
492d1e0a
IS
46732013-09-14 Iain Sandoe <iain@codesourcery.com>
4674
4675 PR target/58269
f77e2d2d 4676 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
492d1e0a
IS
4677 xmm register set described in the psABI.
4678
d13c6b70
EG
46792013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4680
f77e2d2d
JBG
4681 * dwarf2out.c (should_emit_struct_debug): Add check
4682 for type_decl variable is not NULL.
d13c6b70 4683
88b0e79e
JC
46842013-09-13 Jacek Caban <jacek@codeweavers.com>
4685
4686 * config.gcc: Use new winnt-c.c target hooks
4687 * config/t-winnt: New file
4688 * config/winnt-c.c: New file
4689 * doc/tm.texi.in: Document new hook
4690 * doc/tm.texi: Regenerated
4691
19ba6aab
JH
46922013-09-13 Jan Hubicka <jh@suse.cz>
4693
4694 PR middle-end/58094
4695 * ipa-inline.c (check_callers): New function.
4696 (check_caller_edge): Remove.
4697 (want_inline_function_to_all_callers_p): Also permit alises that are
4698 called dirrectly.
4699 (inline_to_all_callers): Terminate the walk when devirtualization
4700 introduce new calls.
4701
a93c18c8
JH
47022013-09-13 Jan Hubicka <jh@suse.cz>
4703
4704 * ipa-inline-analysis.c (struct growth_data): Add node.
4705 (do_estimate_growth_1): Fix detection of recursion.
4706
f3b331d1
JJ
47072013-09-13 Jakub Jelinek <jakub@redhat.com>
4708
4709 PR tree-optimization/58392
4710 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4711 to avoid shadowing of outer loop variable. If
4712 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4713 replace_by_duplicate_decl simduid of loops that have it set and
4714 set dest_cfun->has_simduid_loops and/or
4715 dest_cfun->has_force_vect_loops.
4716 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4717 instead of maybe_lookup_decl.
4718 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4719 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4720 call. Copy over force_vect and copy and remap simduid. Set
4721 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4722 (copy_cfg_body): Remove blocks_to_copy argument. Use
4723 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4724 caller. Don't set cfun->has_simduid_loops and/or
4725 cfun->has_force_vect_loops here.
4726 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4727 caller.
4728 (expand_call_inline, tree_function_versioning): Adjust copy_body
4729 callers.
4730
ae6d0907
MJ
47312013-09-13 Martin Jambor <mjambor@suse.cz>
4732
4733 PR bootstrap/58388
4734 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4735 the assert if the edge was a speculative one.
4736
2fd5894f
RB
47372013-09-13 Richard Biener <rguenther@suse.de>
4738
4739 * tree-data-ref.h (known_dependences_p): Move here ...
4740 * tree-loop-distribution.c (known_dependences_p): ... from here.
4741 (dump_rdg_component, debug_rdg_component): Remove.
4742 (dump_rdg): Adjust.
4743 (generate_loops_for_partition): Use gimple_uid instead of
4744 relying on matching stmt visit order.
4745 (rdg_build_partitions): Take starting stmt vector.
4746 (ldist_gen): Merge into ...
4747 (distribute_loop): ... this function. Do not compute starting
4748 vertices vector.
4749 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
4750
86eb4bd7
KT
47512013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4752
4753 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
4754 Set type attribute correctly. Set predicable_short_it attribute.
4755 (cmpsi_shiftsi): Remove %? from output template.
4756
83a95546
RB
47572013-09-13 Richard Biener <rguenther@suse.de>
4758
4759 * tree-loop-distribution.c (struct rdg_component,
4760 rdg_defs_used_in_other_loops_p, free_rdg_components,
4761 rdg_build_components): Remove.
4762 (stmts_from_loop): Do not record virtual PHIs.
4763 (generate_loops_for_partition): Skip virtual PHIs.
4764 (build_rdg_partition_for_component): Rename to ...
4765 (build_rdg_partition_for_vertex): ... this and adjust.
4766 (rdg_build_partitions): Take a vector of starting vertices
4767 instead of components. Remove unnecessary leftover handling.
4768 (ldist_gen): Do not build components or record other stores.
4769 (distribute_loop): Do not distribute loops containing stmts
4770 with side-effects.
4771
c0ecf7f1
CB
47722013-09-13 Christian Bruel <christian.bruel@st.com>
4773
4774 PR target/58314
4775 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
4776
34f56efc
KT
47772013-09-13 Kai Tietz <ktietz@redhat.com>
4778
740f2bf3
KT
4779 * config.gcc: Separate cases for mingw and cygwin targets,
4780 and add 64-bit cygwin target case.
4781
34f56efc
KT
4782 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
4783 dll-export inline-functions.
4784 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
4785
4786
47872013-09-13 Jeff Law <law@redhat.com>
823b5d6b
JL
4788
4789 PR middle-end/58387
4790 Revert:
4791 2013-09-06 Jeff Law <law@redhat.com>
4792
4793 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4794 edge implied equivalences into successor phis.
4795
aadb52ad
DD
47962013-09-12 DJ Delorie <dj@redhat.com>
4797
4798 * config/rl78/rl78-virt.md: Change from | to \; for asm line
4799 separators.
4800
90ae7019
BM
48012013-09-12 Brooks Moses <bmoses@google.com>
4802
4803 PR driver/42955
4804 * Makefile.in: Do not install driver binaries in $(target)/bin.
4805
17b2557c
DD
48062013-09-12 DJ Delorie <dj@redhat.com>
4807
c5215a9b
DD
4808 * config/rl78/rl78.opt (mrelax): New.
4809 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
4810 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
4811
17b2557c
DD
4812 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
4813 between SP and FP.
4814 (rl78_expand_epilogue): Likewise.
4815
ecafcf05
VM
48162013-09-12 Vladimir Makarov <vmakarov@redhat.com>
4817
4818 PR middle-end/58335
84f6d9fa 4819 * lra-eliminations.c (remove_reg_equal_offset_note): New.
ecafcf05
VM
4820 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
4821 pointer elimination with using remove_reg_equal_offset_note.
4822
8369f38a
DD
48232013-09-12 DJ Delorie <dj@redhat.com>
4824
f6a83b4a
DD
4825 * config/msp430/: New port.
4826 * config.gcc (msp430): Added.
4827 * doc/invoke.texi: Document MSP430 options.
4828 * doc/install.texi: Document msp430-elf
4829 * doc/md.texi: Document msp430-elf
4830 * doc/contrib.texi: Document msp430-elf
4831
8369f38a
DD
4832 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
4833 PSImode.
4834
a854f856
MJ
48352013-09-12 Martin Jambor <mjambor@suse.cz>
4836
4837 PR ipa/58389
4838 * ipa-prop.c (remove_described_reference): Give up if the edge in the
4839 reference descriptor is NULL.
4840 (ipa_edge_removal_hook): If owning a reference descriptor, set its
4841 edge to NULL.
4842
b12ebd96
AM
48432013-09-12 Andrew MacLeod <amacleod@redhat.com>
4844
4845 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
4846 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
4847 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
4848 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
4849 make_temp_ssa_name): move to tree-ssanames.h
4850 * tree-ssa-alias.h: Move prototype.
4851 * tree-ssa.h: Include tree-ssanames.h.
4852 * tree-ssanames.c (FREE_SSANAMES): Move to here.
4853 * tree-ssanames.h: New. Move items from tree-flow*.h
4854 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
4855
e1fb4ad3
RB
48562013-09-12 Richard Biener <rguenther@suse.de>
4857
4858 PR tree-optimization/58404
4859 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
4860 propagate non-invariant addresses into dereferences wrapped
4861 in component references.
4862
154c9188
RB
48632013-09-12 Richard Biener <rguenther@suse.de>
4864
4865 PR tree-optimization/58402
4866 * passes.def: Move pass_late_warn_uninitialized later.
4867
7a300452
AM
48682013-09-12 Andrew MacLeod <amacleod@redhat.com>
4869
4870 * tree-ssa.h: New. Move content from tree-flow.h and
4871 tree-flow-inline.h.
4872 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
4873 Move prototypes belonging to tree-ssa.c.
f77e2d2d 4874 * tree-flow-inline.h (redirect_edge_var_map_def,
7a300452
AM
4875 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
4876 tree-ssa.h.
4877 * gimple.h: Adjust prototypes.
f77e2d2d
JBG
4878 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
4879 to...
4880 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
7a300452
AM
4881 * tree.h: Move prototype to tree-ssa.h.
4882 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
4883 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
4884 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
4885 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
4886 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
4887 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
4888 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
4889 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
f77e2d2d
JBG
4890 graphite-blocking.c, graphite-clast-to-gimple.c,
4891 graphite-dependences.c, graphite-interchange.c,
4892 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
4893 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
4894 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
4895 ipa-reference.c, ipa-split.c, ipa-utils.c,
7a300452
AM
4896 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
4897 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
4898 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
4899 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4900 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
4901 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
4902 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
4903 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
4904 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
4905 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
4906 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
4907 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4908 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
4909 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
4910 value-prof.c, var-tracking.c,
4911 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
4912
97463b2b
RB
49132013-09-12 Richard Biener <rguenther@suse.de>
4914
4915 PR tree-optimization/58396
4916 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
4917 (build_rdg): Take a loop-nest parameter, fix memleaks.
4918 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
4919
80ecd89d
YR
49202013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4921
4922 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
4923 for SLM.
4924
893e85fa
CM
49252013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
4926
4927 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
4928 The FMA instruction names should have a 'v' prefix.
4929
174ec470
RB
49302013-09-12 Richard Biener <rguenther@suse.de>
4931
4932 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
4933 (dot_rdg): Use popen instead of system in optional code.
4934 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
4935 (already_processed_vertex_p): Adjust.
4936 (has_anti_or_output_dependence, predecessor_has_mem_write,
4937 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
4938 rdg_flag_uses): Remove.
4939 (rdg_flag_vertex): Simplify.
4940 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
4941 remove recursion.
4942 (build_rdg_partition_for_component): Process the first vertex
4943 of a component only.
4944 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
4945
38c89763
AM
49462013-09-12 Alan Modra <amodra@gmail.com>
4947
4948 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
4949
3e3d9d17
DD
49502013-09-11 DJ Delorie <dj@redhat.com>
4951 Nick Clifton <nickc@redhat.com>
4952
4953 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
4954 (rl78_stack_based_mem): New.
4955 * config/rl78/constraints.md (Iv08): New.
4956 (Iv16): New.
4957 (Iv24): New.
4958 (Is09): New.
4959 (Is17): New.
4960 (Is25): New.
4961 (ISsi): New.
4962 (IShi): New.
4963 (ISqi): New.
4964 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
4965 (movhi): Likewise.
4966 (movsi): Change from expand to insn-and-split.
4967 (ashrsi3): Clobber AX.
4968 (lshrsi3): New.
4969 (ashlsi3): New.
4970 (cbranchsi4): New.
4971 * config/rl78/rl78.md (CC_REG): Fix.
4972 (addsi3): Allow memory and immediate operands.
4973 (addsi3_internal): Split into...
4974 (addsi3_internal_virt): ...new, and ...
4975 (addsi3_internal_real): ...new.
4976 (subsi): New.
4977 (subsi3_internal_virt): New.
4978 (subsi3_internal_real): New.
4979 (mulsi3): Add memory operand.
4980 (mulsi3_rl78): Likewise.
4981 (mulsi3_g13): Likewise.
4982 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
4983 (cbranchqi4_real): Add more constraint options.
4984 (cbranchhi4_real): Expand pattern.
4985 (cbranchhi4_real_signed): New.
4986 (cbranchhi4_real_inverted): New.
4987 (cbranchsi4_real_lt): New.
4988 (cbranchsi4_real_ge): New.
4989 (cbranchsi4_real_signed): New.
4990 (cbranchsi4_real): New.
4991 (peephole2): New.
893e85fa
CM
4992 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
4993 constant shifts.
3e3d9d17
DD
4994 (lshrsi3_virt): Likewise.
4995 (ashlsi3_virt): Likewise.
4996 (cbranchqi4_virt_signed): New.
4997 (cbranchhi4_virt_signed): New.
4998 (cbranchsi4_virt): New.
4999 * config/rl78/rl78.c: Whitespace fixes throughout.
5000 (move_elim_pass): New.
5001 (pass_data_rl78_move_elim): New.
5002 (pass_rl78_move_elim): New.
5003 (make_pass_rl78_move_elim): New.
5004 (rl78_devirt_info): Run devirt earlier.
5005 (rl78_move_elim_info): New.
5006 (rl78_asm_file_start): Register it.
5007 (rl78_split_movsi): New.
5008 (rl78_as_legitimate_address): Allow virtual base registers when
5009 appropriate.
5010 (rl78_addr_space_convert): Remove spurious debug stuff.
5011 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
5012 (rl78_print_operand): More cases for not printing '#'.
5013 (rl78_expand_compare): Remove most of the logic.
5014 (content_memory): New.
5015 (clear_content_memory): New.
5016 (get_content_index): New.
5017 (get_content_name): New.
5018 (display_content_memory): New.
5019 (update_content): New.
5020 (record_content): New.
5021 (already_contains): New.
5022 (insn_ok_now): Re-recog insns with virtual registers.
5023 (add_postponed_content_update): New.
5024 (process_postponed_content_update): New.
5025 (gen_and_emit_move): New.
893e85fa
CM
5026 (transcode_memory_rtx): Record new location content.
5027 Use gen_and_emit_move.
3e3d9d17
DD
5028 (force_into_acc): New.
5029 (move_to_acc): Use gen_and_emit_move.
5030 (move_from_acc): Likewise.
5031 (move_acc_to_reg): Likewise.
5032 (move_to_x): Likewise.
5033 (move_to_hl): Likewise.
5034 (move_to_de): Likewise.
5035 (rl78_alloc_physical_registers_op1): Record location content.
5036 (has_constraint): New.
5037 (rl78_alloc_physical_registers_op2): Record location content.
5038 Optimize use of HL.
5039 (rl78_alloc_physical_registers_ro1): Likewise.
5040 (rl78_alloc_physical_registers_cmp): Likewise.
5041 (rl78_alloc_physical_registers_umul): Likewise.
5042 (rl78_alloc_address_registers_macax): New.
5043 (rl78_alloc_physical_registers): Initialize and set location
5044 content memory as needed.
5045 (rl78_reorg): Make sure split2 is called.
5046 (rl78_rtx_costs): New.
5047
088845a5
RS
50482013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5049
5050 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5051 for (not (neg ...)) and (neg (not ...)) cases.
5052
5c2961cf
RB
50532013-09-11 Richard Biener <rguenther@suse.de>
5054
5055 PR middle-end/58377
5056 * passes.def: Split critical edges before late uninit warning passes.
5057 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5058
427e6a14
JJ
50592013-09-11 Jakub Jelinek <jakub@redhat.com>
5060
5061 PR tree-optimization/58385
5062 * fold-const.c (build_range_check): If both low and high are NULL,
5063 use omit_one_operand_loc to preserve exp side-effects.
5064
df45c0cc
KT
50652013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5066
5067 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5068
80ab0b19
RB
50692013-09-11 Richard Biener <rguenther@suse.de>
5070
5071 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5072 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5073 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5074 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5075 stmts_from_loop, known_dependences_p, build_empty_rdg,
5076 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5077 * tree-loop-distribution.c: ... here.
5078 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5079 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5080 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5081 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5082 * tree-loop-distribution.c: ... here.
5083 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5084 (struct partition_s): Add loops member.
5085 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5086 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5087 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5088
85a77221
AI
50892013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5090 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5091 Sergey Lega <sergey.s.lega@intel.com>
5092 Anna Tikhonova <anna.tikhonova@intel.com>
5093 Ilya Tocar <ilya.tocar@intel.com>
5094 Andrey Turetskiy <andrey.turetskiy@intel.com>
5095 Ilya Verbin <ilya.verbin@intel.com>
5096 Kirill Yukhin <kirill.yukhin@intel.com>
5097 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5098
5099 * config/i386/constraints.md (k): New.
5100 (Yk): Ditto.
5101 * config/i386/i386.c (const regclass_map): Add new mask registers.
5102 (dbx_register_map): Ditto.
5103 (dbx64_register_map): Ditto.
5104 (svr4_dbx_register_map): Ditto.
5105 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5106 disabled.
5107 (ix86_preferred_reload_class): Disable constants for mask registers.
5108 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5109 (ix86_hard_regno_mode_ok): Support new mask registers.
5110 (x86_order_regs_for_local_alloc): Ditto.
5111 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5112 (FIXED_REGISTERS): Add new mask registers.
5113 (CALL_USED_REGISTERS): Ditto.
5114 (REG_ALLOC_ORDER): Ditto.
5115 (VALID_MASK_REG_MODE): New.
5116 (FIRST_MASK_REG): Ditto.
5117 (LAST_MASK_REG): Ditto.
5118 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5119 (MAYBE_MASK_CLASS_P): New.
5120 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5121 (REG_CLASS_CONTENTS): Ditto.
5122 (MASK_REGNO_P): New.
5123 (ANY_MASK_REG_P): Ditto.
5124 (HI_REGISTER_NAMES): Add new mask registers.
893e85fa
CM
5125 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5126 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5127 mask registers.
85a77221
AI
5128 (attribute "type"): Add mskmov, msklog.
5129 (attribute "length_immediate"): Support them.
5130 (attribute "memory"): Ditto.
5131 (attribute "prefix_0f"): Ditto.
5132 (*movhi_internal): Support new mask registers.
5133 (*movqi_internal): Ditto.
5134 (define_split): Split out clobber pattern is a logic
5135 insn on mask registers.
5136 (*k<logic><mode>): New.
5137 (*andhi_1): Extend to support mask regs.
5138 (*andqi_1): Extend to support mask regs.
5139 (kandn<mode>): New.
5140 (define_split): Split and-not to and and not if operands
5141 are not mask regs.
5142 (*<code><mode>_1): Separate HI mode to new pattern...
5143 (*<code>hi_1): This.
5144 (*<code>qi_1): Extend to support mask regs.
5145 (kxnor<mode>): New.
5146 (kortestzhi): Ditto.
5147 (kortestchi): Ditto.
5148 (kunpckhi): Ditto.
5149 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5150 (*one_cmplhi2_1): ...Here, now with mask registers support.
5151 (*one_cmplqi2_1): Support new mask registers.
893e85fa
CM
5152 (HI/QImode arithmetics splitter): Don't split if mask registers
5153 are used.
85a77221
AI
5154 (HI/QImode not splitter): Ditto.
5155 * config/i386/predicated.md (mask_reg_operand): New.
5156 (general_reg_operand): Ditto.
5157
7e570821
AI
51582013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5159
5160 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5161 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5162
9e1376e9
JL
51632013-09-10 Jeff Law <law@redhat.com>
5164
5165 PR tree-optimization/58380
5166 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5167 of threading through a buried loop header.
5168
5169 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5170 return value for single successor case.
5171
5bbcb888
JH
51722013-09-10 Jan Hubicka <jh@suse.cz>
5173
5174 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5175
d742ff4b
RE
51762013-09-10 Richard Earnshaw <rearnsha@arm.com>
5177
5178 PR target/58361
5179 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5180 support conditional execution.
5181 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5182
b0681c9e
VM
51832013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5184
5185 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5186 constraint pass.
f77e2d2d
JBG
5187 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5188 Check destination too to check move insn.
b0681c9e 5189 (undo_optional_reloads): Add check that the original peudo did not
f77e2d2d
JBG
5190 changed its allocation and the optional reload was inherited on last
5191 inheritance pass. Break loop after deciding to keep optional reload.
b0681c9e
VM
5192 (lra_undo_inheritance): Add check that inherited pseudo still in
5193 memory.
5194
c9b1c957
JG
51952013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5196
5197 * config/aarch64/aarch64.md (generic_sched): New.
5198 * config/aarch64/aarch64-generic.md (load): Make conditional
5199 on generic_sched attribute.
5200 (nonload): Likewise.
5201
82d618d3
JH
52022013-09-10 Jan Hubicka <jh@suse.cz>
5203
5204 * lto-cgraph.c: Include ipa-utils.h.
5205 (compute_ltrans_boundary): Also add possible targets into the boundary.
5206
668b6894
JH
52072013-09-10 Jan Hubicka <jh@suse.cz>
5208
5209 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5210 VAR_DECL of vtable rather than full expression.
5211
4f90d3e0
JH
52122013-09-10 Jan Hubicka <jh@suse.cz>
5213 Paolo Carlini <paolo.carlini@oracle.com>
5214
5215 * cgraphunit.c (analyze_functions): Save input_location, set it
5216 to UNKNOWN_LOCATION and restore it at the end.
5217
39e87baf
MJ
52182013-09-10 Martin Jambor <mjambor@suse.cz>
5219
5220 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5221 represented by a thunk.
5222
581aedec
JL
52232013-09-10 Jeff Law <law@redhat.com>
5224
5225 PR tree-optimization/58343
5226 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5227 jump threading requests through loop headers buried in the middle
5228 of a jump threading path.
5229
5230 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5231 in return value/type.
5232
84cf4ab6
JJ
52332013-09-10 Jakub Jelinek <jakub@redhat.com>
5234
5235 PR rtl-optimization/58365
5236 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5237 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5238 it differs.
5239
447dd906
RB
52402013-09-10 Richard Biener <rguenther@suse.de>
5241
5242 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5243 * tree-data-ref.c (create_rdg_vertices): Collect all data
5244 references, signal failure to the caller, use data-ref API.
893e85fa
CM
5245 (build_rdg): Compute data references only once. Maintain lifetime
5246 of data references and data dependences from within RDG.
447dd906
RB
5247 (free_rdg): Free dependence relations.
5248 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5249 inventing extra dependences.
5250 (distribute_loop): Update for RDG API changes.
5251
f14a7324
KT
52522013-09-10 Kai Tietz <ktietz@redhat.com>
5253
5254 * doc/invoke.texi (fms-extensions): Document changed
5255 behavior for ms-abi targets.
5256 * config/i386/i386.c (ix86_option_override_internal):
5257 Set default value of option -fms-extension for ms-abi targets.
5258
ff131f26
MZ
52592013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5260
5261 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5262
33115a2c
AM
52632013-09-10 Alan Modra <amodra@gmail.com>
5264
5265 PR target/58330
5266 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5267
b6872a13
AM
52682013-09-10 Alan Modra <amodra@gmail.com>
5269
5270 * config/rs6000/predicates.md (add_cint_operand): New.
5271 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5272 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5273 using add_cint_operand.
5274 (largetoc_high_plus_aix): Likewise.
5275
28fd0ba2
JJ
52762013-09-09 Jakub Jelinek <jakub@redhat.com>
5277
5278 PR tree-optimization/58364
5279 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5280 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5281 the current range can't be an unconditional true or false.
5282
552d99fe
JG
52832013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5284
5285 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5286
f43245d1
UB
52872013-09-09 Uros Bizjak <ubizjak@gmail.com>
5288
5289 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5290
c9b0866a
PC
52912013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5292
5293 PR c++/43452
5294 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5295
78d8b9f0
IB
52962013-09-09 Ian Bolton <ian.bolton@arm.com>
5297
5298 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5299 NO_REGS for immediate that can't be moved directly into FP_REGS.
5300
274b2532
KT
53012013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5302
5303 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5304 comparison with negated operand.
c9b0866a
PC
5305 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5306 RTL form.
274b2532 5307
467a3558
RB
53082013-09-09 Richard Biener <rguenther@suse.de>
5309
5310 PR middle-end/58326
5311 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5312 of a subloop record all its block as affecting loop-closed SSA form.
5313
088c5368
RS
53142013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5315
5316 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5317 of an rtx/bitpos pair.
5318 (store_fixed_bit_field): Update accordingly.
5319
69db2d57
RS
53202013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5321
5322 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5323 GEN_INT.
5324 * builtins.c (expand_errno_check): Likewise.
5325 * dwarf2cfi.c (init_return_column_size): Likewise.
5326 * except.c (sjlj_mark_call_sites): Likewise.
5327 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5328 * lra-constraints.c (emit_inc): Likewise.
5329 * ree.c (combine_set_extension): Likewise.
5330 * regmove.c (fixup_match_2): Likewise.
5331 * reload1.c (inc_for_reload): Likewise.
5332
69a59f0f
RS
53332013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5334
5335 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5336 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5337 Use gen_int_mode with the mode of the associated simplify_* call.
5338 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5339 * expmed.c (expand_shift_1): Likewise.
5340 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5341 * loop-iv.c (iv_number_of_iterations): Likewise.
5342 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5343 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5344
2f1cd2eb
RS
53452013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5346
5347 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5348 of the associated expand_* call.
5349 (asan_emit_stack_protection): Likewise.
5350 * builtins.c (round_trampoline_addr): Likewise.
5351 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5352 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5353 (emit_store_flag): Likewise.
5354 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5355 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5356 Likewise.
5357 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5358 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5359 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5360 * modulo-sched.c (generate_prolog_epilog): Likewise.
5361 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5362 (expand_ctz, expand_ffs, expand_unop): Likewise.
5363
4789c0ce
RS
53642013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5365
5366 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5367 of the associated gen_rtx_* call.
5368 * caller-save.c (init_caller_save): Likewise.
5369 * combine.c (find_split_point, make_extraction): Likewise.
5370 (make_compound_operation): Likewise.
5371 * dwarf2out.c (mem_loc_descriptor): Likewise.
5372 * explow.c (plus_constant, probe_stack_range): Likewise.
5373 * expmed.c (expand_mult_const): Likewise.
5374 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5375 * reload1.c (init_reload): Likewise.
5376 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5377 * var-tracking.c (adjust_mems): Likewise.
5378 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5379 rather than gen_rtx_fmt_ee.
5380
a811f7cb
JH
53812013-09-09 Jan Hubicka <jh@suse.cz>
5382
5383 PR middle-end/58294
5384 * value-prof.c (gimple_ic): Copy also abnormal edges.
5385
dcad1dd3
RS
53862013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5387
5388 * asan.c (asan_shadow_cst): Use gen_int_mode.
5389
daf5c770
JH
53902013-09-08 Jan Hubicka <jh@suse.cz>
5391
5392 * ipa-profile.c: Add toplevel comment.
5393 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5394 (contains_hot_call_p): New function.
5395 (ipa_propagate_frequency): Set frequencies based on counts when
5396 profile is read.
5397 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5398 profile; do not tamper with profile after inlining if it is read.
5399
a12cd2db
JH
54002013-09-08 Jan Hubicka <jh@suse.cz>
5401
5402 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5403 speculative edges.
5404
856a1588
JH
54052013-09-08 Jan Hubicka <jh@suse.cz>
5406
5407 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5408 summary generation.
5409
4f4b0b73
JL
54102013-09-08 Jeff Law <law@redhat.com>
5411
5412 PR bootstrap/58340
5413 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5414 of 'found'.
5415
cc4029ee
AK
54162013-09-08 Andi Kleen <ak@linux.intel.com>
5417
5418 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5419
7b395ddd
JH
54202013-09-08 Jan Hubicka <jh@suse.cz>
5421
5422 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5423
c34c46dd
RS
54242013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5425
5426 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5427 for non-debug insns.
5428 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5429 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5430 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5431 accordingly.
5432
3462aa02
JH
54332013-09-08 Jan Hubicka <jh@suse.cz>
5434
5435 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5436 targets and devirtualize to BUILT_IN_UNREACHABLE.
5437 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5438 * ipa.c (walk_polymorphic_call_targets): New function.
5439 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5440 functions; use the new timevar.
5441 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5442 was removed from the program.
5443 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5444 not consider it in the walk when its vtable is dead.
5445 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5446 record_binfo.
5447 (devirt_variable_node_removal_hook): New function.
5448 (possible_polymorphic_call_targets): Also register
5449 devirt_variable_node_removal_hook.
5450 (ipa_devirt): Do not do non-speculative devirtualization.
5451 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5452
26e5b0fd
JH
54532013-09-08 Jan Hubicka <jh@suse.cz>
5454
5455 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5456 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5457 varpool_remove_variable_insertion_hook): Declare.
5458 * varpool.c (varpool_node_hook_list): New structure.
5459 (first_varpool_node_removal_hook,
5460 first_varpool_variable_insertion_hook): New variables.
5461 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5462 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5463 varpool_remove_variable_insertion_hook,
5464 varpool_call_variable_insertion_hooks): New functions.
5465 (varpool_remove_node): Use it.
5466
97f1f314 54672013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
6df2ca09 5468
97f1f314
PC
5469 PR c++/54941
5470 * diagnostic.c (diagnostic_build_prefix): When s.file is
5471 "<built-in>" don't output line and column numbers.
6df2ca09 5472
bcb650cb
JH
54732013-09-06 Jan Hubicka <jh@suse.cz>
5474
5475 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5476 * lto-streamer-out.c: Stream thunks, too.
5477 * lto-streamer-in.c (input_function): Pop cfun here
5478 (lto_read_body): Instead of here.
5479
87f3fea8
CT
54802013-09-06 Caroline Tice <cmtice@google.com>
5481
5482 * doc/install.texi: Add documentation for the --enable-vtable-verify
5483 and the --disable-libvtv configure options.
5484
5e8d9b45
JL
54852013-09-06 Jeff Law <law@redhat.com>
5486
5487 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5488 edge implied equivalences into successor phis.
5489
3a001aff
JR
54902013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5491
5492 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5493
54942013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
c42091c1
CZ
5495
5496 * resource.c (mark_target_live_regs): Compute resources taking
5497 into account if a call is predicated or not.
5498
2c641f8b
EB
54992013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5500
5501 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5502 the compiler in the function names.
5503
a81b0a3d
JH
55042013-09-06 Jan Hubicka <jh@suse.cz>
5505
5506 PR middle-end/58094
5507 * ipa-inline.c (has_caller_p): New function.
5508 (want_inline_function_to_all_callers_p): Use it.
5509 (sum_callers, inline_to_all_callers): Break out from ...
5510 (ipa_inline): ... here.
5511
68e4ca7e
JH
55122013-09-06 Jan Hubicka <jh@suse.cz>
5513
2c641f8b
EB
5514 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5515 only when AVX is enabled.
68e4ca7e 5516
3c73e2fe
JG
55172013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5518
5519 * config/aarch64/aarch64.md
5520 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5521 is fpsimd_<load/store>2.
5522 (load_pair<mode>): Likewise.
5523 (store_pair<mode>): Likewise.
5524
8c48eecd
JG
55252013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5526
5527 * config/arm/types.md (type): Add "mrs" type.
5528 * config/aarch64/aarch64.md
5529 (aarch64_load_tp_hard): Make type "mrs".
5530 * config/arm/arm.md
5531 (load_tp_hard): Make type "mrs".
5532 * config/arm/cortex-a15.md: Update with new attributes.
5533 * config/arm/cortex-a5.md: Update with new attributes.
5534 * config/arm/cortex-a53.md: Update with new attributes.
5535 * config/arm/cortex-a7.md: Update with new attributes.
5536 * config/arm/cortex-a8.md: Update with new attributes.
5537 * config/arm/cortex-a9.md: Update with new attributes.
5538 * config/arm/cortex-m4.md: Update with new attributes.
5539 * config/arm/cortex-r4.md: Update with new attributes.
5540 * config/arm/fa526.md: Update with new attributes.
5541 * config/arm/fa606te.md: Update with new attributes.
5542 * config/arm/fa626te.md: Update with new attributes.
5543 * config/arm/fa726te.md: Update with new attributes.
5544
9ea187a5
JG
55452013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5546
5547 * config/aarch64/aarch64.md
5548 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5549 (*movtf_aarch64): Likewise.
5550 * config/arm/arm.md
5551 (thumb1_movdi_insn): Use "multiple" for type where more than one
5552 instruction is used for a move.
5553 (*arm32_movhf): Likewise.
5554 (*thumb_movdf_insn): Likewise.
5555
292b89b3
JG
55562013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5557
5558 * config/arm/types.md (type): Rename fcpys to fmov.
5559 * config/arm/vfp.md
5560 (*arm_movsi_vfp): Rename type fcpys as fmov.
5561 (*thumb2_movsi_vfp): Likewise
5562 (*movhf_vfp_neon): Likewise
5563 (*movhf_vfp): Likewise
5564 (*movsf_vfp): Likewise
5565 (*thumb2_movsf_vfp): Likewise
5566 (*movsfcc_vfp): Likewise
5567 (*thumb2_movsfcc_vfp): Likewise
5568 * config/aarch64/aarch64-simd.md
5569 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5570 * config/aarch64/aarch64.md
5571 (*movsi_aarch64): Replace type mov_reg with fmovs.
5572 (*movdi_aarch64): Likewise
5573 (*movsf_aarch64): Likewise
5574 (*movdf_aarch64): Likewise
5575 * config/arm/arm.c
5576 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5577 * config/arm/iwmmxt.md
5578 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5579 * config/arm/arm1020e.md: Update with new attributes.
5580 * config/arm/cortex-a15-neon.md: Update with new attributes.
5581 * config/arm/cortex-a5.md: Update with new attributes.
5582 * config/arm/cortex-a53.md: Update with new attributes.
5583 * config/arm/cortex-a7.md: Update with new attributes.
5584 * config/arm/cortex-a8-neon.md: Update with new attributes.
5585 * config/arm/cortex-a9.md: Update with new attributes.
5586 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5587 * config/arm/cortex-r4f.md: Update with new attributes.
5588 * config/arm/marvell-pj4.md: Update with new attributes.
5589 * config/arm/vfp11.md: Update with new attributes.
5590
6a24a83e
JG
55912013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5592
5593 * config/aarch64/aarch64.md
5594 (*madd<mode>): Fix type attribute.
5595 (*maddsi_uxtw): Likewise.
5596 (*msub<mode>): Likewise.
5597 (*msubsi_uxtw): Likewise.
5598 (<su_optab>maddsidi4): Likewise.
5599 (<su_optab>msubsidi4): Likewise.
5600
b86923f0
JG
56012013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5602
5603 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5604 * config/arm/arm.md (core_cycles): Remove fdiv.
5605 * config/arm/vfp.md:
5606 (*sqrtsf2_vfp): Update for attribute changes.
5607 (*sqrtdf2_vfp): Likewise.
5608 * config/aarch64/aarch64.md:
5609 (sqrt<mode>2): Update for attribute changes.
5610 * config/arm/arm1020e.md: Update with new attributes.
5611 * config/arm/cortex-a15-neon.md: Update with new attributes.
5612 * config/arm/cortex-a5.md: Update with new attributes.
5613 * config/arm/cortex-a53.md: Update with new attributes.
5614 * config/arm/cortex-a7.md: Update with new attributes.
5615 * config/arm/cortex-a8-neon.md: Update with new attributes.
5616 * config/arm/cortex-a9.md: Update with new attributes.
5617 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5618 * config/arm/cortex-r4f.md: Update with new attributes.
5619 * config/arm/marvell-pj4.md: Update with new attributes.
5620 * config/arm/vfp11.md: Update with new attributes.
5621
7b49c9e1
JG
56222013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5623
5624 * config/arm/types.md
5625 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5626 * config/aarch64/aarch64.md
5627 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5628 new attributes.
5629 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5630 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5631 (float<GPI:mode><GPF:mode>2): Likewise.
5632 * config/arm/vfp.md
5633 (*truncsisf2_vfp): Update with new attributes.
5634 (*truncsidf2_vfp): Likewise.
5635 (fixuns_truncsfsi2): Likewise.
5636 (fixuns_truncdfsi2): Likewise.
5637 (*floatsisf2_vfp): Likewise.
5638 (*floatsidf2_vfp): Likewise.
5639 (floatunssisf2): Likewise.
5640 (floatunssidf2): Likewise.
5641 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5642 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5643 * config/arm/arm1020e.md: Update with new attributes.
5644 * config/arm/cortex-a15-neon.md: Update with new attributes.
5645 * config/arm/cortex-a5.md: Update with new attributes.
5646 * config/arm/cortex-a53.md: Update with new attributes.
5647 * config/arm/cortex-a7.md: Update with new attributes.
5648 * config/arm/cortex-a8-neon.md: Update with new attributes.
5649 * config/arm/cortex-a9.md: Update with new attributes.
5650 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5651 * config/arm/cortex-r4f.md: Update with new attributes.
5652 * config/arm/marvell-pj4.md: Update with new attributes.
5653 * config/arm/vfp11.md: Update with new attributes.
5654
e7df8af8
JG
56552013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5656
5657 * config/aarch64/arm_neon.h
5658 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5659 (vqtbx<1,2,3,4><q>_s8): Likewise.
5660
594726e4
JG
56612013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5662
f43245d1
UB
5663 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5664 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
594726e4
JG
5665 (add<mode>3): Add type attribute.
5666 (add<mode>3): Likewise.
5667 (usadd<mode>3): Likewise.
5668 (ssadd<mode>3): Likewise.
5669 (sub<mode>3): Likewise.
5670 (sub<mode>3): Likewise.
5671 (ussub<mode>3): Likewise.
5672 (sssub<mode>3): Likewise.
5673 (ssmulsa3): Likewise.
5674 (usmulusa3): Likewise.
5675 (arm_usatsihi): Likewise.
5676 * config/arm/vfp.md
5677 (*movdi_vfp): Add types for all instructions.
5678 (*movdi_vfp_cortexa8): Likewise.
5679 (*movhf_vfp_neon): Likewise.
5680 (*movhf_vfp): Likewise.
5681 (*movdf_vfp): Likewise.
5682 (*thumb2_movdf_vfp): Likewise.
5683 (*thumb2_movdfcc_vfp): Likewise.
5684 * config/arm/arm.md: Add type attribute to all insn patterns.
5685 (*thumb1_adddi3): Add type attribute.
5686 (*arm_adddi3): Likewise.
5687 (*adddi_sesidi_di): Likewise.
5688 (*adddi_zesidi_di): Likewise.
5689 (*thumb1_addsi3): Likewise.
5690 (addsi3_compare0): Likewise.
5691 (*addsi3_compare0_scratch): Likewise.
5692 (*compare_negsi_si): Likewise.
5693 (cmpsi2_addneg): Likewise.
5694 (*addsi3_carryin_<optab>): Likewise.
5695 (*addsi3_carryin_alt2_<optab>): Likewise.
5696 (*addsi3_carryin_clobercc_<optab>): Likewise.
5697 (*subsi3_carryin): Likewise.
5698 (*subsi3_carryin_const): Likewise.
5699 (*subsi3_carryin_compare): Likewise.
5700 (*subsi3_carryin_compare_const): Likewise.
5701 (*arm_subdi3): Likewise.
5702 (*thumb_subdi3): Likewise.
5703 (*subdi_di_zesidi): Likewise.
5704 (*subdi_di_sesidi): Likewise.
5705 (*subdi_zesidi_di): Likewise.
5706 (*subdi_sesidi_di): Likewise.
5707 (*subdi_zesidi_ze): Likewise.
5708 (thumb1_subsi3_insn): Likewise.
5709 (*arm_subsi3_insn): Likewise.
5710 (*anddi3_insn): Likewise.
5711 (*anddi_zesidi_di): Likewise.
5712 (*anddi_sesdi_di): Likewise.
5713 (*ne_zeroextracts): Likewise.
5714 (*ne_zeroextracts): Likewise.
5715 (*ite_ne_zeroextr): Likewise.
5716 (*ite_ne_zeroextr): Likewise.
5717 (*anddi_notdi_di): Likewise.
5718 (*anddi_notzesidi): Likewise.
5719 (*anddi_notsesidi): Likewise.
5720 (andsi_notsi_si): Likewise.
5721 (thumb1_bicsi3): Likewise.
5722 (*iordi3_insn): Likewise.
5723 (*iordi_zesidi_di): Likewise.
5724 (*iordi_sesidi_di): Likewise.
5725 (*thumb1_iorsi3_insn): Likewise.
5726 (*xordi3_insn): Likewise.
5727 (*xordi_zesidi_di): Likewise.
5728 (*xordi_sesidi_di): Likewise.
5729 (*arm_xorsi3): Likewise.
5730 (*andsi_iorsi3_no): Likewise.
5731 (*smax_0): Likewise.
5732 (*smax_m1): Likewise.
5733 (*arm_smax_insn): Likewise.
5734 (*smin_0): Likewise.
5735 (*arm_smin_insn): Likewise.
5736 (*arm_umaxsi3): Likewise.
5737 (*arm_uminsi3): Likewise.
5738 (*minmax_arithsi): Likewise.
5739 (*minmax_arithsi_): Likewise.
5740 (*satsi_<SAT:code>): Likewise.
5741 (arm_ashldi3_1bit): Likewise.
5742 (arm_ashrdi3_1bit): Likewise.
5743 (arm_lshrdi3_1bit): Likewise.
5744 (*arm_negdi2): Likewise.
5745 (*thumb1_negdi2): Likewise.
5746 (*arm_negsi2): Likewise.
5747 (*thumb1_negsi2): Likewise.
5748 (*negdi_extendsid): Likewise.
5749 (*negdi_zero_extend): Likewise.
5750 (*arm_abssi2): Likewise.
5751 (*thumb1_abssi2): Likewise.
5752 (*arm_neg_abssi2): Likewise.
5753 (*thumb1_neg_abss): Likewise.
5754 (one_cmpldi2): Likewise.
5755 (extend<mode>di2): Likewise.
5756 (*compareqi_eq0): Likewise.
5757 (*arm_extendhisi2addsi): Likewise.
5758 (*arm_movdi): Likewise.
5759 (*thumb1_movdi_insn): Likewise.
5760 (*arm_movt): Likewise.
5761 (*thumb1_movsi_insn): Likewise.
5762 (pic_add_dot_plus_four): Likewise.
5763 (pic_add_dot_plus_eight): Likewise.
5764 (tls_load_dot_plus_eight): Likewise.
5765 (*thumb1_movhi_insn): Likewise.
5766 (*thumb1_movsf_insn): Likewise.
5767 (*movdf_soft_insn): Likewise.
5768 (*thumb_movdf_insn): Likewise.
5769 (cbranchsi4_insn): Likewise.
5770 (cbranchsi4_scratch): Likewise.
5771 (*negated_cbranchsi4): Likewise.
5772 (*tbit_cbranch): Likewise.
5773 (*tlobits_cbranch): Likewise.
5774 (*tstsi3_cbranch): Likewise.
5775 (*cbranchne_decr1): Likewise.
5776 (*addsi3_cbranch): Likewise.
5777 (*addsi3_cbranch_scratch): Likewise.
5778 (*arm_cmpdi_insn): Likewise.
5779 (*arm_cmpdi_unsig): Likewise.
5780 (*arm_cmpdi_zero): Likewise.
5781 (*thumb_cmpdi_zero): Likewise.
5782 (*deleted_compare): Likewise.
5783 (*mov_scc): Likewise.
5784 (*mov_negscc): Likewise.
5785 (*mov_notscc): Likewise.
5786 (*cstoresi_eq0_thumb1_insn): Likewise.
5787 (cstoresi_nltu_thumb1): Likewise.
5788 (cstoresi_ltu_thu): Likewise.
5789 (thumb1_addsi3_addgeu): Likewise.
5790 (*arm_jump): Likewise.
5791 (*thumb_jump): Likewise.
5792 (*check_arch2): Likewise.
5793 (arm_casesi_internal): Likewise.
5794 (thumb1_casesi_dispatch): Likewise.
5795 (*arm_indirect_jump): Likewise.
5796 (*thumb1_indirect_jump): Likewise.
5797 (nop): Likewise.
5798 (*and_scc): Likewise.
5799 (*ior_scc): Likewise.
5800 (*compare_scc): Likewise.
5801 (*cond_move): Likewise.
5802 (*cond_arith): Likewise.
5803 (*cond_sub): Likewise.
5804 (*cmp_ite0): Likewise.
5805 (*cmp_ite1): Likewise.
5806 (*cmp_and): Likewise.
5807 (*cmp_ior): Likewise.
5808 (*ior_scc_scc): Likewise.
5809 (*ior_scc_scc_cmp): Likewise.
5810 (*and_scc_scc): Likewise.
5811 (*and_scc_scc_cmp): Likewise.
5812 (*and_scc_scc_nod): Likewise.
5813 (*negscc): Likewise.
5814 (movcond_addsi): Likewise.
5815 (movcond): Likewise.
5816 (*ifcompare_plus_move): Likewise.
5817 (*if_plus_move): Likewise.
5818 (*ifcompare_move_plus): Likewise.
5819 (*if_move_plus): Likewise.
5820 (*ifcompare_arith_arith): Likewise.
5821 (*if_arith_arith): Likewise.
5822 (*ifcompare_arith_move): Likewise.
5823 (*if_arith_move): Likewise.
5824 (*ifcompare_move_arith): Likewise.
5825 (*if_move_arith): Likewise.
5826 (*ifcompare_move_not): Likewise.
5827 (*if_move_not): Likewise.
5828 (*ifcompare_not_move): Likewise.
5829 (*if_not_move): Likewise.
5830 (*ifcompare_shift_move): Likewise.
5831 (*if_shift_move): Likewise.
5832 (*ifcompare_move_shift): Likewise.
5833 (*if_move_shift): Likewise.
5834 (*ifcompare_shift_shift): Likewise.
5835 (*ifcompare_not_arith): Likewise.
5836 (*ifcompare_arith_not): Likewise.
5837 (*if_arith_not): Likewise.
5838 (*ifcompare_neg_move): Likewise.
5839 (*if_neg_move): Likewise.
5840 (*ifcompare_move_neg): Likewise.
5841 (*if_move_neg): Likewise.
5842 (prologue_thumb1_interwork): Likewise.
5843 (*cond_move_not): Likewise.
5844 (*sign_extract_onebit): Likewise.
5845 (*not_signextract_onebit): Likewise.
5846 (stack_tie): Likewise.
5847 (align_4): Likewise.
5848 (align_8): Likewise.
5849 (consttable_end): Likewise.
5850 (consttable_1): Likewise.
5851 (consttable_2): Likewise.
5852 (consttable_4): Likewise.
5853 (consttable_8): Likewise.
5854 (consttable_16): Likewise.
5855 (*thumb1_tablejump): Likewise.
5856 (prefetch): Likewise.
5857 (force_register_use): Likewise.
5858 (thumb_eh_return): Likewise.
5859 (load_tp_hard): Likewise.
5860 (load_tp_soft): Likewise.
5861 (tlscall): Likewise.
5862 (*arm_movtas_ze): Likewise.
5863 (*arm_rev): Likewise.
5864 (*arm_revsh): Likewise.
5865 (*arm_rev16): Likewise.
5866 * config/arm/thumb2.md
5867 (*thumb2_smaxsi3): Likewise.
5868 (*thumb2_sminsi3): Likewise.
5869 (*thumb32_umaxsi3): Likewise.
5870 (*thumb2_uminsi3): Likewise.
5871 (*thumb2_negdi2): Likewise.
5872 (*thumb2_abssi2): Likewise.
5873 (*thumb2_neg_abss): Likewise.
5874 (*thumb2_movsi_insn): Likewise.
5875 (tls_load_dot_plus_four): Likewise.
5876 (*thumb2_movhi_insn): Likewise.
5877 (*thumb2_mov_scc): Likewise.
5878 (*thumb2_mov_negs): Likewise.
5879 (*thumb2_mov_negs): Likewise.
5880 (*thumb2_mov_nots): Likewise.
5881 (*thumb2_mov_nots): Likewise.
5882 (*thumb2_movsicc_): Likewise.
5883 (*thumb2_movsfcc_soft_insn): Likewise.
5884 (*thumb2_indirect_jump): Likewise.
5885 (*thumb2_and_scc): Likewise.
5886 (*thumb2_ior_scc): Likewise.
5887 (*thumb2_ior_scc_strict_it): Likewise.
5888 (*thumb2_cond_move): Likewise.
5889 (*thumb2_cond_arith): Likewise.
5890 (*thumb2_cond_ari): Likewise.
5891 (*thumb2_cond_sub): Likewise.
5892 (*thumb2_negscc): Likewise.
5893 (*thumb2_movcond): Likewise.
5894 (thumb2_casesi_internal): Likewise.
5895 (thumb2_casesi_internal_pic): Likewise.
5896 (*thumb2_alusi3_short): Likewise.
5897 (*thumb2_mov<mode>_shortim): Likewise.
5898 (*thumb2_addsi_short): Likewise.
5899 (*thumb2_subsi_short): Likewise.
5900 (thumb2_addsi3_compare0): Likewise.
5901 (*thumb2_cbz): Likewise.
5902 (*thumb2_cbnz): Likewise.
5903 (*thumb2_one_cmplsi2_short): Likewise.
5904 (*thumb2_negsi2_short): Likewise.
5905 (*orsi_notsi_si): Likewise.
5906 * config/arm/arm1020e.md: Update with new attributes.
5907 * config/arm/arm1026ejs.md: Update with new attributes.
5908 * config/arm/arm1136jfs.md: Update with new attributes.
5909 * config/arm/arm926ejs.md: Update with new attributes.
5910 * config/arm/cortex-a15.md: Update with new attributes.
5911 * config/arm/cortex-a5.md: Update with new attributes.
5912 * config/arm/cortex-a53.md: Update with new attributes.
5913 * config/arm/cortex-a7.md: Update with new attributes.
5914 * config/arm/cortex-a8.md: Update with new attributes.
5915 * config/arm/cortex-a9.md: Update with new attributes.
5916 * config/arm/cortex-m4.md: Update with new attributes.
5917 * config/arm/cortex-r4.md: Update with new attributes.
5918 * config/arm/fa526.md: Update with new attributes.
5919 * config/arm/fa606te.md: Update with new attributes.
5920 * config/arm/fa626te.md: Update with new attributes.
5921 * config/arm/fa726te.md: Update with new attributes.
5922
1c83b673
JG
59232013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5924
5925 * config/aarch64/aarch64-simd.md
5926 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
5927 <vwx> iterator to ensure correct register choice.
5928 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5929 (aarch64_sqdmull_n<mode>): Likewise.
5930 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5931 * config/aarch64/arm_neon.h
5932 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
5933 (vml<as><q>_n_<su>16): Likewise.
5934 (vml<as>l_high_lane<q>_<su>16): Likewise.
5935 (vml<as>l_high_n_<su>16): Likewise.
5936 (vml<as>l_lane<q>_<su>16): Likewise.
5937 (vml<as>l_n_<su>16): Likewise.
5938 (vmul<q>_lane<q>_<su>16): Likewise.
5939 (vmul<q>_n_<su>16): Likewise.
5940 (vmull_lane<q>_<su>16): Likewise.
5941 (vmull_n_<su>16): Likewise.
5942 (vmull_high_lane<q>_<su>16): Likewise.
5943 (vmull_high_n_<su>16): Likewise.
5944 (vqrdmulh<q>_n_s16): Likewise.
5945
bb1ae543
TB
59462013-09-06 Tejas Belagod <tejas.belagod@arm.com>
5947
5948 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
5949 have the correct lane parameter.
5950
f23c0742
RB
59512013-09-06 Richard Biener <rguenther@suse.de>
5952
5953 * cfganal.c (control_dependences::~control_dependences):
5954 Properly free all of the vector.
5955
c1d49770
KY
59562013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
5957
5958 PR target/58269
5959 * config/i386/i386.c (ix86_conditional_register_usage):
5960 Proper initialize extended SSE registers.
5961
6da47f52
JH
59622013-09-06 Jan Hubicka <jh@suse.cz>
5963
5964 PR tree-optimization/58311
5965 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
5966
59672013-09-06 Jan Hubicka <jh@suse.cz>
5968
5969 * Makefile.in (tree-sra.o): Update dependencies.
5970 * tree-sra.c: Include ipa-utils.h
5971 (scan_function): Use recursive_call_p.
5972 (has_caller_p): New function.
5973 (cgraph_for_node_and_aliases): Count also callers of aliases.
5974
fc11f321
JH
59752013-09-06 Jan Hubicka <jh@suse.cz>
5976
5977 PR middle-end/58094
5978 * cgraph.h (symtab_semantically_equivalent_p): Declare.
5979 * tree-tailcall.c: Include ipa-utils.h.
5980 (find_tail_calls): Use it.
5981 * ipa-pure-const.c (check_call): Likewise.
5982 * ipa-utils.c (recursive_call_p): New function.
5983 * ipa-utils.h (recursive_call_p): Dclare.
5984 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
5985 (symtab_semantically_equivalent_p): New function.
5986 * Makefile.in (tree-tailcall.o): Update dependencies.
5987
9a6af450
EB
59882013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5989
5990 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
5991 non-inlinable part.
5992
35f32ad4
RB
59932013-09-06 Richard Biener <rguenther@suse.de>
5994
5995 * lto-streamer.h (lto_global_var_decls): Remove.
5996 * Makefile.in (OBJS): Remove lto-symtab.o.
5997 (lto-symtab.o): Remove.
5998 (GTFILES): Remove lto-symtab.c
5999 * lto-symtab.c: Move to lto/
6000
d12a76f3
AK
60012013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6002
6003 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
6004 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
6005 (UNSPEC_FPINT_RINT): New constant definitions.
6006 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
6007 definition with 2 attributes.
6008 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
6009 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
6010 definitions.
6011
60122013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
6013
f43245d1 6014 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
d12a76f3
AK
6015 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
6016 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
6017 attribute to "z196_alone".
6018 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
6019 "zEC12_simple".
6020
c8e9d8c3
RB
60212013-09-06 Richard Biener <rguenther@suse.de>
6022
6023 * basic-block.h (class control_dependences): New.
6024 * tree-ssa-dce.c (control_dependence_map): Remove.
6025 (cd): New global.
6026 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
6027 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
6028 find_pdom, find_control_dependence, find_all_control_dependences):
6029 Move to cfganal.c.
f43245d1
UB
6030 (mark_control_dependent_edges_necessary,
6031 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
6032 tree_dce_done, perform_tree_ssa_dce): Adjust.
c8e9d8c3
RB
6033 * cfganal.c (set_control_dependence_map_bit,
6034 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6035 find_all_control_dependences): Move from tree-ssa-dce.c and
6036 implement as methods of control_dependences class.
6037 (control_dependences::control_dependences): New.
6038 (control_dependences::~control_dependences): Likewise.
6039 (control_dependences::get_edges_dependent_on): Likewise.
6040 (control_dependences::get_edge): Likewise.
6041
7d2ba471
JH
60422013-09-04 Jan Hubicka <jh@suse.cz>
6043
6044 * tree.c (types_same_for_odr): Drop overactive check.
6045 * ipa-devirt.c (hash_type_name): Likewise.
6046
e18412fc
JH
60472013-09-04 Jan Hubicka <jh@suse.cz>
6048
6049 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6050 (analyze_functions): ... here.
6051
d352b245
JH
60522013-09-04 Jan Hubicka <jh@suse.cz>
6053
6054 PR middle-end/58201
6055 * cgraphunit.c (analyze_functions): Clear AUX fields
6056 after processing; initialize assembler name has.
6057
770da076
JL
60582013-09-05 Jeff Law <law@redhat.com>
6059
6060 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6061 from thread_around_empty_block. Record threading path into PATH.
6062 Recurse if threading through the initial block is successful.
6063 (thread_across_edge): Corresponding changes to slightly simplify.
6064
fe6f68e2
JG
60652013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6066
6067 * config/aarch64/aarch64.md
6068 (type): Remove frecpe, frecps, frecpx.
6069 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6070 fix to be a TARGET_SIMD instruction.
6071 (aarch64_frecps): Remove.
6072 * config/aarch64/aarch64-simd.md
6073 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6074 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6075
6e4150e1
JG
60762013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6077 Sofiane Naci <sofiane.naci@arm.com>
6078
f43245d1 6079 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6e4150e1 6080 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
f43245d1
UB
6081 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6082 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6083 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6084 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6085 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6086 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6087 into "clz, "rbit". Rename "shift" to "shift_imm".
6e4150e1 6088 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
f43245d1 6089 changes. Update for attribute changes all occurrences of arlo_* and
6e4150e1
JG
6090 shift* types.
6091 * config/arm/arm-fixed.md: Update for attribute changes
6092 all occurrences of arlo_* types.
6093 * config/arm/thumb2.md: Update for attribute changes all occurrences
6094 of arlo_* types.
6095 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6096 (cortexa7_older_only): Likewise.
6097 (cortexa7_younger): Likewise.
6098 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6099 (1020alu_shift_op): Likewise.
6100 (1020alu_shift_reg_op): Likewise.
6101 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6102 (alu_shift_op): Likewise.
6103 (alu_shift_reg_op): Likewise.
f43245d1 6104 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6e4150e1
JG
6105 (11_alu_shift_op): Likewise.
6106 (11_alu_shift_reg_op): Likewise.
6107 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6108 (9_alu_shift_reg_op): Likewise.
6109 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6110 attribute changes.
6111 (cortex_a15_alu_shift): Likewise.
6112 (cortex_a15_alu_shift_reg): Likewise.
6113 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6114 attribute changes.
6115 (cortex_a5_alu_shift): Likewise.
f43245d1
UB
6116 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6117 attribute changes.
6e4150e1 6118 (cortex_a53_alu_shift): Likewise.
f43245d1
UB
6119 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6120 attribute changes.
6e4150e1
JG
6121 (cortex_a7_alu_reg): Likewise.
6122 (cortex_a7_alu_shift): Likewise.
f43245d1
UB
6123 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6124 attribute changes.
6e4150e1
JG
6125 (cortex_a8_alu_shift): Likewise.
6126 (cortex_a8_alu_shift_reg): Likewise.
f43245d1 6127 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6e4150e1 6128 (cortex_a9_dp_shift): Likewise.
f43245d1
UB
6129 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6130 attribute changes.
6e4150e1
JG
6131 * config/arm/cortex-r4.md
6132 (cortex_r4_alu): Update for attribute changes.
6133 (cortex_r4_mov): Likewise.
6134 (cortex_r4_alu_shift_reg): Likewise.
f43245d1 6135 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6e4150e1 6136 (526_alu_shift_op): Likewise.
f43245d1
UB
6137 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6138 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6e4150e1 6139 (626te_alu_shift_op): Likewise.
f43245d1 6140 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6e4150e1
JG
6141 (726te_alu_shift_op): Likewise.
6142 (726te_alu_shift_reg_op): Likewise.
6143 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6144 (mp626_alu_shift_op): Likewise.
6145 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6146 (pj4_alu_conds): Likewise.
6147 (pj4_shift): Likewise.
6148 (pj4_shift_conds): Likewise.
6149 (pj4_alu_shift): Likewise.
6150 (pj4_alu_shift_conds): Likewise.
6151 * config/aarch64/aarch64.md: Update for attribute change
6152 all occurrences of arlo_* and shift* types.
6153
e6523306
MS
61542013-09-05 Mike Stump <mikestump@comcast.net>
6155
6156 * tree.h: Move documentation for tree_function_decl to tree-core.h
6157 with the declaration.
6158
80a18161
PB
61592013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6160
6161 PR target/58139
6162 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6163 looking for widest mode.
6164
6994430a
EB
61652013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6166
6167 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6168
e67f39f7
RB
61692013-09-05 Richard Biener <rguenther@suse.de>
6170
6171 PR tree-optimization/58137
6172 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6173 Do not create vectors of pointers.
6174 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6175 types for the components of the vector initializer.
6176 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6177 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6178
568cda29
MJ
61792013-09-05 Martin Jambor <mjambor@suse.cz>
6180
6181 * ipa-prop.c (remove_described_reference): Accept missing references,
6182 return false if that hppens, otherwise return true.
6183 (cgraph_node_for_jfunc): New function.
6184 (try_decrement_rdesc_refcount): Likewise.
6185 (try_make_edge_direct_simple_call): Use them.
6186 (ipa_edge_removal_hook): Remove references from rdescs.
6187 (ipa_edge_duplication_hook): Clone rdescs and their references
6188 when the new edge has the same caller as the old one.
6189 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6190 reference before removing any edges.
6191
5a200acb
RE
61922013-09-05 Richard Earnshaw <rearnsha@arm.com>
6193
6194 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6195 initial store.
6196 * thumb2.md (thumb2_storewb_parisi): New pattern.
6197
5922847b
YZ
61982013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6199
6200 * config/aarch64/aarch64-option-extensions.def: Add
6201 AARCH64_OPT_EXTENSION of 'crc'.
6202 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6203 (AARCH64_ISA_CRC): Ditto.
6204 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6205 description of the CRC extension.
6206
996746aa
AI
62072013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6208
6209 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6210 * config/rs6000/linux.h: Ditto.
6211 * alpha/linux.h: Ditto.
6212 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6213 no_c99_libc_has_function.
6214 * config/c6x/uclinux-elf.h: Ditto.
6215 * config/lm32/uclinux-elf.h: Ditto.
6216 * config/m68k/uclinux.h: Ditto.
6217 * config/moxie/uclinux.h: Ditto.
6218 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6219 (crisv32-*-linux*, cris-*-linux*): Ditto.
6220 * config/bfin/bfin.c: Include "tm_p.h".
6221
c370d99f
RB
62222013-09-05 Richard Biener <rguenther@suse.de>
6223
6224 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6225 check for a definition without a basic-block.
6226
003bb7f3
JG
62272013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6228 Sofiane Naci <sofiane.naci@arm.com>
6229
6230 * config/aarch64/aarch64.md
6231 (*movti_aarch64): Rename r_2_f and f_2_r.
6232 (*movsf_aarch64): Likewise.
6233 (*movdf_aarch64): Likewise.
6234 (*movtf_aarch64): Likewise.
6235 (aarch64_movdi_<mode>low): Likewise.
6236 (aarch64_movdi_<mode>high): Likewise.
6237 (aarch64_mov<mode>high_di): Likewise.
6238 (aarch64_mov<mode>low_di): Likewise.
6239 (aarch64_movtilow_tilow): Likewise.
6240 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6241 values to config/arm/types.md
6242 (attribute "conds"): Update for attribute change.
6243 (anddi3_insn): Likewise.
6244 (iordi3_insn): Likewise.
6245 (xordi3_insn): Likewise.
6246 (one_cmpldi2): Likewise.
6247 * config/arm/types.md (type): Add Neon types.
6248 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6249 use "type" attribute.
6250 (movmisalign<mode>_neon_store): Likewise.
6251 (movmisalign<mode>_neon_load): Likewise.
6252 (vec_set<mode>_internal): Likewise.
6253 (vec_setv2di_internal): Likewise.
6254 (vec_extract<mode>): Likewise.
6255 (vec_extractv2di): Likewise.
6256 (add<mode>3_neon): Likewise.
6257 (adddi3_neon): Likewise.
6258 (sub<mode>3_neon): Likewise.
6259 (subdi3_neon): Likewise.
6260 (mul<mode>3_neon): Likewise.
6261 (mul<mode>3add<mode>_neon): Likewise.
6262 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6263 (fma<VCVTF:mode>4)): Likewise.
6264 (fma<VCVTF:mode>4_intrinsic): Likewise.
6265 (fmsub<VCVTF:mode>4)): Likewise.
6266 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6267 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6268 (ior<mode>3): Likewise.
6269 (and<mode>3): Likewise.
6270 (anddi3_neon): Likewise.
6271 (orn<mode>3_neon): Likewise.
6272 (orndi3_neon): Likewise.
6273 (bic<mode>3_neon): Likewise.
6274 (bicdi3_neon): Likewise.
6275 (xor<mode>3): Likewise.
6276 (one_cmpl<mode>2): Likewise.
6277 (abs<mode>2): Likewise.
6278 (neg<mode>2): Likewise.
6279 (umin<mode>3_neon): Likewise.
6280 (umax<mode>3_neon): Likewise.
6281 (smin<mode>3_neon): Likewise.
6282 (smax<mode>3_neon): Likewise.
6283 (vashl<mode>3): Likewise.
6284 (vashr<mode>3_imm): Likewise.
6285 (vlshr<mode>3_imm): Likewise.
6286 (ashl<mode>3_signed): Likewise.
6287 (ashl<mode>3_unsigned): Likewise.
6288 (neon_load_count): Likewise.
6289 (ashldi3_neon_noclobber): Likewise.
6290 (signed_shift_di3_neon): Likewise.
6291 (unsigned_shift_di3_neon): Likewise.
6292 (ashrdi3_neon_imm_noclobber): Likewise.
6293 (lshrdi3_neon_imm_noclobber): Likewise.
6294 (widen_ssum<mode>3): Likewise.
6295 (widen_usum<mode>3): Likewise.
6296 (quad_halves_<code>v4si): Likewise.
6297 (quad_halves_<code>v4sf): Likewise.
6298 (quad_halves_<code>v8hi): Likewise.
6299 (quad_halves_<code>v16qi): Likewise.
6300 (reduc_splus_v2di): Likewise.
6301 (neon_vpadd_internal<mode>): Likewise.
6302 (neon_vpsmin<mode>): Likewise.
6303 (neon_vpsmax<mode>): Likewise.
6304 (neon_vpumin<mode>): Likewise.
6305 (neon_vpumax<mode>): Likewise.
6306 (ss_add<mode>_neon): Likewise.
6307 (us_add<mode>_neon): Likewise.
6308 (ss_sub<mode>_neon): Likewise.
6309 (us_sub<mode>_neon): Likewise.
6310 (neon_vadd<mode>_unspec): Likewise.
6311 (neon_vaddl<mode>): Likewise.
6312 (neon_vaddw<mode>): Likewise.
6313 (neon_vhadd<mode>): Likewise.
6314 (neon_vqadd<mode>): Likewise.
6315 (neon_vaddhn<mode>): Likewise.
6316 (neon_vmul<mode>): Likewise.
6317 (neon_vmla<mode>): Likewise.
6318 (neon_vmlal<mode>): Likewise.
6319 (neon_vmls<mode>): Likewise.
6320 (neon_vmlsl<mode>): Likewise.
6321 (neon_vqdmulh<mode>): Likewise.
6322 (neon_vqdmlal<mode>): Likewise.
6323 (neon_vqdmlsl<mode>): Likewise.
6324 (neon_vmull<mode>): Likewise.
6325 (neon_vqdmull<mode>): Likewise.
6326 (neon_vsub<mode>_unspec): Likewise.
6327 (neon_vsubl<mode>): Likewise.
6328 (neon_vsubw<mode>): Likewise.
6329 (neon_vqsub<mode>): Likewise.
6330 (neon_vhsub<mode>): Likewise.
6331 (neon_vsubhn<mode>): Likewise.
6332 (neon_vceq<mode>): Likewise.
6333 (neon_vcge<mode>): Likewise.
6334 (neon_vcgeu<mode>): Likewise.
6335 (neon_vcgt<mode>): Likewise.
6336 (neon_vcgtu<mode>): Likewise.
6337 (neon_vcle<mode>): Likewise.
6338 (neon_vclt<mode>): Likewise.
6339 (neon_vcage<mode>): Likewise.
6340 (neon_vcagt<mode>): Likewise.
6341 (neon_vtst<mode>): Likewise.
6342 (neon_vabd<mode>): Likewise.
6343 (neon_vabdl<mode>): Likewise.
6344 (neon_vaba<mode>): Likewise.
6345 (neon_vabal<mode>): Likewise.
6346 (neon_vmax<mode>): Likewise.
6347 (neon_vmin<mode>): Likewise.
6348 (neon_vpaddl<mode>): Likewise.
6349 (neon_vpadal<mode>): Likewise.
6350 (neon_vpmax<mode>): Likewise.
6351 (neon_vpmin<mode>): Likewise.
6352 (neon_vrecps<mode>): Likewise.
6353 (neon_vrsqrts<mode>): Likewise.
6354 (neon_vqabs<mode>): Likewise.
6355 (neon_vqneg<mode>): Likewise.
6356 (neon_vcls<mode>): Likewise.
6357 (clz<mode>2): Likewise.
6358 (popcount<mode>2): Likewise.
6359 (neon_vrecpe): Likewise.
6360 (neon_vrsqrte): Likewise.
6361 (neon_vget_lane<mode>_sext_internal): Likewise.
6362 (neon_vget_lane<mode>_zext_internal): Likewise.
6363 (neon_vdup_n<mode>): Likewise.
6364 (neon_vdup_nv2di): Likewise.
6365 (neon_vdpu_lane<mode>_internal): Likewise.
6366 (neon_vswp<mode>): Likewise.
6367 (float<mode><V_cvtto>2): Likewise.
6368 (floatuns<mode><V_cvtto>2): Likewise.
6369 (fix_trunc<mode><V_cvtto>)2): Likewise
6370 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6371 (neon_vcvt<mode>): Likewise.
6372 (neon_vcvtv4sfv4hf): Likewise.
6373 (neon_vcvtv4hfv4sf): Likewise.
6374 (neon_vcvt_n<mode>): Likewise.
6375 (neon_vmovn<mode>): Likewise.
6376 (neon_vqmovn<mode>): Likewise.
6377 (neon_vqmovun<mode>): Likewise.
6378 (neon_vmovl<mode>): Likewise.
6379 (neon_vmul_lane<mode>): Likewise.
6380 (neon_vmull_lane<mode>): Likewise.
6381 (neon_vqdmull_lane<mode>): Likewise.
6382 (neon_vqdmulh_lane<mode>): Likewise.
6383 (neon_vmla_lane<mode>): Likewise.
6384 (neon_vmlal_lane<mode>): Likewise.
6385 (neon_vqdmlal_lane<mode>): Likewise.
6386 (neon_vmls_lane<mode>): Likewise.
6387 (neon_vmlsl_lane<mode>): Likewise.
6388 (neon_vqdmlsl_lane<mode>): Likewise.
6389 (neon_vext<mode>): Likewise.
6390 (neon_vrev64<mode>): Likewise.
6391 (neon_vrev32<mode>): Likewise.
6392 (neon_vrev16<mode>): Likewise.
6393 (neon_vbsl<mode>_internal): Likewise.
6394 (neon_vshl<mode>): Likewise.
6395 (neon_vqshl<mode>): Likewise.
6396 (neon_vshr_n<mode>): Likewise.
6397 (neon_vshrn_n<mode>): Likewise.
6398 (neon_vqshrn_n<mode>): Likewise.
6399 (neon_vqshrun_n<mode>): Likewise.
6400 (neon_vshl_n<mode>): Likewise.
6401 (neon_vqshl_n<mode>): Likewise.
6402 (neon_vqshlu_n<mode>): Likewise.
6403 (neon_vshll_n<mode>): Likewise.
6404 (neon_vsra_n<mode>): Likewise.
6405 (neon_vsri_n<mode>): Likewise.
6406 (neon_vsli_n<mode>): Likewise.
6407 (neon_vtbl1v8qi): Likewise.
6408 (neon_vtbl2v8qi): Likewise.
6409 (neon_vtbl3v8qi): Likewise.
6410 (neon_vtbl4v8qi): Likewise.
6411 (neon_vtbx1v8qi): Likewise.
6412 (neon_vtbx2v8qi): Likewise.
6413 (neon_vtbx3v8qi): Likewise.
6414 (neon_vtbx4v8qi): Likewise.
6415 (neon_vtrn<mode>_internal): Likewise.
6416 (neon_vzip<mode>_internal): Likewise.
6417 (neon_vuzp<mode>_internal): Likewise.
6418 (neon_vld1<mode>): Likewise.
6419 (neon_vld1_lane<mode>): Likewise.
6420 (neon_vld1_dup<mode>): Likewise.
6421 (neon_vld1_dupv2di): Likewise.
6422 (neon_vst1<mode>): Likewise.
6423 (neon_vst1_lane<mode>): Likewise.
6424 (neon_vld2<mode>): Likewise.
6425 (neon_vld2_lane<mode>): Likewise.
6426 (neon_vld2_dup<mode>): Likewise.
6427 (neon_vst2<mode>): Likewise.
6428 (neon_vst2_lane<mode>): Likewise.
6429 (neon_vld3<mode>): Likewise.
6430 (neon_vld3qa<mode>): Likewise.
6431 (neon_vld3qb<mode>): Likewise.
6432 (neon_vld3_lane<mode>): Likewise.
6433 (neon_vld3_dup<mode>): Likewise.
6434 (neon_vst3<mode>): Likewise.
6435 (neon_vst3qa<mode>): Likewise.
6436 (neon_vst3qb<mode>): Likewise.
6437 (neon_vst3_lane<mode>): Likewise.
6438 (neon_vld4<mode>): Likewise.
6439 (neon_vld4qa<mode>): Likewise.
6440 (neon_vld4qb<mode>): Likewise.
6441 (neon_vld4_lane<mode>): Likewise.
6442 (neon_vld4_dup<mode>): Likewise.
6443 (neon_vst4<mode>): Likewise.
6444 (neon_vst4qa<mode>): Likewise.
6445 (neon_vst4qb<mode>): Likewise.
6446 (neon_vst4_lane<mode>): Likewise.
6447 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6448 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6449 (neon_vec_<US>mult_lo_<mode>): Likewise.
6450 (neon_vec_<US>mult_hi_<mode>): Likewise.
6451 (neon_vec_<US>shiftl_<mode>): Likewise.
6452 (neon_unpack<US>_<mode>): Likewise.
6453 (neon_vec_<US>mult_<mode>): Likewise.
6454 (vec_pack_trunc_<mode>): Likewise.
6455 (neon_vec_pack_trunk_<mode>): Likewise.
6456 (neon_vabd<mode>_2): Likewise.
6457 (neon_vabd<mode>_3): Likewise.
6458 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6459 (thumb2_movsi_vfp): Likewise.
6460 (movdi_vfp): Likewise.
6461 (movdi_vfp_cortexa8): Likewise.
6462 (movhf_vfp_neon): Likewise.
6463 (movhf_vfp): Likewiwse.
6464 (movsf_vfp): Likewiwse.
6465 (thumb2_movsf_vfp): Likewiwse.
6466 (movdf_vfp): Likewise.
6467 (thumb2_movdf_vfp): Likewise.
6468 (movsfcc_vfp): Likewise.
6469 (thumb2_movsfcc_vfp): Likewise.
6470 (movdfcc_vfp): Likewise.
6471 (thumb2_movdfcc_vfp): Likewise.
6472 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6473 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6474 (v10_v2c): Likewise.
6475 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6476 attribute change.
6477 (cortex_a15_neon_int_2): Likewise.
6478 (cortex_a15_neon_int_3): Likewise.
6479 (cortex_a15_neon_int_4): Likewise.
6480 (cortex_a15_neon_int_5): Likewise.
6481 (cortex_a15_neon_vqneg_vqabs): Likewise.
6482 (cortex_a15_neon_vmov): Likewise.
6483 (cortex_a15_neon_vaba): Likewise.
6484 (cortex_a15_neon_vaba_qqq): Likewise.
6485 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6486 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6487 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6488 scalar_64_32_long_scalar): Likewise.
6489 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6490 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6491 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6492 lotype_qdd_64_32_long): Likewise.
6493 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6494 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6495 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6496 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6497 (cortex_a15_neon_shift_1): Likewise.
6498 (cortex_a15_neon_shift_2): Likewise.
6499 (cortex_a15_neon_shift_3): Likewise.
6500 (cortex_a15_neon_vshl_ddd): Likewise.
6501 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6502 (cortex_a15_neon_vsra_vrsra): Likewise.
6503 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6504 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6505 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6506 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6507 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6508 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6509 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6510 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6511 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6512 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6513 (cortex_a15_neon_bp_simple): Likewise.
6514 (cortex_a15_neon_bp_2cycle): Likewise.
6515 (cortex_a15_neon_bp_3cycle): Likewise.
6516 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6517 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6518 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6519 (cortex_a15_neon_vld2_4_regs): Likewise.
6520 (cortex_a15_neon_vld3_vld4): Likewise.
6521 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6522 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6523 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6524 (cortex_a15_neon_vst3_vst4): Likewise.
6525 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6526 (cortex_a15_neon_vld3_vld4_lane" 10
6527 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6528 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6529 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6530 (cortex_a15_neon_ldm_2): Likewise.0
6531 (cortex_a15_neon_stm_2): Likewise.
6532 (cortex_a15_neon_mcr): Likewise.
6533 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6534 (cortex_a15_neon_mrc): Likewise.
6535 (cortex_a15_neon_mrrc): Likewise.
6536 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6537 change.
6538 (cortex_a15_alu_shift): Likewise.
6539 (cortex_a15_alu_shift_reg): Likewise.
6540 (cortex_a15_mult32): Likewise.
6541 (cortex_a15_mult64): Likewise.
6542 (cortex_a15_block): Likewise.
6543 (cortex_a15_branch): Likewise.
6544 (cortex_a15_load1): Likewise.
6545 (cortex_a15_load3): Likewise.
6546 (cortex_a15_store1): Likewise.
6547 (cortex_a15_store3): Likewise.
6548 (cortex_a15_call): Likewise.
f43245d1 6549 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
003bb7f3
JG
6550 (cortex_a5_f2r): Likewise.
6551 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6552 change.
6553 (cortex_a53_f2r): Likewise.
6554 * config/arm/cortex-a7.md
6555 (cortex_a7_branch): Update for attribute change.
6556 (cortex_a7_call): Likewise.
6557 (cortex_a7_alu_imm): Likewise.
6558 (cortex_a7_alu_reg): Likewise.
6559 (cortex_a7_alu_shift): Likewise.
6560 (cortex_a7_mul): Likewise.
6561 (cortex_a7_load1): Likewise.
6562 (cortex_a7_store1): Likewise.
6563 (cortex_a7_load2): Likewise.
6564 (cortex_a7_store2): Likewise.
6565 (cortex_a7_load3): Likewise.
6566 (cortex_a7_store3): Likewise.
6567 (cortex_a7_load4): Likewise.
6568 (cortex_a7_store4): Likewise.
6569 (cortex_a7_fpalu): Likewise.
6570 (cortex_a7_fconst): Likewise.
6571 (cortex_a7_fpmuls): Likewise.
6572 (cortex_a7_neon_mul): Likewise.
6573 (cortex_a7_fpmacs): Likewise.
6574 (cortex_a7_neon_mla: Likewise.
6575 (cortex_a7_fpmuld: Likewise.
6576 (cortex_a7_fpmacd: Likewise.
6577 (cortex_a7_fpfmad: Likewise.
6578 (cortex_a7_fdivs: Likewise.
6579 (cortex_a7_fdivd: Likewise.
6580 (cortex_a7_r2f: Likewise.
6581 (cortex_a7_f2r: Likewise.
6582 (cortex_a7_f_flags: Likewise.
6583 (cortex_a7_f_loads: Likewise.
6584 (cortex_a7_f_loadd: Likewise.
6585 (cortex_a7_f_stores: Likewise.
6586 (cortex_a7_f_stored: Likewise.
6587 (cortex_a7_neon): Likewise.
6588 * config/arm/cortex-a8-neon.md
6589 (cortex_a8_neon_mrc): Update for attribute change.
6590 (cortex_a8_neon_mrrc): Likewise.
6591 (cortex_a8_neon_int_1): Likewise.
6592 (cortex_a8_neon_int_2): Likewise.
6593 (cortex_a8_neon_int_3): Likewise.
6594 (cortex_a8_neon_int_4): Likewise.
6595 (cortex_a8_neon_int_5): Likewise.
6596 (cortex_a8_neon_vqneg_vqabs): Likewise.
6597 (cortex_a8_neon_vmov): Likewise.
6598 (cortex_a8_neon_vaba): Likewise.
6599 (cortex_a8_neon_vaba_qqq): Likewise.
6600 (cortex_a8_neon_vsma): Likewise.
6601 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6602 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
6603 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6604 long_scalar): Likewise.
003bb7f3
JG
6605 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6606 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6607 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6608 long_scalar_qdd_64_32_long): Likewise.
6609 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6610 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6611 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6612 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6613 (cortex_a8_neon_shift_1): Likewise.
6614 (cortex_a8_neon_shift_2): Likewise.
6615 (cortex_a8_neon_shift_3): Likewise.
6616 (cortex_a8_neon_vshl_ddd): Likewise.
6617 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6618 (cortex_a8_neon_vsra_vrsra): Likewise.
6619 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6620 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6621 (cortex_a8_neon_fp_vsum): Likewise.
6622 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6623 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6624 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6625 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6626 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6627 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6628 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6629 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6630 (cortex_a8_neon_bp_simple): Likewise.
6631 (cortex_a8_neon_bp_2cycle): Likewise.
6632 (cortex_a8_neon_bp_3cycle): Likewise.
6633 (cortex_a8_neon_ldr): Likewise.
6634 (cortex_a8_neon_str): Likewise.
6635 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6636 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6637 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6638 (cortex_a8_neon_vld2_4_regs): Likewise.
6639 (cortex_a8_neon_vld3_vld4): Likewise.
6640 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6641 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6642 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6643 (cortex_a8_neon_vst3_vst4): Likewise.
6644 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6645 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6646 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6647 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6648 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6649 (cortex_a8_neon_mcr): Likewise.
6650 (cortex_a8_neon_mcr_2_mcrr): Likewise.
f43245d1 6651 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
003bb7f3
JG
6652 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6653 change.
6654 (ca9_neon_mrrc): Likewise.
6655 (cortex_a9_neon_int_1): Likewise.
6656 (cortex_a9_neon_int_2): Likewise.
6657 (cortex_a9_neon_int_3): Likewise.
6658 (cortex_a9_neon_int_4): Likewise.
6659 (cortex_a9_neon_int_5): Likewise.
6660 (cortex_a9_neon_vqneg_vqabs): Likewise.
6661 (cortex_a9_neon_vmov): Likewise.
6662 (cortex_a9_neon_vaba): Likewise.
6663 (cortex_a9_neon_vaba_qqq): Likewise.
6664 (cortex_a9_neon_vsma): Likewise.
6665 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6666 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
f43245d1
UB
6667 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6668 long_scalar): Likewise.
003bb7f3
JG
6669 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6670 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6671 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6672 long_scalar_qdd_64_32_long): Likewise.
6673 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6674 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6675 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6676 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6677 (cortex_a9_neon_shift_1): Likewise.
6678 (cortex_a9_neon_shift_2): Likewise.
6679 (cortex_a9_neon_shift_3): Likewise.
6680 (cortex_a9_neon_vshl_ddd): Likewise.
6681 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6682 (cortex_a9_neon_vsra_vrsra): Likewise.
6683 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6684 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6685 (cortex_a9_neon_fp_vsum): Likewise.
6686 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6687 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6688 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6689 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6690 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6691 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6692 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6693 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6694 (cortex_a9_neon_bp_simple): Likewise.
6695 (cortex_a9_neon_bp_2cycle): Likewise.
6696 (cortex_a9_neon_bp_3cycle): Likewise.
6697 (cortex_a9_neon_ldr): Likewise.
6698 (cortex_a9_neon_str): Likewise.
6699 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6700 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6701 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6702 (cortex_a9_neon_vld2_4_regs): Likewise.
6703 (cortex_a9_neon_vld3_vld4): Likewise.
6704 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6705 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6706 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6707 (cortex_a9_neon_vst3_vst4): Likewise.
6708 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6709 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6710 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6711 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6712 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6713 (cortex_a9_neon_mcr): Likewise.
6714 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6715 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6716 (cortex_a9_fps): Likewise.
6717 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6718 change.
6719 (cortex_m4_fmuls): Likewise.
6720 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6721 change.
6722 (cortex_r4_mrc): Likewise.
6723 * config/arm/iterators.md: Update comment referring to neon_type.
f43245d1 6724 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
003bb7f3 6725 (iwmmxt_movsi_insn): Likewise.
f43245d1
UB
6726 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6727 attribute change.
003bb7f3
JG
6728 (pj4_core_to_vfp): Likewise.
6729 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6730 attribute change.
6731 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6732 (vfp_fstore): Likewise.
6733 * doc/md.texi: Change references to neon_type to refer to type.
6734
fbe0dc78
DS
67352013-09-04 Dodji Seketeli <dodji@redhat.com>
6736
6737 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6738
c167bc5b
DE
67392013-09-04 David Edelsohn <dje.gcc@gmail.com>
6740
6741 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6742 lglobl if not weak.
6743
035cb59f
ER
67442013-09-04 Easwaran Raman <eraman@google.com>
6745
6746 PR middle-end/57370
6747 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6748 (build_and_add_sum): Use it.
6749 (appears_later_in_bb): Simplify code.
6750
71191083
TJ
67512013-09-04 Teresa Johnson <tejohnson@google.com>
6752
6753 * dumpfile.c (dump_finish): Don't close stderr/stdout.
6754
0c488213
JG
67552013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
6756
6757 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
6758
61a74079
JH
67592013-09-04 Jan Hubicka <jh@suse.cz>
6760
6761 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
6762 * ipa-devirt.c: Include diganostic.h
6763 (odr_type_d): Add types and types_set.
6764 (hash_type_name): Work for types with vtables during LTO.
6765 (odr_hasher::remove): Fix comment; destroy types_set.
6766 (add_type_duplicate): New function,
6767 (get_odr_type): Use it.
6768 (dump_type_inheritance_graph): Dump type duplicates.
6769 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
6770 graph.
6771 * tree.c (types_same_for_odr): Give exact answers on types with
6772 virtual tables.
6773
24ba8195
DS
67742013-09-04 Dodji Seketeli <dodji@redhat.com>
6775
6776 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
7a5911d3 6777 explaining their differences.
24ba8195 6778
0a04869e
SKS
67792013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
6780
6781 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
6782
3b18bc42
JL
67832013-09-03 Jeff Law <law@redhat.com>
6784
6785 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
6786 when not threading through a joiner block. Pass joiner/no joiner
6787 state to register_jump_thread.
6788 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
6789 state from argument rather than implying on path length.
6790 Dump the entire jump thread path into debugging dump.
6791 * tree-flow.h (register_jump_thread): Update prototype.
6792
ae84e151
XDL
67932013-08-29 Xinliang David Li <davidxl@google.com>
6794
6795 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6796 Remove a trivial gcc_assert.
6797
c716e67f
XDL
67982013-08-29 Xinliang David Li <davidxl@google.com>
6799
6800 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
6801 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
6802 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6803 Delay base decl alignment adjustment.
6804 * tree-vectorizer.c (vect_destroy_datarefs): New function.
6805 * tree-vectorizer.h: New data structure.
6806 (set_dr_misalignment): New function.
6807 (dr_misalignment): Ditto.
6808 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
6809 (vectorizable_load): Ditto.
6810 (ensure_base_align): New function.
6811 (vectorize_loops): Add dbg_cnt support.
6812 (execute_vect_slp): Ditto.
6813 * dbgcnt.def: New debug counter.
6814 * Makefile: New dependency.
6815
db8800bc
MI
68162013-09-03 Meador Inge <meadori@codesourcery.com>
6817
6818 Revert:
6819
6820 2013-08-30 Meador Inge <meadori@codesourcery.com>
6821
6822 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
6823
0d9b4f5c
DE
68242013-09-03 David Edelsohn <dje.gcc@gmail.com>
6825
6826 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
6827 function descriptor.
6828
29e810b9
RB
68292013-09-03 Richard Biener <rguenther@suse.de>
6830
6831 * tree-affine.c (add_elt_to_tree): Fix association issue,
6832 avoid useless converts and make sure to always return a
6833 properly typed result.
6834
dbfc2894
RB
68352013-09-03 Richard Biener <rguenther@suse.de>
6836
6837 PR middle-end/57656
6838 * fold-const.c (negate_expr_p): Fix division case.
6839 (negate_expr): Likewise.
6840
fa3bf4ce
RB
68412013-09-03 Richard Biener <rguenther@suse.de>
6842
6843 PR lto/58285
6844 * tree-streamer-out.c: Include tm.h.
6845 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
6846
a64bbb3f
JH
68472013-09-03 Jan Hubicka <jh@suse.cz>
6848
6849 * tree-profile.c (tree_profiling): Cleanup CFG when done.
6850
93ed07e2
AM
68512013-09-03 Alan Modra <amodra@gmail.com>
6852
6853 * config.gcc (powerpc*-*-linux*): Add support for little-endian
6854 multilibs to big-endian target and vice versa.
6855 * config/rs6000/t-linux64: Use := assignment on all vars.
6856 (MULTILIB_EXTRA_OPTS): Remove fPIC.
6857 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
6858 * config/rs6000/t-linux64le: New file.
6859 * config/rs6000/t-linux64bele: New file.
6860 * config/rs6000/t-linux64lebe: New file.
6861
07d6cd64
JH
68622013-09-02 Jan Hubicka <jh@suse.cz>
6863
6864 * ipa-inline-transform.c (inline_transform): Do not
6865 optimize_inline_calls when not optimizing.
6866
8e4c9a10
JH
68672013-09-02 Jan Hubicka <jh@suse.cz>
6868
6869 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
6870 duplicated nodes for assembler names.
6871 * symtab.c (symtab_unregister_node): Do not attempt to unlink
6872 hard registers from assembler name hash.
6873
c91061e6
JH
68742013-09-02 Jan Hubicka <jh@suse.cz>
6875
6876 * ipa-split.c (execute_split_functions): Split externally visible
6877 functions called once.
6878
2fd0985c
MJ
68792013-09-02 Martin Jambor <mjambor@suse.cz>
6880
6881 PR ipa/58106
6882 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
6883 linked list. When finding the correct duplicate, also consider also
6884 the caller in additon to its inlined_to node.
6885
91bd4114
JG
68862013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
6887
6888 * config/aarch64/aarch64-simd-builtins.def
6889 (dup_lane_scalar): Remove.
6890 * config/aarch64/aarch64-simd.md
6891 (aarch64_simd_dup): Add 'w->w' alternative.
6892 (aarch64_dup_lane<mode>): Allow for VALL.
6893 (aarch64_dup_lane_scalar<mode>): Remove.
6894 (aarch64_dup_lane_<vswap_width_name><mode>): New.
6895 (aarch64_get_lane_signed<mode>): Add w->w altenative.
6896 (aarch64_get_lane_unsigned<mode>): Likewise.
6897 (aarch64_get_lane<mode>): Likewise.
6898 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
6899 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
6900 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
6901 (VCON): Change container of V2SF.
6902 (vswap_width_name): Likewise.
6903 * config/aarch64/arm_neon.h
6904 (__aarch64_vdup_lane_any): New.
6905 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6906 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
6907 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6908
d617d2d8
EB
69092013-09-02 Eric Botcazou <ebotcazou@adacore.com>
6910
6911 PR middle-end/56382
6912 * expr.c (emit_move_complex): Do not move complex FP values as parts if
6913 the source or the destination is a single hard register.
6914
63fdb7be
RB
69152013-09-02 Richard Biener <rguenther@suse.de>
6916
6917 PR middle-end/57511
6918 * tree-scalar-evolution.c (instantiate_scev_name): Allow
6919 non-linear SCEVs.
6920
78de2333
RB
69212013-09-02 Richard Biener <rguenther@suse.de>
6922
6923 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
6924 arithmetic to sizetype.
6925
85ff4ec6
BC
69262013-09-02 Bin Cheng <bin.cheng@arm.com>
6927
6928 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
6929 Find auto-increment use both before and after candidate.
6930
576016fe
MP
69312013-09-02 Marek Polacek <polacek@redhat.com>
6932
6933 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
6934
08f835dc
JH
69352013-09-01 Jan Hubicka <jh@suse.cz>
6936
6937 * Makefile.in: Add ipa-profile.o
6938 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
6939 * cgraph.c (struct cgraph_propagate_frequency_data,
6940 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
6941 ipa-profile.c; replace cgraph_ by ipa_ prefix.
6942 * cgraph.h (cgraph_propagate_frequency): Remove.
f43245d1
UB
6943 * ipa-inline-analysis.c: Include ipa-utils.h;
6944 drop duplicated cfgloop.h.
08f835dc
JH
6945 (inline_update_callee_summaries): Update.
6946 * ipa-profile.c: New file.
6947 * ipa-utils.h (ipa_propagate_frequency): Declare.
6948 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
f43245d1 6949 data-streamer.h, value-prof.h.
08f835dc
JH
6950 (symtab_remove_unreachable_nodes): Update profile.
6951 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
6952 account_time_size, cmp_counts, dump_histogram,
6953 ipa_profile_generate_summary, ipa_profile_write_summary,
6954 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
6955 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
f43245d1 6956 Move to ipa-profile.c.
08f835dc 6957
cec8583c
JDA
69582013-09-01 John David Anglin <danglin@gcc.gnu.org>
6959
6960 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
6961
bbc9396b
JH
69622013-09-01 Jan Hubicka <jh@suse.cz>
6963
6964 * common.opt (fdevirtualize-speculatively): New function.
6965 * invoke.texi (fdevirtualize-speculatively): Document.
6966 * ipa-devirt.c: Include ipa-inline.h
6967 (likely_target_p): New function.
6968 (ipa_devirt): New function.
6969 (gate_ipa_devirt): New function.
6970 (pass_data_ipa_devirt): New static var.
6971 (pass_ipa_devirt): Likewise.
6972 (make_pass_ipa_devirt): New function.
6973 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
6974 (common_handle_option): Disable devirtualization when
6975 value range profiling is available.
6976 * passes.def (pass_ipa_devirt): Add.
6977 * timever.def (TV_IPA_DEVIRT): New timevar.
f43245d1 6978 * tree-pass.h (make_pass_ipa_devirt):
bbc9396b 6979
0cea211e
IS
69802013-09-01 Iain Sandoe <iain@codesourcery.com>
6981
6982 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
6983 include sanitize(undefined).
6984
5e351e96
DN
69852013-08-31 Diego Novillo <dnovillo@google.com>
6986
6987 * Makefile.in (TREE_CORE_H): Define.
6988 (TREE_H): Use.
6989 (GTFILES): Add tree-core.h.
6990 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
6991 size the array.
6992 * tree-core.h: New file.
6993 Move all data structures, enum, typedefs, global
6994 declarations and constants from ...
6995 * tree.h: ... here.
6996
63bf9a90
JH
69972013-08-31 Jan Hubicka <jh@suse.cz>
6998
6999 * bulitins.c (expand_builtin): Do not early exit for gcov
7000 instrumented functions.
7001
5c0abd6a
MP
70022013-08-31 Marek Polacek <polacek@redhat.com>
7003
7004 * ubsan.c: Include tm_p.h.
7005
4843f032
JH
70062013-08-31 Jan Hubicka <jh@suse.cz>
7007
8356c89c
JH
7008 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
7009 warning.
7010
4843f032
JH
7011 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
7012 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
5e581212 7013 * tree-cfg.c (verify_gimple_label): ... here.
4843f032
JH
7014 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
7015 (ipa_merge_profiles): New function.
7016 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
7017 (lto_input_function_body): Likewise.
5e581212
JH
7018 * ipa-utils.h (ipa_merge_profiles): Declare.
7019 * lto-streamer.h (lto_input_function_body): Update prototype.
7020 (emit_label_in_global_context_p): Remove.
7021 * lto-symtab.c: Include ipa-utils.h
7022 (lto_cgraph_replace_node): Use ipa_merge_profiles.
4843f032 7023
57292ce9
JH
70242013-08-31 Jan Hubicka <jh@suse.cz>
7025
7026 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
7027
b9cbfeeb
JH
70282013-08-31 Jan Hubicka <jh@suse.cz>
7029
7030 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
7031
41dedebd
UB
70322013-08-31 Uros Bizjak <ubizjak@gmail.com>
7033
7034 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7035 "cmp" RTX before signed_comparison_operator check to account
7036 for "code" changes.
7037
06d65050
JH
70382013-08-30 Jan Hubicka <jh@suse.cz>
7039
41dedebd 7040 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
06d65050
JH
7041 (detect_type_change_1): Rename to ...
7042 (detect_type_change): ... this one; early return on non-polymorphic
7043 types.
41dedebd 7044 (detect_type_change_ssa): Add comp_type parameter; update
06d65050
JH
7045 use of detect_type_change.
7046 (compute_complex_assign_jump_func): Add param_type parameter;
7047 update use of detect_type_change_ssa.
7048 (compute_complex_ancestor_jump_func): Likewise.
7049 (ipa_get_callee_param_type): New function.
7050 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7051 update calls to the jump function computation functions.
7052
600b5b1d
TJ
70532013-08-30 Teresa Johnson <tejohnson@google.com>
7054 Steven Bosscher <steven@gcc.gnu.org>
7055
7056 * cfgrtl.c (fixup_new_cold_bb): New routine.
7057 (commit_edge_insertions): Invoke fixup_partitions.
7058 (find_partition_fixes): New routine.
7059 (fixup_partitions): Ditto.
7060 (verify_hot_cold_block_grouping): Update comments.
7061 (rtl_verify_edges): Invoke find_partition_fixes.
7062 (rtl_verify_bb_pointers): Update comments.
7063 (rtl_verify_bb_layout): Ditto.
7064 * basic-block.h (probably_never_executed_edge_p): Declare.
7065 (fixup_partitions): Ditto.
7066 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7067 * bb-reorder.c (sanitize_hot_paths): New function.
7068 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7069 sanitize_hot_paths.
7070 * predict.c (probably_never_executed_edge_p): New routine.
7071 * cfg.c (check_bb_profile): Add partition insanity warnings.
7072
0a6ea5c9
MI
70732013-08-30 Meador Inge <meadori@codesourcery.com>
7074
7075 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7076
de5a5fa1
MP
70772013-08-30 Marek Polacek <polacek@redhat.com>
7078
7079 * Makefile.in (ubsan.o): Add.
7080 (c-family/c-ubsan.o): Add.
7081 (builtins.o): Add ubsan.h dependency.
7082 * ubsan.h: New file.
7083 * ubsan.c: New file.
7084 * common.opt: Add -fsanitize=undefined option.
7085 (flag_sanitize): Add variable.
7086 (fsanitize=): Add option. Add Driver.
7087 (fsanitize=thread): Remove option.
7088 (fsanitize=address): Likewise.
7089 (static-libubsan): New option.
7090 * doc/invoke.texi: Document the new flag and -static-libubsan.
7091 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7092 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7093 * builtin-attrs.def (ATTR_COLD): Define.
7094 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7095 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7096 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7097 * flag-types.h (sanitize_code): New enum.
7098 * opts.c (common_handle_option): Parse command line arguments
7099 of -fsanitize=. Add -fsanitize=unreachable option.
7100 * varasm.c (get_variable_section): Adjust.
7101 (assemble_noswitch_variable): Likewise.
7102 (assemble_variable): Likewise.
7103 (output_constant_def_contents): Likewise.
7104 (categorize_decl_for_section): Likewise.
7105 (place_block_symbol): Likewise.
7106 (output_object_block): Likewise.
7107 * builtins.def: Likewise.
7108 * toplev.c (compile_file): Likewise.
7109 (process_options): Likewise.
7110 * cppbuiltin.c: Likewise.
7111 * tsan.c (tsan_pass): Likewise.
7112 (tsan_gate): Likewise.
7113 (tsan_gate_O0): Likewise.
7114 * cfgexpand.c (partition_stack_vars): Likewise.
7115 (expand_stack_vars): Likewise.
7116 (defer_stack_allocation): Likewise.
7117 (expand_used_vars): Likewise.
7118 * cfgcleanup.c (old_insns_match_p): Likewise.
7119 * asan.c (asan_finish_file): Likewise.
7120 (asan_instrument): Likewise.
7121 (gate_asan): Likewise.
7122 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7123 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7124 (asan_global_struct): Use pointer_sized_int_node instead
7125 calling build_nonstandard_integer_type.
7126 (initialize_sanitizer_builtins): Likewise.
7127 (asan_finish_file): Likewise.
7128 * gcc.c: Document %{%:function(args):X}.
7129 (static_spec_functions): Add sanitize.
7130 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7131 store funcval != NULL there.
7132 (do_spec_1): Adjust handle_spec_function caller.
7133 (handle_braces): Allow %:function(args) as condition.
7134 (sanitize_spec_function): New function.
7135 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7136 (LIBUBSAN_SPEC): Likewise.
7137 (LIBUBSAN_EARLY_SPEC): Likewise.
7138 (SANITIZER_SPEC): Handle libubsan.
7139 (SANITIZER_EARLY_SPEC): Likewise.
7140 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7141 instead of fsanitize=address.
7142 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7143 instead of fsanitize=address*.
7144 * builtins.c: Include ubsan.h.
7145 (fold_builtin_0): Instrument __builtin_unreachable.
7146 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7147 instead of flag_asan.
7148 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7149 (pointer_sized_int_node): Define.
41dedebd 7150 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
de5a5fa1 7151
f07f30cf
MS
71522013-08-30 Mike Stump <mikestump@comcast.net>
7153
7154 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7155 with RE patterns.
7156
8a41354f
JH
71572013-08-29 Jan Hubicka <jh@suse.cz>
7158
7159 * cgraph.c (cgraph_function_body_availability): Handle weakref
7160 correctly.
7161 * passes.def: Remove pass_fixup_cfg.
7162 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7163 track when we need to remove functions.
7164 (gate_ipa_inline): Execute inlining always; add comment why.
7165 (pass_data_ipa_inline): Remove TODO_remove_functions.
7166 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7167 do not produce summaries.
7168 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7169 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7170 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7171 constant pool and vtable.
7172
dc8d7a0f
TB
71732013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7174
41dedebd
UB
7175 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7176 New arm_neon.h's internal macros to specify 64-bit constants.
7177 Avoid using stdint.h's macros.
dc8d7a0f 7178
ac7eacd2
JR
71792013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7180
7181 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7182 before applying REG_P.
7183
8b29fd4e
JJ
71842013-08-30 Jakub Jelinek <jakub@redhat.com>
7185
7186 PR tree-optimization/58277
7187 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7188 after seeing too many stmts with vdef in between dombb and current
7189 bb, invalidate everything.
7190
26d75703
RB
71912013-08-30 Richard Biener <rguenther@suse.de>
7192
7193 * fold-const.c (fold_single_bit_test): Fix overflow test.
7194
f7e088e7
EB
71952013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7196
7197 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7198 and which can live in a register, always retrieve the value on entry.
7199 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7200 passed by invisible reference specially.
7201 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7202 (vt_add_function_parameter): Correctly deal with a parameter passed by
7203 invisible reference.
7204
72052013-08-30 Jan Hubicka <jh@suse.cz>
47500756
JH
7206
7207 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7208
6e8dad05
RB
72092013-08-30 Richard Biener <rguenther@suse.de>
7210
7211 PR tree-optimization/58228
7212 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7213 allow invariant loads in nested loop vectorization.
7214
062ef2c8
RB
72152013-08-30 Richard Biener <rguenther@suse.de>
7216
7217 PR tree-optimization/58223
7218 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7219 (has_anti_or_output_dependence): ... this and adjust to also
7220 look for output dependences.
7221 (mark_nodes_having_upstream_mem_writes): Adjust.
7222 (rdg_flag_uses): Likewise.
7223
7a764c60
RB
72242013-08-30 Richard Biener <rguenther@suse.de>
7225
7226 PR tree-optimization/58010
7227 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7228 assert that we have a loop-closed PHI.
7229
82e9d642
JH
72302013-08-29 Jan Hubicka <jh@suse.cz>
7231
7232 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7233 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7234 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7235 (lto_free_function_in_decl_state_for_node): New function.
7236
9cc1fb4b
XDL
72372013-08-29 Xinliang David Li <davidxl@google.com>
7238
41dedebd 7239 * loop-unroll.c (report_unroll_peel): Minor message change.
9cc1fb4b
XDL
7240 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7241 Emit alignment peeling message with default -fopt-info.
7242 (vect_loop_versioning): Emit loop version info message.
41dedebd 7243 * tree-vectorizer.c (vectorize_loops): Minor message change.
9cc1fb4b
XDL
7244 (execute_vect_slp): Ditto.
7245
440a5082
EB
72462013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7247
7248 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7249 of the clone from the DECL_NAME of the original function.
7250
3fa3690d
OE
72512013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7252
7253 * passes.c (register_pass): Add overload.
7254 * tree-pass.h (register_pass): Forward declare it. Add comment.
7255
0170f33c
JH
72562013-08-29 Jan Hubicka <jh@suse.cz>
7257
41dedebd
UB
7258 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7259 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7260 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7261 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
7262 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7263 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
41dedebd 7264 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
0170f33c
JH
7265 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7266
103ff0d6
TJ
72672013-08-29 Teresa Johnson <tejohnson@google.com>
7268
7269 * dumpfile.c (dump_loc): Output column number.
7270 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7271 * doc/invoke.texi: Document optall -fopt-info flag.
7272 * profile.c (read_profile_edge_counts): Use new dump framework.
7273 (compute_branch_probabilities): Ditto.
7274 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7275 when pass not in any opt group.
7276 * pass_manager.h (pass_manager::get_pass_profile): New method.
7277 * value-prof.c (check_counter): Use new dump framework.
7278 (check_ic_target): Ditto.
7279 * coverage.c (get_coverage_counts): Ditto.
7280 (coverage_init): Setup new dump framework.
7281
301bbc16
RB
72822013-08-29 Richard Biener <rguenther@suse.de>
7283
7284 PR tree-optimization/58246
7285 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7286 handle the dominance check inside a basic-block.
7287
1b275000
RB
72882013-08-29 Richard Biener <rguenther@suse.de>
7289
7290 PR middle-end/57287
7291 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7292 of default defs that appear in abnormal PHI nodes.
7293
6b1184ba
RB
72942013-08-29 Richard Biener <rguenther@suse.de>
7295
7296 PR tree-optimization/57685
7297 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7298 single-use operands to avoid exponential complexity.
7299
1b08b734
DC
73002013-08-28 Dehao Chen <dehao@google.com>
7301
7302 * ipa-inline.c (edge_badness): Fix integer underflow.
7303
48a3fa69
UB
73042013-08-28 Uros Bizjak <ubizjak@gmail.com>
7305
7306 * gtm-builtins.def (_ITM_free): Declare leaf.
7307
215f73e6
JJ
73082013-08-28 Jakub Jelinek <jakub@redhat.com>
7309
7310 PR target/58067
7311 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7312 (*tls_local_dynamic_base_64_largepic): Likewise.
7313 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7314 Remove predicate from call operand.
7315 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7316 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7317
34554d1a
JL
73182013-08-28 Jeff Law <law@redhat.com>
7319
7320 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7321 checks for the number of predecessors and successors allowed.
7322 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7323 which require copying a joiner block if there is a request which
7324 is a subpath that requires no joiner block copying.
7325
c01c111b
JH
73262013-08-28 Jan Hubicka <jh@suse.cz>
7327
7328 * lto-streamer-out.c (DFS_write_tree_body): Drop
7329 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7330 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7331 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7332 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7333 Do not read DECL_ERROR_ISSUED.
7334 (unpack_ts_decl_with_vis_value_fields): Do not read
7335 DECL_DEFER_OUTPUT.
48a3fa69
UB
7336 (lto_input_ts_binfo_tree_pointers): Do not read
7337 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
c01c111b
JH
7338 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7339 write DECL_ERROR_ISSUED..
7340 (pack_ts_decl_with_vis_value_fields): Do not write
7341 DECL_DEFER_OUTPUT.
7342 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
48a3fa69 7343 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
c01c111b
JH
7344 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7345 * tree.h (tree_decl_common): Update comment.
7346 (DECL_ERROR_ISSUED): Remove.
7347
4b128ece
JJ
73482013-08-28 Jakub Jelinek <jakub@redhat.com>
7349
7350 PR middle-end/58257
7351 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7352
12211b99 73532013-08-28 Jan Hubicka <jh@suse.cz>
9e57846a
JH
7354
7355 * builtins.def (free): Declare leaf.
7356
8dce4dbc
DM
73572013-08-27 David Malcolm <dmalcolm@redhat.com>
7358
7359 * gdbhooks.py: New.
7360 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7361 * configure: Regenerate.
7362
b8f6e610
MJ
73632013-08-27 Martin Jambor <mjambor@suse.cz>
7364
7365 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7366 (ipa_ancestor_jf_data): Likewise.
7367 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7368 (ipa_get_jf_pass_through_type_preserved): New function.
7369 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7370 (ipa_get_jf_ancestor_type_preserved): New function.
12211b99 7371 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
b8f6e610
MJ
7372 (ipa_get_jf_ancestor_result): Likewise.
7373 (propagate_vals_accross_pass_through): Use
7374 ipa_get_jf_pass_through_result to do all the value mappings.
7375 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7376 type_preserved flag.
7377 (ipa_set_jf_cst_copy): New function.
7378 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7379 (ipa_set_jf_arith_pass_through): Likewise.
7380 (ipa_set_ancestor_jf): Likewise.
7381 (compute_complex_assign_jump_func): Set type_preserved instead of
7382 punting.
7383 (ipa_compute_jump_functions_for_edge): Likewise.
7384 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7385 (update_jump_functions_after_inlining): Update type_preserved.
12211b99 7386 Explicitely create jump functions when combining one with pass_through.
b8f6e610
MJ
7387 (ipa_write_jump_function): Stream the type_preserved flags.
7388 (ipa_read_jump_function): Likewise.
7389
74bf76ed
JJ
73902013-08-27 Jakub Jelinek <jakub@redhat.com>
7391 Aldy Hernandez <aldyh@redhat.com>
7392
7393 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7394 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7395 * function.h (struct function): Add has_force_vect_loops and
7396 has_simduid_loops.
12211b99 7397 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
74bf76ed
JJ
7398 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7399 handle it.
7400 * gimple.def: Update CLAUSES comments.
7401 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7402 (gimple_build_omp_for): Add argument to prototype.
7403 (gimple_omp_for_kind): New.
7404 (gimple_omp_for_set_kind): New.
7405 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7406 GOVD_DATA_SHARE_CLASS.
7407 (enum omp_region_type): Add ORT_SIMD.
7408 (gimple_add_tmp_var): Handle ORT_SIMD.
7409 (gimplify_var_or_parm_decl): Same.
7410 (is_gimple_stmt): Same.
7411 (omp_firstprivatize_variable): Same.
7412 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7413 (omp_notice_variable): Handle ORT_SIMD.
12211b99 7414 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
74bf76ed
JJ
7415 (omp_check_private): Handle ORT_SIMD.
7416 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7417 OMP_CLAUSE_SAFELEN.
7418 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7419 Handle OMP_CLAUSE_LASTPRIVATE.
7420 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7421 OMP_CLAUSE_SAFELEN.
7422 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7423 (gimplify_expr): Handle OMP_SIMD.
7424 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7425 (expand_GOMP_SIMD_VF): New.
7426 (expand_GOMP_SIMD_LAST_LANE): New.
7427 * internal-fn.def (GOMP_SIMD_LANE): New.
7428 (GOMP_SIMD_VF): New.
7429 (GOMP_SIMD_LAST_LANE): New.
7430 * omp-low.c: Include target.h.
7431 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7432 OMP_CLAUSE_SAFELEN.
7433 (check_omp_nesting_restrictions): Same.
7434 (omp_max_vf): New.
7435 (lower_rec_simd_input_clauses): New.
7436 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7437 OMP_CLAUSE_LINEAR.
7438 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7439 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7440 (expand_omp_build_assign): New.
7441 (expand_omp_for_init_counts): New.
7442 (expand_omp_for_init_vars): New.
7443 (extract_omp_for_update_vars): New.
7444 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7445 and rewrite accordingly.
7446 (expand_omp_simd): New.
7447 (expand_omp_for): Use expand_omp_simd.
7448 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7449 (lower_omp_for): Do not lower the body.
7450 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7451 in their own loops.
7452 * tree-flow.h (find_omp_clause): Remove prototype.
12211b99
UB
7453 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7454 forcing vectorization of the loop, or if flag_tree_vectorize.
74bf76ed
JJ
7455 (gate_tree_if_conversion): Similarly.
7456 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7457 gimple_build_omp_for.
7458 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7459 * tree-parloops (create_parallel_loop): Pass kind argument to
7460 gimple_build_omp_for.
7461 * tree-pretty-print.c (dump_omp_clause): Add cases for
7462 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7463 OMP_CLAUSE__SIMDUID_.
7464 (dump_generic_node): Handle OMP_SIMD.
7465 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7466 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7467 unroll OMP_SIMD loops here.
12211b99 7468 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
74bf76ed 7469 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
12211b99 7470 loop->safelen.
74bf76ed
JJ
7471 (vect_analyze_data_refs): Handle simd loops.
7472 * tree-vect-loop.c (vectorizable_live_operation): Handle
7473 IFN_GOMP_SIMD*.
12211b99 7474 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
74bf76ed
JJ
7475 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7476 (vectorizable_load): Same.
12211b99 7477 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
74bf76ed
JJ
7478 (struct simduid_to_vf): New.
7479 (simduid_to_vf::hash): New.
7480 (simduid_to-vf::equal): New.
7481 (struct simd_array_to_simduid): New.
7482 (simd_array_to_simduid::hash): New.
7483 (simd_array_to_simduid::equal): New.
7484 (adjust_simduid_builtins): New.
7485 (struct note_simd_array_uses_struct): New.
7486 (note_simd_array_uses_cb): New.
7487 (note_simd_array_uses): New.
7488 (vectorize_loops): Handle simd hints and adjust simd builtins
7489 accordingly.
7490 * tree-vectorizer.h (struct _stmt_vec_info): Add
7491 simd_lane_access_p field.
7492 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7493 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7494 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7495 (omp_clause_code_name): Same.
7496 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7497 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7498 * tree.def (OMP_SIMD): New entry.
12211b99
UB
7499 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7500 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
74bf76ed
JJ
7501 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7502 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7503 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7504 (OMP_CLAUSE_LINEAR_STEP): New.
7505 (OMP_CLAUSE_SAFELEN_EXPR): New.
7506 (OMP_CLAUSE__SIMDUID__DECL): New.
7507 (find_omp_clause): New prototype.
7508
d0cf4e84
L
75092013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7510
7511 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7512 Haswell processor detection.
7513
992592ec
CW
75142013-08-27 Christian Widmer <shadow@umbrox.de>
7515
7516 PR target/57927
7517 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7518 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7519 AVX2 capable processors.
7520
9116eb22
TB
75212013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7522
7523 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7524 of vget_low_* with implementations in terms of other intrinsics.
7525
1c5abb53
MG
75262013-08-27 Marc Glisse <marc.glisse@inria.fr>
7527
7528 PR middle-end/57219
7529 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7530 values to -1, 0 and 1.
7531
2e100703
VP
75322013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7533
7534 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7535 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7536 (<optab><mode>3_insn): Remove.
7537 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7538 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7539 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7540 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7541 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7542 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7543 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7544 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7545 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7546 (ror<mode>3_insn): Likewise.
7547 * config/aarch64/predicates.md (aarch64_simd_register): New.
7548
4ded8276
RB
75492013-08-27 Richard Biener <rguenther@suse.de>
7550
7551 PR tree-optimization/57521
7552 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7553 one edge is non-critical.
7554 (find_phi_replacement_condition): Make sure to use a non-critical
7555 edge. Cleanup and remove old bug workarounds.
7556 (bb_postdominates_preds): Remove.
7557 (if_convertible_loop_p_1): Do not compute post-dominators.
7558 (combine_blocks): Do not free post-dominators.
7559 (main_tree_if_conversion): Likewise.
7560 (pass_data_if_conversion): Add TODO_verify_ssa.
7561
5aa11061
DD
75622013-08-27 DJ Delorie <dj@redhat.com>
7563
7564 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7565
fd91cfe3
YZ
75662013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7567
7568 * function.c (assign_parm_find_data_types): Set passed_mode and
7569 nominal_mode to the TYPE_MODE of nominal_type for the built
7570 pointer type in case of the struct-pass-by-reference.
7571
907555ce
JR
75722013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7573
7574 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7575 (UINT16_TYPE): Change default to "unsigned int".
7576
67518c93
JR
7577 * config/avr/avr.opt (mfract-convert-truncate): New option.
7578 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7579 is set, round negative fractional integers according to n1169
7580 when converting to integer types.
7581
d6d989fc
JH
75822013-08-26 Jan Hubicka <jh@suse.cz>
7583
7584 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7585 methods can not be called indirectly when their address is not taken.
7586
2aa3da06
JH
75872013-08-26 Jan Hubicka <jh@suse.cz>
7588
12211b99
UB
7589 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7590 ctor_for_folding.
2aa3da06 7591
0987ffe7
JH
75922013-08-26 Jan Hubicka <jh@suse.cz>
7593
7594 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7595 can be unshared.
7596
befe8647
JR
75972013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7598
7599 * reload.c (find_valid_class): Allow classes that do not include
7600 FIRST_PSEUDO_REGISTER - 1.
7601
a21e735e
JH
76022013-08-26 Jan Hubicka <jh@suse.cz>
7603
7604 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7605 fix edge count/frequency when speculation failed; fix type check
7606 for the direct call.
7607
e067bd43
JH
76082013-08-26 Jan Hubicka <jh@suse.cz>
7609
7610 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7611
e8aec975
JH
76122013-08-26 Jan Hubicka <jh@suse.cz>
7613
7614 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7615 to be changed by edge redirection.
7616
d0b66480
JH
76172013-08-26 Jan Hubicka <jh@suse.cz>
7618
12211b99
UB
7619 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7620 formating; add sanity check.
d0b66480
JH
7621 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7622 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7623 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7624 (resolve_noninline_speculation): Update callee keys, too.
7625
0f9aaac7
JH
76262013-08-26 Jan Hubicka <jh@suse.cz>
7627
7628 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7629 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7630
4b37444e
JR
76312013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7632
7633 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7634 into proper place.
7635
5339fc0c
UB
76362013-08-26 Uros Bizjak <ubizjak@gmail.com>
7637
7638 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7639 (x86_64_elf_select_section): Ditto.
7640 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7641 arguments.
7642 (ix86_pass_by_reference): Ditto.
5339fc0c
UB
7643 (output_set_got): Ditto.
7644 (ix86_unary_operator_ok): Ditto.
7645 (ix86_expand_builtin): Ditto.
7646
d5c3d3ef
JH
76472013-08-23 Jan Hubicka <jh@suse.cz>
7648
5339fc0c 7649 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
d5c3d3ef 7650
5b1e7435
JH
76512013-08-23 Jan Hubicka <jh@suse.cz>
7652
7653 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7654 (tree_decl_with_vis): Add FINAL field.
7655
ead69dac
JL
76562013-08-23 Jeff Law <law@redhat.com>
7657
7658 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7659 the debugging dump when the expression is fully redundant.
7660
025311c4
GDR
76612013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7662
7663 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7664 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7665 * pretty-print.c (pp_formatted_text_data): Likewise.
7666 (pp_write_text_to_stream): Likewise.
7667 (pp_write_text_as_dot_label_to_stream): Likewise.
7668 (pp_append_r): Likewise.
7669 (pp_format): Likewise.
7670 (pp_flush): Likewise.
7671 (pp_clear_output_area): Likewise.
7672 (pp_append_text): Likewise.
7673 (pp_formatted_text): Likewise.
7674 (pp_remaining_character_count_for_line): Likewise.
7675 (pp_newline): Likewise.
7676 (pp_character): Likewise.
7677 (output_buffer::~output_buffer): Define.
7678 (pretty_printer::~pretty_printer): Destruct output buffer.
7679 * pretty-print.h (output_buffer::~output_buffer): Declare.
7680 (pretty_printer::~pretty_printer): Declare virtual.
7681
0cadbfaa
MG
76822013-08-24 Marc Glisse <marc.glisse@inria.fr>
7683
7684 PR other/57324
7685 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7686 HOST_WIDE_INT_M1U): New macros.
7687 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7688 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7689 unsigned -1 for lshift.
7690 * cse.c (cse_insn): Likewise.
7691 * double-int.c (rshift_double, lshift_double): Likewise.
7692 * builtins.c (fold_builtin_bitop): Likewise.
7693 * combine.c (force_to_mode): Likewise.
7694 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7695 * simplify-rtx.c (simplify_const_unary_operation,
7696 simplify_const_binary_operation): Likewise.
7697 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7698 check_va_list_escapes): Likewise.
7699 * rtlanal.c (nonzero_bits1): Likewise.
7700 * expmed.c (expand_smod_pow2): Likewise.
7701 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7702
009e5353
JH
77032013-08-23 Jan Hubicka <jh@suse.cz>
7704
7705 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7706 as having address taken.
7707
c4be6568
JH
77082013-08-23 Jan Hubicka <jh@suse.cz>
7709
64cbf23d
JH
7710 * ipa-utils.h (method_class_type): Declare.
7711 * ipa-devirt.c (method_class_type): Export.
7712
c4be6568
JH
7713 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7714 do not walk base classes of anonymous types.
7715
97aba8e9
KK
77162013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7717
7718 PR rtl-optimization/58220
7719 PR regression/58221
7720 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7721 handle SEQUENCE insns properly.
7722
0fc80001
GDR
77232013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7724
7725 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7726 definition.
7727 (pp_newline_and_indent): Likewise.
7728 (pp_separate_with): Likewise.
7729 * pretty-print.c (pp_newline_and_flush): Define.
7730 (pp_newline_and_indent): Likewise.
7731 (pp_separate_with): Likewise.
7732
520a5868
JJ
77332013-08-23 Jakub Jelinek <jakub@redhat.com>
7734
7735 PR target/58218
7736 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7737 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7738
e500c62a
KY
77392013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7740
f43245d1
UB
7741 * config/i386/predicates.md (ext_sse_reg_operand): New.
7742 * config/i386/i386.md (*movti_internal): Use
e500c62a
KY
7743 predicate to determine if EVEX is needed.
7744 (*movsi_internal): Ditto.
7745 (*movdf_internal): Ditto.
7746 (*movsf_internal): Ditto.
f43245d1 7747 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
e500c62a 7748
6b00d7dd
JJ
77492013-08-23 Jakub Jelinek <jakub@redhat.com>
7750
7751 PR tree-optimization/58209
7752 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
7753 (find_tail_calls): Give up for pointer result types if m is non-NULL.
7754 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
7755 emit POINTER_PLUS_EXPR.
7756 (create_tailcall_accumulator): For pointer result type accumulate in
7757 sizetype type.
7758
4f2a9f90
PC
77592013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
7760
7761 * configure.ac: Add backslashes missing from the last change.
7762 * configure: Regenerate.
7763
7bf4274e
JH
77642013-08-22 Jan Hubicka <jh@susue.cz>
7765
4f2a9f90
PC
7766 * ipa.c (function_and_variable_visibility): First remember function
7767 was global and then make it local.
7bf4274e 7768
1bea243d
JB
77692013-08-22 Julian Brown <julian@codesourcery.com>
7770
7771 * configure.ac: Add aarch64 to list of arches which use "nop" in
7772 debug_line test.
7773 * configure: Regenerate.
7774
bcba7ecc
AK
77752013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7776
7777 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
7778 gnu_libc_has_function.
7779 * config/s390/tpf.h: Likewise.
7780
4c495b0d
JH
77812013-08-22 Jan Hubicka <jh@susue.cz>
7782
7783 * timevar.c (validate_phases): Add cast.
7784
5e302bca
JH
77852013-08-22 Jan Hubicka <jh@susue.cz>
7786
7787 * timevar.c (validate_phases): Use size_t for memory.
7788 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
7789
da6ca2b5
GDR
77902013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
7791
7792 * pretty-print.h (output_buffer::output_buffer): Declare.
7793 (pretty_printer::pretty_printer): Likewise.
7794 (pp_construct): Remove.
7795 * pretty-print.c (output_buffer::output_buffer): Define.
7796 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
7797 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
7798 (print_gimple_expr): Likewise.
7799 (print_gimple_seq): Likewise.
7800 (gimple_dump_bb): Likewise.
7801 * sched-vis.c (dump_value_slim): Likewise.
7802 (dump_insn_slim): Likewise.
7803 (dump_rtl_slim): Likewise.
7804 (str_pattern_slim): Likewise.
7805 * tree-mudflap.c (mf_varname_tree): Likewise.
7806 * graph.c (print_graph_cfg): Likewise.
7807 (start_graph_dump): Likewise.
7808 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
7809 placement-new.
7810 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
7811 pretty printer initialization.
7812 * coretypes.h (diagnostic_context): Remove superflous type alias
7813 declaration.
7814 (pretty_printer): Likewise. Declare directly as a class.
7815 (pretty_print_info): Remove declaration as class.
7816 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
7817 and pp_clear_output_area.
7818 (asan_add_global): Likewise.
7819
0e1474e5
JH
78202013-08-22 Jan Hubicka <jh@suse.cz>
7821
7822 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
7823 * ipa-utils.h (update_type_inheritance_graph): Declare.
7824 (possible_polymorphic_call_target_p): Declare.
7825 (possible_polymorphic_call_target_p): New.
7826 * ipa-devirt.c: Update toplevel comments.
7827 (cached_polymorphic_call_targets): Move up.
7828 (odr_type_d): Move ID down.
7829 (polymorphic_type_binfo_p): Update comment.
7830 (odr_hasher::remove): Likewise;
7831 (get_odr_type): Set anonymous_namespace.
7832 (dump_odr_type): Dump it.
7833 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
7834 (maybe_record_node): Record node in cached_polymorphic_call_targets.
7835 (record_binfo): Add comment.
5339fc0c
UB
7836 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
7837 built.
0e1474e5
JH
7838 (devirt_node_removal_hook): Do not iCE when cache is freed.
7839 (possible_polymorphic_call_target_p): New predicate.
7840 (update_type_inheritance_graph): New function.
7841
3f97cb0b
AI
78422013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
7843 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7844 Sergey Lega <sergey.s.lega@intel.com>
7845 Anna Tikhonova <anna.tikhonova@intel.com>
7846 Ilya Tocar <ilya.tocar@intel.com>
7847 Andrey Turetskiy <andrey.turetskiy@intel.com>
7848 Ilya Verbin <ilya.verbin@intel.com>
7849 Kirill Yukhin <kirill.yukhin@intel.com>
7850 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7851
7852 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
7853 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
7854 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
7855 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
7856 (OPTION_MASK_ISA_AVX2_UNSET): Update.
7857 (OPTION_MASK_ISA_AVX512F_UNSET): New.
7858 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
7859 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
7860 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
7861 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
7862 OPT_mavx512pf, OPT_mavx512er cases.
7863 * config/i386/constraints.md (v): New constraint.
7864 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
7865 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
7866 (bit_AVX512CD): New.
7867 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7868 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
7869 * config/i386/i386-c.c (ix86_target_macros_internal):
7870 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
7871 __AVX512PF__.
7872 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
7873 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
7874 * config/i386/i386.c (regclass_map, dbx_register_map)
7875 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
7876 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
7877 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
7878 -mavx512pf options.
7879 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
7880 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
7881 -mavx512pf options. Fix formatting.
7882 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
7883 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
7884 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
7885 -mavx512cd, -mavx512pf options.
7886 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
7887 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
7888 (ix86_preferred_output_reload_class): Replace SSE_REGS with
7889 ALL_SSE_REGS.
7890 (ix86_hard_regno_mode_ok): Support 512-bit registers.
7891 (ix86_set_reg_reg_cost): Ditto.
7892 (x86_order_regs_for_local_alloc): Ditto.
7893 (MAX_VECT_LEN): Extend to 64-byte.
7894 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
7895 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
7896 (TARGET_AVX512ER, TARGET_AVX512CD): New.
7897 (BIGGEST_ALIGNMENT): Extend to 512-bits.
7898 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
7899 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
7900 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
7901 (SSE_REG_MODE_P): Support new modes.
7902 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
7903 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
7904 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
7905 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
7906 (REG_CLASS_CONTENTS): Add new registers.
7907 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
7908 (EXT_REX_SSE_REGNO_P): New.
7909 (HI_REGISTER_NAMES): Add new registers.
7910 * config/i386/i386.md: Define constants for new registers.
7911 (mode): Add new 512-bit modes.
7912 (prefix): Support evex prefix.
7913 (isa): Support avx512f, noavx512f, fma_avx512f.
7914 (ssemodesuffix): Add new 512-bit modes.
7915 (movxi): New.
7916 (*movxi_internal_avx512f): Ditto.
7917 (*movdi_internal): Replace constraint "x" with the new constraint "v".
7918 Support MODE_XI.
7919 (*movsi_internal): Likewise.
7920 (*movdf_internal): Likewise.
7921 (*movsf_internal): Likewise.
7922 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
7923 (<code><mode>3): Likewise.
5339fc0c
UB
7924 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
7925 New.
3f97cb0b
AI
7926 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
7927 with the new constraint "v".
7928 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
7929 modes.
7930 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
7931 with the new constraint "v".
7932 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
7933 (<sse2>_storedqu<avxsizesuffix>): Likewise.
7934 (*<plusminus_insn><mode>3): Likewise.
7935 (<sse>_vm<plusminus_insn><mode>3): Likewise.
7936 (*mul<mode>3): Likewise.
7937 (<sse>_vmmul<mode>3): Likewise.
7938 (<sse>_div<mode>3): Likewise.
7939 (<sse>_vmdiv<mode>3): Likewise.
7940 (<sse>_sqrt<mode>2): Likewise.
7941 (<sse>_vmsqrt<mode>2): Likewise.
7942 (*<code><mode>3_finite): Likewise.
7943 (*<code><mode>3) <smaxmin>: Likewise.
7944 (<sse>_vm<code><mode>3): Likewise.
7945 (*<code><mode>3) <any_logic>: Likewise.
7946 (*fma_fmadd_<mode>): Likewise.
7947 (*fma_fmsub_<mode>): Likewise.
7948 (*fma_fnmadd_<mode>): Likewise.
7949 (*fma_fnmsub_<mode>): Likewise.
7950 (*fma_fmaddsub_<mode>): Likewise.
7951 (*fma_fmsubadd_<mode>): Likewise.
7952 (*fmai_fmadd_<mode>): Likewise.
7953 (*fmai_fmsub_<mode>): Likewise.
7954 (*fmai_fnmadd_<mode>): Likewise.
7955 (*fmai_fnmsub_<mode>): Likewise.
7956 (sse_cvtsi2ss): Likewise.
7957 (sse_cvtsi2ssq): Likewise.
7958 (sse_cvtss2si): Likewise.
7959 (sse_cvtss2si_2): Likewise.
7960 (sse_cvtss2siq): Likewise.
7961 (sse_cvtss2siq_2): Likewise.
7962 (sse_cvttss2si): Likewise.
7963 (sse_cvtss2siq_2): Likewise.
7964 (float<sseintvecmodelower><mode>2): Likewise.
7965 (sse2_cvtsd2si_2): Likewise.
7966 (sse2_cvtsd2siq_2): Likewise.
7967 (*<plusminus_insn><mode>3): Likewise.
7968 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
7969 (*<sse4_1_avx2>_mul<mode>3): Likewise.
7970 (ashr<mode>3): Likewise.
7971 (<shift_insn><mode>3): Likewise.
7972 (avx2_<code><mode>3): Likewise.
7973 (*avx2_<code><mode>3): Likewise.
7974 (*andnot<mode>3): Likewise.
7975 (*<code><mode>3) <any_logic>: Likewise.
7976 (abs<mode>2): Likewise.
7977 (avx2_permvar<mode>): Likewise.
7978 (avx2_perm<mode>_1): Likewise.
7979 (*avx_vpermilp<mode>): Likewise.
7980 (avx_vpermilvar<mode>3): Likewise.
7981 (avx2_ashrv<mode>): Likewise.
7982 (avx2_<shift_insn>v<mode>): Likewise.
7983 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
7984 -mavx512cd.
7985 * doc/rtl.texi: Document XImode.
7986
dd5e8423
JL
79872013-08-21 Jeff Law <law@redhat.com>
7988
b9ebee5d
JL
7989 * tree-flow.h (register_jump_thread): Pass vector of edges
7990 instead of each important edge.
7991 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
7992 thread path into a vector and pass that to register_jump_thread.
7993 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
7994 passed in edge vector to the current 3-edge form.
7995
dd5e8423
JL
7996 Revert:
7997 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7998
9bb6628e 7999 * dce.c (fini_dce): Call df_analyze again just in case
dd5e8423
JL
8000 delete_unmarked_insns removed anything.
8001
445dc8df
JR
80022013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
8003
8004 * reload.h (struct reg_equivs): Rename to ..
8005 (struct reg_equivs_s): .. this.
8006
e0df53dd
ML
80072013-08-20 Martin Liska <marxin.liska@gmail.com>
8008
8009 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
8010
e7d1d3eb
RO
80112013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
8012
5339fc0c 8013 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
e7d1d3eb 8014
5562e26e
JL
80152013-08-21 Jeff Law <law@redhat.com>
8016
8017 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
8018 simplify assignments too. If the RHS collapses to a singleton
8019 range, then return the value for the range.
8020
c7ecdec6
KY
80212013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
8022
7a5911d3
UB
8023 * config/i386/sse.md (V16): Rename to...
8024 (VMOVE): this.
8025 (mov<mode>): Update iterator name.
8026 (*mov<mode>_internal): Ditto.
8027 (push<mode>1): Ditto.
8028 (movmisalign<mode>): Ditto.
c7ecdec6 8029
bfa3b50a
JH
80302013-08-20 Jan Hubicka <jh@suse.cz>
8031
8032 PR bootstrap/58186
8033 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8034 entry for direct edges.
8035 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8036
6cd4d135
DM
80372013-08-20 David Malcolm <dmalcolm@redhat.com>
8038
8039 Revert my last two changes, r201865 and r201864:
8040
8041 Revert r201865:
8042 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8043
8044 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8045 instances can own GC refs.
8046
8047 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8048 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8049 (gcc::context::gt_pch_nx): Likewise.
8050 (gcc::context::gt_pch_nx): Likewise.
8051 * ggc.h (gt_ggc_mx <T>): New.
8052 (gt_pch_nx_with_op <T>): New.
8053 (gt_pch_nx <T>): New.
8054 * passes.c (opt_pass::gt_ggc_mx): New.
8055 (opt_pass::gt_pch_nx): New.
8056 (opt_pass::gt_pch_nx_with_op): New.
8057 (pass_manager::gt_ggc_mx): New.
8058 (pass_manager::gt_pch_nx): New.
8059 (pass_manager::gt_pch_nx_with_op): New.
8060 (pass_manager::operator new): Use
8061 ggc_internal_cleared_alloc_stat rather than xcalloc.
8062 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8063 (pass_manager::gt_ggc_mx): New.
8064 (pass_manager::gt_pch_nx): New.
8065 (pass_manager::gt_pch_nx_with_op): New.
8066 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8067 (opt_pass::operator new): New.
8068 (opt_pass::gt_ggc_mx): New.
8069 (opt_pass::gt_pch_nx): New.
8070 (opt_pass::gt_pch_nx_with_op): New.
8071
8072 Revert r201864:
8073 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8074
8075 * Makefile.in (GTFILES): Add context.h.
8076 * context.c (gcc::context::operator new): New.
8077 (gcc::context::gt_ggc_mx): New.
8078 (gcc::context::gt_pch_nx): New.
8079 (gcc::context::gt_pch_nx): New.
8080 * context.h (gcc::context): Add GTY((user)) marking.
8081 (gcc::context::operator new): New.
8082 (gcc::context::gt_ggc_mx): New.
8083 (gcc::context::gt_pch_nx): New.
8084 (gcc::context::gt_pch_nx): New.
8085 (g): Add GTY marking.
8086 (gt_ggc_mx (gcc::context *)): New.
8087 (gt_pch_nx (gcc::context *)): New.
8088 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8089 void *cookie)): New.
8090 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8091
a046136a
AM
80922013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8093
8094 * dce.c (fini_dce): Call df_analyze again just in case
8095 delete_unmarked_insns removed anything.
8096
67598720
TJ
80972013-08-20 Teresa Johnson <tejohnson@google.com>
8098
8099 PR rtl-optimizations/57451
8100 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8101 from crossing split section boundaries.
8102
f8693faf
MGD
81032013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8104
8105 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8106 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8107 with MULTLIB_DEFAULTS.
8108
7d0b9a9c
NC
81092013-08-20 Nick Clifton <nickc@redhat.com>
8110
8111 * target.def (narrow_volatile_bitfield): Note that the default
8112 value is false, not !TARGET_STRICT_ALIGN.
8113 * doc/tm.texi: Regenerate.
8114
8edb8dc8
PC
81152013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8116
8117 Fix LIB_SPEC for systems without libpthread.
8118
8119 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8120 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8121 for Android.
8122 * config/i386/linux-common.h: Likewise.
8123 * config/mips/linux-common.h: Likewise.
8124
12211b99 81252013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
d71ff3fa
ZZ
8126
8127 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8128 checks.
8129
e42c64cb
DM
81302013-08-20 David Malcolm <dmalcolm@redhat.com>
8131
8132 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8133 instances can own GC refs.
8134
8135 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8136 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8137 (gcc::context::gt_pch_nx): Likewise.
8138 (gcc::context::gt_pch_nx): Likewise.
8139 * ggc.h (gt_ggc_mx <T>): New.
8140 (gt_pch_nx_with_op <T>): New.
8141 (gt_pch_nx <T>): New.
8142 * passes.c (opt_pass::gt_ggc_mx): New.
8143 (opt_pass::gt_pch_nx): New.
8144 (opt_pass::gt_pch_nx_with_op): New.
8145 (pass_manager::gt_ggc_mx): New.
8146 (pass_manager::gt_pch_nx): New.
8147 (pass_manager::gt_pch_nx_with_op): New.
8148 (pass_manager::operator new): Use
8149 ggc_internal_cleared_alloc_stat rather than xcalloc.
8150 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8151 (pass_manager::gt_ggc_mx): New.
8152 (pass_manager::gt_pch_nx): New.
8153 (pass_manager::gt_pch_nx_with_op): New.
8154 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8155 (opt_pass::operator new): New.
8156 (opt_pass::gt_ggc_mx): New.
8157 (opt_pass::gt_pch_nx): New.
8158 (opt_pass::gt_pch_nx_with_op): New.
8159
910c02a0
DM
81602013-08-20 David Malcolm <dmalcolm@redhat.com>
8161
8162 * Makefile.in (GTFILES): Add context.h.
8163 * context.c (gcc::context::operator new): New.
8164 (gcc::context::gt_ggc_mx): New.
8165 (gcc::context::gt_pch_nx): New.
8166 (gcc::context::gt_pch_nx): New.
8167 * context.h (gcc::context): Add GTY((user)) marking.
8168 (gcc::context::operator new): New.
8169 (gcc::context::gt_ggc_mx): New.
8170 (gcc::context::gt_pch_nx): New.
8171 (gcc::context::gt_pch_nx): New.
8172 (g): Add GTY marking.
8173 (gt_ggc_mx (gcc::context *)): New.
8174 (gt_pch_nx (gcc::context *)): New.
8175 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8176 void *cookie)): New.
8177 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8178
af4b4236
AM
81792013-08-20 Alan Modra <amodra@gmail.com>
8180
8181 PR target/57865
8182 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8183 (rs6000_emit_epilogue): Likewise.
8184
12211b99 81852013-08-19 Dehao Chen <dehao@google.com>
0d633222
DC
8186
8187 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8188
e2323f5b
PB
81892013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8190 Jakub Jelinek <jakub@redhat.com>
8191
8192 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8193 (BUILT_IN_FABSD64): Likewise.
8194 (BUILT_IN_FABSD128): Likewise.
8195 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8196 (fold_builtin_1): Likewise.
5339fc0c
UB
8197 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8198 destination and source operands.
e2323f5b
PB
8199 (*abstd2_fpr): Likewise.
8200 (*nabstd2_fpr): Likewise.
8201
cfbf3ee8
RS
82022013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8203
8204 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8205 JUMP_P and INSN_P.
8206
d3136aeb
AH
82072013-08-19 Aldy Hernandez <aldyh@redhat.com>
8208
8209 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8210 incomplete.
8211
d33d9e47
AI
82122013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8213
8214 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8215 * builtins.c (default_libc_has_function): New.
8216 (gnu_libc_has_function): Ditto.
8217 (no_c99_libc_has_function): Ditto.
8218 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8219 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8220 (fold_builtin_sincos): Likewise.
8221 (fold_builtin_cexp): Likewise.
8222 * builtins.def (DEF_C94_BUILTIN): Likewise.
8223 (DEF_C99_BUILTIN): Likewise.
8224 (DEF_C99_C90RES_BUILTIN): Likewise.
8225 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8226 definitions to using this define.
8227 * config/darwin-protos.h (darwin_libc_has_function): New.
5339fc0c 8228 * config/darwin.c (darwin_libc_has_function): Ditto.
d33d9e47
AI
8229 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8230 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8231 * config/darwin.h: Ditto.
8232 * config/elfos.h: Ditto.
8233 * config/freebsd.h: Ditto.
8234 * config/i386/cygming.h: Ditto.
8235 * config/i386/djgpp.h: Ditto.
8236 * config/i386/i386-interix.h: Ditto.
8237 * config/microblaze/microblaze.h: Ditto.
8238 * config/mmix/mmix.h: Ditto.
8239 * config/gnu-user.h: Ditto.
8240 * config/ia64/hpux.h: Ditto.
8241 * config/pa/pa-hpux.h: Ditto.
8242 * config/pdp11/pdp11.h: Ditto.
8243 * config/picochip/picochip.h: Ditto.
8244 * config/linux.h: Ditto.
8245 * config/netbsd.h: Ditto.
8246 * config/openbsd.h: Ditto.
8247 * config/rs6000/aix43.h: Ditto.
8248 * config/rs6000/aix51.h: Ditto.
8249 * config/rs6000/aix52.h: Ditto.
8250 * config/rs6000/aix53.h: Ditto.
8251 * config/rs6000/aix61.h: Ditto.
8252 * config/rs6000/darwin.h: Ditto.
8253 * config/rs6000/linux.h: Ditto.
8254 * config/rs6000/linux64.h: Ditto.
8255 * config/s390/tpf.h: Ditto.
8256 * config/sol2-10.h: Ditto.
8257 * config/sol2.h: Ditto.
8258 * config/vms/vms.h: Ditto.
8259 * config/vxworks.h: Ditto.
8260 * config/linux-android.c (linux_android_libc_has_function):
8261 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8262 * config/linux-protos.h (linux_android_libc_has_function):
8263 New declaration.
8264 * config/i386/i386.c (ix86_libc_has_function): New.
8265 * config/i386/i386-protos.h
8266 (ix86_libc_has_function): New declaration.
8267 * config/i386/i386.md
8268 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8269 ("isinf<mode>2): Likewise.
8270 * convert.c (convert_to_integer): Using new target hook
8271 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8272 TARGET_C99_FUNCTIONS.
8273 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8274 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8275 * coretypes.h (function_class): New enum for different
8276 classes of functions.
8277 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8278 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8279 (TARGET_HAS_SINCOS): Likewise.
8280 (TARGET_LIBC_HAS_FUNCTION): New.
8281 * doc/tm.texi: Regenerated.
8282 * targhooks.h (default_libc_has_function): New declaration.
8283 (no_c99_libc_has_function): Ditto.
8284 (gnu_libc_has_function): Ditto.
8285 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8286 and TARGET_HAS_SINCOS.
8287
eefe9a99
JH
82882013-08-18 Jan Hubicka <jh@suse.cz>
8289
8290 * Makeifle-in (ipa-devirt.o): New.
8291 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8292 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8293 (analyze_functions): Look into possible targets of polymorphic call.
8294 * dumpfile.c (dump_files): Add type-inheritance dump.
8295 * dumpfile.h (TDI_inheritance): New.
8296 * ipa-devirt.c: New file.
8297 * ipa-utils.h (odr_type_d): Forward declare.
8298 (odr_type): New type.
8299 (build_type_inheritance_graph): Declare.
8300 (possible_polymorphic_call_targets): Declare and introduce inline
8301 variant when only edge is pased.
8302 (dump_possible_polymorphic_call_targets): Likewise.
8303 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8304 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8305 (types_same_for_odr): ... here.
8306 * tree.h (type_in_anonymous_namespace_p): Declare.
8307
29b89442
JJ
83082013-08-18 Jakub Jelinek <jakub@redhat.com>
8309
8310 PR tree-optimization/58006
8311 * tree-parloops.c (take_address_of): Don't ICE if get_name
8312 returns NULL.
8313 (eliminate_local_variables_stmt): Remove clobber stmts.
8314
4f219961
EB
83152013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8316
8317 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8318 error message is issued for an alias to undefined symbol.
8319
1d5755ef
JH
83202013-08-18 Jan Hubicka <jh@suse.cz>
8321
8322 * cgraph.c (cgraph_create_indirect_edge): Discover
8323 polymorphic calls and record basic info into indirect_info.
8324 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8325 devirtualization, ignore objc function calls.
8326 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8327 call with no parm index info.
8328 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8329 * tree.c (virtual_method_call_p): New function.
8330 * tree.h (virtual_method_call_p): Declare.
8331
5945bebf
JH
83322013-08-16 Jan Hubicka <jh@suse.cz>
8333
8334 PR middle-end/58179
8335 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8336
47fa5d23
DE
83372013-08-16 David Edelsohn <dje.gcc@gmail.com>
8338
8339 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8340 attribute.
8341
c74f54a0
DM
83422013-08-16 David Malcolm <dmalcolm@redhat.com>
8343
8344 * gengtype.c (type_for_name): Add special-case support for
8345 locating types within the "gcc::" namespace.
8346 (open_base_files): Emit a "using namespace gcc" directive.
8347
f758f299
MM
83482013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8349
8350 PR target/58160
8351 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8352 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8353
8354 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8355 array instead of each individual operand as a separate argument.
8356 (emit_fusion_gpr_load): Likewise.
8357 (expand_fusion_gpr_load): Add new function declaration.
8358
8359 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8360 signature to have the operands passed as an array, instead of as
8361 separate arguments. Allow ZERO_EXTEND to be in the memory
8362 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8363 depend on the register live/dead flags when peepholes are run.
8364 (expand_fusion_gpr_load): New function to be called from the
8365 peephole2 pass, to change the register that addis sets to be the
8366 target register.
8367 (emit_fusion_gpr_load): Change the calling signature to have the
8368 operands passed as an array, instead of as separate arguments.
8369 Allow ZERO_EXTEND to be in the memory address, and also
8370 SIGN_EXTEND if -mpower8-fusion-sign.
8371
8372 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8373 unspec enumeration.
8374 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8375 adjust the register addis loads up in the peephole2 pass. Do not
5339fc0c 8376 depend on the register live/dead state when the peephole pass is done.
f758f299 8377
158f4e4f
DM
83782013-08-16 David Malcolm <dmalcolm@redhat.com>
8379
8380 * gengtype.c (create_user_defined_type): Ensure that the kind
8381 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8382 declaration is seen before the GTY((user)) marking.
8383
22f8cea5
BE
83842013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8385
8386 PR target/58105
8387 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8388
c49bdb2e
JH
83892013-08-16 Jan Hubicka <jh@suse.cz>
8390
8391 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8392 arugment expected_type.
8393 (gimple_fold_call): Use it.
8394 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8395 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
22f8cea5 8396 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
c49bdb2e
JH
8397 (try_make_edge_direct_virtual_call): Likewise.
8398 * tree.c (obj_type_ref_class): New.
8399 * tree.h (obj_type_ref_class): Use it.
8400
4042dca9
GDR
84012013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8402
8403 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8404 (rtl_slim_pp): Likewise.
8405 (init_rtl_slim_pretty_print): Likewise.
8406 (dump_value_slim): Don't call it. Use local pretty printer.
8407 (dump_insn_slim): Likewise.
8408 (dump_rtl_slim): Likewise.
8409 (str_pattern_slim): Likewise.
8410 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8411 Simplify.
8412
7a460594
JJ
84132013-08-16 Jakub Jelinek <jakub@redhat.com>
8414
639dc669
JJ
8415 PR tree-optimization/58164
8416 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8417 walk gimple_goto_dest of GIMPLE_GOTO.
8418
7a460594
JJ
8419 PR tree-optimization/58165
8420 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8421 bi_call must be the last stmt in a bb, don't split_block, instead
8422 use fallthru edge from it and give up if there is none.
8423 Release conds vector when returning early.
8424
e3c02cc7
XDL
84252013-08-14 Xinliang David Li <davidxl@google.com>
8426
8427 * config/i386/i386.c (ix86_option_override_internal):
8428 Remove unused variable and field.
8429
6469da2c
BS
84302013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8431
8432 PR target/57949
22f8cea5 8433 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
6469da2c
BS
8434 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8435 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8436 and Linux, correct BLKmode alignment when 128-bit alignment is
8437 required and compatibility flag is not set.
22f8cea5
BE
8438 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8439 for zero-size arguments when compatibility flag is not set.
6469da2c 8440
4ca890e2
JJ
84412013-08-14 Jakub Jelinek <jakub@redhat.com>
8442
8443 PR tree-optimization/58145
8444 * tree-sra.c (build_ref_for_offset): If prev_base has
8445 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8446
daec1759 84472013-08-14 Xinliang David Li <davidxl@google.com>
4ca890e2 8448
daec1759 8449 * config/i386/i386.c (ix86_option_override_internal):
e3c02cc7 8450 Fix uninitialized variable error.
daec1759 8451
3ad20bd4
XDL
84522013-08-14 Xinliang David Li <davidxl@google.com>
8453
8454 * config/i386/i386.opt: Define two new options.
8455 * config/i386/x86-tune.def: Add arch selector field in macros.
8456 * config/i386/i386.h: Adjust macro definition.
8457 * config/i386/i386.c (ix86_option_override_internal):
8458 Refactor the code.
8459 (parse_mtune_ctrl_str): New function.
8460 (set_ix86_tune_features): New function.
8461 (ix86_function_specific_restore): Call the new helper function.
8462
e839e2a9
AB
84632013-08-14 Andrey Belevantsev <abel@ispras.ru>
8464
8465 PR rtl-optimization/57662
8466 * sel-sched.c (code_motion_process_successors): When the current insn
8467 is removed after the recursive traversal, break from the loop.
8468 Add comments and debug printouts.
8469
27ed665f
JJ
84702013-08-14 Jakub Jelinek <jakub@redhat.com>
8471 Alexandre Oliva <aoliva@redhat.com>
8472
8473 PR target/58067
8474 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8475 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8476 there also UNSPEC_PLTOFF.
8477
61c1a609
MP
84782013-08-14 Marek Polacek <polacek@redhat.com>
8479
8480 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8481 NUM_CONDITIONS bit positions.
8482
0e901c67
CC
84832013-08-13 Cary Coutant <ccoutant@google.com>
8484
8485 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8486 (attr_checksum): Hash vector contents instead of pointer.
8487 (attr_checksum_ordered): Likewise.
8488
061eff6d
UB
84892013-08-13 Uros Bizjak <ubizjak@gmail.com>
8490
8491 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8492 when Pmode != word_mode. Add length_address attribute.
8493 (sse3_monitor_<mode>): Merge from sse3_monitor and
8494 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8495 Pmode != word_mode. Update insn length attribute.
8496 * config/i386/i386.c (ix86_option_override_internal): Update
8497 ix86_gen_monitor selection for merged sse3_monitor insn.
8498
8ec9249b
JB
84992013-08-13 Julian Brown <julian@codesourcery.com>
8500
8501 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8502 perform invalid legitimization on greater-than-word-size modes for
8503 TARGET_E500_DOUBLE.
8504
761a8eb7
VM
85052013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8506
8507 * ira.c (setup_class_translate_array): Use aclass instead of cl
8508 for classes not fully covered by allocno classes.
8509
99c2bd54
JJ
85102013-08-13 Jakub Jelinek <jakub@redhat.com>
8511
4029a5e0
JJ
8512 PR tree-optimization/57661
8513 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8514 * tree-inline.c (tree_function_versioning): Initialize it.
8515 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8516 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8517 that is not being copied.
8518
99c2bd54
JJ
8519 PR sanitizer/56417
8520 * asan.c (instrument_strlen_call): Fix typo in comment.
8521 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8522
b5f8f063
SE
85232013-08-13 Steve Ellcey <sellcey@mips.com>
8524
8525 * config/mips/mips.md (prefetch): Use lw instead of ld on
8526 loongson in 32bit mode.
8527
4de80584
NC
85282013-08-13 Nick Clifton <nickc@redhat.com>
8529
8530 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8531
09ce3660
JH
85322013-08-13 Jan Hubicka <jh@suse.cz>
8533
8534 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8535 introduced edge; fix typo in sanity check.
8536 (cgraph_resolve_speculation): Export; improve diagnostic.
8537 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8538 speculation at type mismatch.
8539 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8540 (cgraph_resolve_speculation): Declare.
8541 (symtab_can_be_discarded): New function.
8542 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8543 * ipa-inline-transform.c (speculation_removed): New global var.
8544 (clone_inlined_nodes): See if speculation can be removed.
8545 (inline_call): If speculations was removed, we growths may not match.
8546 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8547 (speculation_useful_p): New function.
8548 (resolve_noninline_speculation): New function.
8549 (inline_small_functions): Resolve useless speculations.
8550 * ipa-inline.h (speculation_useful_p): Declare
8551 * ipa.c (can_replace_by_local_alias): Simplify.
8552 (ipa_profile): Produce speculative calls in non-lto, too;
8553 add simple cost model; produce local aliases.
8554
f971dc24
DM
85552013-08-13 David Malcolm <dmalcolm@redhat.com>
8556
8557 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
061eff6d 8558 PASS_MANAGER_H.
f971dc24 8559
2e507b9a
PC
85602013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8561
8562 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8563
c97fd4b8
UB
85642013-08-12 Uros Bizjak <ubizjak@gmail.com>
8565
8566 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8567 iterator instead of X87MODEF.
8568
85692013-08-12 Perez Read <netfirewall@gmail.com>
8570
8571 PR target/58132
8572 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8573 operand 0 for intel asm alternative.
8574 (*movabs<mode>_2): Ditto for operand 1.
8575
f4f4204c
JG
85762013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8577
8578 * config/aarch64/arm_none.h
8579 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8580
133b1a8e
NC
85812013-08-12 Nick Clifton <nickc@redhat.com>
8582
8583 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8584
15dd8b3a
YR
85852013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8586
8587 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8588 expand for QI/HImode operand to produce more effictive code for
8589 unsigned char(short) --> float(double) conversion.
8590
94109a6a
AM
85912013-08-12 Alexander Monakov <amonakov@ispras.ru>
8592
8593 * doc/invoke.texi: Mention that -ftls-model does not force the final
8594 model.
8595
4e115102
MP
85962013-08-12 Marek Polacek <polacek@redhat.com>
8597 Marc Glisse <marc.glisse@inria.fr>
8598
8599 PR tree-optimization/57980
8600 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8601 when creating -1 constant.
8602
748d71f3
JH
86032013-08-10 Jan Hubicka <jh@suse.cz>
8604
1f97e193 8605 Workaround binutils PR14342.
748d71f3
JH
8606 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8607 (gimple_init_edge_profiler): Likewise.
8608 (gimple_gen_ic_func_profiler): Likewise.
8609
5979aa54
JH
86102013-08-09 Jan Hubicka <jh@suse.cz>
8611
8612 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8613
ad83025e
XDL
86142013-08-09 Xinliang David Li <davidxl@google.com>
8615
8616 * config/i386/stringop.def: New file.
8617 * config/i386/stringop.opt: New file.
8618 * config/i386/i386-opts.h: Include stringopt.def.
8619 * config/i386/i386.opt: Include stringopt.opt.
8620 * config/i386/i386.c (ix86_option_override_internal):
c97fd4b8 8621 Override default size based stringop inline strategies with options.
ad83025e
XDL
8622 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8623 New function.
8624
e59df5fd
JH
86252013-08-09 Jan Hubicka <jh@suse.cz>
8626
c97fd4b8 8627 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
e59df5fd 8628
634ab819
JH
86292013-08-09 Jan Hubicka <jh@suse.cz>
8630
8631 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8632 CGRAPH_FREQ_MAX.
8633 (dump_cgraph_node): Dump profile-id.
8634 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8635 and common_target_probability.
8636 * lto-cgraph.c (lto_output_edge): Stream common targets.
8637 (lto_output_node): Stream profile ids.
8638 (input_node): Stream profile ids.
8639 (input_edge): Stream common targets.
8640 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8641 * ipa.c: Include value-prof.h
8642 (ipa_profile_generate_summary): Turn indirect call statement histograms
8643 into common targets.
8644 (ipa_profile): Turn common targets into speculative edges.
8645
2fa3d31b
JH
86462013-08-09 Jan Hubicka <jh@suse.cz>
8647
8648 * cgraph.h (cgraph_node): Add profile_id.
8649 * value-prof.c (cgraph_node_map): Turn into pointer_map.
c97fd4b8 8650 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
2fa3d31b
JH
8651 (del_node_map): Update.
8652 (find_func_by_funcdef_no): Replace by ...
8653 (find_func_by_profile_id): ... this one.
8654 (gimple_ic_transform): Do not remove useful histograms when
8655 speculation is not done; dump info when indirect call removal
8656 can happen at LTO.
8657 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8658 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8659 (__gcov_indirect_call_profiler_v2): .. this one.
8660 * profile.h (init_node_map): Update.
8661 * coverage.c (coverage_compute_profile_id): New function.
8662 * coverage.h (coverage_compute_profile_id): Declare.
8663 * tree-profile.c (init_ic_make_global_vars): Make
8664 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8665 (gimple_init_edge_profiler): Update prototype of
8666 __gcov_indirect_call_profiler.
8667 (gimple_gen_ic_func_profiler): Simplify.
8668 (tree_profiling): Use init_node_map
8669
042ae7d2
JH
86702013-08-09 Jan Hubicka <jh@suse.cz>
8671
c97fd4b8
UB
8672 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8673 non-speculative refs.
042ae7d2
JH
8674 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8675 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8676 (cgraph_set_call_stmt): Likewise.
8677 (cgraph_create_edge_1): Fix release checking compilatoin;
8678 clear lto_stmt_uid.
8679 (cgraph_free_edge): Free indirect info.
8680 (cgraph_turn_edge_to_speculative): New function.
8681 (cgraph_speculative_call_info): New function.
8682 (cgraph_make_edge_direct): Return direct edge; handle speculation.
c97fd4b8 8683 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
042ae7d2
JH
8684 (dump_cgraph_node): Dump speculation.
8685 (verify_edge_count_and_frequency): Accept speculative edges.
8686 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8687 (verify_cgraph_node): Handle speculation.
8688 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8689 (cgraph_set_call_stmt): Update prototype.
8690 (cgraph_make_edge_direct): Update prototype.
8691 (cgraph_speculative_call_info): Declare.
8692 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8693 to change; update call of ipa_find_references.
8694 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8695 lto_stmt_uid and speculative flags.
8696 (ipa_dump_references): Dump speculation.
8697 (ipa_clone_references): Clone speculative flag.
8698 (ipa_clone_referring): Likewise.
8699 (ipa_clone_ref): New function.
8700 (ipa_find_reference): Look into lto_stmt_uids
8701 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8702 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8703 (ipa_find_reference): Update declaration.
8704 (ipa_clone_ref): Declare.
8705 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8706 stream speculative flag.
8707 (lto_output_ref): Stream statements uids and speculation.
8708 (input_ref): Likewise.
8709 (input_edge): Stream speuclation.
8710 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8711 (cgraph_set_call_stmt_including_clones): Handle speculation.
8712 * ipa-inline.c (heap_edge_removal_hook): New function.
8713 (inline_small_functions): Register it.
8714 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8715 also initialize refs.
8716 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8717 edge to change.
8718 (try_make_edge_direct_simple_call): Likewise.
8719 (try_make_edge_direct_simple_call): Likewise.
8720 (update_indirect_edges_after_inlining): Likewise.
8721 (remove_described_reference): Look proper lto_stmt_uid.
8722 (propagate_controlled_uses): Likewise.
8723 (propagate_controlled_uses): Liekwise.
8724 * tree-inline.c (copy_bb): Copy speculative edges.
8725 (redirect_all_calls): New function.
c97fd4b8
UB
8726 (copy_cfg_body): Do redirection after loop info is updated.
8727 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
042ae7d2 8728
aa1e10cc
JH
87292013-08-09 Jan Hubicka <jh@suse.cz>
8730
8731 * lto-streamer-out.c (output_function): Renumber PHIs.
8732 * lto-streamer-in.c (input_function): Likewise.
8733
66adb8eb
JG
87342013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8735
8736 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8737 (get_lane_unsigned): Likewise.
8738 (dup_lane_scalar): Likewise.
8739 (get_lane): enable for VALL.
8740 * config/aarch64/aarch64-simd.md
8741 (aarch64_dup_lane_scalar<mode>): Remove.
8742 (aarch64_get_lane_signed<mode>): Likewise.
8743 (aarch64_get_lane_unsigned<mode>): Likewise.
8744 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8745 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8746 (aarch64_get_lane<mode>): Enable for all vector modes.
8747 (aarch64_get_lanedi): Remove misleading constraints.
8748 * config/aarch64/arm_neon.h
8749 (__aarch64_vget_lane_any): Define.
8750 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
8751 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
8752 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
8753 * config/aarch64/iterators.md (VDQQH): New.
8754 (VDQQHS): Likewise.
8755 (vwcore): Likewise.
8756
1f65ae7a
EB
87572013-08-09 Eric Botcazou <ebotcazou@adacore.com>
8758
8759 * configure.ac: Add GAS check for LEON instructions on SPARC.
8760 * configure: Regenerate.
8761 * config.in: Likewise.
8762 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
8763 sparc*-*-* block.
8764 * config/sparc/sparc.opt (LEON, LEON3): New masks.
8765 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
8766 for LEON or LEON3.
8767 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
8768 (AS_LEON_FLAG): New macro.
8769 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
8770 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
8771 Deal with LEON and LEON3 for the memory model.
23a6cb78 8772 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
1f65ae7a
EB
8773 (atomic_compare_and_swap<mode>_1): Likewise.
8774 (*atomic_compare_and_swap<mode>_1): Likewise.
8775
ff522f7f
ZC
87762013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8777
8778 * config/arm/neon.md (vcond): Fix floating-point vector
8779 comparisons against 0.
8780
cb1cca12
VM
87812013-08-08 Vladimir Makarov <vmakarov@redhat.com>
8782
8783 * lra-constraints.c (emit_spill_move): Remove assert.
8784 (process_alt_operands): Add more debugging
8785 output. Increase reject for spilling into memory. Decrease
8786 reject for reloading scratch.
8787 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
8788
a3719e31
SE
87892013-08-08 Steve Ellcey <sellcey@mips.com>
8790
8791 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
8792 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
8793 micromips incompatible. Add nan2008.
8794 (MULTILIB_DIRNAMES): Add nan2008.
8795 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8796 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
8797 and micromips incompatible. Add nan2008.
8798 (MULTILIB_DIRNAMES): Add nan2008.
8799 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8800
f40423e2
RS
88012013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
8802
8803 PR rtl-optimization/58079
8804 * combine.c (combine_simplify_rtx): Avoid using SUBST if
8805 simplify_comparison has widened a comparison with an integer.
8806
869b9125
KT
88072013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8808
8809 * config/arm/neon.md (movmisalign<mode>): Disable when we
8810 don't allow unaligned accesses.
8811 (*movmisalign<mode>_neon_store): Likewise.
8812 (*movmisalign<mode>_neon_load): Likewise.
8813 (*movmisalign<mode>_neon_store): Likewise.
8814 (*movmisalign<mode>_neon_load): Likewise.
8815
71cafea9
JH
88162013-08-08 Jan Hubicka <jh@suse.cz>
8817
8818 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
8819 (make_pass_rebuild_cgraph_edges): Also clear references.
8820 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
8821 * ipa-inline-transform.c (inline_transform): Remove all references
8822 after inlining.
c97fd4b8
UB
8823 * cgraphunit.c (expand_function): Remove all references after
8824 expansion.
71cafea9
JH
8825 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
8826 (ipa_find_reference): Rewrite to iterator.
8827 (remove_stmt_references): Likewise.
8828 (ipa_clear_stmts_in_references): New function.
8829 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
03b0ee0a
UB
8830 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
8831 clear references.
71cafea9
JH
8832 * ipa-split.c (split_function): Remove references in split function.
8833
27d2e612
RE
88342013-08-08 Richard Earnshaw <rearnsha@arm.com>
8835
8836 PR target/57431
1f65ae7a 8837 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
27d2e612
RE
8838 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
8839
5efc661c
RE
88402013-08-08 Richard Earnshaw <rearnsha@arm.com>
8841
8842 PR target/56979
1f65ae7a 8843 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
5efc661c
RE
8844 suggested mode for the assignment isn't compatible with the
8845 registers required.
8846
f276d31d
BE
88472013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8848
8849 PR target/58065
8850 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
8851
4b8bc035
XDL
88522013-08-07 Xinliang David Li <davidxl@google.com>
8853
8854 * config/i386/i386.opt: New option -mtune-ctrl=.
8855 * config/i386/x86-tune.def: New file.
8856 * config/i386/i386.h: include x86-tune.def.
8857 * config/i386/i386.c (ix86_option_override_internal):
8858 Parsing -mtune-ctrl= option and set tune features.
8859
12211b99 88602013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
a5b38740
OE
8861
8862 PR other/12081
03b0ee0a 8863 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
a5b38740
OE
8864 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
8865 to gen_2arg_fn_t.
8866
5cf6635b
EB
88672013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8868
8869 * rtl.h (update_alignments): Declare.
8870 * final.c (grow_label_align): New function extracted from...
8871 (shorten_branches): ...here. Call it.
8872 (update_alignments): New function.
8873 * reorg.c (sibling_labels): New variable.
8874 (get_label_before): Add SIBLING parameter. If it is non-zero, push
8875 the new label along with it onto the sibling_labels vector.
8876 (fill_simple_delay_slots): Adjust call to get_label_before.
8877 (fill_slots_from_thread): Likewise.
8878 (relax_delay_slots): Likewise.
8879 (make_return_insns): Likewise.
8880 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
8881
3c8ca1ab
EB
88822013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8883
8884 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
8885 document its semantics.
8886 (diagnostic_report_diagnostic): Adjust accordingly.
8887
5ee5b32c
DM
88882013-08-07 David Malcolm <dmalcolm@redhat.com>
8889
03b0ee0a 8890 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
5ee5b32c
DM
8891 (sparc_option_override): ...and port to new C++ pass API.
8892 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
8893
abd566fa
PB
88942013-08-07 Peter Bergner <bergner@vnet.ibm.com>
8895
8896 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
8897
2077db1b
CT
88982013-08-06 Caroline Tice <cmtice@google.com>
8899
8900 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
8901 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
8902 * tree-pass.h: Add pass_vtable_verify.
8903 * varasm.c (assemble_variable): Add code to properly set the comdat
8904 section and name for the .vtable_map_vars section.
8905 (assemble_vtyv_preinit_initializer): New function.
8906 (default_sectin_type_flags): Make sure .vtable_map_vars section has
8907 LINK_ONCE flag.
8908 * output.h: Add function decl for assemble_vtv_preinit_initializer.
8909 * vtable-verify.c: New file.
8910 * vtable-verify.h: New file.
8911 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
8912 initialiation levels.
8913 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
8914 * passes.def: Insert pass_vtable_verify.
8915 * aclocal.m4: Reorder includes.
8916 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
8917 -fvtv-counts options.
8918 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
8919 as appropriate, if -fvtable-verify=... is used.
8920 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
8921 -fvtable-verify=... is used.
8922 * Makefile.in (OBJS): Add vtable-verify.o to list.
8923 (vtable-verify.o): Add new build rule.
8924 (GTFILES): Add vtable-verify.c to list.
8925 * common.opt (fvtable-verify=): New flag.
8926 (vtv_priority): Values for fvtable-verify= flag.
8927 (fvtv-counts): New flag.
8928 (fvtv-debug): New flag.
8929 * tree.h (save_vtable_map_decl): New extern function decl.
8930
03085d1c
DM
89312013-08-07 David Malcolm <dmalcolm@redhat.com>
8932
8933 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
8934 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
8935 (pass_data_rl78_devirt): ...new pass_data instance and...
8936 (make_pass_rl78_devirt): ...new function.
8937 (rl78_asm_file_start): Port pass registration to new C++ API.
8938
05555c4a
DM
89392013-08-07 David Malcolm <dmalcolm@redhat.com>
8940
8941 * coretypes.h (rtl_opt_pass): Add.
8942 (gcc::context): Add.
8943 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
8944 (epiphany_init): Port to new C++ pass API.
8945 (epiphany_optimize_mode_switching): Likewise.
8946 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
8947 (pass_manager::get_pass_mode_switching): New.
8948 (pass_manager::get_pass_peephole2): New.
8949 * mode-switching.c (pass_mode_switching): Add clone method.
8950 * recog.c (pass_peephole2): Add clone method.
8951 (pass_split_all_insns): Add clone method.
8952
c5a12e26
DM
89532013-08-06 David Malcolm <dmalcolm@redhat.com>
8954
03b0ee0a
UB
8955 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
8956 (mips_option_override): ...here, porting to new C++ API for passes.
c5a12e26 8957
a2e2a668
JH
89582013-08-06 Jan Hubicka <jh@suse.cz>
8959
8960 * cgraph.c (cgraph_get_body): New function based on lto.c
8961 implementation.
8962 * cgraph.h (cgraph_get_body): Declare.
03b0ee0a
UB
8963 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
8964 LTO paths.
a2e2a668
JH
8965 * cgraphunit.c (expand_function): Get body prior expanding.
8966 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
03b0ee0a
UB
8967 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
8968 really need.
a2e2a668
JH
8969 * passes.c (do_per_function_toporder): Get body.
8970 * tree-inline.c (expand_call_inline): Get body prior inlining it.
8971 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
8972
15682f24
MJ
89732013-08-06 Martin Jambor <mjambor@suse.cz>
8974
8975 PR fortran/57987
8976 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
8977 is not re-finalized. Rename second parameter to no_collect.
8978
78f6dd68
MJ
89792013-08-06 Martin Jambor <mjambor@suse.cz>
8980
8981 PR middle-end/58041
8982 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
8983 MEM_REF has proper alignment information.
8984
12211b99 89852013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
eb923e54
OE
8986
8987 PR other/12081
8988 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
8989 class insn_gen_fn.
8990 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
8991 rtx (*) (rtx, ...) with insn_gen_fn.
8992 * genoutput.c (output_insn_data): Cast gen_? function pointers to
8993 insn_gen_fn::stored_funcptr. Add initializer braces.
8994
8ac69a6c
DM
89952013-08-05 David Malcolm <dmalcolm@redhat.com>
8996
8997 Rewrite how instances of passes are cloned to remove assumptions
8998 about their sizes (thus allowing pass subclasses to have
8999 additional data fields, albeit non-GC-managed ones at this point).
9000
9001 * passes.c (make_pass_instance): Now that passes have clone
9002 methods, rewrite this function to eliminate XNEW and memcpy
9003 calls that used hardcoded sizes. Since this function no longer
9004 creates pass instances, rename it to...
03b0ee0a
UB
9005 (add_pass_instance): ...this. Document the old way that passes were
9006 numbered and flagged, and rework this function to continue using it.
8ac69a6c
DM
9007 (next_pass_1): Add an initial_pass argument for use by
9008 add_pass_instance.
9009 (position_pass): When adding multiple instances of a pass, use
9010 the pass's clone method, rather than relying on the XNEW/memcpy
9011 within the former make_pass_instance (now add_pass_instance).
03b0ee0a
UB
9012 (pass_manager::pass_manager): When invoking next_pass_1, also supply
9013 the initial instance of the current pass within the pass manager.
8ac69a6c 9014
27a4cd48
DM
90152013-08-05 David Malcolm <dmalcolm@redhat.com>
9016
9017 This is the automated part of the conversion of passes from C
9018 structs to C++ classes.
9019
9020 Patch autogenerated by refactor_passes.py from
9021 https://github.com/davidmalcolm/gcc-refactoring-scripts
9022 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
9023
9024 * asan.c (pass_asan): Convert from a global struct to a subclass of
9025 gimple_opt_pass along with...
9026 (pass_data_asan): ...new pass_data instance and...
9027 (make_pass_asan): ...new function.
9028 (pass_asan_O0): Convert from a global struct to a subclass of
9029 gimple_opt_pass along with...
9030 (pass_data_asan_O0): ...new pass_data instance and...
9031 (make_pass_asan_O0): ...new function.
9032 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
9033 subclass of rtl_opt_pass along with...
9034 (pass_data_inc_dec): ...new pass_data instance and...
9035 (make_pass_inc_dec): ...new function.
9036 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9037 a subclass of rtl_opt_pass along with...
9038 (pass_data_reorder_blocks): ...new pass_data instance and...
9039 (make_pass_reorder_blocks): ...new function.
9040 (pass_duplicate_computed_gotos): Convert from a global struct to a
9041 subclass of rtl_opt_pass along with...
9042 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9043 (make_pass_duplicate_computed_gotos): ...new function.
9044 (pass_partition_blocks): Convert from a global struct to a subclass of
9045 rtl_opt_pass along with...
9046 (pass_data_partition_blocks): ...new pass_data instance and...
9047 (make_pass_partition_blocks): ...new function.
9048 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9049 struct to a subclass of rtl_opt_pass along with...
9050 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9051 and...
9052 (make_pass_branch_target_load_optimize1): ...new function.
9053 (pass_branch_target_load_optimize2): Convert from a global struct to a
9054 subclass of rtl_opt_pass along with...
9055 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9056 and...
9057 (make_pass_branch_target_load_optimize2): ...new function.
9058 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9059 of rtl_opt_pass along with...
9060 (pass_data_jump): ...new pass_data instance and...
9061 (make_pass_jump): ...new function.
9062 (pass_jump2): Convert from a global struct to a subclass of
9063 rtl_opt_pass along with...
9064 (pass_data_jump2): ...new pass_data instance and...
9065 (make_pass_jump2): ...new function.
9066 * cfgexpand.c (pass_expand): Convert from a global struct to a
9067 subclass of rtl_opt_pass along with...
9068 (pass_data_expand): ...new pass_data instance and...
9069 (make_pass_expand): ...new function.
9070 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9071 of rtl_opt_pass along with...
9072 (pass_data_free_cfg): ...new pass_data instance and...
9073 (make_pass_free_cfg): ...new function.
9074 (pass_into_cfg_layout_mode): Convert from a global struct to a
9075 subclass of rtl_opt_pass along with...
9076 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9077 (make_pass_into_cfg_layout_mode): ...new function.
9078 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9079 subclass of rtl_opt_pass along with...
9080 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9081 (make_pass_outof_cfg_layout_mode): ...new function.
9082 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9083 struct to a subclass of gimple_opt_pass along with...
9084 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9085 (make_pass_build_cgraph_edges): ...new function.
9086 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9087 subclass of gimple_opt_pass along with...
9088 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9089 (make_pass_rebuild_cgraph_edges): ...new function.
9090 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9091 subclass of gimple_opt_pass along with...
9092 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9093 and...
9094 (make_pass_remove_cgraph_callee_edges): ...new function.
9095 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9096 struct to a subclass of rtl_opt_pass along with...
9097 (pass_data_stack_adjustments): ...new pass_data instance and...
9098 (make_pass_stack_adjustments): ...new function.
9099 * combine.c (pass_combine): Convert from a global struct to a subclass
9100 of rtl_opt_pass along with...
9101 (pass_data_combine): ...new pass_data instance and...
9102 (make_pass_combine): ...new function.
9103 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9104 global struct to a subclass of rtl_opt_pass along with...
9105 (pass_data_compare_elim_after_reload): ...new pass_data instance
9106 and...
9107 (make_pass_compare_elim_after_reload): ...new function.
9108 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9109 of rtl_opt_pass along with...
9110 (pass_data_rtl_cprop): ...new pass_data instance and...
9111 (make_pass_rtl_cprop): ...new function.
9112 * cse.c (pass_cse): Convert from a global struct to a subclass of
9113 rtl_opt_pass along with...
9114 (pass_data_cse): ...new pass_data instance and...
9115 (make_pass_cse): ...new function.
9116 (pass_cse2): Convert from a global struct to a subclass of
9117 rtl_opt_pass along with...
9118 (pass_data_cse2): ...new pass_data instance and...
9119 (make_pass_cse2): ...new function.
9120 (pass_cse_after_global_opts): Convert from a global struct to a
9121 subclass of rtl_opt_pass along with...
9122 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9123 (make_pass_cse_after_global_opts): ...new function.
9124 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9125 of rtl_opt_pass along with...
9126 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9127 (make_pass_ud_rtl_dce): ...new function.
9128 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9129 rtl_opt_pass along with...
9130 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9131 (make_pass_fast_rtl_dce): ...new function.
9132 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9133 a subclass of rtl_opt_pass along with...
9134 (pass_data_df_initialize_opt): ...new pass_data instance and...
9135 (make_pass_df_initialize_opt): ...new function.
9136 (pass_df_initialize_no_opt): Convert from a global struct to a
9137 subclass of rtl_opt_pass along with...
9138 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9139 (make_pass_df_initialize_no_opt): ...new function.
9140 (pass_df_finish): Convert from a global struct to a subclass of
9141 rtl_opt_pass along with...
9142 (pass_data_df_finish): ...new pass_data instance and...
9143 (make_pass_df_finish): ...new function.
9144 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9145 rtl_opt_pass along with...
9146 (pass_data_rtl_dse1): ...new pass_data instance and...
9147 (make_pass_rtl_dse1): ...new function.
9148 (pass_rtl_dse2): Convert from a global struct to a subclass of
9149 rtl_opt_pass along with...
9150 (pass_data_rtl_dse2): ...new pass_data instance and...
9151 (make_pass_rtl_dse2): ...new function.
9152 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9153 subclass of rtl_opt_pass along with...
9154 (pass_data_dwarf2_frame): ...new pass_data instance and...
9155 (make_pass_dwarf2_frame): ...new function.
9156 * except.c (pass_set_nothrow_function_flags): Convert from a global
9157 struct to a subclass of rtl_opt_pass along with...
9158 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9159 and...
9160 (make_pass_set_nothrow_function_flags): ...new function.
9161 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9162 subclass of rtl_opt_pass along with...
9163 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9164 and...
9165 (make_pass_convert_to_eh_region_ranges): ...new function.
9166 * final.c (pass_compute_alignments): Convert from a global struct to a
9167 subclass of rtl_opt_pass along with...
9168 (pass_data_compute_alignments): ...new pass_data instance and...
9169 (make_pass_compute_alignments): ...new function.
9170 (pass_final): Convert from a global struct to a subclass of
9171 rtl_opt_pass along with...
9172 (pass_data_final): ...new pass_data instance and...
9173 (make_pass_final): ...new function.
9174 (pass_shorten_branches): Convert from a global struct to a subclass of
9175 rtl_opt_pass along with...
9176 (pass_data_shorten_branches): ...new pass_data instance and...
9177 (make_pass_shorten_branches): ...new function.
9178 (pass_clean_state): Convert from a global struct to a subclass of
9179 rtl_opt_pass along with...
9180 (pass_data_clean_state): ...new pass_data instance and...
9181 (make_pass_clean_state): ...new function.
9182 * function.c (pass_instantiate_virtual_regs): Convert from a global
9183 struct to a subclass of rtl_opt_pass along with...
9184 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9185 (make_pass_instantiate_virtual_regs): ...new function.
9186 (pass_leaf_regs): Convert from a global struct to a subclass of
9187 rtl_opt_pass along with...
9188 (pass_data_leaf_regs): ...new pass_data instance and...
9189 (make_pass_leaf_regs): ...new function.
9190 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9191 subclass of rtl_opt_pass along with...
9192 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9193 and...
9194 (make_pass_thread_prologue_and_epilogue): ...new function.
9195 (pass_match_asm_constraints): Convert from a global struct to a
9196 subclass of rtl_opt_pass along with...
9197 (pass_data_match_asm_constraints): ...new pass_data instance and...
9198 (make_pass_match_asm_constraints): ...new function.
9199 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9200 subclass of rtl_opt_pass along with...
9201 (pass_data_rtl_fwprop): ...new pass_data instance and...
9202 (make_pass_rtl_fwprop): ...new function.
9203 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9204 rtl_opt_pass along with...
9205 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9206 (make_pass_rtl_fwprop_addr): ...new function.
9207 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9208 rtl_opt_pass along with...
9209 (pass_data_rtl_pre): ...new pass_data instance and...
9210 (make_pass_rtl_pre): ...new function.
9211 (pass_rtl_hoist): Convert from a global struct to a subclass of
9212 rtl_opt_pass along with...
9213 (pass_data_rtl_hoist): ...new pass_data instance and...
9214 (make_pass_rtl_hoist): ...new function.
9215 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9216 subclass of gimple_opt_pass along with...
9217 (pass_data_lower_cf): ...new pass_data instance and...
9218 (make_pass_lower_cf): ...new function.
9219 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9220 from a global struct to a subclass of gimple_opt_pass along with...
9221 (pass_data_strength_reduction): ...new pass_data instance and...
9222 (make_pass_strength_reduction): ...new function.
9223 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9224 of rtl_opt_pass along with...
9225 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9226 (make_pass_rtl_ifcvt): ...new function.
9227 (pass_if_after_combine): Convert from a global struct to a subclass of
9228 rtl_opt_pass along with...
9229 (pass_data_if_after_combine): ...new pass_data instance and...
9230 (make_pass_if_after_combine): ...new function.
9231 (pass_if_after_reload): Convert from a global struct to a subclass of
9232 rtl_opt_pass along with...
9233 (pass_data_if_after_reload): ...new pass_data instance and...
9234 (make_pass_if_after_reload): ...new function.
9235 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9236 a subclass of rtl_opt_pass along with...
9237 (pass_data_initialize_regs): ...new pass_data instance and...
9238 (make_pass_initialize_regs): ...new function.
9239 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9240 of ipa_opt_pass_d along with...
9241 (pass_data_ipa_cp): ...new pass_data instance and...
9242 (make_pass_ipa_cp): ...new function.
9243 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9244 global struct to a subclass of gimple_opt_pass along with...
9245 (pass_data_inline_parameters): ...new pass_data instance and...
9246 (make_pass_inline_parameters): ...new function.
9247 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9248 subclass of gimple_opt_pass along with...
9249 (pass_data_early_inline): ...new pass_data instance and...
9250 (make_pass_early_inline): ...new function.
9251 (pass_ipa_inline): Convert from a global struct to a subclass of
9252 ipa_opt_pass_d along with...
9253 (pass_data_ipa_inline): ...new pass_data instance and...
9254 (make_pass_ipa_inline): ...new function.
9255 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9256 struct to a subclass of gimple_opt_pass along with...
9257 (pass_data_local_pure_const): ...new pass_data instance and...
9258 (make_pass_local_pure_const): ...new function.
9259 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9260 ipa_opt_pass_d along with...
9261 (pass_data_ipa_pure_const): ...new pass_data instance and...
9262 (make_pass_ipa_pure_const): ...new function.
9263 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9264 to a subclass of ipa_opt_pass_d along with...
9265 (pass_data_ipa_reference): ...new pass_data instance and...
9266 (make_pass_ipa_reference): ...new function.
9267 * ipa-split.c (pass_split_functions): Convert from a global struct to
9268 a subclass of gimple_opt_pass along with...
9269 (pass_data_split_functions): ...new pass_data instance and...
9270 (make_pass_split_functions): ...new function.
9271 (pass_feedback_split_functions): Convert from a global struct to a
9272 subclass of gimple_opt_pass along with...
9273 (pass_data_feedback_split_functions): ...new pass_data instance and...
9274 (make_pass_feedback_split_functions): ...new function.
9275 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9276 global struct to a subclass of simple_ipa_opt_pass along with...
9277 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9278 instance and...
9279 (make_pass_ipa_function_and_variable_visibility): ...new function.
9280 (pass_ipa_free_inline_summary): Convert from a global struct to a
9281 subclass of simple_ipa_opt_pass along with...
9282 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9283 (make_pass_ipa_free_inline_summary): ...new function.
9284 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9285 subclass of ipa_opt_pass_d along with...
9286 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9287 and...
9288 (make_pass_ipa_whole_program_visibility): ...new function.
9289 (pass_ipa_profile): Convert from a global struct to a subclass of
9290 ipa_opt_pass_d along with...
9291 (pass_data_ipa_profile): ...new pass_data instance and...
9292 (make_pass_ipa_profile): ...new function.
9293 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9294 ipa_opt_pass_d along with...
9295 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9296 (make_pass_ipa_cdtor_merge): ...new function.
9297 * ira.c (pass_ira): Convert from a global struct to a subclass of
9298 rtl_opt_pass along with...
9299 (pass_data_ira): ...new pass_data instance and...
9300 (make_pass_ira): ...new function.
9301 (pass_reload): Convert from a global struct to a subclass of
9302 rtl_opt_pass along with...
9303 (pass_data_reload): ...new pass_data instance and...
9304 (make_pass_reload): ...new function.
9305 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9306 subclass of rtl_opt_pass along with...
9307 (pass_data_cleanup_barriers): ...new pass_data instance and...
9308 (make_pass_cleanup_barriers): ...new function.
9309 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9310 of rtl_opt_pass along with...
9311 (pass_data_loop2): ...new pass_data instance and...
9312 (make_pass_loop2): ...new function.
9313 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9314 rtl_opt_pass along with...
9315 (pass_data_rtl_loop_init): ...new pass_data instance and...
9316 (make_pass_rtl_loop_init): ...new function.
9317 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9318 rtl_opt_pass along with...
9319 (pass_data_rtl_loop_done): ...new pass_data instance and...
9320 (make_pass_rtl_loop_done): ...new function.
9321 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9322 subclass of rtl_opt_pass along with...
9323 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9324 (make_pass_rtl_move_loop_invariants): ...new function.
9325 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9326 rtl_opt_pass along with...
9327 (pass_data_rtl_unswitch): ...new pass_data instance and...
9328 (make_pass_rtl_unswitch): ...new function.
9329 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9330 subclass of rtl_opt_pass along with...
9331 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9332 and...
9333 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9334 (pass_rtl_doloop): Convert from a global struct to a subclass of
9335 rtl_opt_pass along with...
9336 (pass_data_rtl_doloop): ...new pass_data instance and...
9337 (make_pass_rtl_doloop): ...new function.
9338 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9339 a subclass of rtl_opt_pass along with...
9340 (pass_data_lower_subreg): ...new pass_data instance and...
9341 (make_pass_lower_subreg): ...new function.
9342 (pass_lower_subreg2): Convert from a global struct to a subclass of
9343 rtl_opt_pass along with...
9344 (pass_data_lower_subreg2): ...new pass_data instance and...
9345 (make_pass_lower_subreg2): ...new function.
9346 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9347 struct to a subclass of ipa_opt_pass_d along with...
9348 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9349 (make_pass_ipa_lto_gimple_out): ...new function.
9350 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9351 of ipa_opt_pass_d along with...
9352 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9353 (make_pass_ipa_lto_finish_out): ...new function.
9354 * mode-switching.c (pass_mode_switching): Convert from a global struct
9355 to a subclass of rtl_opt_pass along with...
9356 (pass_data_mode_switching): ...new pass_data instance and...
9357 (make_pass_mode_switching): ...new function.
9358 * modulo-sched.c (pass_sms): Convert from a global struct to a
9359 subclass of rtl_opt_pass along with...
9360 (pass_data_sms): ...new pass_data instance and...
9361 (make_pass_sms): ...new function.
9362 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9363 subclass of gimple_opt_pass along with...
9364 (pass_data_expand_omp): ...new pass_data instance and...
9365 (make_pass_expand_omp): ...new function.
9366 (pass_lower_omp): Convert from a global struct to a subclass of
9367 gimple_opt_pass along with...
9368 (pass_data_lower_omp): ...new pass_data instance and...
9369 (make_pass_lower_omp): ...new function.
9370 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9371 of gimple_opt_pass along with...
9372 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9373 (make_pass_diagnose_omp_blocks): ...new function.
9374 * passes.c (pass_early_local_passes): Convert from a global struct to
9375 a subclass of simple_ipa_opt_pass along with...
9376 (pass_data_early_local_passes): ...new pass_data instance and...
9377 (make_pass_early_local_passes): ...new function.
9378 (pass_all_early_optimizations): Convert from a global struct to a
9379 subclass of gimple_opt_pass along with...
9380 (pass_data_all_early_optimizations): ...new pass_data instance and...
9381 (make_pass_all_early_optimizations): ...new function.
9382 (pass_all_optimizations): Convert from a global struct to a subclass
9383 of gimple_opt_pass along with...
9384 (pass_data_all_optimizations): ...new pass_data instance and...
9385 (make_pass_all_optimizations): ...new function.
9386 (pass_all_optimizations_g): Convert from a global struct to a subclass
9387 of gimple_opt_pass along with...
9388 (pass_data_all_optimizations_g): ...new pass_data instance and...
9389 (make_pass_all_optimizations_g): ...new function.
9390 (pass_rest_of_compilation): Convert from a global struct to a subclass
9391 of rtl_opt_pass along with...
9392 (pass_data_rest_of_compilation): ...new pass_data instance and...
9393 (make_pass_rest_of_compilation): ...new function.
9394 (pass_postreload): Convert from a global struct to a subclass of
9395 rtl_opt_pass along with...
9396 (pass_data_postreload): ...new pass_data instance and...
9397 (make_pass_postreload): ...new function.
9398 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9399 subclass of rtl_opt_pass along with...
9400 (pass_data_gcse2): ...new pass_data instance and...
9401 (make_pass_gcse2): ...new function.
9402 * postreload.c (pass_postreload_cse): Convert from a global struct to
9403 a subclass of rtl_opt_pass along with...
9404 (pass_data_postreload_cse): ...new pass_data instance and...
9405 (make_pass_postreload_cse): ...new function.
9406 * predict.c (pass_profile): Convert from a global struct to a subclass
9407 of gimple_opt_pass along with...
9408 (pass_data_profile): ...new pass_data instance and...
9409 (make_pass_profile): ...new function.
9410 (pass_strip_predict_hints): Convert from a global struct to a subclass
9411 of gimple_opt_pass along with...
9412 (pass_data_strip_predict_hints): ...new pass_data instance and...
9413 (make_pass_strip_predict_hints): ...new function.
9414 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9415 of rtl_opt_pass along with...
9416 (pass_data_peephole2): ...new pass_data instance and...
9417 (make_pass_peephole2): ...new function.
9418 (pass_split_all_insns): Convert from a global struct to a subclass of
9419 rtl_opt_pass along with...
9420 (pass_data_split_all_insns): ...new pass_data instance and...
9421 (make_pass_split_all_insns): ...new function.
9422 (pass_split_after_reload): Convert from a global struct to a subclass
9423 of rtl_opt_pass along with...
9424 (pass_data_split_after_reload): ...new pass_data instance and...
9425 (make_pass_split_after_reload): ...new function.
9426 (pass_split_before_regstack): Convert from a global struct to a
9427 subclass of rtl_opt_pass along with...
9428 (pass_data_split_before_regstack): ...new pass_data instance and...
9429 (make_pass_split_before_regstack): ...new function.
9430 (pass_split_before_sched2): Convert from a global struct to a subclass
9431 of rtl_opt_pass along with...
9432 (pass_data_split_before_sched2): ...new pass_data instance and...
9433 (make_pass_split_before_sched2): ...new function.
9434 (pass_split_for_shorten_branches): Convert from a global struct to a
9435 subclass of rtl_opt_pass along with...
9436 (pass_data_split_for_shorten_branches): ...new pass_data instance
9437 and...
9438 (make_pass_split_for_shorten_branches): ...new function.
9439 * ree.c (pass_ree): Convert from a global struct to a subclass of
9440 rtl_opt_pass along with...
9441 (pass_data_ree): ...new pass_data instance and...
9442 (make_pass_ree): ...new function.
9443 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9444 subclass of rtl_opt_pass along with...
9445 (pass_data_stack_regs): ...new pass_data instance and...
9446 (make_pass_stack_regs): ...new function.
9447 (pass_stack_regs_run): Convert from a global struct to a subclass of
9448 rtl_opt_pass along with...
9449 (pass_data_stack_regs_run): ...new pass_data instance and...
9450 (make_pass_stack_regs_run): ...new function.
9451 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9452 subclass of rtl_opt_pass along with...
9453 (pass_data_cprop_hardreg): ...new pass_data instance and...
9454 (make_pass_cprop_hardreg): ...new function.
9455 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9456 subclass of rtl_opt_pass along with...
9457 (pass_data_reginfo_init): ...new pass_data instance and...
9458 (make_pass_reginfo_init): ...new function.
9459 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9460 of rtl_opt_pass along with...
9461 (pass_data_regmove): ...new pass_data instance and...
9462 (make_pass_regmove): ...new function.
9463 * regrename.c (pass_regrename): Convert from a global struct to a
9464 subclass of rtl_opt_pass along with...
9465 (pass_data_regrename): ...new pass_data instance and...
9466 (make_pass_regrename): ...new function.
9467 * reorg.c (pass_delay_slots): Convert from a global struct to a
9468 subclass of rtl_opt_pass along with...
9469 (pass_data_delay_slots): ...new pass_data instance and...
9470 (make_pass_delay_slots): ...new function.
9471 (pass_machine_reorg): Convert from a global struct to a subclass of
9472 rtl_opt_pass along with...
9473 (pass_data_machine_reorg): ...new pass_data instance and...
9474 (make_pass_machine_reorg): ...new function.
9475 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9476 of rtl_opt_pass along with...
9477 (pass_data_sched): ...new pass_data instance and...
9478 (make_pass_sched): ...new function.
9479 (pass_sched2): Convert from a global struct to a subclass of
9480 rtl_opt_pass along with...
9481 (pass_data_sched2): ...new pass_data instance and...
9482 (make_pass_sched2): ...new function.
9483 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9484 to a subclass of rtl_opt_pass along with...
9485 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9486 (make_pass_stack_ptr_mod): ...new function.
9487 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9488 to a subclass of rtl_opt_pass along with...
9489 (pass_data_rtl_store_motion): ...new pass_data instance and...
9490 (make_pass_rtl_store_motion): ...new function.
9491 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9492 of gimple_opt_pass along with...
9493 (pass_data_tracer): ...new pass_data instance and...
9494 (make_pass_tracer): ...new function.
9495 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9496 to a subclass of gimple_opt_pass along with...
9497 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9498 (make_pass_diagnose_tm_blocks): ...new function.
9499 (pass_lower_tm): Convert from a global struct to a subclass of
9500 gimple_opt_pass along with...
9501 (pass_data_lower_tm): ...new pass_data instance and...
9502 (make_pass_lower_tm): ...new function.
9503 (pass_tm_init): Convert from a global struct to a subclass of
9504 gimple_opt_pass along with...
9505 (pass_data_tm_init): ...new pass_data instance and...
9506 (make_pass_tm_init): ...new function.
9507 (pass_tm_mark): Convert from a global struct to a subclass of
9508 gimple_opt_pass along with...
9509 (pass_data_tm_mark): ...new pass_data instance and...
9510 (make_pass_tm_mark): ...new function.
9511 (pass_tm_edges): Convert from a global struct to a subclass of
9512 gimple_opt_pass along with...
9513 (pass_data_tm_edges): ...new pass_data instance and...
9514 (make_pass_tm_edges): ...new function.
9515 (pass_tm_memopt): Convert from a global struct to a subclass of
9516 gimple_opt_pass along with...
9517 (pass_data_tm_memopt): ...new pass_data instance and...
9518 (make_pass_tm_memopt): ...new function.
9519 (pass_ipa_tm): Convert from a global struct to a subclass of
9520 simple_ipa_opt_pass along with...
9521 (pass_data_ipa_tm): ...new pass_data instance and...
9522 (make_pass_ipa_tm): ...new function.
9523 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9524 subclass of gimple_opt_pass along with...
9525 (pass_data_call_cdce): ...new pass_data instance and...
9526 (make_pass_call_cdce): ...new function.
9527 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9528 subclass of gimple_opt_pass along with...
9529 (pass_data_build_cfg): ...new pass_data instance and...
9530 (make_pass_build_cfg): ...new function.
9531 (pass_split_crit_edges): Convert from a global struct to a subclass of
9532 gimple_opt_pass along with...
9533 (pass_data_split_crit_edges): ...new pass_data instance and...
9534 (make_pass_split_crit_edges): ...new function.
9535 (pass_warn_function_return): Convert from a global struct to a
9536 subclass of gimple_opt_pass along with...
9537 (pass_data_warn_function_return): ...new pass_data instance and...
9538 (make_pass_warn_function_return): ...new function.
9539 (pass_warn_function_noreturn): Convert from a global struct to a
9540 subclass of gimple_opt_pass along with...
9541 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9542 (make_pass_warn_function_noreturn): ...new function.
9543 (pass_warn_unused_result): Convert from a global struct to a subclass
9544 of gimple_opt_pass along with...
9545 (pass_data_warn_unused_result): ...new pass_data instance and...
9546 (make_pass_warn_unused_result): ...new function.
9547 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9548 a subclass of gimple_opt_pass along with...
9549 (pass_data_merge_phi): ...new pass_data instance and...
9550 (make_pass_merge_phi): ...new function.
9551 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9552 a subclass of gimple_opt_pass along with...
9553 (pass_data_lower_complex): ...new pass_data instance and...
9554 (make_pass_lower_complex): ...new function.
9555 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9556 gimple_opt_pass along with...
9557 (pass_data_lower_complex_O0): ...new pass_data instance and...
9558 (make_pass_lower_complex_O0): ...new function.
9559 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9560 subclass of gimple_opt_pass along with...
9561 (pass_data_lower_eh): ...new pass_data instance and...
9562 (make_pass_lower_eh): ...new function.
9563 (pass_refactor_eh): Convert from a global struct to a subclass of
9564 gimple_opt_pass along with...
9565 (pass_data_refactor_eh): ...new pass_data instance and...
9566 (make_pass_refactor_eh): ...new function.
9567 (pass_lower_resx): Convert from a global struct to a subclass of
9568 gimple_opt_pass along with...
9569 (pass_data_lower_resx): ...new pass_data instance and...
9570 (make_pass_lower_resx): ...new function.
9571 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9572 of gimple_opt_pass along with...
9573 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9574 (make_pass_lower_eh_dispatch): ...new function.
9575 (pass_cleanup_eh): Convert from a global struct to a subclass of
9576 gimple_opt_pass along with...
9577 (pass_data_cleanup_eh): ...new pass_data instance and...
9578 (make_pass_cleanup_eh): ...new function.
9579 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9580 to a subclass of simple_ipa_opt_pass along with...
9581 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9582 (make_pass_ipa_lower_emutls): ...new function.
9583 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9584 a subclass of gimple_opt_pass along with...
9585 (pass_data_if_conversion): ...new pass_data instance and...
9586 (make_pass_if_conversion): ...new function.
9587 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9588 subclass of gimple_opt_pass along with...
9589 (pass_data_build_ssa): ...new pass_data instance and...
9590 (make_pass_build_ssa): ...new function.
9591 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9592 global struct to a subclass of gimple_opt_pass along with...
9593 (pass_data_loop_distribution): ...new pass_data instance and...
9594 (make_pass_loop_distribution): ...new function.
9595 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9596 subclass of gimple_opt_pass along with...
9597 (pass_data_mudflap_1): ...new pass_data instance and...
9598 (make_pass_mudflap_1): ...new function.
9599 (pass_mudflap_2): Convert from a global struct to a subclass of
9600 gimple_opt_pass along with...
9601 (pass_data_mudflap_2): ...new pass_data instance and...
9602 (make_pass_mudflap_2): ...new function.
9603 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9604 subclass of gimple_opt_pass along with...
9605 (pass_data_mudflap_1): ...new pass_data instance and...
9606 (make_pass_mudflap_1): ...new function.
9607 (pass_mudflap_2): Convert from a global struct to a subclass of
9608 gimple_opt_pass along with...
9609 (pass_data_mudflap_2): ...new pass_data instance and...
9610 (make_pass_mudflap_2): ...new function.
9611 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9612 gimple_opt_pass along with...
9613 (pass_data_nrv): ...new pass_data instance and...
9614 (make_pass_nrv): ...new function.
9615 (pass_return_slot): Convert from a global struct to a subclass of
9616 gimple_opt_pass along with...
9617 (pass_data_return_slot): ...new pass_data instance and...
9618 (make_pass_return_slot): ...new function.
9619 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9620 to a subclass of gimple_opt_pass along with...
9621 (pass_data_object_sizes): ...new pass_data instance and...
9622 (make_pass_object_sizes): ...new function.
9623 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9624 global struct to a subclass of gimple_opt_pass along with...
9625 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9626 and...
9627 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9628 (pass_fixup_cfg): Convert from a global struct to a subclass of
9629 gimple_opt_pass along with...
9630 (pass_data_fixup_cfg): ...new pass_data instance and...
9631 (make_pass_fixup_cfg): ...new function.
9632 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9633 (make_pass_mudflap_1): ...new function.
9634 (pass_mudflap_2): Replace declaration with that of...
9635 (make_pass_mudflap_2): ...new function.
9636 (pass_asan): Replace declaration with that of...
9637 (make_pass_asan): ...new function.
9638 (pass_asan_O0): Replace declaration with that of...
9639 (make_pass_asan_O0): ...new function.
9640 (pass_tsan): Replace declaration with that of...
9641 (make_pass_tsan): ...new function.
9642 (pass_tsan_O0): Replace declaration with that of...
9643 (make_pass_tsan_O0): ...new function.
9644 (pass_lower_cf): Replace declaration with that of...
9645 (make_pass_lower_cf): ...new function.
9646 (pass_refactor_eh): Replace declaration with that of...
9647 (make_pass_refactor_eh): ...new function.
9648 (pass_lower_eh): Replace declaration with that of...
9649 (make_pass_lower_eh): ...new function.
9650 (pass_lower_eh_dispatch): Replace declaration with that of...
9651 (make_pass_lower_eh_dispatch): ...new function.
9652 (pass_lower_resx): Replace declaration with that of...
9653 (make_pass_lower_resx): ...new function.
9654 (pass_build_cfg): Replace declaration with that of...
9655 (make_pass_build_cfg): ...new function.
9656 (pass_early_tree_profile): Replace declaration with that of...
9657 (make_pass_early_tree_profile): ...new function.
9658 (pass_cleanup_eh): Replace declaration with that of...
9659 (make_pass_cleanup_eh): ...new function.
9660 (pass_sra): Replace declaration with that of...
9661 (make_pass_sra): ...new function.
9662 (pass_sra_early): Replace declaration with that of...
9663 (make_pass_sra_early): ...new function.
9664 (pass_early_ipa_sra): Replace declaration with that of...
9665 (make_pass_early_ipa_sra): ...new function.
9666 (pass_tail_recursion): Replace declaration with that of...
9667 (make_pass_tail_recursion): ...new function.
9668 (pass_tail_calls): Replace declaration with that of...
9669 (make_pass_tail_calls): ...new function.
9670 (pass_tree_loop): Replace declaration with that of...
9671 (make_pass_tree_loop): ...new function.
9672 (pass_tree_loop_init): Replace declaration with that of...
9673 (make_pass_tree_loop_init): ...new function.
9674 (pass_lim): Replace declaration with that of...
9675 (make_pass_lim): ...new function.
9676 (pass_tree_unswitch): Replace declaration with that of...
9677 (make_pass_tree_unswitch): ...new function.
9678 (pass_predcom): Replace declaration with that of...
9679 (make_pass_predcom): ...new function.
9680 (pass_iv_canon): Replace declaration with that of...
9681 (make_pass_iv_canon): ...new function.
9682 (pass_scev_cprop): Replace declaration with that of...
9683 (make_pass_scev_cprop): ...new function.
9684 (pass_empty_loop): Replace declaration with that of...
9685 (make_pass_empty_loop): ...new function.
9686 (pass_record_bounds): Replace declaration with that of...
9687 (make_pass_record_bounds): ...new function.
9688 (pass_graphite): Replace declaration with that of...
9689 (make_pass_graphite): ...new function.
9690 (pass_graphite_transforms): Replace declaration with that of...
9691 (make_pass_graphite_transforms): ...new function.
9692 (pass_if_conversion): Replace declaration with that of...
9693 (make_pass_if_conversion): ...new function.
9694 (pass_loop_distribution): Replace declaration with that of...
9695 (make_pass_loop_distribution): ...new function.
9696 (pass_vectorize): Replace declaration with that of...
9697 (make_pass_vectorize): ...new function.
9698 (pass_slp_vectorize): Replace declaration with that of...
9699 (make_pass_slp_vectorize): ...new function.
9700 (pass_complete_unroll): Replace declaration with that of...
9701 (make_pass_complete_unroll): ...new function.
9702 (pass_complete_unrolli): Replace declaration with that of...
9703 (make_pass_complete_unrolli): ...new function.
9704 (pass_parallelize_loops): Replace declaration with that of...
9705 (make_pass_parallelize_loops): ...new function.
9706 (pass_loop_prefetch): Replace declaration with that of...
9707 (make_pass_loop_prefetch): ...new function.
9708 (pass_iv_optimize): Replace declaration with that of...
9709 (make_pass_iv_optimize): ...new function.
9710 (pass_tree_loop_done): Replace declaration with that of...
9711 (make_pass_tree_loop_done): ...new function.
9712 (pass_ch): Replace declaration with that of...
9713 (make_pass_ch): ...new function.
9714 (pass_ccp): Replace declaration with that of...
9715 (make_pass_ccp): ...new function.
9716 (pass_phi_only_cprop): Replace declaration with that of...
9717 (make_pass_phi_only_cprop): ...new function.
9718 (pass_build_ssa): Replace declaration with that of...
9719 (make_pass_build_ssa): ...new function.
9720 (pass_build_alias): Replace declaration with that of...
9721 (make_pass_build_alias): ...new function.
9722 (pass_build_ealias): Replace declaration with that of...
9723 (make_pass_build_ealias): ...new function.
9724 (pass_dominator): Replace declaration with that of...
9725 (make_pass_dominator): ...new function.
9726 (pass_dce): Replace declaration with that of...
9727 (make_pass_dce): ...new function.
9728 (pass_dce_loop): Replace declaration with that of...
9729 (make_pass_dce_loop): ...new function.
9730 (pass_cd_dce): Replace declaration with that of...
9731 (make_pass_cd_dce): ...new function.
9732 (pass_call_cdce): Replace declaration with that of...
9733 (make_pass_call_cdce): ...new function.
9734 (pass_merge_phi): Replace declaration with that of...
9735 (make_pass_merge_phi): ...new function.
9736 (pass_split_crit_edges): Replace declaration with that of...
9737 (make_pass_split_crit_edges): ...new function.
9738 (pass_pre): Replace declaration with that of...
9739 (make_pass_pre): ...new function.
9740 (pass_profile): Replace declaration with that of...
9741 (make_pass_profile): ...new function.
9742 (pass_strip_predict_hints): Replace declaration with that of...
9743 (make_pass_strip_predict_hints): ...new function.
9744 (pass_lower_complex_O0): Replace declaration with that of...
9745 (make_pass_lower_complex_O0): ...new function.
9746 (pass_lower_complex): Replace declaration with that of...
9747 (make_pass_lower_complex): ...new function.
9748 (pass_lower_vector): Replace declaration with that of...
9749 (make_pass_lower_vector): ...new function.
9750 (pass_lower_vector_ssa): Replace declaration with that of...
9751 (make_pass_lower_vector_ssa): ...new function.
9752 (pass_lower_omp): Replace declaration with that of...
9753 (make_pass_lower_omp): ...new function.
9754 (pass_diagnose_omp_blocks): Replace declaration with that of...
9755 (make_pass_diagnose_omp_blocks): ...new function.
9756 (pass_expand_omp): Replace declaration with that of...
9757 (make_pass_expand_omp): ...new function.
9758 (pass_expand_omp_ssa): Replace declaration with that of...
9759 (make_pass_expand_omp_ssa): ...new function.
9760 (pass_object_sizes): Replace declaration with that of...
9761 (make_pass_object_sizes): ...new function.
9762 (pass_strlen): Replace declaration with that of...
9763 (make_pass_strlen): ...new function.
9764 (pass_fold_builtins): Replace declaration with that of...
9765 (make_pass_fold_builtins): ...new function.
9766 (pass_stdarg): Replace declaration with that of...
9767 (make_pass_stdarg): ...new function.
9768 (pass_early_warn_uninitialized): Replace declaration with that of...
9769 (make_pass_early_warn_uninitialized): ...new function.
9770 (pass_late_warn_uninitialized): Replace declaration with that of...
9771 (make_pass_late_warn_uninitialized): ...new function.
9772 (pass_cse_reciprocals): Replace declaration with that of...
9773 (make_pass_cse_reciprocals): ...new function.
9774 (pass_cse_sincos): Replace declaration with that of...
9775 (make_pass_cse_sincos): ...new function.
9776 (pass_optimize_bswap): Replace declaration with that of...
9777 (make_pass_optimize_bswap): ...new function.
9778 (pass_optimize_widening_mul): Replace declaration with that of...
9779 (make_pass_optimize_widening_mul): ...new function.
9780 (pass_warn_function_return): Replace declaration with that of...
9781 (make_pass_warn_function_return): ...new function.
9782 (pass_warn_function_noreturn): Replace declaration with that of...
9783 (make_pass_warn_function_noreturn): ...new function.
9784 (pass_cselim): Replace declaration with that of...
9785 (make_pass_cselim): ...new function.
9786 (pass_phiopt): Replace declaration with that of...
9787 (make_pass_phiopt): ...new function.
9788 (pass_forwprop): Replace declaration with that of...
9789 (make_pass_forwprop): ...new function.
9790 (pass_phiprop): Replace declaration with that of...
9791 (make_pass_phiprop): ...new function.
9792 (pass_tree_ifcombine): Replace declaration with that of...
9793 (make_pass_tree_ifcombine): ...new function.
9794 (pass_dse): Replace declaration with that of...
9795 (make_pass_dse): ...new function.
9796 (pass_nrv): Replace declaration with that of...
9797 (make_pass_nrv): ...new function.
9798 (pass_rename_ssa_copies): Replace declaration with that of...
9799 (make_pass_rename_ssa_copies): ...new function.
9800 (pass_sink_code): Replace declaration with that of...
9801 (make_pass_sink_code): ...new function.
9802 (pass_fre): Replace declaration with that of...
9803 (make_pass_fre): ...new function.
9804 (pass_check_data_deps): Replace declaration with that of...
9805 (make_pass_check_data_deps): ...new function.
9806 (pass_copy_prop): Replace declaration with that of...
9807 (make_pass_copy_prop): ...new function.
9808 (pass_vrp): Replace declaration with that of...
9809 (make_pass_vrp): ...new function.
9810 (pass_uncprop): Replace declaration with that of...
9811 (make_pass_uncprop): ...new function.
9812 (pass_return_slot): Replace declaration with that of...
9813 (make_pass_return_slot): ...new function.
9814 (pass_reassoc): Replace declaration with that of...
9815 (make_pass_reassoc): ...new function.
9816 (pass_rebuild_cgraph_edges): Replace declaration with that of...
9817 (make_pass_rebuild_cgraph_edges): ...new function.
9818 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
9819 (make_pass_remove_cgraph_callee_edges): ...new function.
9820 (pass_build_cgraph_edges): Replace declaration with that of...
9821 (make_pass_build_cgraph_edges): ...new function.
9822 (pass_local_pure_const): Replace declaration with that of...
9823 (make_pass_local_pure_const): ...new function.
9824 (pass_tracer): Replace declaration with that of...
9825 (make_pass_tracer): ...new function.
9826 (pass_warn_unused_result): Replace declaration with that of...
9827 (make_pass_warn_unused_result): ...new function.
9828 (pass_diagnose_tm_blocks): Replace declaration with that of...
9829 (make_pass_diagnose_tm_blocks): ...new function.
9830 (pass_lower_tm): Replace declaration with that of...
9831 (make_pass_lower_tm): ...new function.
9832 (pass_tm_init): Replace declaration with that of...
9833 (make_pass_tm_init): ...new function.
9834 (pass_tm_mark): Replace declaration with that of...
9835 (make_pass_tm_mark): ...new function.
9836 (pass_tm_memopt): Replace declaration with that of...
9837 (make_pass_tm_memopt): ...new function.
9838 (pass_tm_edges): Replace declaration with that of...
9839 (make_pass_tm_edges): ...new function.
9840 (pass_split_functions): Replace declaration with that of...
9841 (make_pass_split_functions): ...new function.
9842 (pass_feedback_split_functions): Replace declaration with that of...
9843 (make_pass_feedback_split_functions): ...new function.
9844 (pass_strength_reduction): Replace declaration with that of...
9845 (make_pass_strength_reduction): ...new function.
9846 (pass_ipa_lower_emutls): Replace declaration with that of...
9847 (make_pass_ipa_lower_emutls): ...new function.
9848 (pass_ipa_function_and_variable_visibility): Replace declaration with
9849 that of...
9850 (make_pass_ipa_function_and_variable_visibility): ...new function.
9851 (pass_ipa_tree_profile): Replace declaration with that of...
9852 (make_pass_ipa_tree_profile): ...new function.
9853 (pass_early_local_passes): Replace declaration with that of...
9854 (make_pass_early_local_passes): ...new function.
9855 (pass_ipa_whole_program_visibility): Replace declaration with that
9856 of...
9857 (make_pass_ipa_whole_program_visibility): ...new function.
9858 (pass_ipa_lto_gimple_out): Replace declaration with that of...
9859 (make_pass_ipa_lto_gimple_out): ...new function.
9860 (pass_ipa_increase_alignment): Replace declaration with that of...
9861 (make_pass_ipa_increase_alignment): ...new function.
9862 (pass_ipa_inline): Replace declaration with that of...
9863 (make_pass_ipa_inline): ...new function.
9864 (pass_ipa_free_lang_data): Replace declaration with that of...
9865 (make_pass_ipa_free_lang_data): ...new function.
9866 (pass_ipa_free_inline_summary): Replace declaration with that of...
9867 (make_pass_ipa_free_inline_summary): ...new function.
9868 (pass_ipa_cp): Replace declaration with that of...
9869 (make_pass_ipa_cp): ...new function.
9870 (pass_ipa_reference): Replace declaration with that of...
9871 (make_pass_ipa_reference): ...new function.
9872 (pass_ipa_pure_const): Replace declaration with that of...
9873 (make_pass_ipa_pure_const): ...new function.
9874 (pass_ipa_pta): Replace declaration with that of...
9875 (make_pass_ipa_pta): ...new function.
9876 (pass_ipa_lto_finish_out): Replace declaration with that of...
9877 (make_pass_ipa_lto_finish_out): ...new function.
9878 (pass_ipa_tm): Replace declaration with that of...
9879 (make_pass_ipa_tm): ...new function.
9880 (pass_ipa_profile): Replace declaration with that of...
9881 (make_pass_ipa_profile): ...new function.
9882 (pass_ipa_cdtor_merge): Replace declaration with that of...
9883 (make_pass_ipa_cdtor_merge): ...new function.
9884 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
9885 of...
9886 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9887 (pass_init_datastructures): Replace declaration with that of...
9888 (make_pass_init_datastructures): ...new function.
9889 (pass_fixup_cfg): Replace declaration with that of...
9890 (make_pass_fixup_cfg): ...new function.
9891 (pass_expand): Replace declaration with that of...
9892 (make_pass_expand): ...new function.
9893 (pass_instantiate_virtual_regs): Replace declaration with that of...
9894 (make_pass_instantiate_virtual_regs): ...new function.
9895 (pass_rtl_fwprop): Replace declaration with that of...
9896 (make_pass_rtl_fwprop): ...new function.
9897 (pass_rtl_fwprop_addr): Replace declaration with that of...
9898 (make_pass_rtl_fwprop_addr): ...new function.
9899 (pass_jump): Replace declaration with that of...
9900 (make_pass_jump): ...new function.
9901 (pass_jump2): Replace declaration with that of...
9902 (make_pass_jump2): ...new function.
9903 (pass_lower_subreg): Replace declaration with that of...
9904 (make_pass_lower_subreg): ...new function.
9905 (pass_cse): Replace declaration with that of...
9906 (make_pass_cse): ...new function.
9907 (pass_fast_rtl_dce): Replace declaration with that of...
9908 (make_pass_fast_rtl_dce): ...new function.
9909 (pass_ud_rtl_dce): Replace declaration with that of...
9910 (make_pass_ud_rtl_dce): ...new function.
9911 (pass_rtl_dce): Replace declaration with that of...
9912 (make_pass_rtl_dce): ...new function.
9913 (pass_rtl_dse1): Replace declaration with that of...
9914 (make_pass_rtl_dse1): ...new function.
9915 (pass_rtl_dse2): Replace declaration with that of...
9916 (make_pass_rtl_dse2): ...new function.
9917 (pass_rtl_dse3): Replace declaration with that of...
9918 (make_pass_rtl_dse3): ...new function.
9919 (pass_rtl_cprop): Replace declaration with that of...
9920 (make_pass_rtl_cprop): ...new function.
9921 (pass_rtl_pre): Replace declaration with that of...
9922 (make_pass_rtl_pre): ...new function.
9923 (pass_rtl_hoist): Replace declaration with that of...
9924 (make_pass_rtl_hoist): ...new function.
9925 (pass_rtl_store_motion): Replace declaration with that of...
9926 (make_pass_rtl_store_motion): ...new function.
9927 (pass_cse_after_global_opts): Replace declaration with that of...
9928 (make_pass_cse_after_global_opts): ...new function.
9929 (pass_rtl_ifcvt): Replace declaration with that of...
9930 (make_pass_rtl_ifcvt): ...new function.
9931 (pass_into_cfg_layout_mode): Replace declaration with that of...
9932 (make_pass_into_cfg_layout_mode): ...new function.
9933 (pass_outof_cfg_layout_mode): Replace declaration with that of...
9934 (make_pass_outof_cfg_layout_mode): ...new function.
9935 (pass_loop2): Replace declaration with that of...
9936 (make_pass_loop2): ...new function.
9937 (pass_rtl_loop_init): Replace declaration with that of...
9938 (make_pass_rtl_loop_init): ...new function.
9939 (pass_rtl_move_loop_invariants): Replace declaration with that of...
9940 (make_pass_rtl_move_loop_invariants): ...new function.
9941 (pass_rtl_unswitch): Replace declaration with that of...
9942 (make_pass_rtl_unswitch): ...new function.
9943 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
9944 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9945 (pass_rtl_doloop): Replace declaration with that of...
9946 (make_pass_rtl_doloop): ...new function.
9947 (pass_rtl_loop_done): Replace declaration with that of...
9948 (make_pass_rtl_loop_done): ...new function.
9949 (pass_web): Replace declaration with that of...
9950 (make_pass_web): ...new function.
9951 (pass_cse2): Replace declaration with that of...
9952 (make_pass_cse2): ...new function.
9953 (pass_df_initialize_opt): Replace declaration with that of...
9954 (make_pass_df_initialize_opt): ...new function.
9955 (pass_df_initialize_no_opt): Replace declaration with that of...
9956 (make_pass_df_initialize_no_opt): ...new function.
9957 (pass_reginfo_init): Replace declaration with that of...
9958 (make_pass_reginfo_init): ...new function.
9959 (pass_inc_dec): Replace declaration with that of...
9960 (make_pass_inc_dec): ...new function.
9961 (pass_stack_ptr_mod): Replace declaration with that of...
9962 (make_pass_stack_ptr_mod): ...new function.
9963 (pass_initialize_regs): Replace declaration with that of...
9964 (make_pass_initialize_regs): ...new function.
9965 (pass_combine): Replace declaration with that of...
9966 (make_pass_combine): ...new function.
9967 (pass_if_after_combine): Replace declaration with that of...
9968 (make_pass_if_after_combine): ...new function.
9969 (pass_ree): Replace declaration with that of...
9970 (make_pass_ree): ...new function.
9971 (pass_partition_blocks): Replace declaration with that of...
9972 (make_pass_partition_blocks): ...new function.
9973 (pass_match_asm_constraints): Replace declaration with that of...
9974 (make_pass_match_asm_constraints): ...new function.
9975 (pass_regmove): Replace declaration with that of...
9976 (make_pass_regmove): ...new function.
9977 (pass_split_all_insns): Replace declaration with that of...
9978 (make_pass_split_all_insns): ...new function.
9979 (pass_fast_rtl_byte_dce): Replace declaration with that of...
9980 (make_pass_fast_rtl_byte_dce): ...new function.
9981 (pass_lower_subreg2): Replace declaration with that of...
9982 (make_pass_lower_subreg2): ...new function.
9983 (pass_mode_switching): Replace declaration with that of...
9984 (make_pass_mode_switching): ...new function.
9985 (pass_sms): Replace declaration with that of...
9986 (make_pass_sms): ...new function.
9987 (pass_sched): Replace declaration with that of...
9988 (make_pass_sched): ...new function.
9989 (pass_ira): Replace declaration with that of...
9990 (make_pass_ira): ...new function.
9991 (pass_reload): Replace declaration with that of...
9992 (make_pass_reload): ...new function.
9993 (pass_clean_state): Replace declaration with that of...
9994 (make_pass_clean_state): ...new function.
9995 (pass_branch_prob): Replace declaration with that of...
9996 (make_pass_branch_prob): ...new function.
9997 (pass_value_profile_transformations): Replace declaration with that
9998 of...
9999 (make_pass_value_profile_transformations): ...new function.
10000 (pass_postreload_cse): Replace declaration with that of...
10001 (make_pass_postreload_cse): ...new function.
10002 (pass_gcse2): Replace declaration with that of...
10003 (make_pass_gcse2): ...new function.
10004 (pass_split_after_reload): Replace declaration with that of...
10005 (make_pass_split_after_reload): ...new function.
10006 (pass_branch_target_load_optimize1): Replace declaration with that
10007 of...
10008 (make_pass_branch_target_load_optimize1): ...new function.
10009 (pass_thread_prologue_and_epilogue): Replace declaration with that
10010 of...
10011 (make_pass_thread_prologue_and_epilogue): ...new function.
10012 (pass_stack_adjustments): Replace declaration with that of...
10013 (make_pass_stack_adjustments): ...new function.
10014 (pass_peephole2): Replace declaration with that of...
10015 (make_pass_peephole2): ...new function.
10016 (pass_if_after_reload): Replace declaration with that of...
10017 (make_pass_if_after_reload): ...new function.
10018 (pass_regrename): Replace declaration with that of...
10019 (make_pass_regrename): ...new function.
10020 (pass_cprop_hardreg): Replace declaration with that of...
10021 (make_pass_cprop_hardreg): ...new function.
10022 (pass_reorder_blocks): Replace declaration with that of...
10023 (make_pass_reorder_blocks): ...new function.
10024 (pass_branch_target_load_optimize2): Replace declaration with that
10025 of...
10026 (make_pass_branch_target_load_optimize2): ...new function.
10027 (pass_leaf_regs): Replace declaration with that of...
10028 (make_pass_leaf_regs): ...new function.
10029 (pass_split_before_sched2): Replace declaration with that of...
10030 (make_pass_split_before_sched2): ...new function.
10031 (pass_compare_elim_after_reload): Replace declaration with that of...
10032 (make_pass_compare_elim_after_reload): ...new function.
10033 (pass_sched2): Replace declaration with that of...
10034 (make_pass_sched2): ...new function.
10035 (pass_stack_regs): Replace declaration with that of...
10036 (make_pass_stack_regs): ...new function.
10037 (pass_stack_regs_run): Replace declaration with that of...
10038 (make_pass_stack_regs_run): ...new function.
10039 (pass_df_finish): Replace declaration with that of...
10040 (make_pass_df_finish): ...new function.
10041 (pass_compute_alignments): Replace declaration with that of...
10042 (make_pass_compute_alignments): ...new function.
10043 (pass_duplicate_computed_gotos): Replace declaration with that of...
10044 (make_pass_duplicate_computed_gotos): ...new function.
10045 (pass_variable_tracking): Replace declaration with that of...
10046 (make_pass_variable_tracking): ...new function.
10047 (pass_free_cfg): Replace declaration with that of...
10048 (make_pass_free_cfg): ...new function.
10049 (pass_machine_reorg): Replace declaration with that of...
10050 (make_pass_machine_reorg): ...new function.
10051 (pass_cleanup_barriers): Replace declaration with that of...
10052 (make_pass_cleanup_barriers): ...new function.
10053 (pass_delay_slots): Replace declaration with that of...
10054 (make_pass_delay_slots): ...new function.
10055 (pass_split_for_shorten_branches): Replace declaration with that of...
10056 (make_pass_split_for_shorten_branches): ...new function.
10057 (pass_split_before_regstack): Replace declaration with that of...
10058 (make_pass_split_before_regstack): ...new function.
10059 (pass_convert_to_eh_region_ranges): Replace declaration with that
10060 of...
10061 (make_pass_convert_to_eh_region_ranges): ...new function.
10062 (pass_shorten_branches): Replace declaration with that of...
10063 (make_pass_shorten_branches): ...new function.
10064 (pass_set_nothrow_function_flags): Replace declaration with that of...
10065 (make_pass_set_nothrow_function_flags): ...new function.
10066 (pass_dwarf2_frame): Replace declaration with that of...
10067 (make_pass_dwarf2_frame): ...new function.
10068 (pass_final): Replace declaration with that of...
10069 (make_pass_final): ...new function.
10070 (pass_rtl_seqabstr): Replace declaration with that of...
10071 (make_pass_rtl_seqabstr): ...new function.
10072 (pass_release_ssa_names): Replace declaration with that of...
10073 (make_pass_release_ssa_names): ...new function.
10074 (pass_early_inline): Replace declaration with that of...
10075 (make_pass_early_inline): ...new function.
10076 (pass_inline_parameters): Replace declaration with that of...
10077 (make_pass_inline_parameters): ...new function.
10078 (pass_update_address_taken): Replace declaration with that of...
10079 (make_pass_update_address_taken): ...new function.
10080 (pass_convert_switch): Replace declaration with that of...
10081 (make_pass_convert_switch): ...new function.
10082 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10083 to a subclass of simple_ipa_opt_pass along with...
10084 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10085 (make_pass_ipa_tree_profile): ...new function.
10086 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10087 subclass of gimple_opt_pass along with...
10088 (pass_data_sra_early): ...new pass_data instance and...
10089 (make_pass_sra_early): ...new function.
10090 (pass_sra): Convert from a global struct to a subclass of
10091 gimple_opt_pass along with...
10092 (pass_data_sra): ...new pass_data instance and...
10093 (make_pass_sra): ...new function.
10094 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10095 gimple_opt_pass along with...
10096 (pass_data_early_ipa_sra): ...new pass_data instance and...
10097 (make_pass_early_ipa_sra): ...new function.
10098 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10099 subclass of gimple_opt_pass along with...
10100 (pass_data_ccp): ...new pass_data instance and...
10101 (make_pass_ccp): ...new function.
10102 (pass_fold_builtins): Convert from a global struct to a subclass of
10103 gimple_opt_pass along with...
10104 (pass_data_fold_builtins): ...new pass_data instance and...
10105 (make_pass_fold_builtins): ...new function.
10106 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10107 subclass of gimple_opt_pass along with...
10108 (pass_data_copy_prop): ...new pass_data instance and...
10109 (make_pass_copy_prop): ...new function.
10110 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10111 global struct to a subclass of gimple_opt_pass along with...
10112 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10113 (make_pass_rename_ssa_copies): ...new function.
10114 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10115 subclass of gimple_opt_pass along with...
10116 (pass_data_dce): ...new pass_data instance and...
10117 (make_pass_dce): ...new function.
10118 (pass_dce_loop): Convert from a global struct to a subclass of
10119 gimple_opt_pass along with...
10120 (pass_data_dce_loop): ...new pass_data instance and...
10121 (make_pass_dce_loop): ...new function.
10122 (pass_cd_dce): Convert from a global struct to a subclass of
10123 gimple_opt_pass along with...
10124 (pass_data_cd_dce): ...new pass_data instance and...
10125 (make_pass_cd_dce): ...new function.
10126 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10127 subclass of gimple_opt_pass along with...
10128 (pass_data_dominator): ...new pass_data instance and...
10129 (make_pass_dominator): ...new function.
10130 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10131 gimple_opt_pass along with...
10132 (pass_data_phi_only_cprop): ...new pass_data instance and...
10133 (make_pass_phi_only_cprop): ...new function.
10134 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10135 subclass of gimple_opt_pass along with...
10136 (pass_data_dse): ...new pass_data instance and...
10137 (make_pass_dse): ...new function.
10138 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10139 a subclass of gimple_opt_pass along with...
10140 (pass_data_forwprop): ...new pass_data instance and...
10141 (make_pass_forwprop): ...new function.
10142 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10143 struct to a subclass of gimple_opt_pass along with...
10144 (pass_data_tree_ifcombine): ...new pass_data instance and...
10145 (make_pass_tree_ifcombine): ...new function.
10146 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10147 subclass of gimple_opt_pass along with...
10148 (pass_data_ch): ...new pass_data instance and...
10149 (make_pass_ch): ...new function.
10150 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10151 subclass of gimple_opt_pass along with...
10152 (pass_data_tree_loop): ...new pass_data instance and...
10153 (make_pass_tree_loop): ...new function.
10154 (pass_tree_loop_init): Convert from a global struct to a subclass of
10155 gimple_opt_pass along with...
10156 (pass_data_tree_loop_init): ...new pass_data instance and...
10157 (make_pass_tree_loop_init): ...new function.
10158 (pass_lim): Convert from a global struct to a subclass of
10159 gimple_opt_pass along with...
10160 (pass_data_lim): ...new pass_data instance and...
10161 (make_pass_lim): ...new function.
10162 (pass_tree_unswitch): Convert from a global struct to a subclass of
10163 gimple_opt_pass along with...
10164 (pass_data_tree_unswitch): ...new pass_data instance and...
10165 (make_pass_tree_unswitch): ...new function.
10166 (pass_predcom): Convert from a global struct to a subclass of
10167 gimple_opt_pass along with...
10168 (pass_data_predcom): ...new pass_data instance and...
10169 (make_pass_predcom): ...new function.
10170 (pass_vectorize): Convert from a global struct to a subclass of
10171 gimple_opt_pass along with...
10172 (pass_data_vectorize): ...new pass_data instance and...
10173 (make_pass_vectorize): ...new function.
10174 (pass_graphite): Convert from a global struct to a subclass of
10175 gimple_opt_pass along with...
10176 (pass_data_graphite): ...new pass_data instance and...
10177 (make_pass_graphite): ...new function.
10178 (pass_graphite_transforms): Convert from a global struct to a subclass
10179 of gimple_opt_pass along with...
10180 (pass_data_graphite_transforms): ...new pass_data instance and...
10181 (make_pass_graphite_transforms): ...new function.
10182 (pass_check_data_deps): Convert from a global struct to a subclass of
10183 gimple_opt_pass along with...
10184 (pass_data_check_data_deps): ...new pass_data instance and...
10185 (make_pass_check_data_deps): ...new function.
10186 (pass_iv_canon): Convert from a global struct to a subclass of
10187 gimple_opt_pass along with...
10188 (pass_data_iv_canon): ...new pass_data instance and...
10189 (make_pass_iv_canon): ...new function.
10190 (pass_scev_cprop): Convert from a global struct to a subclass of
10191 gimple_opt_pass along with...
10192 (pass_data_scev_cprop): ...new pass_data instance and...
10193 (make_pass_scev_cprop): ...new function.
10194 (pass_record_bounds): Convert from a global struct to a subclass of
10195 gimple_opt_pass along with...
10196 (pass_data_record_bounds): ...new pass_data instance and...
10197 (make_pass_record_bounds): ...new function.
10198 (pass_complete_unroll): Convert from a global struct to a subclass of
10199 gimple_opt_pass along with...
10200 (pass_data_complete_unroll): ...new pass_data instance and...
10201 (make_pass_complete_unroll): ...new function.
10202 (pass_complete_unrolli): Convert from a global struct to a subclass of
10203 gimple_opt_pass along with...
10204 (pass_data_complete_unrolli): ...new pass_data instance and...
10205 (make_pass_complete_unrolli): ...new function.
10206 (pass_parallelize_loops): Convert from a global struct to a subclass
10207 of gimple_opt_pass along with...
10208 (pass_data_parallelize_loops): ...new pass_data instance and...
10209 (make_pass_parallelize_loops): ...new function.
10210 (pass_loop_prefetch): Convert from a global struct to a subclass of
10211 gimple_opt_pass along with...
10212 (pass_data_loop_prefetch): ...new pass_data instance and...
10213 (make_pass_loop_prefetch): ...new function.
10214 (pass_iv_optimize): Convert from a global struct to a subclass of
10215 gimple_opt_pass along with...
10216 (pass_data_iv_optimize): ...new pass_data instance and...
10217 (make_pass_iv_optimize): ...new function.
10218 (pass_tree_loop_done): Convert from a global struct to a subclass of
10219 gimple_opt_pass along with...
10220 (pass_data_tree_loop_done): ...new pass_data instance and...
10221 (make_pass_tree_loop_done): ...new function.
10222 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10223 struct to a subclass of gimple_opt_pass along with...
10224 (pass_data_cse_reciprocals): ...new pass_data instance and...
10225 (make_pass_cse_reciprocals): ...new function.
10226 (pass_cse_sincos): Convert from a global struct to a subclass of
10227 gimple_opt_pass along with...
10228 (pass_data_cse_sincos): ...new pass_data instance and...
10229 (make_pass_cse_sincos): ...new function.
10230 (pass_optimize_bswap): Convert from a global struct to a subclass of
10231 gimple_opt_pass along with...
10232 (pass_data_optimize_bswap): ...new pass_data instance and...
10233 (make_pass_optimize_bswap): ...new function.
10234 (pass_optimize_widening_mul): Convert from a global struct to a
10235 subclass of gimple_opt_pass along with...
10236 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10237 (make_pass_optimize_widening_mul): ...new function.
10238 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10239 subclass of gimple_opt_pass along with...
10240 (pass_data_phiopt): ...new pass_data instance and...
10241 (make_pass_phiopt): ...new function.
10242 (pass_cselim): Convert from a global struct to a subclass of
10243 gimple_opt_pass along with...
10244 (pass_data_cselim): ...new pass_data instance and...
10245 (make_pass_cselim): ...new function.
10246 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10247 subclass of gimple_opt_pass along with...
10248 (pass_data_phiprop): ...new pass_data instance and...
10249 (make_pass_phiprop): ...new function.
10250 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10251 subclass of gimple_opt_pass along with...
10252 (pass_data_pre): ...new pass_data instance and...
10253 (make_pass_pre): ...new function.
10254 (pass_fre): Convert from a global struct to a subclass of
10255 gimple_opt_pass along with...
10256 (pass_data_fre): ...new pass_data instance and...
10257 (make_pass_fre): ...new function.
10258 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10259 subclass of gimple_opt_pass along with...
10260 (pass_data_reassoc): ...new pass_data instance and...
10261 (make_pass_reassoc): ...new function.
10262 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10263 subclass of gimple_opt_pass along with...
10264 (pass_data_sink_code): ...new pass_data instance and...
10265 (make_pass_sink_code): ...new function.
10266 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10267 subclass of gimple_opt_pass along with...
10268 (pass_data_strlen): ...new pass_data instance and...
10269 (make_pass_strlen): ...new function.
10270 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10271 struct to a subclass of gimple_opt_pass along with...
10272 (pass_data_build_alias): ...new pass_data instance and...
10273 (make_pass_build_alias): ...new function.
10274 (pass_build_ealias): Convert from a global struct to a subclass of
10275 gimple_opt_pass along with...
10276 (pass_data_build_ealias): ...new pass_data instance and...
10277 (make_pass_build_ealias): ...new function.
10278 (pass_ipa_pta): Convert from a global struct to a subclass of
10279 simple_ipa_opt_pass along with...
10280 (pass_data_ipa_pta): ...new pass_data instance and...
10281 (make_pass_ipa_pta): ...new function.
10282 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10283 subclass of gimple_opt_pass along with...
10284 (pass_data_uncprop): ...new pass_data instance and...
10285 (make_pass_uncprop): ...new function.
10286 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10287 global struct to a subclass of gimple_opt_pass along with...
10288 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10289 (make_pass_late_warn_uninitialized): ...new function.
10290 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10291 to a subclass of gimple_opt_pass along with...
10292 (pass_data_init_datastructures): ...new pass_data instance and...
10293 (make_pass_init_datastructures): ...new function.
10294 (pass_early_warn_uninitialized): Convert from a global struct to a
10295 subclass of gimple_opt_pass along with...
10296 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10297 (make_pass_early_warn_uninitialized): ...new function.
10298 (pass_update_address_taken): Convert from a global struct to a
10299 subclass of gimple_opt_pass along with...
10300 (pass_data_update_address_taken): ...new pass_data instance and...
10301 (make_pass_update_address_taken): ...new function.
10302 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10303 struct to a subclass of gimple_opt_pass along with...
10304 (pass_data_release_ssa_names): ...new pass_data instance and...
10305 (make_pass_release_ssa_names): ...new function.
10306 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10307 subclass of gimple_opt_pass along with...
10308 (pass_data_stdarg): ...new pass_data instance and...
10309 (make_pass_stdarg): ...new function.
10310 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10311 global struct to a subclass of gimple_opt_pass along with...
10312 (pass_data_convert_switch): ...new pass_data instance and...
10313 (make_pass_convert_switch): ...new function.
10314 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10315 to a subclass of gimple_opt_pass along with...
10316 (pass_data_tail_recursion): ...new pass_data instance and...
10317 (make_pass_tail_recursion): ...new function.
10318 (pass_tail_calls): Convert from a global struct to a subclass of
10319 gimple_opt_pass along with...
10320 (pass_data_tail_calls): ...new pass_data instance and...
10321 (make_pass_tail_calls): ...new function.
10322 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10323 struct to a subclass of gimple_opt_pass along with...
10324 (pass_data_lower_vector): ...new pass_data instance and...
10325 (make_pass_lower_vector): ...new function.
10326 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10327 gimple_opt_pass along with...
10328 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10329 (make_pass_lower_vector_ssa): ...new function.
10330 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10331 to a subclass of gimple_opt_pass along with...
10332 (pass_data_slp_vectorize): ...new pass_data instance and...
10333 (make_pass_slp_vectorize): ...new function.
10334 (pass_ipa_increase_alignment): Convert from a global struct to a
10335 subclass of simple_ipa_opt_pass along with...
10336 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10337 (make_pass_ipa_increase_alignment): ...new function.
10338 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10339 gimple_opt_pass along with...
10340 (pass_data_vrp): ...new pass_data instance and...
10341 (make_pass_vrp): ...new function.
10342 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10343 subclass of simple_ipa_opt_pass along with...
10344 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10345 (make_pass_ipa_free_lang_data): ...new function.
10346 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10347 gimple_opt_pass along with...
10348 (pass_data_tsan): ...new pass_data instance and...
10349 (make_pass_tsan): ...new function.
10350 (pass_tsan_O0): Convert from a global struct to a subclass of
10351 gimple_opt_pass along with...
10352 (pass_data_tsan_O0): ...new pass_data instance and...
10353 (make_pass_tsan_O0): ...new function.
10354 * var-tracking.c (pass_variable_tracking): Convert from a global
10355 struct to a subclass of rtl_opt_pass along with...
10356 (pass_data_variable_tracking): ...new pass_data instance and...
10357 (make_pass_variable_tracking): ...new function.
10358 * web.c (pass_web): Convert from a global struct to a subclass of
10359 rtl_opt_pass along with...
10360 (pass_data_web): ...new pass_data instance and...
10361 (make_pass_web): ...new function.
10362 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10363 declaration with that of...
10364 (make_pass_mode_switch_use): ...new function.
10365 (pass_resolve_sw_modes): Replace declaration with that of...
10366 (make_pass_resolve_sw_modes): ...new function.
10367 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10368 from a global struct to a subclass of rtl_opt_pass along with...
10369 (pass_data_mode_switch_use): ...new pass_data instance and...
10370 (make_pass_mode_switch_use): ...new function.
10371 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10372 from a global struct to a subclass of rtl_opt_pass along with...
10373 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10374 (make_pass_resolve_sw_modes): ...new function.
10375 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10376 struct to a subclass of rtl_opt_pass along with...
10377 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10378 (make_pass_insert_vzeroupper): ...new function.
10379 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10380 global struct to a subclass of rtl_opt_pass along with...
10381 (pass_data_work_around_errata): ...new pass_data instance and...
10382 (make_pass_work_around_errata): ...new function.
10383 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10384 struct to a subclass of rtl_opt_pass along with...
10385 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10386 (make_pass_mips_machine_reorg2): ...new function.
10387
6a389ed5
DM
103882013-08-05 David Malcolm <dmalcolm@redhat.com>
10389
10390 * passes.c (pass_manager::operator new): New.
10391
f7695dbf
DM
103922013-08-05 David Malcolm <dmalcolm@redhat.com>
10393
10394 Handwritten part of conversion of passes to C++ classes.
10395
10396 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10397 (toplev.o): Add dep on PASS_MANAGER_H.
10398 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10399 of early local pases to reflect this moving from a global to a
10400 member of gcc::pass_manager.
10401 (cgraph_add_new_function): Likewise.
10402 * lto-cgraph.c (lto_output_node): Update for conversion of
10403 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10404 * passes.c (opt_pass::clone): New.
10405 (opt_pass::gate): New.
10406 (opt_pass::execute): New.
10407 (opt_pass::opt_pass): New.
10408 (pass_manager::execute_early_local_passes): New.
10409 (pass_manager::execute_pass_mode_switching): new.
10410 (finish_optimization_passes): Convert to...
10411 (pass_manager::finish_optimization_passes): ...this.
10412 (finish_optimization_passes): Update for conversion of passes to
10413 C++ classes.
10414 (register_dump_files_1): Use has_gate since we cannot portably
10415 check a vtable entry against NULL.
10416 (dump_one_pass): Likewise.
10417 (ipa_write_summaries_2): Likewise.
10418 (ipa_write_optimization_summaries_1): Likewise.
10419 (ipa_read_summaries_1): Likewise.
10420 (ipa_read_optimization_summaries_1): Likewise.
10421 (execute_ipa_stmt_fixups): Likewise.
10422 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10423 pass-creation functions rather than wiring up globals, and
10424 storing the results in fields of pass_manager generated using
10425 pass-instances.def.
10426 (pass_manager::dump_profile_report): Update for conversion of
10427 passes to C++ classes.
10428 (pass_manager::execute_ipa_summary_passes): Likewise.
10429 (execute_one_ipa_transform_pass): Likewise.
10430 (execute_one_pass): Use has_gate and has_execute since we cannot
10431 portably check a vtable entry against NULL.
10432 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10433 (pass_manager): Use pass-instances.def to add fields for the
10434 various pass instances.
10435 * toplev.c (finalize): Update for move of
10436 finish_optimization_passes to a method of gcc::pass_manager.
10437 * toplev.h (finish_optimization_passes): Move to method of class
10438 pass_manager.
10439 * tree-pass.h (struct pass_data): New.
03b0ee0a 10440 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
f7695dbf
DM
10441 (opt_pass::gate): Convert to virtual function.
10442 (opt_pass::~opt_pass): New.
10443 (opt_pass::clone): New.
10444 (opt_pass::execute): Convert to virtual function.
10445 (opt_pass::opt_pass): New.
10446 (opt_pass::ctxt_): new.
10447 (gimple_opt_pass): Convert to subclass of opt_pass.
10448 (gimple_opt_pass::gimple_opt_pass): New.
10449 (rtl_opt_pass): Convert to subclass of opt_pass.
10450 (rtl_opt_pass::rtl_opt_pass): New.
10451 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10452 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10453 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10454 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10455 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10456 invocation of pass_mode_switching to reflect this moving from a
10457 global to a member of gcc::pass_manager.
10458 (ix86_option_override): Rework how pass_insert_vzeroupper is
10459 added to the pass_manager to reflect autogenerated changes.
03b0ee0a 10460 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
f7695dbf 10461
d0b2f831
RE
104622013-08-05 Richard Earnshaw <rearnsha@arm.com>
10463
10464 PR rtl-optimization/57708
10465 * recog.c (peep2_find_free_register): Validate all regs in a
10466 multi-reg mode.
10467
51a5c0c2
JH
104682013-08-05 Jan Hubicka <jh@suse.cz>
10469
10470 PR lto/57602
03b0ee0a
UB
10471 * cgraph.c (verify_cgraph_node): Accept local flags from other
10472 partitions.
51a5c0c2
JH
10473 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10474 (function_and_variable_visibility): Likewise.
10475 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10476
65f0a120
GDR
104772013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10478
10479 * graph.c (init_graph_slim_pretty_print): Remove.
10480 (print_graph_cfg): Do not call it. Use local pretty printer.
10481 (start_graph_dump): Likewise.
10482
b3f80694
GDR
104832013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10484
10485 * gimple-pretty-print.c (buffer): Remove.
10486 (initialized): Likewise.
10487 (maybe_init_pretty_print): Likewise.
10488 (print_gimple_stmt): Do not call it. Use non-static local
10489 pretty_printer variable.
10490 (print_gimple_expr): Likewise.
10491 (print_gimple_seq): Likewise.
10492 (gimple_dump_bb): Likewise.
10493
11a877b3
GDR
104942013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10495
10496 * asan.c (asan_pp): Remove.
10497 (asan_pp_initialized): Likewise.
10498 (asan_pp_initialize): Likewise.
10499 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10500 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10501 (asan_add_global): Likewise.
10502
b066401f
GDR
105032013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10504
10505 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10506 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10507 * pretty-print.c (pp_formatted_text_data): Likewise.
10508 (pp_indent): Rename from pp_base_indent.
10509 (pp_format): Rename from pp_base_format.
10510 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10511 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10512 (pp_flush): Rename from pp_base_flush.
10513 (pp_set_line_maximum_length): Rename from
10514 pp_base_set_line_maximum_length.
10515 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10516 (pp_set_prefix): Rename from pp_base_set_prefix.
10517 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10518 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10519 (pp_append_text): Rename from pp_base_append_text.
10520 (pp_formatted_text): Rename from pp_base_formatted_text.
10521 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10522 (pp_remaining_character_count_for_line): Rename from
10523 pp_base_remaining_character_count_for_line.
10524 (pp_newline): Rename from pp_base_newline.
10525 (pp_character): Rename from pp_base_character.
10526 (pp_string): Rename from pp_base_string.
10527 (pp_maybe_space): Rename from pp_base_maybe_space.
10528 * asan.c (asan_pp_string): Adjust.
10529 (asan_emit_stack_protection): Likewise.
10530 (asan_add_global): Likewise.
10531 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10532 * tree-mudflap.c (mf_varname_tree): Likewise.
10533 * tree-pretty-print.c (pp_tree_identifier): Rename from
10534 pp_base_tree_identifier.
10535 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10536 Declare as function.
10537
137a1a27
GDR
105382013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10539
10540 * pretty-print.h (pp_bar_bar): New.
10541 (pp_ampersand_ampersand): Likewise.
10542 (pp_less_equal): Likewise.
10543 (pp_greater_equal): Likewise.
10544 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10545 printer functions instead of pp_string or operators and punctuators.
10546 (dump_gimple_call): Likewise.
10547 (dump_gimple_omp_for): Likewise.
10548 (dump_gimple_transaction): Likewise.
10549 (dump_gimple_phi): Likewise.
10550 (pp_gimple_stmt_1): Likewise.
10551 * sched-vis.c (print_insn): Likewise.
10552 * tree-mudflap.c (mf_varname_tree): Likewise.
10553 * tree-pretty-print.c (dump_block_node): Likewise.
10554 (dump_generic_node): Likewise.
10555
815effe1
JH
105562013-08-02 Jan Hubicka <jh@suse.cz>
10557
10558 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10559 boundaries.
10560 * lto-streamer-out.c (tree_is_indexable): Results decls and
10561 parm decls are not indexable.
10562 (DFS_write_tree_body): Do not follow args and results.
10563 (hash_tree): Likewise.
10564 (output_functions): Rearrange so struct function is needed
10565 only when real body is output; be able to also ouptut abstract
10566 functions; output DECL_ARGUMENTS and DECL_RESULT.
10567 (lto_output): When not in WPA, ale store abstract functions.
10568 (write_symbol): Do not care about RESULT_DECL.
10569 (output_symbol_p): Handle correctly sbtract decls.
10570 * lto-streamer-in.c (input_function): Rearrange so struct
10571 function can be NULL at entry; allow streaming of
10572 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10573 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10574 sanity check during LTO.
10575 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10576 RESULT_DECl and DECL_ARGUMENTS.
10577 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10578 Likewise.
10579
07838b13
GDR
105802013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10581
10582 * pretty-print.h (pp_underscore): New.
10583 (pp_comma): Tidy.
10584 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10585 printer functions instead of pp_character.
10586 (dump_binary_rhs): Likewise.
10587 (dump_ternary_rhs): Likewise.
10588 (dump_gimple_call_args): Likewise.
10589 (pp_points_to_solution): Likewise.
10590 (dump_gimple_call): Likewise.
10591 (dump_gimple_switch): Likewise.
10592 (dump_gimple_cond): Likewise.
10593 (dump_gimple_bind): Likewise.
10594 (dump_gimple_try): Likewise.
10595 (dump_gimple_omp_for): Likewise.
10596 (dump_gimple_omp_continue): Likewise.
10597 (dump_gimple_omp_single): Likewise.
10598 (dump_gimple_omp_sections): Likewise.
10599 (dump_gimple_omp_block): Likewise.
10600 (dump_gimple_omp_critical): Likewise.
10601 (dump_gimple_transaction): Likewise.
10602 (dump_gimple_asm): Likewise.
10603 (dump_gimple_phi): Likewise.
10604 (dump_gimple_omp_parallel): Likewise.
10605 (dump_gimple_omp_task): Likewise.
10606 (dump_gimple_omp_atomic_load): Likewise.
10607 (dump_gimple_omp_atomic_store): Likewise.
10608 (dump_gimple_mem_ops): Likewise.
10609 (pp_gimple_stmt_1): Likewise.
10610 (pp_cfg_jump): Likewise.
10611 (dump_implicit_edges): Likewise.
10612 (gimple_dump_bb_for_graph): Likewise.
10613 * graph.c (draw_cfg_node): Likewise.
10614 * langhooks.c (lhd_print_error_function): Likewise.
10615 * sched-vis.c (print_exp): Likewise.
10616 (print_value): Likewise.
10617 (print_pattern): Likewise.
10618 (print_insn): Likewise.
10619 (rtl_dump_bb_for_graph): Likewise.
10620 * tree-pretty-print.c (dump_function_declaration): Likewise.
10621 (dump_array_domain): Likewise.
10622 (dump_omp_clause): Likewise.
10623 (dump_location): Likewise.
10624 (dump_generic_node): Likewise.
10625 (print_struct_decl): Likewise.
10626 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10627
bb0d2039
BS
106282013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10629
10630 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10631 candidate table when replacing a candidate statement.
10632 (replace_rhs_if_not_dup): Likewise.
10633 (replace_one_candidate): Likewise.
10634
bc0ec027 106352013-08-02 Jan Hubicka <jh@suse.cz>
cdaa808a 10636 Martin Liska <marxin.liska@gmail.com>
bc0ec027
JH
10637
10638 * cgraphunit.c (add_new_function): Fix logic when adding from
10639 late IPA pass.
10640 (assemble_thunk): Rename to ...
03b0ee0a
UB
10641 (expand_thunk); .. this one; export; get it working with
10642 general functions; make produced gimple valid.
bc0ec027
JH
10643 * cgraph.h (expand_thunk): Declare.
10644
0e8853ee
JH
106452013-08-02 Jan Hubicka <jh@suse.cz>
10646
03b0ee0a
UB
10647 * ipa-cp.c (gather_context_independent_values): Use
10648 ipa_get_param_move_cost.
10649 (get_replacement_map): Remove PARAM; move parameter folding
10650 into tree-inline.c
0e8853ee
JH
10651 (create_specialized_node): Update.
10652 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10653 assert that we have gimple body; update move_cost.
10654 (count_formal_params): Assert that we have gimple body.
10655 (ipa_dump_param): New function.
10656 (ipa_alloc_node_params): Break out from ...
10657 (ipa_initialize_node_params): ... here.
10658 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10659 (ipa_write_node_info): Stream move costs.
10660 (ipa_read_node_info): Read move costs.
10661 (ipa_update_after_lto_read): Do not recompute node params.
10662 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10663 (ipa_get_param): Check we are not in WPA.
10664 (ipa_get_param_move_cost): New.
10665 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10666 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10667 parm numbers to be present.
10668
ecee672b
VM
106692013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10670
10671 PR rtl-optimization/58048
10672 * lra-constraints.c (process_alt_operands): Don't check asm
10673 operand on register.
10674
13ec0527
EB
106752013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10676
10677 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10678 the implied StoreLoad barrier for atomic operations if before.
10679
af15184a
JH
106802013-08-02 Jan Hubicka <jh@suse.cz>
10681 Martin Liska <marxin.liska@gmail.com>
10682
03b0ee0a
UB
10683 * cgraph.c (cgraph_function_body_availability): Do not check
10684 cgraph flags.
af15184a
JH
10685 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10686 symtab_node_availability): Declare.
10687 * ipa.c (can_replace_by_local_alias): New.
10688 (function_and_variable_visibility): Use it.
03b0ee0a
UB
10689 * symtab.c (symtab_for_node_and_aliases,
10690 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
af15184a 10691
01e54ef8
VM
106922013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10693
10694 PR rtl-optimization/57963
03b0ee0a 10695 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
01e54ef8
VM
10696 (lra_constraints): Use them.
10697
e2fd7ca7
SN
106982013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10699
03b0ee0a
UB
10700 * config/arm/types.md (define_attr "type"): Add "load_acq"
10701 and "store_rel".
e2fd7ca7
SN
10702 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10703 changes.
10704 (cortex_a53_store1): Likewise.
10705
0dcc26c3
JH
107062013-08-01 Jan Hubicka <jh@suse.cz>
10707
03b0ee0a
UB
10708 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10709 partitions are not needed.
0dcc26c3 10710
0bd72901
UB
107112013-08-01 Uros Bizjak <ubizjak@gmail.com>
10712
10713 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10714 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10715 MAYBE_NON_Q_CLASS_P where appropriate.
10716
12123452
JH
107172013-08-01 Jan Hubicka <jh@suse.cz>
10718
10719 * cgraph.h (release_function_body): Declare.
10720 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10721 of unused delcarations.
10722
ce852f9c
KT
107232013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10724
03b0ee0a
UB
10725 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10726 RTL form when subtracting a constant.
ce852f9c 10727
c743b246
KT
107282013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10729
10730 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10731 Generate canonical plus rtx with negated immediate instead of minus
10732 where appropriate.
10733 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10734
c0c123ef
JH
107352013-08-01 Jan Hubicka <jh@suse.cz>
10736
10737 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10738 (cgraph_release_function_body): Likewise.
10739 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10740 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10741 to used_as_abstract_origin.
10742 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10743 symbols used as abstract origins.
10744 * cgraphunit.c (analyze_functions): Update.
0bd72901
UB
10745 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10746 used_as_abstract_origin.
c0c123ef 10747 * tree-inline.c (tree_function_versioning): Update
0bd72901
UB
10748 used_as_abstract_origin; be ready for DECL_RESULT and
10749 DECL_ARGUMENTS to be NULL.
c0c123ef 10750
0bd72901
UB
10751 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
10752 for abstract functions.
10753 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
10754 real symbols.
c0c123ef 10755
1f029433
JH
107562013-08-01 Jan Hubicka <jh@suse.cz>
10757
10758 * profile.c (compute_value_histograms): Fix thinko.
10759
c451f4d6
SN
107602013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10761
10762 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
10763 aarch-common-protos.h to extra_headers.
10764 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10765 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
10766 * config/aarch64/t-aarch64 (aarch-common.o): Define.
10767
34b512f5
SN
107682013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10769
10770 * config/aarch64/aarch64.md (define_attr "type"): Delete.
10771 Include "../arm/types.md". Define "type" attribute for all patterns.
10772 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
10773 attribute changes.
10774
d86e633a
MM
107752013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
10776
10777 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
10778 to support power8 load fusion.
10779 (fusion_gpr_mem_load): Likewise.
10780
10781 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
10782
10783 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
10784 declarations for power8 load fusion.
10785 (emit_fusion_gpr_load): Likewise.
10786
10787 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
10788 tuning for power8, turn on fusion mode by default. Turn on sign
10789 extending fusion mode if normal fusion mode is on, and we are at
10790 -O2 or -O3.
10791 (fusion_gpr_load_p): New function, return true if we can fuse an
10792 addis instruction with a dependent load to a GPR.
10793 (emit_fusion_gpr_load): Emit the instructions for power8 load
10794 fusion to GPRs.
10795
0bd72901 10796 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
d86e633a
MM
10797 (VSX load fusion peepholes): New peepholes to fuse together an
10798 addi instruction with a VSX load instruction.
10799
10800 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
10801 peepholes to fuse an addis instruction with a load to a GPR base
10802 register. If we are supporting sign extending fusions, convert
10803 sign extending loads to zero extending loads and add an explicit
10804 sign extension.
10805
c3f35647
SN
108062013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10807
10808 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
10809 aarch-common-protos.h to extra_headers.
10810 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10811 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
10812 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
10813 (arm_no_early_alu_shift_dep): Likewise.
10814 (arm_no_early_alu_shift_value_dep): Likewise.
10815 (arm_no_early_mul_dep): Likewise.
10816 (arm_no_early_store_addr_dep): Likewise.
10817 (arm_mac_accumulator_is_mul_result): Likewise.
10818 (arm_mac_accumulator_is_result): Likewise.
c3f35647 10819 * config/arm/aarch-common.c: ... here. New file.
0bd72901
UB
10820 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
10821 here to ...
c3f35647 10822 (arm_early_store_addr_dep): Likewise.
0bd72901
UB
10823 (arm_no_early_alu_shift_dep): Likewise.
10824 (arm_no_early_alu_shift_value_dep): Likewise.
10825 (arm_no_early_mul_dep): Likewise.
10826 (arm_no_early_store_addr_dep): Likewise.
10827 (arm_mac_accumulator_is_mul_result): Likewise.
10828 (arm_mac_accumulator_is_result): Likewise.
c3f35647
SN
10829 * config/arm/aarch-common-protos.h: ... here. New file.
10830 * config/arm/t-arm (aarch-common.o): Define.
10831
108322013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10833
10834 * config/arm/arm.md: Include new file "types.md".
10835 (define_attr "type"): Move from here to ...
10836 (define_attr "mul32"): Likewise.
10837 (define_attr "mul64"): Likewise.
10838 * config/arm/types.md: ... here. New file.
10839
228c1313
SH
108402013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
10841
10842 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
10843 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
10844
d6e9046d 108452013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
8ac499bb
JBG
10846
10847 * gen-pass-instances.awk: Fix offset of substr().
10848
a167b052
DM
108492013-07-31 David Malcolm <dmalcolm@redhat.com>
10850
10851 * Makefile.in (pass-instances.def): New.
10852 (passes.o): Replace dependency on passes.def with one on
10853 pass-instances.def
10854
10855 * gen-pass-instances.awk: New.
10856
10857 * passes.c (pass_manager::pass_manager): Use pass-instances.def
10858 rather than passes.def, updating local definition of NEXT_PASS
10859 macro to add an extra NUM parameter (currently unused).
10860
315f8c0e
DM
108612013-07-30 David Malcolm <dmalcolm@redhat.com>
10862
10863 * Makefile.in (PASS_MANAGER_H): New.
10864 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
10865 (passes.o): Likewise.
10866 (statistics.o): Likewise.
10867 (cgraphunit.o): Likewise.
10868 (context.o): Depend on PASS_MANAGER_H.
10869
10870 * pass_manager.h: New.
10871
10872 * cgraphunit.c (cgraph_add_new_function): Update for moves
10873 of globals to fields of pass_manager.
10874 (analyze_function): Likewise.
10875 (expand_function): Likewise.
10876 (ipa_passes): Likewise.
10877 (compile): Likewise.
10878
10879 * context.c (context::context): New.
10880 * context.h (context::context): New.
10881 (context::get_passes): New.
10882 (context::passes_): New.
10883
10884 * lto-cgraph.c (input_node): Update for moves of globals to
10885 fields of pass_manager.
10886
10887 * passes.c (all_passes): Remove, in favor of a field of the
10888 same name within the new class pass_manager.
10889 (all_small_ipa_passes): Likewise.
10890 (all_lowering_passes): Likewise.
10891 (all_regular_ipa_passes): Likewise.
10892 (all_late_ipa_passes): Likewise.
10893 (all_lto_gen_passes): Likewise.
10894 (passes_by_id): Likewise.
10895 (passes_by_id_size): Likewise.
10896 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10897 the new class pass_manager.
10898 (set_pass_for_id): Convert to...
10899 (pass_manager::set_pass_for_id): ...method.
10900 (get_pass_for_id): Convert to...
10901 (pass_manager::get_pass_for_id): ...method.
10902 (register_one_dump_file): Move body of implementation into...
10903 (pass_manager::register_one_dump_file): ...here.
10904 (register_dump_files_1): Convert to...
10905 (pass_manager::register_dump_files_1): ...method.
10906 (register_dump_files): Convert to...
10907 (pass_manager::register_dump_files): ...method.
10908 (create_pass_tab): Update for moves of globals to fields of
10909 pass_manager.
10910 (dump_passes): Move body of implementation into...
10911 (pass_manager::dump_passes): ...here.
10912 (register_pass): Move body of implementation into...
10913 (pass_manager::register_pass): ...here.
10914 (init_optimization_passes): Convert into...
10915 (pass_manager::pass_manager): ...constructor for new
10916 pass_manager class, and initialize the pass_lists array.
10917 (check_profile_consistency): Update for moves of globals to
10918 fields of pass_manager.
10919 (dump_profile_report): Move body of implementation into...
10920 (pass_manager::dump_profile_report): ...here.
10921 (ipa_write_summaries_1): Update for moves of pass lists from
10922 being globals to fields of pass_manager.
10923 (ipa_write_optimization_summaries): Likewise.
10924 (ipa_read_summaries): Likewise.
10925 (ipa_read_optimization_summaries): Likewise.
10926 (execute_all_ipa_stmt_fixups): Likewise.
10927
10928 * statistics.c (statistics_fini): Update for moves of globals to
10929 fields of pass_manager.
10930
10931 * toplev.c (general_init): Replace call to
10932 init_optimization_passes with construction of the pass_manager
10933 instance.
10934
10935 * tree-pass.h (all_passes): Remove, in favor of a field of the
10936 same name within the new class pass_manager.
10937 (all_small_ipa_passes): Likewise.
10938 (all_lowering_passes): Likewise.
10939 (all_regular_ipa_passes): Likewise.
10940 (all_lto_gen_passes): Likewise.
10941 (all_late_ipa_passes): Likewise.
10942 (passes_by_id): Likewise.
10943 (passes_by_id_size): Likewise.
10944 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10945 the new class pass_manager.
10946 (get_pass_for_id): Remove.
10947
02cab1c1
RE
109482013-07-30 Richard Earnshaw <rearnsha@arm.com>
10949
10950 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
10951 configs.
10952
b2d3f886
RE
109532013-07-30 Richard Earnshaw <rearnsha@arm.com>
10954
10955 * arm.md (mulhi3): New expand pattern.
10956
605e86fa
JH
109572013-07-30 Jan Hubicka <jh@suse.cz>
10958 Martin Liska <marxin.liska@gmail.com>
10959
10960 * profile.c (compute_value_histograms): Do not ICE when
10961 there is mismatch only on some counters.
10962
14a87636
ZC
109632013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10964
10965 PR rtl-optimization/57637
10966 * function.c (move_insn_for_shrink_wrap): Also check the
10967 GEN set of the LIVE problem for the liveness analysis
10968 if it exists, otherwise give up.
10969
0100cd3f
BS
109702013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10971
10972 PR tree-optimization/57993
10973 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
10974 replaced statement in the candidate table.
10975 (phi_add_costs): Return infinite cost when the hidden basis does
10976 not dominate all phis on which the candidate is dependent.
10977 (replace_one_candidate): Record replaced statement in the
10978 candidate table.
10979
f0d811f7
JR
109802013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
10981
10982 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
10983 (ashlv2si3): New expander.
10984 (*ashlv2si3_i): New define_insn_and_split.
10985 * predicates.md (float_operation): Allow patterns with three
10986 basic sub-patterns.
10987
bba33211
JR
10988 PR rtl-optimization/58021
10989 * mode-switching.c (create_pre_exit): Always split off preceding
10990 insns if we are not at the basic block head.
10991
ff3f3951
MR
109922013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
10993
10994 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
10995 (UCLIBC_DYNAMIC_LINKER): New macro.
10996 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
10997 `-mnan=2008'.
10998 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
10999 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
11000 `-mnan=2008'.
11001 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
11002 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
11003 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
11004 for SF and DF modes. Use ieee_quad_format for TF mode.
11005 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
11006 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
11007 (mips_option_override): Handle `-mnan=legacy'.
11008 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
11009 `-mabs=2008' and `-mnan=2008'.
11010 (OPTION_DEFAULT_SPECS): Add "nan" default.
11011 (ASM_SPEC): Handle `-mnan='.
11012 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
11013 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
11014 comment accordingly.
11015 (neg<mode>2): Likewise.
11016 * config/mips/mips.opt (mabs, mnan): New options.
11017 * doc/install.texi (Configuration): Document `--with-nan=' option.
11018 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
11019 `-mnan=' options.
11020 (MIPS Options): Document them.
11021 * config.gcc <mips*-*-*>: Handle `--with-nan='.
11022 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
11023 * configure: Regenerate.
11024 * config.in: Regenerate.
11025
576e2f37
UB
110262013-07-29 Uros Bizjak <ubizjak@gmail.com>
11027
11028 * config/i386/i386.md (float post-reload splitters): Do not check
11029 for subregs of SSE registers.
11030
110312013-07-29 Uros Bizjak <ubizjak@gmail.com>
11032 H.J. Lu <hongjiu.lu@intel.com>
11033
11034 PR target/57954
11035 PR target/57988
11036 * config/i386/i386.md (post-reload splitter
11037 to avoid partial SSE reg dependency stalls): New pattern.
11038
a71f0749
DV
110392013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11040
11041 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11042
4c97f1cc
JR
110432013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11044
11045 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11046 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11047
b0c733d8
CC
110482013-07-26 Cary Coutant <ccoutant@google.com>
11049
11050 * dwarf2out.c (die_checksum_ordered): Don't include template
11051 instantiations in signature.
11052 (is_template_parameter): New function.
11053 (is_template_instantiation): New function.
11054 (generate_skeleton_bottom_up): Don't include template instantiations
11055 in type unit DIE.
11056 (generate_skeleton): Likewise.
11057 (break_out_comdat_types): Move recursive call to break out nested
11058 types earlier.
11059 (prune_unused_types_mark_generic_parms_dies): Call
11060 is_template_parameter.
11061
040d8a1c
IB
110622013-07-26 Ian Bolton <ian.bolton@arm.com>
11063
11064 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11065 uses vector registers.
89fdc743 11066 * config/aarch64/iterators.md: Add attributes rtn and vas.
040d8a1c 11067
85bd4ac6 110682013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
576e2f37 11069 Richard Earnshaw <richard.earnshaw@arm.com>
85bd4ac6
KT
11070
11071 * combine.c (simplify_comparison): Re-canonicalize operands
576e2f37 11072 where appropriate.
85bd4ac6
KT
11073 * config/arm/arm.md (movcond_addsi): New splitter.
11074
b6051207
SA
110752013-07-25 Sterling Augustine <saugustine@google.com>
11076
11077 * dwarf2out.c (size_of_pubnames): Move code to...
11078 (include_pubname_in_output): ...here. New.
11079 (want_pubnames): Rearrange.
11080 (output_pubnames): Call include_pubname_in_output. Move assertion.
11081
41a7c215
CM
110822013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11083
11084 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11085
110862013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11087
11088 PR target/38836
11089 * doc/extend.texi: Remove obsolete builtins. Fix
11090 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11091
49bde175
JH
110922013-07-25 Jan Hubicka <jh@suse.cz>
11093
11094 * cgraph.c (release_function_body): Break out from ...
11095 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11096 and DECL_ARGUMENTS.
11097 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11098 old_tree in the map.
11099 (create_specialized_node): Update.
11100 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11101 into index.
41a7c215
CM
11102 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11103 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
49bde175
JH
11104 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11105 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11106 DECL_RESULT.
11107
4870352d
KT
111082013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11109
11110 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11111 addsi3_carryin_alt2_<optab>): Correct output template.
11112
1d6aee1c
KT
111132013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11114
11115 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11116 Adjust for arm_restrict_it.
11117 Remove trailing whitespace.
11118
c7cafd75
MK
111192013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11120
017d38f5
MK
11121 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11122 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11123
c7cafd75
MK
11124 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11125
00b7527b
VM
111262013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11127
11128 PR rtl-optimization/57960
11129 * lra-constraints.c (process_alt_operands): Use the right mode
11130 when checking strict_low.
11131
4f63dfc6
JH
111322013-07-25 Jan Hubicka <jh@suse.cz>
11133
11134 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11135 * cgraph.c (cgraph_remove_node): Do not release function body
11136 when in cgraph streaming.
11137 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
41a7c215
CM
11138 in other partitions are not considered reachable; fix handling of
11139 clones.
4f63dfc6 11140
e6dcfa18
RR
111412013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11142
11143 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11144
0b93d3b6
RR
111452013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11146
11147 PR target/19599
11148 PR target/57731
bb80c2eb 11149 PR target/57837
e6dcfa18 11150 * config/arm/arm.md ("*sibcall_insn): Replace use of
0b93d3b6 11151 Ss with US. Adjust output for v5 and v4t.
41a7c215 11152 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
e6dcfa18 11153
0b93d3b6
RR
11154 * config/arm/constraints.md ("Ss"): Rename to US.
11155
690688b3
TG
111562013-07-25 Terry Guo <terry.guo@arm.com>
11157
11158 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11159 shift_add/shift_sub0/shift_sub1 RTXs.
11160
a4ad093b 111612013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
12211b99 11162 Anton Blanchard <anton@au1.ibm.com>
a4ad093b
BS
11163
11164 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11165 (altivec_vpks<VI_char>ss): Likewise.
11166 (altivec_vpks<VI_char>us): Likewise.
11167 (altivec_vpku<VI_char>us): Likewise.
11168 (altivec_vpku<VI_char>um): Likewise.
11169
8c5005ce
DM
111702013-07-24 David Malcolm <dmalcolm@redhat.com>
11171
11172 Introduce context class.
11173
11174 * Makefile.in (CONTEXT_H): New.
11175 (OBJS): Add context.o.
11176 (toplev.o): Add CONTEXT_H to dependencies.
11177 (context.o): New.
11178
41a7c215 11179 * toplev.c (general_init): Create the singleton gcc::context instance.
8c5005ce
DM
11180
11181 * context.c: New.
11182
11183 * context.h: New.
11184
ce4a9422
JR
111852013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11186
11187 PR rtl-optimization/57968
11188 * mode-switching.c (create_pre_exit): Allow instructions that
11189 don't set a return register to need a non-exit mode.
11190
493f4c9e 111912013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11192 Anton Blanchard <anton@au1.ibm.com>
493f4c9e 11193
b5860fd3
WS
11194 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11195 operands to vperm for little endian.
11196 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11197 of lvsl to create the control mask for a vperm for little endian.
493f4c9e 11198
73792b92 111992013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11200 Anton Blanchard <anton@au1.ibm.com>
73792b92
BS
11201
11202 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11203 two operands for little-endian.
11204
14c21302
SE
112052013-07-23 Steve Ellcey <sellcey@mips.com>
11206
11207 * config/mips/mips.c (mips_case_values_threshold): New.
11208 (TARGET_CASE_VALUES_THRESHOLD): Define.
11209
a6056198 112102013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 11211 Anton Blanchard <anton@au1.ibm.com>
a6056198
BS
11212
11213 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11214 selection of field for vector splat in little endian mode.
11215
dd7a40e1
MM
112162013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11217
11218 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11219 expanders to rs6000.md.
11220 (ior<mode>3): Likewise.
11221 (and<mode>3): Likewise.
11222 (one_cmpl<mode>2): Likewise.
11223 (nor<mode>3): Likewise.
11224 (andc<mode>3): Likewise.
11225 (eqv<mode>3): Likewise.
11226 (nand<mode>3): Likewise.
11227 (orc<mode>3): Likewise.
11228
11229 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11230 declaration.
11231
11232 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11233 to split multi-word logical operations.
11234 (rs6000_split_logical_di): Likewise.
11235 (rs6000_split_logical): Likewise.
11236
11237 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11238 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11239 and allow TImode operations in 32-bit.
11240 (vsx_and<mode>3_64bit): Likewise.
11241 (vsx_ior<mode>3_32bit): Likewise.
11242 (vsx_ior<mode>3_64bit): Likewise.
11243 (vsx_xor<mode>3_32bit): Likewise.
11244 (vsx_xor<mode>3_64bit): Likewise.
11245 (vsx_one_cmpl<mode>2_32bit): Likewise.
11246 (vsx_one_cmpl<mode>2_64bit): Likewise.
11247 (vsx_nor<mode>3_32bit): Likewise.
11248 (vsx_nor<mode>3_64bit): Likewise.
11249 (vsx_andc<mode>3_32bit): Likewise.
11250 (vsx_andc<mode>3_64bit): Likewise.
11251 (vsx_eqv<mode>3_32bit): Likewise.
11252 (vsx_eqv<mode>3_64bit): Likewise.
11253 (vsx_nand<mode>3_32bit): Likewise.
11254 (vsx_nand<mode>3_64bit): Likewise.
11255 (vsx_orc<mode>3_32bit): Likewise.
11256 (vsx_orc<mode>3_64bit): Likewise.
11257
11258 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11259 logical types in GPRs.
11260
11261 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11262 logical insns to rs6000.md, and allow TImode operations in
11263 32-bit.
11264 (altivec_ior<mode>3): Likewise.
11265 (altivec_xor<mode>3): Likewise.
11266 (altivec_one_cmpl<mode>2): Likewise.
11267 (altivec_nor<mode>3): Likewise.
11268 (altivec_andc<mode>3): Likewise.
11269
11270 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11271 attributes for moving the 128-bit logical operations into
11272 rs6000.md.
11273 (BOOL_REGS_OUTPUT): Likewise.
11274 (BOOL_REGS_OP1): Likewise.
11275 (BOOL_REGS_OP2): Likewise.
11276 (BOOL_REGS_UNARY): Likewise.
11277 (BOOL_REGS_AND_CR0): Likewise.
11278 (one_cmpl<mode>2): Add support for DI logical operations on
11279 32-bit, splitting the operations to 32-bit.
11280 (anddi3): Likewise.
11281 (iordi3): Likewise.
11282 (xordi3): Likewise.
11283 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11284 changes to combine the 32/64-bit code, allow logical operations on
11285 TI mode in 32-bit, and to use similar match_operator patterns like
11286 scalar mode uses. Combine the Altivec and VSX code for logical
11287 operations, and move it here.
11288 (ior<mode>3, 128-bit types): Likewise.
11289 (xor<mode>3, 128-bit types): Likewise.
11290 (one_cmpl<mode>3, 128-bit types): Likewise.
11291 (nor<mode>3, 128-bit types): Likewise.
11292 (andc<mode>3, 128-bit types): Likewise.
11293 (eqv<mode>3, 128-bit types): Likewise.
11294 (nand<mode>3, 128-bit types): Likewise.
11295 (orc<mode>3, 128-bit types): Likewise.
11296 (and<mode>3_internal): Likewise.
11297 (bool<mode>3_internal): Likewise.
11298 (boolc<mode>3_internal1): Likewise.
11299 (boolc<mode>3_internal2): Likewise.
11300 (boolcc<mode>3_internal1): Likewise.
11301 (boolcc<mode>3_internal2): Likewise.
11302 (eqv<mode>3_internal1): Likewise.
11303 (eqv<mode>3_internal2): Likewise.
11304 (one_cmpl1<mode>3_internal): Likewise.
11305
12211b99 113062013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
9b69d4b4
DH
11307
11308 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11309 Rename flag_stack_usage to flag_stack_usage_info.
11310
12211b99 113112013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
450b0ebb 11312
a5f257fd
ME
11313 * config/microblaze/sync.md: New file.
11314 * config/microblaze/microblaze.md: Include sync.md
11315 * config/microblaze/microblaze.c: Add print_operand 'y'.
11316 * config/microblaze/constraints.md: Add memory_contraint
11317 'Q' which is a single register.
450b0ebb 11318
49fba14f
EB
113192013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11320
11321 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11322
c7b3b99f
PCC
113232013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11324
11325 * reload.c (find_reloads): Exit loop once we find this operand
11326 cannot be reloaded somehow for this alternative.
11327
11328 * reload.c (find_reloads): Exit loop once we find a hard register.
11329
11330 * rtlanal.c (computed_jump_p): Exit loop once we find label
11331 reference is used.
11332
11333 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11334
11335 * cfgloopmanip.c (remove_path): Exit loop after setting
11336 irred_invalidated.
11337
11338 * gensupport.c (subst_dup): Avoid loop if code is not
11339 MATCH_DUP nor MATCH_OP_DUP.
11340
c8fbf1fa
NBJ
113412013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11342
11343 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11344
9259db42
YZ
113452013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11346
11347 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11348 true for SP_REGNUM if mode == ptr_mode.
11349 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11350 with value R0_REGNUM + 31.
11351
00edcfbe
YZ
113522013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11353
11354 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11355 pad pointer-typed argument downward.
11356
43be9a95
YZ
113572013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11358
11359 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11360 and __ILP32__ when the ILP32 model is in use.
11361
28514dda
YZ
113622013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11363
11364 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11365 (aarch64_load_symref_appropriately): In the case of
11366 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11367 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11368 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11369 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11370 if 'mode' doesn't equal to 'ptr_mode'.
11371 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11372 'vcall_offset'; change to call aarch64_emit_move differently depending
11373 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11374 to calculate the upper bound of 'vcall_offset'.
11375 (aarch64_cannot_force_const_mem): Change to also return true if
11376 mode != ptr_mode.
11377 (aarch64_legitimize_reload_address): In the case of large
11378 displacements, add new local variable 'xmode' and an assertion
11379 based on it; change to use 'xmode' to generate the new rtx and
11380 reload.
11381 (aarch64_asm_trampoline_template): Change to generate the template
11382 differently depending on TARGET_ILP32 or not; change to use
11383 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11384 (aarch64_trampoline_size): Removed.
11385 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11386 and replace immediate literals with it. Change to use 'ptr_mode'
11387 instead of 'DImode' and call convert_memory_address if the mode
11388 of 'fnaddr' doesn't equal to 'ptr_mode'.
11389 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11390 to output symbol.
11391 (aarch64_elf_asm_destructor): Likewise.
11392 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11393 on TARGET_ILP32 instead of aarch64_trampoline_size.
11394 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11395 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11396 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11397 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11398 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11399 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11400 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11401 depending on the value of 'mode'.
11402 (add_losym_<mode>): New.
11403 (ldr_got_small_<mode>): New, based on ldr_got_small.
11404 (ldr_got_small): Remove.
11405 (ldr_got_small_sidi): New.
11406 * config/aarch64/iterators.md (P): New.
11407 (PTR): Change to 'ptr_mode' in the condition.
11408
17a819cb
YZ
114092013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11410
11411 * config.gcc (aarch64*-*-*): Support --with-abi.
11412 (aarch64*-*-elf): Support --with-multilib-list.
11413 (aarch64*-*-linux*): Likewise.
11414 (supported_defaults): Add abi to aarch64*-*-*.
11415 * configure.ac: Mention AArch64 for --with-multilib-list.
11416 * configure: Re-generated.
11417 * config/aarch64/biarchilp32.h: New file.
11418 * config/aarch64/biarchlp64.h: New file.
11419 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11420 (ABI_SPEC): Ditto.
11421 (MULTILIB_DEFAULTS): Ditto.
11422 (DRIVER_SELF_SPECS): Ditto.
11423 (ASM_SPEC): Update to also substitute -mabi.
11424 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11425 file whose name depends on -mabi= and -mbig-endian.
11426 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11427 TARGET_ILP32.
11428 (POINTER_SIZE): New define.
11429 (POINTERS_EXTEND_UNSIGNED): Ditto.
11430 (enum aarch64_abi_type): New enumeration tag.
11431 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11432 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11433 (TARGET_ILP32): New define.
11434 * config/aarch64/aarch64.opt (mabi): New.
11435 (aarch64_abi): New.
11436 (ilp32, lp64): New values for -mabi.
11437 * config/aarch64/t-aarch64 (comma): New define.
11438 (MULTILIB_OPTIONS): Ditto.
11439 (MULTILIB_DIRNAMES): Ditto.
11440 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11441 * doc/invoke.texi: Document -mabi for AArch64.
11442
9f9cbdce
GJL
114432013-07-23 Georg-Johann Lay <avr@gjlay.de>
11444
11445 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11446
7061977b 114472013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
12211b99 11448 Anton Blanchard <anton@au1.ibm.com>
7061977b
BS
11449
11450 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11451 endianness when selecting field to splat.
11452
75a41878
EC
114532013-07-22 Eric Christopher <echristo@gmail.com>
11454
11455 * dwarf2out.c (die_odr_checksum): New function to use
11456 CHECKSUM_ macros and ULEB128 for DIE tag.
11457 (generate_type_signature): Use.
41a7c215 11458
38ae58ca
EB
114592013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11460
11461 * config.gcc (sparc*-*-*): Accept leon3 processor.
11462 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11463 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11464 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11465 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11466 (mfix-ut699): Adjust comment.
11467 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11468 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11469 (CPP_CPU_SPEC): Likewise.
11470 (ASM_CPU_SPEC): Likewise.
11471 * config/sparc/sparc.c (leon3_cost): New constant.
11472 (sparc_option_override): Add leon3 support.
11473 (mem_ref): New function.
11474 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11475 (sparc_do_work_around_errata): Look into the instruction in the delay
11476 slot and adjust accordingly. Add fix for the data cache nullify issues
11477 of the UT699. Change insertion position for the NOP.
11478 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11479 (leon3_load): New reservation.
11480 (leon_store): Bump latency to 2.
11481 (grfpu): New automaton.
11482 (grfpu_alu): New unit.
11483 (grfpu_ds): Likewise.
11484 (leon_fp_alu): Adjust.
11485 (leon_fp_mult): Delete.
11486 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11487 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11488 * config/sparc/sparc.md (cpu): Add leon3.
11489 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11490 (swapsi): Likewise.
11491 (atomic_test_and_set): Likewise.
11492 (ldstub): Likewise.
11493
04dfc6df
JU
114942013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11495
11496 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11497 default for R5900 targets.
11498 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11499 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11500 * config/mips/mips.c (mips_option_override): Report an error for
11501 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11502 for -march=r5900 -mhard-float.
11503
c7b3b99f 115042013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
ae382ebd
PCC
11505
11506 * df-problems.c (can_move_insns_across): Exit loop once we
11507 find a non-fixed, non-global register.
11508
11509 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11510 setting can_throw.
11511
11512 * omega.c (omega_eliminate_red): Break after setting red_found.
11513 (omega_problem_has_red_equations): Similarly after setting found.
11514 (omega_query_variable): Similarly after setting coupled.
11515
9b6e6981
MP
115162013-07-22 Marek Polacek <polacek@redhat.com>
11517
11518 * gimplify.c: Don't include gimple.h twice.
11519
5e5f7673
KT
115202013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11521
11522 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11523 instead of TARGET_THUMB1.
11524 (Pz): New constraint.
11525 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11526 encodings.
11527 (compare_negsi_si): Likewise.
11528 (compare_addsi2_op0): Likewise.
11529 (compare_addsi2_op1): Likewise.
11530 (addsi3_carryin_<optab>): Likewise.
11531 (addsi3_carryin_alt2_<optab>): Likewise.
11532 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11533 for arm_restrict_it.
11534 (subsi3_carryin): Likewise.
11535 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11536 (minmax_arithsi): Disable for arm_restrict_it.
11537 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11538 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11539 (satsi_<SAT:code>_shift): Likewise.
11540 (arm_shiftsi3): Add alternative for 16-bit encoding.
11541 (arm32_movhf): Disable for arm_restrict_it.
11542 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11543 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11544
859abddd
SN
115452013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11546
11547 * config/arm/arm.md (attribute "insn"): Delete.
11548 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11549 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11550 (not_shiftsi): Update for attribute change.
11551 (not_shiftsi_compare0): Likewise.
11552 (not_shiftsi_compare0_scratch): Likewise.
11553 (arm_one_cmplsi2): Likewise.
11554 (thumb1_one_cmplsi2): Likewise.
11555 (notsi_compare0): Likewise.
11556 (notsi_compare0_scratch): Likewise.
11557 (thumb1_movdi_insn): Likewise.
11558 (arm_movsi_insn): Likewise.
11559 (movhi_insn_arch4): Likewise.
11560 (movhi_bytes): Likewise.
11561 (arm_movqi_insn): Likewise.
11562 (thumb1_movqi_insn): Likewise.
11563 (arm32_movhf): Likewise.
11564 (thumb1_movhf): Likewise.
11565 (arm_movsf_soft_insn): Likewise.
11566 (thumb1_movsf_insn): Likewise.
11567 (thumb_movdf_insn): Likewise.
11568 (movsicc_insn): Likewise.
11569 (movsfcc_soft_insn): Likewise.
11570 (and_scc): Likewise.
11571 (cond_move): Likewise.
11572 (if_move_not): Likewise.
11573 (if_not_move): Likewise.
11574 (if_shift_move): Likewise.
11575 (if_move_shift): Likewise.
11576 (if_shift_shift): Likewise.
11577 (if_not_arith): Likewise.
11578 (if_arith_not): Likewise.
11579 (cond_move_not): Likewise.
11580 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11581 (neon_mov<mode>): Likewise.
11582 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11583 (thumb2_movsi_vfp): Likewise.
11584 (movsf_vfp): Likewise.
11585 (thumb2_movsf_vfp): Likewise.
371e77e3
UB
11586 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11587 change.
859abddd
SN
11588 (cortexa7_older_only): Likewise.
11589 (cortexa7_younger): Likewise.
11590 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11591 (1020alu_shift_op): Likewise.
11592 (1020alu_shift_reg_op): Likewise.
11593 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11594 (alu_shift_op): Likewise.
11595 (alu_shift_reg_op): Likewise.
11596 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11597 (11_alu_shift_op): Likewise.
11598 (11_alu_shift_reg_op): Likewise.
11599 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11600 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
11601 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11602 change.
859abddd
SN
11603 (cortex_a15_alu_shift): Likewise.
11604 (cortex_a15_alu_shift_reg): Likewise.
11605 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11606 (cortex_a5_alu_shift): Likewise.
371e77e3
UB
11607 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11608 change.
859abddd 11609 (cortex_a53_alu_shift): Likewise.
371e77e3
UB
11610 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11611 change.
859abddd
SN
11612 (cortex_a7_alu_reg): Likewise.
11613 (cortex_a7_alu_shift): Likewise.
11614 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11615 (cortex_a8_alu_shift): Likewise.
11616 (cortex_a8_alu_shift_reg): Likewise.
11617 (cortex_a8_mov): Likewise.
11618 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11619 (cortex_a9_dp_shift): Likewise.
11620 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11621 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11622 (cortex_r4_mov): Likewise.
11623 (cortex_r4_alu_shift): Likewise.
11624 (cortex_r4_alu_shift_reg): Likewise.
11625 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11626 (526_alu_shift_op): Likewise.
11627 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11628 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11629 (626te_alu_shift_op): Likewise.
11630 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11631 (726te_alu_op): Likewise.
11632 (726te_alu_shift_op): Likewise.
11633 (726te_alu_shift_reg_op): Likewise.
11634 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11635 (mp626_alu_shift_op): Likewise.
11636 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11637 (pj4_alu_e1_conds): Likewise.
11638 (pj4_alu): Likewise.
11639 (pj4_alu_conds): Likewise.
11640 (pj4_shift): Likewise.
11641 (pj4_shift_conds): Likewise.
11642 (pj4_alu_shift): Likewise.
11643 (pj4_alu_shift_conds): Likewise.
11644
95b97fac
KT
116452013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11646
11647 * config/arm/predicates.md (shiftable_operator_strict_it):
11648 New predicate.
11649 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11650 Disable cond_exec version for arm_restrict_it.
11651 (thumb2_smaxsi3): Convert to generate cond_exec.
11652 (thumb2_sminsi3): Likewise.
11653 (thumb32_umaxsi3): Likewise.
11654 (thumb2_uminsi3): Likewise.
11655 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11656 (thumb2_neg_abssi2): Likewise.
11657 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11658 (thumb2_movsicc_insn): Adjust alternatives.
11659 (thumb2_mov_negscc): Disable for arm_restrict_it.
11660 (thumb2_mov_negscc_strict_it): New pattern.
11661 (thumb2_mov_notscc_strict_it): New pattern.
11662 (thumb2_mov_notscc): Disable for arm_restrict_it.
11663 (thumb2_ior_scc): Likewise.
11664 (thumb2_ior_scc_strict_it): New pattern.
11665 (thumb2_cond_move): Adjust for arm_restrict_it.
11666 (thumb2_cond_arith): Disable for arm_restrict_it.
11667 (thumb2_cond_arith_strict_it): New pattern.
11668 (thumb2_cond_sub): Adjust for arm_restrict_it.
11669 (thumb2_movcond): Likewise.
11670 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11671 (thumb2_zero_extendhisi2_v6): Likewise.
11672 (thumb2_zero_extendqisi2_v6): Likewise.
11673 (orsi_notsi_si): Likewise.
11674 (orsi_not_shiftsi_si): Likewise.
11675
d5a6ef82
GJL
116762013-07-22 Georg-Johann Lay <avr@gjlay.de>
11677
11678 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11679 instruction sequence is 1 byte shorter.
11680
d9e71154
UB
116812013-07-22 Uros Bizjak <ubizjak@gmail.com>
11682
11683 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11684 it is not needed after split.
11685
8c6381ec
IS
116862013-07-20 Iain Sandoe <iain@codesourcery.com>
11687
11688 PR target/51784
11689 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11690 second label for nonlocal goto receivers. Don't output pic base labels
11691 unless we're producing PIC; mark that action unreachable().
11692 (ix86_save_reg): If the function contains a nonlocal label, save the
11693 PIC base reg.
11694 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
f43245d1 11695 * config/darwin.c (emitted_pic_label_num): New GTY.
8c6381ec
IS
11696 (update_pic_label_number_if_needed): New.
11697 (machopic_output_function_base_name): Adjust for nonlocal receiver
11698 case.
11699 (machopic_should_output_picbase_label): New.
11700 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11701 (nonlocal_goto_receiver): New insn and split.
11702
d05d0709
JG
117032013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11704
11705 * config/aarch64/aarch64-builtins.c
11706 (aarch64_fold_builtin): Fold abs in all modes.
11707 * config/aarch64/aarch64-simd-builtins.def
11708 (abs): Enable for all modes.
11709 * config/aarch64/arm_neon.h
11710 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11711 (vabs_f64): Add missing intrinsic.
11712
ba590e63
IB
117132013-07-19 Ian Bolton <ian.bolton@arm.com>
11714
11715 * config/aarch64/arm_neon.h (vabs_s64): New function
11716
02371798
GJL
117172013-07-19 Georg-Johann Lay <avr@gjlay.de>
11718
11719 PR target/57516
11720 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11721 * config/avr/avr.md (adjust_len): Add `round'.
11722 * config/avr/avr-protos.h (avr_out_round): New prototype.
11723 (avr_out_plus): Add `out_label' argument.
11724 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11725 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11726 Handle the case where `insn' is just a pattern.
11727 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11728 (avr_out_round): New function.
11729 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11730
12211b99 117312013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
b950ac25
DH
11732
11733 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11734 Add check for flag_stack_usage to handle -fstack-usage support
11735
12211b99 117362013-07-18 Pat Haugen <pthaugen@us.ibm.com>
ed1fd999
PH
11737
11738 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11739 interaction for new Power8 flags and VSX.
11740
df9dda2d
ST
117412013-07-18 Sriraman Tallam <tmsriram@google.com>
11742
11743 PR middle-end/57698
11744 * tree-inline.c (expand_call_inline): Emit errors during
11745 early_inlining only if optimization is not turned on.
11746
0f9cdef4
DM
117472013-07-18 David Malcolm <dmalcolm@redhat.com>
11748
11749 * passes.def: New.
11750
11751 * passes.c (init_optimization_passes): Move the construction of
11752 the pass hierarchy into a new passes.def file.
11753
11754 * Makefile.in (passes.o): Add dependency on passes.def.
11755
2efa4087
DM
117562013-07-18 David Malcolm <dmalcolm@redhat.com>
11757
11758 * passes.c (init_optimization_passes): Introduce macros for
11759 constructing the tree of passes (INSERT_PASSES_AFTER,
11760 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
11761 TERMINATE_PASS_LIST).
11762
bc404e1b
VM
117632013-07-18 Vladimir Makarov <vmakarov@redhat.com>
11764 Wei Mi <wmi@google.com>
11765
11766 PR rtl-optimization/57878
11767 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
11768 top.
11769 (reload_pseudo_compare_func): Check nregs first for reload
11770 pseudos.
11771
d675843f
DM
117722013-07-18 David Malcolm <dmalcolm@redhat.com>
11773
11774 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
11775
c7b3b99f 117762013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
bcb97fbd
PCC
11777
11778 * read-rtl.c (validate_const_int): Once an invalid character is
11779 seen, quit the loop.
11780
11781 * gengtype.c (write_roots): Similarly once we find the "deletable"
11782 or "if_marked" option.
11783
52fceb44
SN
117842013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11785
11786 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
11787 "xtab" and "sat". Move value "clz" from here to ...
11788 (attriubte "type"): ... here.
11789 (satsi_<SAT:code>): Delete "insn" attribute.
11790 (satsi_<SAT:code>_shift): Likewise.
11791 (arm_zero_extendqisi2addsi): Likewise.
11792 (arm_extendqisi2addsi): Likewise.
11793 (clzsi2): Update for attribute changes.
11794 (rbitsi2): Likewise.
371e77e3
UB
11795 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
11796 attribute.
52fceb44
SN
11797 (arm_usatsihi): Likewise.
11798 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11799
006bd006
SN
118002013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11801
11802 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
371e77e3
UB
11803 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
11804 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
006bd006
SN
11805 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
11806 in alphabetical order.
11807 (attribute "core_cycles"): Update for attribute changes.
11808 (arm_addsi3): Likewise.
11809 (addsi3_compare0): Likewise.
11810 (addsi3_compare0_scratch): Likewise.
11811 (addsi3_compare_op1): Likewise.
11812 (addsi3_compare_op2): Likewise.
11813 (compare_addsi2_op0): Likewise.
11814 (compare_addsi2_op1): Likewise.
11815 (addsi3_carryin_shift_<optab>): Likewise.
11816 (subsi3_carryin_shift): Likewise.
11817 (rsbsi3_carryin_shift): Likewise.
11818 (arm_subsi3_insn): Likewise.
11819 (subsi3_compare0): Likewise.
11820 (subsi3_compare): Likewise.
11821 (arm_andsi3_insn): Likewise.
11822 (thumb1_andsi3_insn): Likewise.
11823 (andsi3_compare0): Likewise.
11824 (andsi3_compare0_scratch): Likewise.
11825 (zeroextractsi_compare0_scratch
11826 (andsi_not_shiftsi_si): Likewise.
11827 (iorsi3_insn): Likewise.
11828 (iorsi3_compare0): Likewise.
11829 (iorsi3_compare0_scratch): Likewise.
11830 (arm_xorsi3): Likewise.
11831 (thumb1_xorsi3_insn): Likewise.
11832 (xorsi3_compare0): Likewise.
11833 (xorsi3_compare0_scratch): Likewise.
11834 (satsi_<SAT:code>_shift): Likewise.
11835 (rrx): Likewise.
11836 (arm_shiftsi3): Likewise.
11837 (shiftsi3_compare0): Likewise.
11838 (not_shiftsi): Likewise.
11839 (not_shiftsi_compare0): Likewise.
11840 (not_shiftsi_compare0_scratch): Likewise.
11841 (arm_one_cmplsi2): Likewise.
11842 (thumb_one_complsi2): Likewise.
11843 (notsi_compare0): Likewise.
11844 (notsi_compare0_scratch): Likewise.
11845 (thumb1_zero_extendhisi2): Likewise.
11846 (arm_zero_extendhisi2): Likewise.
11847 (arm_zero_extendhisi2_v6): Likewise.
11848 (arm_zero_extendhisi2addsi): Likewise.
11849 (thumb1_zero_extendqisi2): Likewise.
11850 (thumb1_zero_extendqisi2_v6): Likewise.
11851 (arm_zero_extendqisi2): Likewise.
11852 (arm_zero_extendqisi2_v6): Likewise.
11853 (arm_zero_extendqisi2addsi): Likewise.
11854 (thumb1_extendhisi2): Likewise.
11855 (arm_extendhisi2): Likewise.
11856 (arm_extendhisi2_v6): Likewise.
11857 (arm_extendqisi): Likewise.
11858 (arm_extendqisi_v6): Likewise.
11859 (arm_extendqisi2addsi): Likewise.
11860 (thumb1_extendqisi2): Likewise.
11861 (thumb1_movdi_insn): Likewise.
11862 (arm_movsi_insn): Likewise.
11863 (movsi_compare0): Likewise.
11864 (movhi_insn_arch4): Likewise.
11865 (movhi_bytes): Likewise.
11866 (arm_movqi_insn): Likewise.
11867 (thumb1_movqi_insn): Likewise.
11868 (arm32_movhf): Likewise.
11869 (thumb1_movhf): Likewise.
11870 (arm_movsf_soft_insn): Likewise.
11871 (thumb1_movsf_insn): Likewise.
11872 (movdf_soft_insn): Likewise.
11873 (thumb_movdf_insn): Likewise.
11874 (arm_cmpsi_insn): Likewise.
11875 (cmpsi_shiftsi): Likewise.
11876 (cmpsi_shiftsi_swp): Likewise.
11877 (arm_cmpsi_negshiftsi_si): Likewise.
11878 (movsicc_insn): Likewise.
11879 (movsfcc_soft_insn): Likewise.
11880 (arith_shiftsi): Likewise.
11881 (arith_shiftsi_compare0
11882 (arith_shiftsi_compare0_scratch
11883 (sub_shiftsi): Likewise.
11884 (sub_shiftsi_compare0
11885 (sub_shiftsi_compare0_scratch
11886 (and_scc): Likewise.
11887 (cond_move): Likewise.
11888 (if_plus_move): Likewise.
11889 (if_move_plus): Likewise.
11890 (if_move_not): Likewise.
11891 (if_not_move): Likewise.
11892 (if_shift_move): Likewise.
11893 (if_move_shift): Likewise.
11894 (if_shift_shift): Likewise.
11895 (if_not_arith): Likewise.
11896 (if_arith_not): Likewise.
11897 (cond_move_not): Likewise.
11898 (thumb1_ashlsi3): Set type attribute.
11899 (thumb1_ashrsi3): Likewise.
11900 (thumb1_lshrsi3): Likewise.
11901 (thumb1_rotrsi3): Likewise.
11902 (shiftsi3_compare0_scratch): Likewise.
11903 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
11904 (neon_mov<mode>): Likewise.
371e77e3
UB
11905 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
11906 attribute changes.
006bd006
SN
11907 (thumb2_movsi_insn): Likewise.
11908 (thumb2_cmpsi_neg_shiftsi): Likewise.
11909 (thumb2_extendqisi_v6): Likewise.
11910 (thumb2_zero_extendhisi2_v6): Likewise.
11911 (thumb2_zero_extendqisi2_v6): Likewise.
11912 (thumb2_shiftsi3_short): Likewise.
11913 (thumb2_addsi3_compare0_scratch): Likewise.
11914 (orsi_not_shiftsi_si): Likewise.
11915 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11916 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
11917 changes.
11918 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11919 (1020alu_shift_op): Likewise.
11920 (1020alu_shift_reg_op): Likewise.
11921 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11922 (alu_shift_op): Likewise.
11923 (alu_shift_reg_op): Likewise.
11924 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11925 (11_alu_shift_op): Likewise.
11926 (11_alu_shift_reg_op): Likewise.
11927 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11928 (9_alu_shift_reg_op): Likewise.
371e77e3
UB
11929 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11930 changes.
006bd006
SN
11931 (cortex_a15_alu_shift): Likewise.
11932 (cortex_a15_alu_shift_reg): Likewise.
371e77e3
UB
11933 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
11934 changes.
006bd006
SN
11935 (cortex_a5_alu_shift): Likewise.
11936 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
11937 changes.
11938 (cortex_a53_alu_shift): Likewise.
11939 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11940 changes.
11941 (cortex_a7_alu_reg): Likewise.
11942 (cortex_a7_alu_shift): Likewise.
371e77e3
UB
11943 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
11944 changes.
006bd006
SN
11945 (cortex_a8_alu_shift): Likewise.
11946 (cortex_a8_alu_shift_reg): Likewise.
11947 (cortex_a8_mov): Likewise.
11948 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11949 (cortex_a9_dp_shift): Likewise.
371e77e3
UB
11950 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
11951 changes.
11952 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
11953 changes.
006bd006
SN
11954 (cortex_r4_mov): Likewise.
11955 (cortex_r4_alu_shift): Likewise.
11956 (cortex_r4_alu_shift_reg): Likewise.
11957 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11958 (526_alu_shift_op): Likewise.
11959 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11960 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11961 (626te_alu_shift_op): Likewise.
11962 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
11963 (726te_alu_op): Likewise.
11964 (726te_alu_shift_op): Likewise.
11965 (726te_alu_shift_reg_op): Likewise.
11966 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11967 (mp626_alu_shift_op): Likewise.
11968 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
11969 (pj4_alu_e1_conds): Likewise.
11970 (pj4_alu): Likewise.
11971 (pj4_alu_conds): Likewise.
11972 (pj4_shift): Likewise.
11973 (pj4_shift_conds): Likewise.
11974 (pj4_alu_shift): Likewise.
11975 (pj4_alu_shift_conds): Likewise.
371e77e3
UB
11976 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11977 changes.
006bd006
SN
11978 (cortexa7_older_only): Likewise.
11979 (cortexa7_younger): Likewise.
11980
651df1b2
DM
119812013-07-18 David Malcolm <dmalcolm@redhat.com>
11982
11983 * ipa-pure-const.c (generate_summary): Rename to...
11984 (pure_const_generate_summary): ... this.
11985
ea717bd7
IS
119862013-07-17 Iain Sandoe <iain@codesourcery.com>
11987
11988 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
11989
a1259a13
YR
119902013-07-17 Yvan Roux <yvan.roux@linaro.org>
11991
11992 PR target/57909
11993 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
11994 usage in HI mode.
11995
167f68ed
AK
119962013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11997
11998 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
11999 enabled without -march=zEC12.
12000 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
12001 flags to be set.
12002
f900a982
MR
120032013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12004
12005 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
12006 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
12007 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
12008 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
12009 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
12010 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
12011 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
12012 ISA_HAS_FP4.
12013 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
12014 and ISA_HAS_NMADD3_NMSUB3.
12015 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
12016 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
12017 (nmsub4<mode>, nmsub3<mode>): Likewise.
12018 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
12019
d4ed27eb
MR
120202013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12021
12022 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
12023 TARGET_MIPS5400 checking.
12024
7a5add18
PB
120252013-07-16 Jakub Jelinek <jakub@redhat.com>
12026 Peter Bergner <bergner@vnet.ibm.com>
12027
12028 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
12029 registers in the comment.
12030 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
12031 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
12032 rather than FIRST_PSEUDO_REGISTERS.
12033
120342013-07-16 Peter Bergner <bergner@vnet.ibm.com>
c694dc01
PB
12035
12036 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12037 enable extra ISA flags with TARGET_HTM.
12038
d2ab0929
MR
120392013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12040
12041 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12042 Fix comment typos.
12043
839c74bc
CH
120442013-07-15 Cong Hou <congh@google.com>
12045
12046 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12047 in compare function for sorting.
12048
7a5add18 120492013-07-15 Peter Bergner <bergner@vnet.ibm.com>
0258b6e4
PB
12050
12051 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12052 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12053 * config/rs6000/rs6000.opt: Add -mhtm option.
12054 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12055 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12056 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12057 __HTM__ if the HTM instructions are available.
12058 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12059 htm_spr_reg_operand): New define_predicates.
12060 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12061 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12062 Include htm.md.
12063 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12064 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12065 HTM builtin functions.
12066 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12067 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12068 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12069 (rs6000_builtin_mask_calculate): Likewise.
12070 (rs6000_option_override_internal): Likewise.
12071 (bdesc_htm): Add new HTM builtin support.
12072 (htm_spr_num): New function.
12073 (htm_spr_regno): Likewise.
12074 (rs6000_htm_spr_icode): Likewise.
12075 (htm_expand_builtin): Likewise.
12076 (htm_init_builtins): Likewise.
12077 (rs6000_expand_builtin): Add support for HTM builtin functions.
12078 (rs6000_init_builtins): Likewise.
371e77e3
UB
12079 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12080 option.
0258b6e4
PB
12081 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12082 (TARGET_HTM, MASK_HTM): Define macros.
12083 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12084 (FIXED_REGISTERS): Likewise.
12085 (CALL_USED_REGISTERS): Likewise.
12086 (CALL_REALLY_USED_REGISTERS): Likewise.
12087 (REG_ALLOC_ORDER): Likewise.
12088 (enum reg_class): Likewise.
12089 (REG_CLASS_NAMES): Likewise.
12090 (REG_CLASS_CONTENTS): Likewise.
12091 (REGISTER_NAMES): Likewise.
12092 (ADDITIONAL_REGISTER_NAMES): Likewise.
12093 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12094 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12095 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12096 * config/rs6000/htm.md: New file.
12097 * config/rs6000/htmintrin.h: New file.
12098 * config/rs6000/htmxlintrin.h: New file.
12099
87dd8ab0
MS
121002013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12101
12102 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12103 Define SYMBOL_TINY_GOT, update comment.
12104 * config/aarch64/aarch64.c
12105 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12106 (aarch64_expand_mov_immediate): Likewise.
12107 (aarch64_print_operand): Likewise.
12108 (aarch64_classify_symbol): Likewise.
12109 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12110 (ldr_got_tiny): Define.
12111
52d676b6
TG
121122013-07-13 Tobias Grosser <tobias@grosser.es>
12113
12114 PR tree-optimization/54094
12115 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12116 scheduling dimension for the parallelism check from the polyhedral
12117 information in the AST.
12118 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12119 in 2D + 1 form.
12120
286e8fc1
JM
121212013-07-13 Jason Merrill <jason@redhat.com>
12122
12123 * print-tree.c (debug_vec_tree): Use debug_raw.
12124 (debug_raw (vec<tree, va_gc> &)): New.
12125 (debug_raw (vec<tree, va_gc> *)): New.
12126 * tree.h: Declare them.
12127
e43257e8
BC
121282013-07-13 Bin Cheng <bin.cheng@arm.com>
12129
12130 * ifcvt.c (ifcvt_after_combine): New static variable.
12131 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12132 for size.
12133 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12134 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12135 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12136
da582d46
MR
121372013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12138
12139 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12140
18b0ea8f
MM
121412013-07-12 Michael Matz <matz@suse.de>
12142
12143 PR middle-end/55771
12144 * convert.c (convert_to_real): Reject non-float inner types.
12145
e4f0f84d
TB
121462013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12147
12148 * config/aarch64/aarch64-protos.h
12149 (aarch64_simd_immediate_valid_for_move): Remove.
12150 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12151 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12152 cases.
12153 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12154
76500022
SE
121552013-07-11 Steve Ellcey <sellcey@mips.com>
12156
12157 * config/mips/mips.c (mips_conditional_register_usage): Do not
12158 use t[0-7] registers in MIPS16 mode when optimizing for size.
12159
78039734
ST
121602013-07-11 Sriraman Tallam <tmsriram@google.com>
12161
12162 * config/i386/i386.c (dispatch_function_versions): Fix array
12163 indexing of function_version_info to match actual_versions.
12164
7b24b675
TJ
121652013-07-11 Teresa Johnson <tejohnson@google.com>
12166
12167 * vec.h (struct va_gc): Move release out-of-line.
12168 (va_gc::release): Call ggc_free on released vec.
12169
85d4cbb8
UW
121702013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12171
12172 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12173 Require GOT register as additional operand in UNSPEC.
12174 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12175 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12176 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12177 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12178 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12179 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12180 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12181
52b3f9e2
GJL
121822013-07-11 Georg-Johann Lay <avr@gjlay.de>
12183
12184 PR target/57631
12185 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12186 name seen by assembler/linker rather if available.
12187
3f3e266e
AS
121882013-07-11 Andreas Schwab <schwab@suse.de>
12189
12190 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12191
21231ca6
VM
121922013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12193
371e77e3 12194 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
21231ca6 12195
d331d374
JM
121962013-07-10 Joseph Myers <joseph@codesourcery.com>
12197
673c2f63
JM
12198 * doc/tm.texi.in: Move hook documentation to ....
12199 * target.def: ... here.
12200
d331d374
JM
12201 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12202 text on @hook line.
12203 * doc/tm.texi: Regenerate.
12204
dfeadaa0
PC
122052013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12206
12207 PR c++/57869
12208 * doc/invoke.texi: Document -Wconditionally-supported.
12209
bb6f2bac
GJL
122102013-07-10 Georg-Johann Lay <avr@gjlay.de>
12211
12212 PR target/57844
12213 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12214 of my_fp.
12215
5219b232
GJL
122162013-07-10 Georg-Johann Lay <avr@gjlay.de>
12217
12218 PR target/57506
12219 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12220 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12221 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12222 Remove duplicate devices.
12223 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12224 * config/avr/t-multilib: Regenerate.
12225 * config/avr/avr-tables.opt: Regenerate.
12226 * doc/avr-mmcu.texi: Regenerate.
12227
283045c0
GJL
122282013-07-10 Georg-Johann Lay <avr@gjlay.de>
12229
12230 PR target/56987
12231 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12232
ceb2d59f 122332013-07-10 Graham Stott <graham.stott@btinternet.com>
6a2fa4b2 12234
ceb2d59f
GS
12235 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12236 the cost of MULT when optimizing for size.
12237
055e0a99
JBG
122382013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12239
12240 * config/cr16/cr16-protos.h: Don't include target.h.
12241
ef43e868
JM
122422013-07-09 Joseph Myers <joseph@codesourcery.com>
12243
12244 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12245 adjust register size for TDmode and TFmode for VSX registers.
12246
4a283fdf
KT
122472013-07-08 Kai Tietz <ktietz@redhat.com>
12248
12249 PR target/56892
12250 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12251 hook_bool_const_tree_true.
12252
2cf4c39e
AK
122532013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12254
12255 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12256 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12257 * config/s390/s390.md: Define FPR*_REGNUM constants.
12258 Fix FPR2_REGNUM constant (18 -> 17).
12259 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12260 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12261 ("trunc<BFP:mode><DFP_ALL:mode>2")
12262 ("trunc<DFP_ALL:mode><BFP:mode>2")
12263 ("*extend<BFP:mode><DFP_ALL:mode>2")
12264 ("*extend<DFP_ALL:mode><BFP:mode>2")
12265 ("extend<BFP:mode><DFP_ALL:mode>2")
12266 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12267 FPR4_REGNUM.
12268
e6ac0270 122692013-07-08 Graham Stott <graham.stott@btinternet.com>
e6ac0270 12270
6a2fa4b2 12271 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
64a92610 12272
b89b22fc
AK
122732013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12274
12275 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12276 and cfun_fpr_bit_p to cfun_fpr_save_p.
12277 (s390_frame_area, s390_register_info, s390_frame_info)
12278 (s390_emit_prologue, s390_emit_epilogue)
12279 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12280 register numbers.
12281 * config/s390/s390.h: Define *_REGNUM macros for floating point
12282 register numbers.
12283
07f398aa
EB
122842013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12285
12286 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12287
75fefa91
PCC
122882013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12289
12290 PR rtl-optimization/57786
12291 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12292 and break out of the loop when it is set to false.
12293
046f1eee
JJ
122942013-07-08 Jakub Jelinek <jakub@redhat.com>
12295
8140c065
JJ
12296 PR target/57819
12297 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12298 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12299 (const_int 63)) 0)).
12300 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12301 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12302 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12303
046f1eee
JJ
12304 PR rtl-optimization/57829
12305 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12306 mask bits outside of mode are just sign-extension from mode to HWI.
12307
930b700b
MZ
123082013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12309
12310 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12311 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12312 adjust_address instead of change_address to keep info about alignment.
12313 (emit_strmov): Remove.
12314 (emit_memmov): New function.
12315 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12316 (expand_movmem_epilogue): Likewise and return updated rtx for
12317 destination.
12318 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12319 destination and source.
12320 (decide_alignment): Refactor, handle vector_loop.
12321 (ix86_expand_movmem): Likewise.
12322 (ix86_expand_setmem): Likewise.
12323 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12324
7b9d1bd8
UB
123252013-07-07 Uros Bizjak <ubizjak@gmail.com>
12326
12327 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12328 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12329
eabb5f48
UB
123302013-07-06 Uros Bizjak <ubizjak@gmail.com>
12331
12332 * config/i386/sse.md (sse_movlhps): Change alternative 3
12333 of operand 2 to "m".
12334
123352013-07-06 Uros Bizjak <ubizjak@gmail.com>
12336
12337 PR target/57807
12338 * config/i386/sse.md (iptr): New mode attribute.
12339 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12340 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12341 (<sse>_vmmul<mode>3): Ditto.
12342 (<sse>_vmdiv<mode>3): Ditto.
12343 (sse_vmrcpv4sf2): Ditto.
12344 (<sse>_vmsqrt<mode>2): Ditto.
12345 (sse_vmrsqrtv4sf2): Ditto.
12346 (<sse>_vm<code><mode>3): Ditto.
12347 (avx_vmcmp<mode>3): Ditto.
12348 (<sse>_vmmaskcmp<mode>3): Ditto.
12349 (<sse>_comi): Ditto.
12350 (<sse>_ucomi): Ditto.
12351 (*xop_vmfrcz_<mode>): Ditto.
12352 (*fmai_fmadd_<mode>): Ditto.
12353 (*fmai_fmsub_<mode>): Ditto.
12354 (*fmai_fnmadd_<mode>): Ditto.
12355 (*fmai_fnmsub_<mode>): Ditto.
12356 (*fma4i_vmfmadd_<mode>): Ditto.
12357 (*fma4i_vmfmsub_<mode>): Ditto.
12358 (*fma4i_vmfnmadd_<mode>): Ditto.
12359 (*fma4i_vmfnmsub_<mode>): Ditto.
12360 (*xop_vmfrcz_<mode>): Ditto.
12361 (sse_cvtps2pi): Ditto.
12362 (sse_cvttps2pi): Ditto.
12363 (sse_cvtss2si): Ditto.
12364 (sse_cvtss2si_2): Ditto.
12365 (sse_cvtss2siq_2): Ditto.
12366 (sse_cvttss2si): Ditto.
12367 (sse_cvttss2siq): Ditto.
12368 (sse_cvtsd2si): Ditto.
12369 (sse_cvtsd2si_2): Ditto.
12370 (sse_cvtsd2siq_2): Ditto.
12371 (sse_cvttsd2si): Ditto.
12372 (sse_cvttsd2siq): Ditto.
12373 (sse_cvtsd2ss): Ditto.
12374 (sse_cvtss2sd): Ditto.
12375 (avx2_pbroadcast<mode>): Ditto.
12376 (avx2_pbroadcast<mode>_1): Ditto.
12377 (*avx_vperm_broadcast_v4sf): Ditto.
12378
12379 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12380 (sse_movlhps): Ditto.
12381 (sse_storehps): Ditto.
12382 (sse_loadhps): Ditto.
12383 (sse_storelps): Ditto.
12384 (sse_loadlps): Ditto.
12385 (*vec_concatv4sf): Ditto.
12386 (*vec_interleave_highv2df): Ditto.
12387 (*vec_interleave_lowv2df): Ditto.
12388 (*vec_extractv2df_1_sse): Ditto.
12389 (*vec_extractv2df_0_sse): Ditto.
12390 (sse2_storelpd): Ditto.
12391 (sse2_loadlpd): Ditto.
12392 (sse2_movsd): Ditto.
12393 (*vec_concatv4si): Ditto.
12394 (vec_concatv2di): Ditto.
12395
12396 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12397 for Intel asm dialect.
12398 (mmx_punpcklwd): Ditto.
12399 (mmx_punpckldq): Ditto.
12400
12401 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12402 for intel assembler dialect.
12403
1f6eac90
JJ
124042013-07-06 Jakub Jelinek <jakub@redhat.com>
12405
12406 PR target/29776
12407 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12408 for BUILT_IN_C{LZ,LRSB}*.
12409 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12410 * tree-vrp.c (extract_range_basic): Handle
12411 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12412 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12413 fall thru to code calling set_value*.
12414 * builtins.c (expand_builtin): Remove *IMAX cases.
12415 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12416 if width is bigger than 2*HWI.
12417
2b778c9d
VM
124182013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12419
12420 PR rtl-optimization/55342
12421 * lra-int.h (lra_subreg_reload_pseudos): New.
12422 * lra.c: Add undoing optional reloads to the block diagram.
12423 (lra_subreg_reload_pseudos): New.
12424 (lra_optional_reload_pseudos): Change comments.
12425 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12426 lra_optional_reload_pseudos after undo transformations.
12427 * lra-assigns.c (pseudo_prefix_title): New.
12428 (lra_setup_reg_renumber): Use it.
12429 (spill_for): Ditto. Check subreg reload pseudos too.
12430 (assign_by_spills): Consider subreg reload pseudos too.
12431 * lra-constraints.c (simplify_operand_subreg): Use
12432 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12433 (curr_insn_transform): Recognize and do optional reloads.
12434 (undo_optional_reloads): New.
12435 (lra_undo_inheritance): Call undo_optional_reloads.
12436
207156e4
TQ
124372013-07-05 Thomas Quinot <quinot@adacore.com>
12438
12439 * tree-complex.c (expand_complex_operations_1): Fix typo.
12440
8990e73a
TB
124412013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12442
12443 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12444 (tune_params): New member 'const vec_costs'.
12445 * config/aarch64/aarch64.c (generic_vector_cost): New.
12446 (generic_tunings): New member 'generic_vector_cost'.
12447 (aarch64_builtin_vectorization_cost): New.
12448 (aarch64_add_stmt_cost): New.
12449 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12450 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12451
f7bc421d
JJ
124522013-07-03 Jakub Jelinek <jakub@redhat.com>
12453
12454 PR target/57777
12455 * config/i386/predicates.md (vsib_address_operand): Disallow
12456 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12457
e90d1568
HPN
124582013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12459
12460 PR middle-end/55030
12461 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12462 expand_builtin_setjmp_receiver.
12463 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12464 with NULL for the label parameter.
12465 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12466 the frame-pointer. Adjust comments.
12467 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12468 only if LABEL is non-NULL.
12469
cbd7413b
YZ
124702013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12471
12472 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12473 (ARM_ABI_AAPCS64): Ditto.
12474 (arm_abi): Ditto.
12475 (ARM_DEFAULT_ABI): Ditto.
12476
dec11868
JG
124772013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12478
12479 * config/aarch64/aarch64-builtins.c
12480 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12481 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12482 (st1): Likewise.
12483 * config/aarch64/aarch64-simd.md
12484 (aarch64_ld1<VALL:mode>): New.
12485 (aarch64_st1<VALL:mode>): Likewise.
12486 * config/aarch64/arm_neon.h
12487 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12488
786fb9b6
ST
124892013-07-02 Sriraman Tallam <tmsriram@google.com>
12490
0ba6aff3
UB
12491 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12492 target ISA is AVX.
786fb9b6
ST
12493 (ix86_option_override_internal):Turn on all -mavx target flags by
12494 default as they are dependent on AVX anyway.
12495
4b043553
CC
124962013-07-02 Cary Coutant <ccoutant@google.com>
12497
12498 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12499 deterministic hash.
12500 (loc_checksum_ordered): Likewise.
12501 (hash_loc_operands): Remove inline keyword.
12502
c4e87a13
JJ
125032013-07-02 Jakub Jelinek <jakub@redhat.com>
12504
12505 PR tree-optimization/57741
12506 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12507 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12508 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12509 Allow REAL_CST step_exprs if flag_associative_math.
12510 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12511
26366d28
IB
125122013-07-02 Ian Bolton <ian.bolton@arm.com>
12513
0ba6aff3 12514 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
26366d28 12515
2879bb2b
IB
125162013-07-02 Ian Bolton <ian.bolton@arm.com>
12517
12518 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12519
afc5e8a6
KT
125202013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12521
12522 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12523 encoding.
12524 (iorsi3_insn): Likewise.
12525 (arm_xorsi3): Likewise.
12526
31f8442b
SN
125272013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12528
12529 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12530 to ...
12531 (attribute "type"): ... here, and prefix with "wmmx_".
12532 (attribute "core_cycles"): Update for attribute changes.
12533 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12534 (tbcstv4hi): Likewise.
12535 (tbcstv2si): Likewise.
12536 (iwmmxt_iordi3): Likewise.
12537 (iwmmxt_xordi3): Likewise.
12538 (iwmmxt_anddi3): Likewise.
12539 (iwmmxt_nanddi3): Likewise.
12540 (iwmmxt_arm_movdi): Likewise.
12541 (iwmmxt_movsi_insn): Likewise.
12542 (mov<mode>_internal): Likewise.
12543 (and<mode>3_iwmmxt): Likewise.
12544 (ior<mode>3_iwmmxt): Likewise.
12545 (xor<mode>3_iwmmxt): Likewise.
12546 (add<mode>3_iwmmxt): Likewise.
12547 (ssaddv8qi3): Likewise.
12548 (ssaddv4hi3): Likewise.
12549 (ssaddv2si3): Likewise.
12550 (usaddv8qi3): Likewise.
12551 (usaddv4hi3): Likewise.
12552 (usaddv2si3): Likewise.
12553 (sub<mode>3_iwmmxt): Likewise.
12554 (sssubv8qi3): Likewise.
12555 (sssubv4hi3): Likewise.
12556 (sssubv2si3): Likewise.
12557 (ussubv8qi3): Likewise.
12558 (ussubv4hi3): Likewise.
12559 (ussubv2si3): Likewise.
12560 (mulv4hi3_iwmmxt): Likewise.
12561 (smulv4hi3_highpart): Likewise.
12562 (umulv4hi3_highpart): Likewise.
12563 (iwmmxt_wmacs): Likewise.
12564 (iwmmxt_wmacsz): Likewise.
12565 (iwmmxt_wmacu): Likewise.
12566 (iwmmxt_wmacuz): Likewise.
12567 (iwmmxt_clrdi): Likewise.
12568 (iwmmxt_clrv8qi): Likewise.
12569 (iwmmxt_clr4hi): Likewise.
12570 (iwmmxt_clr2si): Likewise.
12571 (iwmmxt_uavgrndv8qi3): Likewise.
12572 (iwmmxt_uavgrndv4hi3): Likewise.
12573 (iwmmxt_uavgv8qi3): Likewise.
12574 (iwmmxt_uavgv4hi3): Likewise.
12575 (iwmmxt_tinsrb): Likewise.
12576 (iwmmxt_tinsrh): Likewise.
12577 (iwmmxt_tinsrw): Likewise.
12578 (iwmmxt_textrmub): Likewise.
12579 (iwmmxt_textrmsb): Likewise.
12580 (iwmmxt_textrmuh): Likewise.
12581 (iwmmxt_textrmsh): Likewise.
12582 (iwmmxt_textrmw): Likewise.
12583 (iwmxxt_wshufh): Likewise.
12584 (eqv8qi3): Likewise.
12585 (eqv4hi3): Likewise.
12586 (eqv2si3): Likewise.
12587 (gtuv8qi3): Likewise.
12588 (gtuv4hi3): Likewise.
12589 (gtuv2si3): Likewise.
12590 (gtv8qi3): Likewise.
12591 (gtv4hi3): Likewise.
12592 (gtv2si3): Likewise.
12593 (smax<mode>3_iwmmxt): Likewise.
12594 (umax<mode>3_iwmmxt): Likewise.
12595 (smin<mode>3_iwmmxt): Likewise.
12596 (umin<mode>3_iwmmxt): Likewise.
12597 (iwmmxt_wpackhss): Likewise.
12598 (iwmmxt_wpackwss): Likewise.
12599 (iwmmxt_wpackdss): Likewise.
12600 (iwmmxt_wpackhus): Likewise.
12601 (iwmmxt_wpackwus): Likewise.
12602 (iwmmxt_wpackdus): Likewise.
12603 (iwmmxt_wunpckihb): Likewise.
12604 (iwmmxt_wunpckihh): Likewise.
12605 (iwmmxt_wunpckihw): Likewise.
12606 (iwmmxt_wunpckilb): Likewise.
12607 (iwmmxt_wunpckilh): Likewise.
12608 (iwmmxt_wunpckilw): Likewise.
12609 (iwmmxt_wunpckehub): Likewise.
12610 (iwmmxt_wunpckehuh): Likewise.
12611 (iwmmxt_wunpckehuw): Likewise.
12612 (iwmmxt_wunpckehsb): Likewise.
12613 (iwmmxt_wunpckehsh): Likewise.
12614 (iwmmxt_wunpckehsw): Likewise.
12615 (iwmmxt_wunpckelub): Likewise.
12616 (iwmmxt_wunpckeluh): Likewise.
12617 (iwmmxt_wunpckeluw): Likewise.
12618 (iwmmxt_wunpckelsb): Likewise.
12619 (iwmmxt_wunpckelsh): Likewise.
12620 (iwmmxt_wunpckelsw): Likewise.
12621 (ror<mode>3): Likewise.
12622 (ashr<mode>3_iwmmxt): Likewise.
12623 (lshr<mode>3_iwmmxt): Likewise.
12624 (ashl<mode>3_iwmmxt): Likewise.
12625 (ror<mode>3_di): Likewise.
12626 (ashr<mode>3_di): Likewise.
12627 (lshr<mode>3_di): Likewise.
12628 (ashl<mode>3_di): Likewise.
12629 (iwmmxt_wmadds): Likewise.
12630 (iwmmxt_wmaddu): Likewise.
12631 (iwmmxt_tmia): Likewise.
12632 (iwmmxt_tmiaph): Likewise.
12633 (iwmmxt_tmiabb): Likewise.
12634 (iwmmxt_tmiatb): Likewise.
12635 (iwmmxt_tmiabt): Likewise.
12636 (iwmmxt_tmiatt): Likewise.
12637 (iwmmxt_tmovmskb): Likewise.
12638 (iwmmxt_tmovmskh): Likewise.
12639 (iwmmxt_tmovmskw): Likewise.
12640 (iwmmxt_waccb): Likewise.
12641 (iwmmxt_wacch): Likewise.
12642 (iwmmxt_waccw): Likewise.
12643 (iwmmxt_waligni): Likewise.
12644 (iwmmxt_walignr): Likewise.
12645 (iwmmxt_walignr0): Likewise.
12646 (iwmmxt_walignr1): Likewise.
12647 (iwmmxt_walignr2): Likewise.
12648 (iwmmxt_walignr3): Likewise.
12649 (iwmmxt_wsadb): Likewise.
12650 (iwmmxt_wsadh): Likewise.
12651 (iwmmxt_wsadbz): Likewise.
12652 (iwmmxt_wsadhz): Likewise.
12653 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12654 (iwmmxt_wabsdiffb): Likewise.
12655 (iwmmxt_wabsdiffh): Likewise.
12656 (iwmmxt_wabsdiffw): Likewise.
12657 (iwmmxt_waddsubhx): Likewise
12658 (iwmmxt_wsubaddhx): Likewise.
12659 (addc<mode>3): Likewise.
12660 (iwmmxt_avg4): Likewise.
12661 (iwmmxt_avg4r): Likewise.
12662 (iwmmxt_wmaddsx): Likewise.
12663 (iwmmxt_wmaddux): Likewise.
12664 (iwmmxt_wmaddsn): Likewise.
12665 (iwmmxt_wmaddun): Likewise.
12666 (iwmmxt_wmulwsm): Likewise.
12667 (iwmmxt_wmulwum): Likewise.
12668 (iwmmxt_wmulsmr): Likewise.
12669 (iwmmxt_wmulumr): Likewise.
12670 (iwmmxt_wmulwsmr): Likewise.
12671 (iwmmxt_wmulwumr): Likewise.
12672 (iwmmxt_wmulwl): Likewise.
12673 (iwmmxt_wqmulm): Likewise.
12674 (iwmmxt_wqmulwm): Likewise.
12675 (iwmmxt_wqmulmr): Likewise.
12676 (iwmmxt_wqmulwmr): Likewise.
12677 (iwmmxt_waddbhusm): Likewise.
12678 (iwmmxt_waddbhusl): Likewise.
12679 (iwmmxt_wqmiabb): Likewise.
12680 (iwmmxt_wqmiabt): Likewise.
12681 (iwmmxt_wqmiatb): Likewise.
12682 (iwmmxt_wqmiatt): Likewise.
12683 (iwmmxt_wqmiabbn): Likewise.
12684 (iwmmxt_wqmiabtn): Likewise.
12685 (iwmmxt_wqmiatbn): Likewise.
12686 (iwmmxt_wqmiattn): Likewise.
12687 (iwmmxt_wmiabb): Likewise.
12688 (iwmmxt_wmiabt): Likewise.
12689 (iwmmxt_wmiatb): Likewise.
12690 (iwmmxt_wmiatt): Likewise.
12691 (iwmmxt_wmiabbn): Likewise.
12692 (iwmmxt_wmiabtn): Likewise.
12693 (iwmmxt_wmiatbn): Likewise.
12694 (iwmmxt_wmiattn): Likewise.
12695 (iwmmxt_wmiawbb): Likewise.
12696 (iwmmxt_wmiawbt): Likewise.
12697 (iwmmxt_wmiawtb): Likewise.
12698 (iwmmxt_wmiawtt): Likewise.
12699 (iwmmxt_wmiawbbn): Likewise.
12700 (iwmmxt_wmiawbtn): Likewise.
12701 (iwmmxt_wmiawtbn): Likewise.
12702 (iwmmxt_wmiawttn): Likewise.
12703 (iwmmxt_wmerge): Likewise.
12704 (iwmmxt_tandc<mode>3): Likewise.
12705 (iwmmxt_torc<mode>3): Likewise.
12706 (iwmmxt_torvsc<mode>3): Likewise.
12707 (iwmmxt_textrc<mode>3): Likewise.
12708 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12709 (wmmxt_pack): Likewise.
12710 (wmmxt_mult_c1): Likewise.
12711 (wmmxt_mult_c2): Likewise.
12712 (wmmxt_alu_c1): Likewise.
12713 (wmmxt_alu_c2): Likewise.
12714 (wmmxt_alu_c3): Likewise.
12715 (wmmxt_transfer_c1): Likewise.
12716 (wmmxt_transfer_c2): Likewise.
12717 (wmmxt_transfer_c3): Likewise.
12718 (marvell_f_iwmmxt_wstr): Likewise.
12719 (marvell_f_iwmmxt_wldr): Likewise.
12720
bdb7bf8a
YZ
127212013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12722
0ba6aff3 12723 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
bdb7bf8a 12724
e6631d38
VM
127252013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12726
12727 Revert:
12728 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12729 * lra-constraints.c (need_for_split_p): Check call used hard regs
12730 living through calls.
12731
12732 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12733 call used regs for call insn.
12734
53c1275b
JJ
127352013-06-28 Jakub Jelinek <jakub@redhat.com>
12736
12737 PR target/57736
0ba6aff3
UB
12738 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12739 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
53c1275b 12740
1141ed3f
BI
127412013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12742
12743 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
e6631d38 12744
e32e4c4a
VM
127452013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12746
12747 * lra-constraints.c (need_for_split_p): Check call used hard regs
12748 living through calls.
12749
f161bfd3
MM
127502013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
12751
12752 PR target/57744
12753 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
12754 to tie with any other modes. Eliminate Altivec vector mode tests,
12755 since these are a subset of ALTIVEC or VSX vector modes. Simplify
12756 code, to return 0 if testing MODE2 for a condition, if we've
12757 already tested MODE1 for the same condition.
12758
7eda14e1
MS
127592013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12760
12761 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
12762 layout.
12763
c822f852
MS
127642013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12765
12766 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12767 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
12768
da4f13a4
MS
127692013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12770
0ba6aff3
UB
12771 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
12772 Define.
da4f13a4
MS
12773 (aarch64_symbolic_constant_p): Remove.
12774 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
12775 static. Fix line length and white space.
12776 (aarch64_symbolic_constant_p): Remove.
12777 * config/aarch64/predicates.md (aarch64_valid_symref):
12778 Use aarch64_classify_symbol_expression.
12779
956a95a5
KT
127802013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12781
12782 * config/arm/constraints.md (Ts): New constraint.
12783 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
12784 16-bit encodings.
12785 (compare_scc): Use "Ts" constraint for operand 0.
12786 (ior_scc_scc): Likewise.
12787 (and_scc_scc): Likewise.
12788 (and_scc_scc_nodom): Likewise.
12789 (ior_scc_scc_cmp): Likewise for operand 7.
12790 (and_scc_scc_cmp): Likewise.
12791 * config/arm/thumb2.md (thumb2_movsi_insn):
12792 Add alternatives for 16-bit encodings.
12793 (thumb2_movhi_insn): Likewise.
12794 (thumb2_movsicc_insn): Likewise.
12795 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
12796 (thumb2_negscc): Use "Ts" constraint.
12797 Move mvn instruction outside cond_exec block.
12798 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
12799 for 16-bit encodings.
12800
31dfce10
KT
128012013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12802
12803 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
12804 encoding.
12805 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
12806 (mulsi3subsi): Likewise.
12807 (mulsidi3adddi): Likewise.
12808 (mulsidi3_v6): Likewise.
12809 (umulsidi3_v6): Likewise.
12810 (umulsidi3adddi_v6): Likewise.
12811 (smulsi3_highpart_v6): Likewise.
12812 (umulsi3_highpart_v6): Likewise.
12813 (mulhisi3tb): Likewise.
12814 (mulhisi3bt): Likewise.
12815 (mulhisi3tt): Likewise.
12816 (maddhisi4): Likewise.
12817 (maddhisi4tb): Likewise.
12818 (maddhisi4tt): Likewise.
12819 (maddhidi4): Likewise.
12820 (maddhidi4tb): Likewise.
12821 (maddhidi4tt): Likewise.
12822 (zeroextractsi_compare0_scratch): Likewise.
12823 (insv_zero): Likewise.
12824 (insv_t2): Likewise.
12825 (anddi_notzesidi_di): Likewise.
12826 (anddi_notsesidi_di): Likewise.
12827 (andsi_notsi_si): Likewise.
12828 (iordi_zesidi_di): Likewise.
12829 (xordi_zesidi_di): Likewise.
12830 (andsi_iorsi3_notsi): Likewise.
12831 (smax_0): Likewise.
12832 (smax_m1): Likewise.
12833 (smin_0): Likewise.
12834 (not_shiftsi): Likewise.
12835 (unaligned_loadsi): Likewise.
12836 (unaligned_loadhis): Likewise.
12837 (unaligned_loadhiu): Likewise.
12838 (unaligned_storesi): Likewise.
12839 (unaligned_storehi): Likewise.
12840 (extv_reg): Likewise.
12841 (extzv_t2): Likewise.
12842 (divsi3): Likewise.
12843 (udivsi3): Likewise.
12844 (arm_zero_extendhisi2addsi): Likewise.
12845 (arm_zero_extendqisi2addsi): Likewise.
12846 (compareqi_eq0): Likewise.
12847 (arm_extendhisi2_v6): Likewise.
12848 (arm_extendqisi2addsi): Likewise.
12849 (arm_movt): Likewise.
12850 (thumb2_ldrd): Likewise.
12851 (thumb2_ldrd_base): Likewise.
12852 (thumb2_ldrd_base_neg): Likewise.
12853 (thumb2_strd): Likewise.
12854 (thumb2_strd_base): Likewise.
12855 (thumb2_strd_base_neg): Likewise.
12856 (arm_negsi2): Add alternative for 16-bit encoding.
12857 (arm_one_cmplsi2): Likewise.
12858
1572e697
KT
128592013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12860
12861 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
12862 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
12863 (movdfcc): Likewise.
12864 * config/arm/vfp.md (*thumb2_movsf_vfp):
12865 Disable predication for arm_restrict_it.
12866 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
12867 (*thumb2_movdfcc_vfp): Likewise.
12868 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
12869 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
12870 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
12871 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
12872 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
12873 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
12874 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
12875 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
12876 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
12877 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
12878 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
12879 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
12880 Disable predication for arm_restrict_it.
12881
ebf8f0ea
KY
128822013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
12883
12884 * config/i386/bmiintrin.h (_bextr_u32): New.
12885 (_bextr_u64): Ditto.
12886
a290fcda
RS
128872013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
12888
12889 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
12890 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
12891 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
12892 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
12893 * config/mips/n32-elf.h: ...this new file.
12894
ea169996
MG
128952013-06-27 Marc Glisse <marc.glisse@inria.fr>
12896
12897 PR target/57224
12898 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
12899 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
12900
12211b99 129012013-06-27 Catherine Moore <clm@codesourcery.com>
8e932114
CM
12902
12903 * config/mips/mips-tables.opt: Regenerate.
12904 * config/mips/mips-cpus.def: Add m14ke and m14kec.
0ba6aff3 12905 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
8e932114 12906 * doc/invoke.texi: Add -m14kc.
ebf8f0ea 12907
2444244e
JJ
129082013-06-27 Jakub Jelinek <jakub@redhat.com>
12909
cd06d2a2
JJ
12910 PR target/57623
12911 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
12912 constraints of operand 1 and 2.
12913
2444244e
JJ
12914 PR target/57623
12915 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
12916 to match RTL canonicalization. Swap predicates and
12917 constraints of operand 1 and 2.
12918
2f259720
VM
129192013-06-27 Vladimir Makarov <vmakarov@redhat.com>
12920
0ba6aff3
UB
12921 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
12922 Process OP_INOUT regs for splitting too.
2f259720 12923
fef4d2b3
JJ
129242013-06-27 Jakub Jelinek <jakub@redhat.com>
12925
12926 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
12927 decl before the loop, initialize to NULL.
12928 (vectorizable_load): Initialize ptr_incr to NULL.
12929
79ee9826
MJ
129302013-06-27 Martin Jambor <mjambor@suse.cz>
12931
12932 PR lto/57208
12933 * ipa-ref.h (ipa_maybe_record_reference): Declare.
12934 * ipa-ref.c (ipa_maybe_record_reference): New function.
12935 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
12936 * ipa-cp.c (create_specialized_node): Record potential references from
12937 aggvals.
12938 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
12939
e18b4a81
YZ
129402013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12941
12942 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
12943 parameter 'mode' of type 'enum machine_mode mode'; change to pass
12944 'mode' to force_reg.
12945 (aarch64_add_offset): Update calls to aarch64_force_temporary.
12946 (aarch64_expand_mov_immediate): Likewise.
12947
9c023bf0
YZ
129482013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12949
12950 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
12951 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
12952
5a3fe9b6
AK
129532013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12954
12955 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
12956 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12957 (struct machine_function): Add tbegin_p.
12958 (s390_canonicalize_comparison): Fold CC mode compares to
12959 conditional jump if possible.
12960 (s390_emit_jump): Return the emitted jump.
12961 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
12962 Handle CCRAWmode compares.
12963 (s390_option_override): Default to -mhtm if available.
12964 (s390_reg_clobbered_rtx): Handle floating point regs as well.
12965 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
12966 FPRs instead of df_regs_ever_live_p.
12967 (s390_optimize_nonescaping_tx): New function.
12968 (s390_init_frame_layout): Extend clobbered_regs array to cover
12969 FPRs as well.
12970 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
12971 (s390_expand_tbegin): New function.
12972 (enum s390_builtin): New enum definition.
12973 (code_for_builtin): New array definition.
12974 (s390_init_builtins): New function.
12975 (s390_expand_builtin): New function.
12976 (TARGET_INIT_BUILTINS): Define.
12977 (TARGET_EXPAND_BUILTIN): Define.
0ba6aff3 12978 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
5a3fe9b6
AK
12979 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
12980 (s390_alc_comparison): Likewise.
12981 * config/s390/s390-modes.def: Add CCRAWmode.
12982 * config/s390/s390.h (processor_flags): Add PF_TX.
12983 (TARGET_CPU_HTM): Define macro.
12984 (TARGET_HTM): Define macro.
12985 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
12986 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
12987 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12988 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
0ba6aff3 12989 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
5a3fe9b6
AK
12990 (TBEGIN_MASK, TBEGINC_MASK): New constants.
12991 ("*cc_to_int"): Move up.
12992 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
12993 constants other than 0.
12994 ("*ccraw_to_int"): New insn and splitter definition.
12995 ("tbegin", "tbegin_nofloat", "tbegin_retry")
12996 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
12997 ("tx_assist"): New expander.
12998 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
12999 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
13000 * config/s390/s390.opt: Add -mhtm option.
13001 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
13002 * config/s390/htmxlintrin.h: New file.
13003 * config/s390/htmintrin.h: New file.
13004 * config/s390/s390intrin.h: New file.
13005 * doc/extend.texi: Document htm builtins.
13006 * config.gcc: Add the new header files to extra_headers.
13007
9e216629
TS
130082013-06-26 Thomas Schwinge <thomas@codesourcery.com>
13009
13010 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
0ba6aff3 13011 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
9e216629 13012
fd8c65e7 130132013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
13014 Pat Haugen <pthaugen@us.ibm.com>
13015 Peter Bergner <bergner@vnet.ibm.com>
fd8c65e7
MM
13016
13017 * config/rs6000/power8.md: New.
13018 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
13019 setting for power8 entry.
13020 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
13021 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
13022 test for Power4/Power5 only.
13023 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
13024 support.
13025 (force_new_group): Adjust comment.
13026 * config/rs6000/rs6000.md: Include power8.md.
13027
b24a2ce5
GY
130282013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
13029
13030 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
13031 * config/arm/arm-protos.h (arm_max_conditional_execute): New
13032 declaration.
13033 (tune_params): Update comment.
13034 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13035 (arm_max_conditional_execute): New function.
13036 (thumb2_final_prescan_insn): Use max_insn_skipped and
13037 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13038
f6fdeb62
JJ
130392013-06-25 Jakub Jelinek <jakub@redhat.com>
13040
13041 PR tree-optimization/57705
13042 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13043 SSA_NAME step, provided that it is not defined inside the loop.
0ba6aff3 13044 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
f6fdeb62
JJ
13045 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13046
c13bc3d9
MJ
130472013-06-25 Martin Jambor <mjambor@suse.cz>
13048
13049 PR middle-end/57670
13050 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13051 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13052 calls in the dump.
13053 (ipa_note_param_call): Initialize member_ptr flag.
13054 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13055 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13056 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13057 (ipa_read_indirect_edge_info): Likewise.
13058
8a845901
RB
130592013-06-25 Richard Biener <rguenther@suse.de>
13060
13061 PR middle-end/56977
13062 * passes.c (init_optimization_passes): Move pass_fold_builtins
13063 and pass_dce earlier with -Og.
13064
33be0bec
EB
130652013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13066
13067 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13068 <BIT_FIELD_REF>: Remove trailing TAB.
13069 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13070 remove blank line.
13071
5fe8e757
MJ
130722013-06-24 Martin Jambor <mjambor@suse.cz>
13073
13074 PR tree-optimization/57358
13075 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13076 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13077 (ipa_analyze_params_uses): Generate pessimistic info when true.
13078
44a60244
MJ
130792013-06-24 Martin Jambor <mjambor@suse.cz>
13080
13081 PR tree-optimization/57539
13082 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13083 global.inlined_to of the new node to it. All callers changed.
13084 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13085 inlining_into, pass it to cgraph_clone_node.
13086 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13087 ipa_free_edge_args_substructures.
13088 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13089 rdesc linked list. Do not assert rdesc edges have inlined caller.
13090 Assert we have found an rdesc in the rdesc list.
13091
7c5848b8
RB
130922013-06-24 Richard Biener <rguenther@suse.de>
13093
0ba6aff3 13094 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
7c5848b8
RB
13095 (pointer_set_lookup): Declare.
13096 (class pointer_map): New template class implementing a
13097 generic pointer to T map.
13098 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13099 pointer_map<T>::contains, pointer_map<T>::insert,
13100 pointer_map<T>::traverse): New functions.
13101 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13102 (pointer_set_lookup): New function.
13103 (pointer_set_contains): Use pointer_set_lookup.
13104 (pointer_set_insert): Likewise.
13105 (insert_aux): Remove.
13106 (struct pointer_map_t): Embed a pointer_set_t.
13107 (pointer_map_create): Adjust.
13108 (pointer_map_destroy): Likewise.
13109 (pointer_map_contains): Likewise.
13110 (pointer_map_insert): Likewise.
13111 (pointer_map_traverse): Likewise.
13112 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13113 pointer_map<unsigned> instead of a pointer_map_t.
13114 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13115 (streamer_tree_cache_lookup): Likewise.
13116 (streamer_tree_cache_create): Likewise.
13117 (streamer_tree_cache_delete): Likewise.
13118 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13119 pointer_map<unsigned> instead of a pointer_map_t.
13120 (lto_init_tree_ref_encoder): Adjust.
13121 (lto_destroy_tree_ref_encoder): Likewise.
13122 * lto-section-out.c (lto_output_decl_index): Likewise.
13123 (lto_record_function_out_decl_state): Likewise.
13124 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13125
3dbc97a9
RB
131262013-06-24 Richard Biener <rguenther@suse.de>
13127
13128 PR tree-optimization/57488
13129 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13130
8345965a
AM
131312013-06-24 Alan Modra <amodra@gmail.com>
13132
13133 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13134 (gen_easy_altivec_constant): Likewise.
13135 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13136 easy_vector_constant_msb): Likewise.
13137
b7df379f
JJ
131382013-06-23 Jakub Jelinek <jakub@redhat.com>
13139
13140 PR target/57688
13141 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13142 add missing return true.
13143
12211b99 131442013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
0bcf9a09
OE
13145
13146 PR target/52483
13147 * config/sh/predicates.md (general_extend_operand): Invoke
13148 general_movsrc_operand for memory operands.
13149 (general_movsrc_operand): Allow reg+reg addressing, do not use
13150 general_operand for memory operands.
13151
97db2bf7
ST
131522013-06-23 Sriraman Tallam <tmsriram@google.com>
13153
13154 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13155 when current target options does not apply.
13156 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13157 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
0ba6aff3
UB
13158 * config/i386/bmiintrin.h: Pass appropriate target
13159 attributes to header.
97db2bf7
ST
13160 * config/i386/mmintrin.h: Ditto.
13161 * config/i386/nmmintrin.h: Ditto.
13162 * config/i386/avx2intrin.h: Ditto.
13163 * config/i386/fxsrintrin.h: Ditto.
13164 * config/i386/tbmintrin.h: Ditto.
13165 * config/i386/xsaveintrin.h: Ditto.
13166 * config/i386/f16cintrin.h: Ditto.
13167 * config/i386/xtestintrin.h: Ditto.
13168 * config/i386/xsaveoptintrin.h: Ditto.
13169 * config/i386/bmi2intrin.h: Ditto.
13170 * config/i386/lzcntintrin.h: Ditto.
13171 * config/i386/smmintrin.h: Ditto.
13172 * config/i386/wmmintrin.h: Ditto.
13173 * config/i386/x86intrin.h: Remove all header include guards.
13174 * config/i386/prfchwintrin.h: Ditto.
13175 * config/i386/pmmintrin.h: Ditto.
13176 * config/i386/tmmintrin.h: Ditto.
13177 * config/i386/xmmintrin.h: Ditto.
13178 * config/i386/popcntintrin.h: Ditto.
13179 * config/i386/rdseedintrin.h: Ditto.
13180 * config/i386/ammintrin.h: Ditto.
13181 * config/i386/emmintrin.h: Ditto.
13182 * config/i386/immintrin.h: Remove all header include guards.
13183 * config/i386/fma4intrin.h: Ditto.
13184 * config/i386/lwpintrin.h: Ditto.
13185 * config/i386/xopintrin.h: Ditto.
13186 * config/i386/ia32intrin.h: Ditto.
13187 * config/i386/avxintrin.h: Ditto.
13188 * config/i386/rtmintrin.h: Ditto.
13189 * config/i386/fmaintrin.h: Ditto.
13190 * config/i386/mm3dnow.h: Ditto.
13191
495e6879
ST
131922013-06-22 Sriraman Tallam <tmsriram@google.com>
13193
13194 * common/config/i386/i386-common.c: Handle LZCNT.
13195
11e8f687
AK
131962013-06-22 Andi Kleen <ak@linux.intel.com>
13197
13198 * doc/extend.texi: Use __atomic_store_n instead of
13199 __atomic_store in HLE example.
13200
12211b99 132012013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
664dd721
OE
13202
13203 * config/sh/sh.c: Remove <cstdlib> workaround.
13204
927f908b
AK
132052013-06-21 Andi Kleen <ak@linux.intel.com>
13206
0ba6aff3 13207 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
927f908b 13208
cca410b6
AK
132092013-06-21 Andi Kleen <ak@linux.intel.com>
13210
13211 * doc/extend.texi: Document that __atomic_clear and
13212 __atomic_test_and_set should only be used with bool.
13213
d84db1ed
JH
132142013-06-20 Jan Hubicka <jh@suse.cz>
13215
13216 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13217 types_same_for_odr.
13218 * tree.c (decls_same_for_odr): New function.
13219 (same_for_edr): New function.
13220 (types_same_for_odr): New function.
13221 (get_binfo_at_offset): Use it.
13222 * tree.h (types_same_for_odr): Declare.
13223
12211b99 132242013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
469a31a9
OE
13225 Jason Merrill <jason@redhat.com>
13226
13227 * system.h: Include <cstdlib> as well as <stdlib.h>.
13228
49b4ff63
UB
132292013-06-20 Uros Bizjak <ubizjak@gmail.com>
13230
13231 PR target/57655
13232 * config/i386/i386.c (construct_container): Report error if
13233 long double is used with disabled x87 float returns.
13234
7a89ad00
JH
132352013-06-20 Jan Hubicka <jh@suse.cz>
13236
13237 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13238
12211b99 132392013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
bd280792
JR
13240
13241 PR rtl-optimization/57425
13242 PR rtl-optimization/57569
13243 * alias.c (write_dependence_p): Remove parameters mem_mode and
13244 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13245 Changed all callers.
13246 (canon_anti_dependence): Get comments and semantics in sync.
13247 Add parameter mem_canonicalized. Changed all callers.
13248 * rtl.h (canon_anti_dependence): Update prototype.
13249
d16e9a99
RB
132502013-06-20 Richard Biener <rguenther@suse.de>
13251
13252 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13253 byte case, inline streamer_read_uchar and defer section
13254 overrun check.
13255
c3a9b91b
RB
132562013-06-20 Richard Biener <rguenther@suse.de>
13257
13258 PR tree-optimization/57584
13259 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13260 SSA names into the expanded expression that take part in
13261 abnormal coalescing.
13262
371e77e3 132632013-06-19 Sharad Singhai <singhai@google.com>
fc8a650e
SS
13264
13265 * gcov.c (print_usage): Handle new option.
13266 (process_args): Ditto.
13267 (get_gcov_intermediate_filename): New function.
13268 (output_intermediate_file): New function.
13269 (output_gcov_file): New function
13270 (generate_results): Handle new option.
13271 (release_function): Relase demangled name.
13272 (read_graph_file): Handle demangled name.
13273 (output_lines): Ditto.
13274 * doc/gcov.texi: Document gcov intermediate format.
13275
11f5b71a
VM
132762013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13277
13278 PR bootstrap/57604
13279 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13280 (lra_emit_add): Use the functions. Add comment about Y as an
13281 address segment.
13282
e1f9340b
DE
132832013-06-19 David Edelsohn <dje.gcc@gmail.com>
13284
13285 PR driver/57652
13286 * collect2.c (collect_atexit): New.
13287 (collect_exit): Delete.
13288 (main): Register collect_atexit with atexit.
13289 (collect_wait): Change collect_exit to exit.
13290 (do_wait): Same.
13291 * collect2.h (collect_exit): Delete.
13292 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13293
e3f9e0ac
WM
132942013-06-19 Wei Mi <wmi@google.com>
13295
13296 PR rtl-optimization/57518
13297 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13298 if regno is used in paradoxical subreg.
13299 (update_equiv_regs): Check pdx_subregs[regno] before
13300 set a reg to be equivalent with a mem.
13301
a09f1a76
MK
133022013-06-19 Matthias Klose <doko@ubuntu.com>
13303
13304 PR driver/57651
13305 * file-find.h (find_a_file): Add a mode parameter.
13306 * file-find.c (find_a_file): Likewise.
13307 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13308 with X_OK for the executables.
13309 * collect2.c (main): Call find_a_file with X_OK.
13310
7d18b0ad
SE
133112013-06-19 Steve Ellcey <sellcey@mips.com>
13312
13313 PR target/56942
13314 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13315 Use NEXT_INSN instead of next_real_insn.
13316
6a6dac52
JH
133172013-06-19 Jan Hubicka <jh@suse.cz>
13318
13319 * cgraph.h (const_value_known_p): Replace by ...
13320 (ctor_for_folding): .. this one.
13321 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13322 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13323 * expr.c (expand_expr_real_1): Likewise.
13324 (string_constant): Likewise.
13325 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13326 * ipa.c (process_references): Likewise.
13327 (symtab_remove_unreachable_nodes): Likewise.
13328 * ipa-inline-analysis.c (param_change_prob): Likewise.
13329 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13330 (get_base_constructor): Likwise.
13331 * varpool.c (varpool_remove_node): Likewise.
13332 (varpool_remove_initializer): LIkewise.
13333 (dump_varpool_node): LIkwise.
13334 (const_value_known_p): Rewrite to ...
13335 (ctor_for_folding): ... this one.
13336
216c12ab
JJ
133372013-06-19 Jakub Jelinek <jakub@redhat.com>
13338
13339 PR driver/57651
13340 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13341 PERSONALITY in $PATH derived prefixes.
13342
d53e2f99
JL
133432013-06-19 Jeff Law <law@redhat.com>
13344
ecdbf306
JL
13345 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13346 in comment.
13347
d53e2f99
JL
13348 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13349 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13350 booleans.
13351
89e7fc04
SN
133522013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13353
13354 * config/arm/vfp.md: Move VFP instruction classification documentation
13355 to ...
13356 * config/arm/arm.md: ... here. Update instruction classification
13357 documentation.
13358
5e7f6aaa
RE
133592013-06-19 Richard Earnshaw <rearnsha@arm.com>
13360
13361 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13362 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13363 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13364
5fdc1e5d
SB
133652013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13366
13367 PR target/57609
13368 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13369 with NEXT_INSN. Use tablejump_p to check for jump table data
13370 insns.
13371
e952afba
PC
133722013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13373
13374 PR c++/56544
13375 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13376 that now in C++ the value is correct per the C++ standards.
13377
3fd0303a
RB
133782013-06-19 Richard Biener <rguenther@suse.de>
13379
13380 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13381 for global context.
13382
b04e4b08
AK
133832013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13384
13385 Revert:
13386 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13387
13388 PR target/57609
13389 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13390 with next_active_insn.
13391
bef8491a
ST
133922013-06-18 Sriraman Tallam <tmsriram@google.com>
13393
13394 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13395 functions are inlined during failures to flag an error.
13396 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13397 in early inline pass.
22fc7d3d 13398
c21b257d
L
133992013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13400
13401 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13402 in comments.
13403
33255ae3
JB
134042013-06-18 Julian Brown <julian@codesourcery.com>
13405
13406 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13407 Permit virtual register pre-reload if !strict.
13408 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13409 change.
13410 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13411 prototype.
13412 * config/arm/neon.md (movmisalign<mode>): Use
13413 neon_perm_struct_or_reg_operand instead of
13414 neon_struct_or_register_operand.
13415 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13416 neon_permissive_struct_operand instead of neon_struct_operand.
13417 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13418 neon_vector_mem_operand.
13419 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13420 neon_vector_mem_operand.
13421 (neon_permissive_struct_operand): New.
13422 (neon_struct_or_register_operand): Rename to...
13423 (neon_perm_struct_or_reg_operand): This. Adjust call to
13424 neon_vector_mem_operand.
13425
d579fcda
RB
134262013-06-18 Richard Biener <rguenther@suse.de>
13427
13428 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13429 * lto-streamer.h: Include pointer-set.h.
13430 (struct lto_decl_slot): Remove.
13431 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13432 Remove next_index entry.
13433 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13434 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13435 (lto_init_tree_ref_encoder): Adjust.
13436 (lto_destroy_tree_ref_encoder): Likewise.
13437 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13438 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13439 (lto_output_decl_index): Adjust.
13440 (lto_new_out_decl_state): Likewise.
13441 (lto_record_function_out_decl_state): Likewise.
13442 * lto-streamer-out.c (copy_function): Likewise.
13443
31ee20ba
RB
134442013-06-18 Richard Biener <rguenther@suse.de>
13445
13446 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13447 * cgraphunit.c: Include cfgloop.h.
13448 (init_lowered_empty_function): Initialize the loop tree.
13449 (assemble_thunk): Insert new BBs into loops.
13450
ac0511f2
RB
134512013-06-18 Richard Biener <rguenther@suse.de>
13452
13453 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13454 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13455 the map from cache entry to cache index optional.
13456 (streamer_tree_cache_replace_tree): Adjust accordingly.
13457 (streamer_tree_cache_append): Likewise.
13458 (streamer_tree_cache_delete): Likewise.
13459 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13460 streamer cache map from cache entry to cache index.
13461 * lto-streamer-out.c (create_output_block): Adjust.
13462
09485a08
SN
134632013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13464
22fc7d3d
UB
13465 * config/arm/arm.md (attribute "insn"): Move multiplication and
13466 division attributes to...
09485a08
SN
13467 (attribute "type"): ... here. Remove mult.
13468 (attribute "mul32"): New attribute.
13469 (attribute "mul64"): Add umaal.
13470 (*arm_mulsi3): Update attributes.
13471 (*arm_mulsi3_v6): Likewise.
13472 (*thumb_mulsi3): Likewise.
13473 (*thumb_mulsi3_v6): Likewise.
13474 (*mulsi3_compare0): Likewise.
13475 (*mulsi3_compare0_v6): Likewise.
13476 (*mulsi_compare0_scratch): Likewise.
13477 (*mulsi_compare0_scratch_v6): Likewise.
13478 (*mulsi3addsi): Likewise.
13479 (*mulsi3addsi_v6): Likewise.
13480 (*mulsi3addsi_compare0): Likewise.
13481 (*mulsi3addsi_compare0_v6): Likewise.
13482 (*mulsi3addsi_compare0_scratch): Likewise.
13483 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13484 (*mulsi3subsi): Likewise.
13485 (*mulsidi3adddi): Likewise.
13486 (*mulsi3addsi_v6): Likewise.
13487 (*mulsidi3adddi_v6): Likewise.
13488 (*mulsidi3_nov6): Likewise.
13489 (*mulsidi3_v6): Likewise.
13490 (*umulsidi3_nov6): Likewise.
13491 (*umulsidi3_v6): Likewise.
13492 (*umulsidi3adddi): Likewise.
13493 (*umulsidi3adddi_v6): Likewise.
13494 (*smulsi3_highpart_nov6): Likewise.
13495 (*smulsi3_highpart_v6): Likewise.
13496 (*umulsi3_highpart_nov6): Likewise.
13497 (*umulsi3_highpart_v6): Likewise.
13498 (mulhisi3): Likewise.
13499 (*mulhisi3tb): Likewise.
13500 (*mulhisi3bt): Likewise.
13501 (*mulhisi3tt): Likewise.
13502 (maddhisi4): Likewise.
13503 (*maddhisi4tb): Likewise.
13504 (*maddhisi4tt): Likewise.
13505 (maddhidi4): Likewise.
13506 (*maddhidi4tb): Likewise.
13507 (*maddhidi4tt): Likewise.
13508 (divsi3): Likewise.
13509 (udivsi3): Likewise.
13510 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13511 (thumb2_mulsi_short_compare0): Likewise.
13512 (thumb2_mulsi_short_compare0_scratch): Likewise.
13513 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13514 (1020mult2): Likewise.
13515 (1020mult3): Likewise.
13516 (1020mult4): Likewise.
13517 (1020mult5): Likewise.
13518 (1020mult6): Likewise.
22fc7d3d
UB
13519 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13520 change.
09485a08
SN
13521 (cortex_a15_mult64): Likewise.
13522 (cortex_a15_sdiv): Likewise.
13523 (cortex_a15_udiv): Likewise.
13524 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13525 (mult2): Likewise.
13526 (mult3): Likewise.
13527 (mult4): Likewise.
13528 (mult5): Likewise.
13529 (mult6): Likewise.
13530 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13531 (pj4_ir_div): Likewise.
13532 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13533 (11_mult2): Likewise.
13534 (11_mult3): Likewise.
13535 (11_mult4): Likewise.
13536 (11_mult5): Likewise.
13537 (11_mult6): Likewise.
13538 (11_mult7): Likewise.
13539 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13540 (cortex_a8_mla): Likewise.
13541 (cortex_a8_mull): Likewise.
13542 (cortex_a8_smulwy): Likewise.
13543 (cortex_a8_smlald): Likewise.
13544 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13545 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13546 (cortex_r4_mul_3): Likewise.
13547 (cortex_r4_mla_4): Likewise.
13548 (cortex_r4_mla_3): Likewise.
13549 (cortex_r4_smlald): Likewise.
13550 (cortex_r4_mull): Likewise.
13551 (cortex_r4_sdiv): Likewise.
13552 (cortex_r4_udiv): Likewise.
13553 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13554 (cortex_a7_idiv): Likewise.
13555 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13556 (9_mult2): Likewise.
13557 (9_mult3): Likewise.
13558 (9_mult4): Likewise.
13559 (9_mult5): Likewise.
13560 (9_mult6): Likewise.
13561 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13562 (cortex_a53_sdiv): Likewise.
13563 (cortex_a53_udiv): Likewise.
13564 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13565 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13566 (mp626_mult2): Likewise.
13567 (mp626_mult3): Likewise.
13568 (mp626_mult4): Likewise.
13569 * config/arm/fa526.md (526_mult1): Update attribute change.
13570 (526_mult2): Likewise.
13571 * config/arm/arm-generic.md (mult): Update attribute change.
13572 (mult_ldsched_strongarm): Likewise.
13573 (mult_ldsched): Likewise.
13574 (multi_cycle): Likewise.
13575 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13576 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13577 (606te_mult2): Likewise.
13578 (606te_mult3): Likewise.
13579 (606te_mult4): Likewise.
13580 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13581 (cortex_a9_mac16): Likewise.
13582 (cortex_a9_multiply): Likewise.
13583 (cortex_a9_mac): Likewise.
13584 (cortex_a9_multiply_long): Likewise.
13585 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13586 (626te_mult2): Likewise.
13587 (626te_mult3): Likewise.
13588 (626te_mult4): Likewise.
13589
8951f345
RB
135902013-06-18 Richard Biener <rguenther@suse.de>
13591
13592 PR lto/57334
13593 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13594
d8101d05
AK
135952013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13596
13597 PR target/57609
13598 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13599 with next_active_insn.
13600
69eff9da
AM
136012013-06-18 Alan Modra <amodra@gmail.com>
13602
13603 * config/rs6000/rs6000.h (enum data_align): New.
13604 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13605 (DATA_ABI_ALIGNMENT): Define.
13606 (CONSTANT_ALIGNMENT): Correct comment.
13607 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13608 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13609
8ed00d76
DM
136102013-06-17 David Malcolm <dmalcolm@redhat.com>
13611
13612 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13613 ATTRIBUTE_UNUSED marking.
13614
3d33d06b
SN
136152013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13616
13617 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13618 alternative and update.
13619 (aarch64_dup_lanedi): Delete.
13620 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13621 * config/aarch64/aarch64-simd-builtins.def: Update.
13622
ee03e71d
RB
136232013-06-17 Richard Biener <rguenther@suse.de>
13624
13625 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13626 (lto_input_scc): Declare.
13627 (lto_input_tree_1): Likewise.
13628 (struct lto_stats_d): Add num_tree_bodies_output and
13629 num_pickle_refs_output.
13630 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13631 (lto_read_tree_1): Split out from ...
13632 (lto_read_tree): ... this.
13633 (lto_input_scc): New function.
13634 (lto_input_tree_1): Split out from ...
13635 (lto_input_tree): ... this. Handle LTO_tree_scc.
13636 (lto_data_in_create): Create the streamer cache without hashes.
13637 * lto-streamer-out.c (create_output_block): Create the streamer
13638 cache with hashes when not doing WPA.
13639 (lto_write_tree_1): Split out from ...
13640 (lto_write_tree): ... this.
13641 (get_symbol_initial_value): New function.
13642 (lto_output_tree_1): Split out from ...
13643 (lto_output_tree): ... this. Write trees as series of SCCs
13644 using a DFS walk via DFS_write_tree.
13645 (struct sccs, struct scc_entry): New types.
13646 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13647 (DFS_write_tree_body): New function.
13648 (DFS_write_tree): Likewise.
13649 (hash_tree): Likewise.
13650 (scc_entry_compare): Likewise.
13651 (hash_scc): Likewise.
13652 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13653 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13654 TREE_CHAIN as regular reference.
13655 (streamer_read_integer_cst): Remove.
13656 (streamer_get_pickled_tree): Adjust.
13657 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13658 of DECL_EXTERNALs in BLOCK_VARS for now.
13659 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13660 reference.
13661 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13662 Add hash value argument and record that if hashes are recorded
13663 in the cache.
13664 (streamer_tree_cache_insert_1): Adjust.
13665 (streamer_tree_cache_insert): Likewise.
13666 (streamer_tree_cache_insert_at): Rename to ...
13667 (streamer_tree_cache_replace_tree): ... this and adjust.
13668 (streamer_tree_cache_append): Adjust.
13669 (record_common_node): Likewise.
13670 (streamer_tree_cache_create): Add argument whether to
13671 record hash values together with trees.
13672 (streamer_tree_cache_delete): Adjust.
13673 * tree-streamer.h (struct streamer_tree_cache_d): Add
13674 vector of hashes.
13675 (streamer_read_integer_cst): Remove.
13676 (streamer_tree_cache_insert): Adjust.
13677 (streamer_tree_cache_append): Likewise.
13678 (streamer_tree_cache_insert_at): Rename to ...
13679 (streamer_tree_cache_replace_tree): ... this and adjust.
13680 (streamer_tree_cache_create): Add argument whether to record hashes.
13681 (streamer_tree_cache_get): Rename to ...
13682 (streamer_tree_cache_get_tree): ... this.
13683 (streamer_tree_cache_get_hash): New function.
13684 * tree.c (cache_integer_cst): New function.
13685 * tree.h (cache_integer_cst): Declare.
13686 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13687 * lto-symtab.c (lto_varpool_replace_node): Only release
13688 DECL_INITIAL of non-prevailing decls.
13689 * varpool.c (varpool_remove_initializer): Do not release
13690 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13691
0de86a92
JU
136922013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13693
13694 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13695 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13696 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13697 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13698 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13699 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13700 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13701 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13702 instead of TARGET_64BIT.
13703 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13704 Require ISA_HAS_<D>DIV.
13705
74240413
RS
137062013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13707
13708 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13709 (mips*-*-linux*): Move default with_llsc setting to where other
13710 defaults are set.
13711 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13712 with_arch block.
13713 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13714 Likewise. Remove default with_tune setting. Move default float
13715 setting to its own block. Handle with_llsc in the same block as above.
13716
12211b99 137172013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
393f9fed
JR
13718
13719 PR rtl-optimization/57425
13720 PR rtl-optimization/57569
299a5f6a 13721 * alias.c (write_dependence_p): Add new parameters mem_mode,
393f9fed
JR
13722 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13723 Changed all callers.
13724 (canon_anti_dependence): New function.
13725 * cse.c (check_dependence): Use canon_anti_dependence.
13726 * cselib.c (cselib_invalidate_mem): Likewise.
13727 * rtl.h (canon_anti_dependence): Declare.
13728
b259d352
JU
137292013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13730
13731 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13732 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13733 ".set mips3" for 64-bit targets.
13734
4de09b85
DC
137352013-06-15 Dehao Chen <dehao@google.com>
13736
13737 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13738 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13739 (gimple_check_call_args): Likewise.
13740 * value-prof.c (check_ic_target): Likewise.
13741 * ipa-inline.c (early_inliner): Likewise.
13742 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13743 * cgraph.c (cgraph_create_edge_1): Likewise.
13744 (cgraph_make_edge_direct): Likewise.
13745
d1aa4a76
MM
137462013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
13747
13748 PR target/57615
13749 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
13750 rs6000_output_move_128bit to handle emitting quad memory
13751 operations. Set attribute length to 8 bytes.
13752
12211b99 137532013-06-14 Vidya Praveen <vidyapraveen@arm.com>
24244e4d
VP
13754
13755 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
13756 New pattern.
13757 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
13758 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
13759 (aarch64_<su>mlsl<mode>): Likewise.
13760
f5273574
MS
137612013-06-14 Mike Stump <mikestump@comcast.net>
13762
13763 * Makefile.in (TARGET_H): Add insn-codes.h.
13764
ef7e7ab2
AM
137652013-06-14 Alan Modra <amodra@gmail.com>
13766
13767 PR middle-end/57134
13768 PR middle-end/57586
13769 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
13770 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
13771 bitfield expansion when EXPAND_MEMORY.
13772 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
13773
135faab6
MM
137742013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
13775
13776 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
13777 test for clearing quad memory on 32-bit later.
13778
948a1fd9
MG
137792013-06-13 Marc Glisse <marc.glisse@inria.fr>
13780
13781 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
13782 (fold_negate_expr): Likewise.
13783 (fold_real_zero_addition_p): Handle vectors.
13784 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
13785
0597fb9c
AM
137862013-06-14 Alan Modra <amodra@gmail.com>
13787
13788 * varasm.c (force_const_mem): Revert 2013-06-07 change.
13789
e5b962d0
JH
137902013-06-13 Jan Hubicka <jh@suse.cz>
13791
13792 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13793 Local comdats are not externally visible.
13794 * symtab.c (dump_symtab_base): Dump externally visible.
13795 (verify_symtab_base): Verify back links in the symtab hash.
13796
a8355e51
BC
137972013-06-13 Bin Cheng <bin.cheng@arm.com>
13798
13799 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
13800 CONVERT_EXPR as equal nodes.
13801
8f7e6e33
BC
138022013-06-13 Bin Cheng <bin.cheng@arm.com>
13803
22fc7d3d 13804 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
8f7e6e33 13805
5d418483
MG
138062013-06-13 Marc Glisse <marc.glisse@inria.fr>
13807
13808 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
13809 Generalize to complex and vector.
13810 * tree.c (build_all_ones_cst): New function.
13811 * tree.h (build_all_ones_cst): Declare it.
13812
50751417
AM
138132013-06-13 Alan Modra <amodra@gmail.com>
13814
13815 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
13816 * config/rs6000/rs6000.md (signbittf2): New insn.
13817 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
13818 (abstf2_internal, cmptf_internal2): Likewise.
13819 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
13820
4b02c962 138212013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
13822 Pat Haugen <pthaugen@us.ibm.com>
13823 Peter Bergner <bergner@vnet.ibm.com>
4b02c962
MM
13824
13825 * config/rs6000/rs6000.c (emit_load_locked): Add support for
13826 power8 byte, half-word, and quad-word atomic instructions.
13827 (emit_store_conditional): Likewise.
13828 (rs6000_expand_atomic_compare_and_swap): Likewise.
13829 (rs6000_expand_atomic_op): Likewise.
13830
13831 * config/rs6000/sync.md (larx): Add new modes for power8.
13832 (stcx): Likewise.
13833 (AINT): New mode iterator to include TImode as well as normal
13834 integer modes on power8.
13835 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
13836 that VSX registers are not considered. Use AINT mode iterator
13837 instead of INT1 to allow inclusion of quad word atomic operations
13838 on power8.
13839 (load_locked<mode>): Likewise.
13840 (store_conditional<mode>): Likewise.
13841 (atomic_compare_and_swap<mode>): Likewise.
13842 (atomic_exchange<mode>): Likewise.
13843 (atomic_nand<mode>): Likewise.
13844 (atomic_fetch_<fetchop_name><mode>): Likewise.
13845 (atomic_nand_fetch<mode>): Likewise.
13846 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
13847 each type.
13848 (ATOMIC): On power8, add QImode, HImode modes.
13849 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
13850 modes that promote to SImode.
13851 (load_lockedti): Convert TImode arguments to PTImode, so that we
13852 get a guaranteed even/odd register pair.
13853 (load_lockedpti): Likewise.
13854 (store_conditionalti): Likewise.
13855 (store_conditionalpti): Likewise.
13856
13857 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
13858 atomic load/store instructions.
13859 (HSI): Likewise.
13860
710b6cc1
RS
138612013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
13862
13863 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
13864 loads.
13865 (insn_count): New attribute, with most cases extracted from...
13866 (length): ...here. Redefine most cases in terms of insn_count.
13867 (single_insn): Delete.
13868 (can_delay): Use insn_count to check for single instructions.
13869 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
13870 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
13871 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
13872 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
13873 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
13874 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
13875 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
13876 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
13877 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
13878 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
13879 rather than "length".
13880 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
13881 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
13882 Use "insn_count" rather than "length".
13883 * config/mips/mips-dsp.md
13884 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
13885 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
13886 length attributes.
13887
1951f101
MG
138882013-06-12 Marc Glisse <marc.glisse@inria.fr>
13889
13890 PR tree-optimization/57361
13891 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
13892
8b033a8a
SN
138932013-06-12 Sofiane Naci <sofiane.naci@arm.com>
13894
22fc7d3d
UB
13895 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
13896 to split.
8b033a8a
SN
13897 (aarch64_simd_combine<mode>): New instruction expansion.
13898 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
13899 function prototype.
13900 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
13901 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
13902
ca0f62a8
JH
139032013-06-12 Jan Hubicka <jh@suse.cz>
13904
13905 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
13906 decl has when in streaming stage.
13907 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
13908 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
13909
12211b99 139102013-06-12 Roland Stigge <stigge@antcom.de>
cddd69eb
RS
13911
13912 PR target/57578
13913 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
13914
bfdeda2c
JJ
139152013-06-12 Jakub Jelinek <jakub@redhat.com>
13916
13917 PR tree-optimization/57537
13918 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
13919 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
13920
a4fa02d1
RB
139212013-06-12 Richard Biener <rguenther@suse.de>
13922
13923 * data-streamer.h (streamer_write_char_stream): CSE
13924 obs->current_pointer.
13925 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
13926 streamer_write_char_stream manually and optimize the resulting loop.
13927 (streamer_write_hwi_stream): Likewise.
13928
bbf9ad07
JH
139292013-06-12 Jan Hubicka <jh@suse.cz>
13930
13931 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
13932 * cgraph.h (varpool_create_empty_node): Declare.
13933 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
13934 duplicated nodes.
13935 * symtab.c (symtab_unregister_node): Be lax about missin entries
13936 in node hash.
13937 (symtab_get_node): Update comment.
13938 * varpool.c (varpool_create_empty_node): Break out from ...
13939 (varpool_node_for_decl): ... here.
13940 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
13941
a53f90ad
EB
139422013-06-12 Eric Botcazou <ebotcazou@adacore.com>
13943
13944 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
13945 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
13946 part. Use straight-line flow at the end.
13947 <COMPONENT_REF>: Remove superfluous else.
13948 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
13949
946f9306
JJ
139502013-06-12 Jakub Jelinek <jakub@redhat.com>
13951
13952 PR target/56564
13953 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
13954 target hook even for !TREE_PUBLIC decls. If no resolution info
13955 is available, return false for common and external decls.
13956
a7fca6f0
KP
139572013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13958
13959 * config/rl78/constraints.md (U): New constraint.
13960 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
13961 valloc attribute.
13962
c7d8f446
MM
139632013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
13964
13965 PR target/57589
13966 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
13967 to allow returning address to AT_PLATFORM name.
13968
08346abd
JH
139692013-06-11 Jan Hubicka <jh@suse.cz>
13970
13971 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
13972 * cgraph.h (symtab_node_base): Add weakref flag.
13973 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
13974 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
13975 (output_weakrefs): Use weakref flag.
13976 * fold-const.c (simple_operand_p): Handle WEAK.
13977 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
13978 * ipa.c (varpool_externally_visible_p): Drop weakref.
13979 (function_and_variable_visibility): Update comment; fix weakref
13980 sanity checks; do not clear DECL_WEAK on them.
13981 * lto-cgraph.c (lto_output_node): update.
13982 (lto_output_varpool_node): Update.
13983 (input_overwrite_node): Update.
13984 (input_node): Update.
13985 (input_varpool_node): Update.
13986 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
13987 (lto_symtab_merge_symbols): Add sanity check.
13988 (lto_symtab_prevailing_decl): Do not special case weakrefs.
13989 * passes.c (rest_of_decl_compilation): Set static flag, too.
13990 * symtab.c (dump_symtab_base): Dump weakref.
13991 (verify_symtab_base): Sanity check weakrefs.
13992 (symtab_make_decl_local): Remove duplicated code.
13993 (symtab_alias_ultimate_target): Simplify.
13994 * varpool.c (varpool_create_variable_alias): Set weakref flag.
c7d8f446 13995
2e6a2f88
TV
139962013-06-11 Tom de Vries <tom@codesourcery.com>
13997
13998 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
13999 sequence_vect == NULL.
14000
d9af3652
DD
140012013-06-11 DJ Delorie <dj@redhat.com>
14002
14003 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
14004 (rl78_unwind_word_mode): New.
14005
ffa66012
DM
140062013-06-11 David Malcolm <dmalcolm@redhat.com>
14007
14008 * final.c (debug_prefix_maps): Make static.
14009
783f2d35
DM
140102013-06-11 David Malcolm <dmalcolm@redhat.com>
14011
14012 * function.c (initial_trampoline): Remove stray copy.
14013
7af26def
SN
140142013-06-11 Sofiane Naci <sofiane.naci@arm.com>
14015
14016 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
14017
97756c0e
MJ
140182013-06-11 Martin Jambor <mjambor@suse.cz>
14019
14020 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
14021 within bounds at the beginning of the function.
14022
2c7b8bf6
AM
140232013-06-11 Alan Modra <amodra@gmail.com>
14024
14025 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
14026 reporting.
14027 (get_named_section): Don't NULL !DECL_P decl.
14028
77cea46e
IZ
140292013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
14030
14031 * doc/invoke.texi (core-avx2): Document.
14032 (slm): Likewise.
14033 (atom): Updated with MOVBE.
14034
888f0920
RB
140352013-06-11 Richard Biener <rguenther@suse.de>
14036
22fc7d3d 14037 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
888f0920 14038
6b438d58
AB
140392013-06-11 Anton Blanchard <anton@samba.org>
14040
14041 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14042 correct shift value in little-endian mode.
14043
3bc66938
JJ
140442013-06-11 Jakub Jelinek <jakub@redhat.com>
14045
14046 PR target/56564
14047 * varasm.c (get_variable_align): Move #endif to the right place.
14048
a5dc7f88
CC
140492013-06-10 Cary Coutant <ccoutant@google.com>
14050
14051 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14052 for hash so that hash table traversal order is deterministic.
14053
19be72ab 140542013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
14055 Pat Haugen <pthaugen@us.ibm.com>
14056 Peter Bergner <bergner@vnet.ibm.com>
19be72ab
MM
14057
14058 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14059 of vectors in GPRS if they are direct moves or quad word load or
14060 store moves.
14061
14062 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14063 declaration.
14064 (direct_move_p): Likewise.
14065 (quad_load_store_p): Likewise.
14066
14067 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14068 classes into bins based on the physical register type.
14069 (reg_class_to_reg_type): Likewise.
14070 (IS_STD_REG_TYPE): Likewise.
14071 (IS_FP_VECT_REG_TYPE): Likewise.
14072 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14073 use direct move instructions.
14074 (reload_gpr_vsx): Likewise.
14075 (reload_vsx_gpr): Likewise.
14076 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14077 information that is a simplification of register classes. Also
14078 precalculate direct move reload helpers.
14079 (direct_move_p): New function to return true if the operation can
14080 be done as a direct move instruciton.
14081 (quad_load_store_p): New function to return true if the operation
14082 is a quad memory operation.
14083 (rs6000_legitimize_address): If quad memory, only allow register
14084 indirect for TImode addresses.
14085 (rs6000_legitimate_address_p): Likewise.
14086 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14087 (rs6000_reload_register_type): Likewise.
14088 (register_to_reg_type): Return register type.
14089 (rs6000_secondary_reload_simple_move): New helper function for
14090 secondary reload and secondary memory needed to identify anything
14091 that is a simple move, and does not need reloading.
14092 (rs6000_secondary_reload_direct_move): New helper function for
14093 secondary reload to identify cases that can be done with several
14094 instructions via the direct move instructions.
14095 (rs6000_secondary_reload_move): New helper function for secondary
14096 reload to identify moves between register types that can be done.
14097 (rs6000_secondary_reload): Add support for quad memory operations
14098 and for direct move.
14099 (rs6000_secondary_memory_needed): Likewise.
14100 (rs6000_debug_secondary_memory_needed): Change argument names.
14101 (rs6000_output_move_128bit): New function to return the move to
14102 use for 128-bit moves, including knowing about the various
14103 limitations of quad memory operations.
14104
14105 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14106 memory operations. call rs6000_output_move_128bit for the actual
14107 instruciton(s) to generate.
14108 (vsx_movti_64bit): Likewise.
14109
14110 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14111 (UNSPEC_P8V_MTVSRWZ): Likewise.
14112 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14113 (UNSPEC_P8V_MTVSRD): Likewise.
14114 (UNSPEC_P8V_XXPERMDI): Likewise.
14115 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14116 (UNSPEC_FUSION_GPR): Likewise.
14117 (FMOVE128_GPR): New iterator for direct move.
22fc7d3d 14118 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
19be72ab
MM
14119 (f32_sv): Likewise.
14120 (f32_dm): Likewise.
14121 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14122 loads and direct move instructions.
14123 (zero_extendsidi2_lfiwzx): Likewise.
14124 (extendsidi2_lfiwax): Likewise.
14125 (extendsidi2_nocell): Likewise.
14126 (floatsi<mode>2_lfiwax): Likewise.
14127 (lfiwax): Likewise.
14128 (floatunssi<mode>2_lfiwzx): Likewise.
14129 (lfiwzx): Likewise.
14130 (fix_trunc<mode>_stfiwx): Likewise.
14131 (fixuns_trunc<mode>_stfiwx): Likewise.
14132 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14133 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14134 (parity<mode>2_cmpb): Set length/type attr.
14135 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14136 for 'mr.' to fast_compare.
14137 (bpermd_<mode>): Change type attr to popcnt.
14138 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14139 (p8_mtvsrwz_1): Likewise.
14140 (p8_mtvsrwz_2): Likewise.
14141 (reload_fpr_from_gpr<mode>): Likewise.
14142 (p8_mtvsrd_1): Likewise.
14143 (p8_mtvsrd_2): Likewise.
14144 (p8_xxpermdi_<mode>): Likewise.
14145 (reload_vsx_from_gpr<mode>): Likewise.
14146 (reload_vsx_from_gprsf): Likewise.
14147 (p8_mfvsrd_3_<mode>): LIkewise.
14148 (reload_gpr_from_vsx<mode>): Likewise.
14149 (reload_gpr_from_vsxsf): Likewise.
14150 (p8_mfvsrd_4_disf): Likewise.
14151 (multi-word GPR splits): Do not split direct moves or quad memory
14152 operations.
14153
16876bdc
DM
141542013-06-10 David Malcolm <dmalcolm@redhat.com>
14155
14156 * tree-into-ssa.c (interesting_blocks): Make static.
14157
df8a1d28
JJ
141582013-06-10 Jakub Jelinek <jakub@redhat.com>
14159
14160 PR target/56564
14161 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14162 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14163 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14164 (get_variable_align): New function.
14165 (get_variable_section, emit_bss, emit_common,
14166 assemble_variable_contents, place_block_symbol): Use
14167 get_variable_align instead of DECL_ALIGN.
14168 (assemble_noswitch_variable): Add align argument, use it
14169 instead of DECL_ALIGN.
14170 (assemble_variable): Adjust caller. Use get_variable_align
14171 instead of DECL_ALIGN.
14172 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14173 caller.
14174 (DATA_ABI_ALIGNMENT): Define.
14175 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14176 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14177 opt is false, only return the psABI mandated alignment increase.
14178 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14179 (DATA_ABI_ALIGNMENT): ... this.
14180 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14181 (DATA_ABI_ALIGNMENT): ... this.
14182 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14183 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14184 (DATA_ABI_ALIGNMENT): ... this.
14185 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14186 * doc/tm.texi: Regenerated.
14187
57b29ca6
UB
141882013-06-10 Uros Bizjak <ubizjak@gmail.com>
14189
14190 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14191 cmp_code to construct REG_EQUAL note.
14192
3e56ed50
JJ
141932013-06-09 Jakub Jelinek <jakub@redhat.com>
14194
14195 PR target/57568
14196 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14197 that operands[2] doesn't overlap with operands[0].
14198
74fe2a1d
DE
141992013-06-09 David Edelsohn <dje.gcc@gmail.com>
14200 Jan Hubicka <jh@suse.cz>
14201
14202 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14203 hack to mark symbols as used.
14204
ed52a84e
VM
142052013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14206
14207 PR rtl-optimization/57559
14208 * lra-constraints.c (process_alt_operands): Don't discourage
14209 memory with known offset for offsetable memory constraint.
14210 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14211
299316ed
EB
142122013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14213
14214 * varasm.c (struct oc_local_state): Reorder fields.
14215 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14216 and adjust accordingly.
14217 (output_constructor): Reorder initialization code and adjust call to
14218 output_constructor_bitfield.
14219
88ac60d3
JH
142202013-06-07 Jan Hubicka <jh@suse.cz>
14221
14222 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14223
4cdce1a8
DM
142242013-06-07 David Malcolm <dmalcolm@redhat.com>
14225
14226 * tree-object-size.c (unknown): Make const.
14227
b5e0425c
AK
142282013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14229
14230 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14231 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14232 for last alternative in the cpu_facility attribute.
14233
79678d04
KT
142342013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14235
14236 PR target/56315
14237 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14238 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14239 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14240 * config/arm/constraints.md (Dg): New constraint.
14241 * config/arm/neon.md (xordi3_neon): Remove.
14242 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14243 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14244
b31ddbdb
KT
142452013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14246
14247 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14248 Clean up alternatives.
14249
8f90b7d4
AM
142502013-06-07 Alan Modra <amodra@gmail.com>
14251
14252 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14253 va_list_gpr_size.
14254
961ce119
AM
142552013-06-07 Alan Modra <amodra@gmail.com>
14256
14257 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14258
33e49835
KT
142592013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14260
14261 * config/arm/constraints.md (Df): New constraint.
14262 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14263 Correct length attribute for last two alternatives.
14264
acd17ae6
AM
142652013-06-07 Alan Modra <amodra@gmail.com>
14266
14267 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14268 override user -mfp-in-toc.
14269 (offsettable_ok_by_alignment): Consider just the current access
14270 rather than the whole object, unless BLKmode. Handle
14271 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14272 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14273 for -mcmodel=medium.
14274 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14275 override user -mfp-in-toc or -msum-in-toc. Default to
14276 -mno-fp-in-toc for -mcmodel=medium.
14277
73310b0e
DD
142782013-06-06 DJ Delorie <dj@redhat.com>
14279
14280 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14281 TARGET_VALID_POINTER_MODE.
14282
0bd62dca 142832013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
14284 Pat Haugen <pthaugen@us.ibm.com>
14285 Peter Bergner <bergner@vnet.ibm.com>
0bd62dca
MM
14286
14287 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14288 Document new power8 builtins.
14289
14290 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14291 condition code register, to allow 128-bit logical operations to be
14292 done in the VSX or GPR registers.
14293 (nor<mode>3): Use the canonical form for nor.
14294 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14295 vclz*, and vpopcnt* vector instructions.
14296 (nand<mode>3): Likewise.
14297 (orc<mode>3): Likewise.
14298 (clz<mode>2): LIkewise.
14299 (popcount<mode>2): Likewise.
14300
14301 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14302 that only the GPRs are recognized.
14303
14304 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14305 support for new power8 builtins.
14306
14307 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14308 builtin functions.
14309 (xscvdpspn): Likewise.
14310 (vclz): Likewise.
14311 (vclzb): Likewise.
14312 (vclzh): Likewise.
14313 (vclzw): Likewise.
14314 (vclzd): Likewise.
14315 (vpopcnt): Likewise.
14316 (vpopcntb): Likewise.
14317 (vpopcnth): Likewise.
14318 (vpopcntw): Likewise.
14319 (vpopcntd): Likewise.
14320 (vgbbd): Likewise.
14321 (vmrgew): Likewise.
14322 (vmrgow): Likewise.
14323 (eqv): Likewise.
14324 (eqv_v16qi3): Likewise.
14325 (eqv_v8hi3): Likewise.
14326 (eqv_v4si3): Likewise.
14327 (eqv_v2di3): Likewise.
14328 (eqv_v4sf3): Likewise.
14329 (eqv_v2df3): Likewise.
14330 (nand): Likewise.
14331 (nand_v16qi3): Likewise.
14332 (nand_v8hi3): Likewise.
14333 (nand_v4si3): Likewise.
14334 (nand_v2di3): Likewise.
14335 (nand_v4sf3): Likewise.
14336 (nand_v2df3): Likewise.
14337 (orc): Likewise.
14338 (orc_v16qi3): Likewise.
14339 (orc_v8hi3): Likewise.
14340 (orc_v4si3): Likewise.
14341 (orc_v2di3): Likewise.
14342 (orc_v4sf3): Likewise.
14343 (orc_v2df3): Likewise.
14344
14345 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14346 allow power8 quad mode in 64-bit.
14347 (rs6000_builtin_vectorized_function): Add support to vectorize
14348 ISA 2.07 count leading zeros, population count builtins.
14349 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14350 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14351 (builtin_function_type): Add vgbbd builtin function which takes an
14352 unsigned argument.
14353 (altivec_expand_vec_perm_const): Add support for new power8 merge
14354 instructions.
14355
14356 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14357 that does not include TImdoe for use with 32-bit.
14358 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14359 instructions.
14360 (UNSPEC_VSX_CVDPSPN): Likewise.
14361 (vsx_xscvdpspn): Likewise.
14362 (vsx_xscvspdpn): Likewise.
14363 (vsx_xscvdpspn_scalar): Likewise.
14364 (vsx_xscvspdpn_directmove): Likewise.
14365 (vsx_and<mode>3): Split logical operations into 32-bit and
14366 64-bit. Add support to do logical operations on TImode as well as
14367 VSX vector types. Allow logical operations to be done in either
14368 VSX registers or in general purpose registers in 64-bit mode. Add
14369 splitters if GPRs were used. For AND, add clobber of CCmode to
14370 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14371 encoding.
14372 (vsx_and<mode>3_32bit): Likewise.
14373 (vsx_and<mode>3_64bit): Likewise.
14374 (vsx_ior<mode>3): Likewise.
14375 (vsx_ior<mode>3_32bit): Likewise.
14376 (vsx_ior<mode>3_64bit): Likewise.
14377 (vsx_xor<mode>3): Likewise.
14378 (vsx_xor<mode>3_32bit): Likewise.
14379 (vsx_xor<mode>3_64bit): Likewise.
14380 (vsx_one_cmpl<mode>2): Likewise.
14381 (vsx_one_cmpl<mode>2_32bit): Likewise.
14382 (vsx_one_cmpl<mode>2_64bit): Likewise.
14383 (vsx_nor<mode>3): Likewise.
14384 (vsx_nor<mode>3_32bit): Likewise.
14385 (vsx_nor<mode>3_64bit): Likewise.
14386 (vsx_andc<mode>3): Likewise.
14387 (vsx_andc<mode>3_32bit): Likewise.
14388 (vsx_andc<mode>3_64bit): Likewise.
14389 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14390 and xxlorc instructions.
14391 (vsx_eqv<mode>3_64bit): Likewise.
14392 (vsx_nand<mode>3_32bit): Likewise.
14393 (vsx_nand<mode>3_64bit): Likewise.
14394 (vsx_orc<mode>3_32bit): Likewise.
14395 (vsx_orc<mode>3_64bit): Likewise.
14396
14397 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14398
14399 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14400 instruction.
14401 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14402 (p8_vmrgow): Likewise.
14403 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14404 GPRs to be split under VSX.
14405 (p8v_clz<mode>2): Add power8 count leading zero support.
14406 (p8v_popcount<mode>2): Add power8 population count support.
14407 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14408 support.
14409
14410 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14411 instruction.
14412
14413 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14414 builtin functions.
14415 (vec_nand): Likewise.
14416 (vec_vclz): Likewise.
14417 (vec_vclzb): Likewise.
14418 (vec_vclzd): Likewise.
14419 (vec_vclzh): Likewise.
14420 (vec_vclzw): Likewise.
14421 (vec_vgbbd): Likewise.
14422 (vec_vmrgew): Likewise.
14423 (vec_vmrgow): Likewise.
14424 (vec_vpopcnt): Likewise.
14425 (vec_vpopcntb): Likewise.
14426 (vec_vpopcntd): Likewise.
14427 (vec_vpopcnth): Likewise.
14428 (vec_vpopcntw): Likewise.
14429
37684c46
VM
144302013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14431
14432 PR rtl-optimization/57468
14433 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14434 spilled pseudos.
14435
d89ae27c
VM
144362013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14437
14438 PR rtl-optimization/57459
14439 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14440 type when setting live regs.
14441
3597e113
VM
144422013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14443
14444 * config/s390/s390.opt (mlra): New option.
14445 * config/s390/s390.c (s390_decompose_address): Check displacement
14446 for all registers for LRA.
14447 (s390_secondary_reload): Don't used secondary reloads for LRA.
14448 (s390_lra_p): New function.
14449 (TARGET_LRA_P): Define.
14450 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14451 of attribute cpu_facility to zarch for the last alternative.
14452 (*cmpmem_short): Ditto.
14453
01037aeb
EB
144542013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14455
14456 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14457 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14458 (arm_expand_prologue): Likewise.
14459
3371a64f
TJ
144602013-06-06 Teresa Johnson <tejohnson@google.com>
14461
14462 PR c++/53743
14463 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14464 as this is now done by redirect_edge_and_branch_force.
14465 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14466 barriers, and fix interaction with splitting.
14467 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14468 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14469 reflect changes made in the routine.
14470 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14471 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14472 since this is called in cfglayout mode, and replace partition fixup
14473 with assert as that is now done by force_nonfallthru_and_redirect.
14474 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14475 already be marked with region crossing note.
14476 (insert_section_boundary_note): Make non-static, gate on flag
14477 has_bb_partition, rewrite to also check for multiple partitions.
14478 (rest_of_handle_reorder_blocks): Remove call to
14479 insert_section_boundary_note, now done later during free_cfg.
14480 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14481 * bb-reorder.h (insert_section_boundary_note): Declare.
14482 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14483 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14484 invoke insert_section_boundary_note.
14485 (try_redirect_by_replacing_jump): Remove unnecessary
14486 check for region crossing note.
14487 (fixup_partition_crossing): New function.
14488 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14489 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14490 in non-cfglayout mode.
14491 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14492 remove old code that tried to do this. Emit barrier correctly
14493 when we are in cfglayout mode.
14494 (last_bb_in_partition): New function.
14495 (rtl_split_edge): Correctly fixup partition boundaries.
14496 (commit_one_edge_insertion): Remove old code that tried to
14497 fixup region crossing edge since this is now handled in
14498 split_block, and set up insertion point correctly since
14499 block may now end in a jump.
14500 (verify_hot_cold_block_grouping): Guard against checking when not in
14501 linearized RTL mode.
14502 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14503 notes.
14504 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14505 rtl_verify_flow_info, so not called in cfglayout mode.
14506 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14507 (fixup_reorder_chain): Remove old code that attempted to fixup region
14508 crossing note as this is now handled in force_nonfallthru_and_redirect.
14509 (duplicate_insn_chain): Don't duplicate switch section notes.
14510 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14511 note.
14512 * basic-block.h (emit_barrier_after_bb): Declare.
14513
66071e10
KT
145142013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14515
14516 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14517 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14518 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14519
0e26bf3d
KT
145202013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14521
14522 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14523 where appropriate.
14524 * config/arm/ldmstm.md: Regenerate.
14525
12b4e7ef
KT
145262013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14527
14528 * config/arm/sync.md (atomic_loaddi_1):
14529 Disable predication for arm_restrict_it.
14530 (arm_load_exclusive<mode>): Likewise.
14531 (arm_load_exclusivesi): Likewise.
14532 (arm_load_exclusivedi): Likewise.
14533 (arm_load_acquire_exclusive<mode>): Likewise.
14534 (arm_load_acquire_exclusivesi): Likewise.
14535 (arm_load_acquire_exclusivedi): Likewise.
14536 (arm_store_exclusive<mode>): Likewise.
14537 (arm_store_exclusive<mode>): Likewise.
14538 (arm_store_release_exclusivedi): Likewise.
14539 (arm_store_release_exclusive<mode>): Likewise.
14540
15d16c8a
RB
145412013-06-06 Richard Biener <rguenther@suse.de>
14542
14543 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14544 after LTO_null.
14545 (lto_tag_is_tree_code_p): Adjust.
14546 (lto_tag_is_gimple_code_p): Likewise.
14547 (lto_gimple_code_to_tag): Likewise.
14548 (lto_tag_to_gimple_code): Likewise.
14549 (lto_tree_code_to_tag): Likewise.
14550 (lto_tag_to_tree_code): Likewise.
14551 * data-streamer.h (streamer_write_hwi_in_range): Use
14552 uhwi streaming to stream the normalized range.
14553 (streamer_read_hwi_in_range): Likewise.
14554
17a7fc37
KT
145552013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14556
14557 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14558 (predicable_short_it): Likewise.
14559 (predicated): Likewise.
14560 (enabled): Handle above.
14561 (define_cond_exec): Set predicated attribute to yes.
14562
b57ca59b
MS
145632013-06-05 Mike Stump <mikestump@comcast.net>
14564
14565 * gdbinit.in (__FUNCTION__): Add.
14566
c1e183a9
UB
145672013-06-05 Uros Bizjak <ubizjak@gmail.com>
14568
14569 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14570 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14571
0378bddb
JH
145722013-06-05 Jan Hubicka <jh@suse.cz>
14573
14574 * varasm.c (mark_decl_referenced): Revert the removal until targets
14575 are fixed.
14576
8f8a2057
DE
145772013-06-05 David Edelsohn <dje.gcc@gmail.com>
14578
14579 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14580 instead of mark_decl_referenced.
14581
edb983b2
JH
145822013-06-05 Jan Hubicka <jh@suse.cz>
14583
14584 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14585 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14586 and symtab_used_from_object_file_p.
14587 (cgraph_make_node_local_1): Clear forced_by_abi.
14588 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14589 * cgraph.h (symtab_node_base): Add forced_by_abi.
14590 (decide_is_variable_needed): Remove.
14591 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14592 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14593 (decide_is_symbol_needed): ... this one; handle symbols in general;
14594 always analyze virtuals; honnor forced_by_abi.
14595 (cgraph_finalize_function): Update.
14596 (varpool_finalize_decl): Update.
14597 (symbol_defined_and_needed): Remove.
14598 (analyze_functions): Update.
14599 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14600 output_refs, input_overwrite_node): Handle forced_by_abi.
14601 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14602 (address_taken_from_non_vtable_p): ... this one.
14603 (comdat_can_be_unshared_p_1): New function.
14604 (cgraph_comdat_can_be_unshared_p): Rename to ...
14605 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14606 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14607 (function_and_variable_visibility): Clear forced_by_abi as needed.
14608 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14609 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
edb983b2
JH
14610 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14611 * varpool.c (decide_is_variable_needed): Remove.
14612
9912dbe5
KT
146132013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14614
14615 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14616 (arm_option_override): Override arm_restrict_it where appropriate.
14617 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14618 * config/arm/arm.opt (mrestrict-it): New command-line option.
14619 * doc/invoke.texi: Document -mrestrict-it.
14620
34c136b6
DM
146212013-06-05 David Malcolm <dmalcolm@redhat.com>
14622
14623 * tsan.c (tsan_atomic_table): Make const.
14624
a367df53
RB
146252013-06-05 Richard Biener <rguenther@suse.de>
14626
14627 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14628 index associated with the tree we are supposed to replace.
14629 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14630 TREE_ASM_WRITTEN as zero for everything but SSA names.
14631
70d51a19
DM
146322013-06-05 David Malcolm <dmalcolm@redhat.com>
14633
14634 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14635
b9bd2075
JH
146362013-06-04 Jan Hubicka <jh@suse.cz>
14637
14638 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14639 (input_node, input_varpool_node): Handle correctly external same
14640 body aliases.
14641 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14642 nodes at ltrans stage.
14643
aaae719d
JH
146442013-06-04 Jan Hubicka <jh@suse.cz>
14645
14646 * ipa-inline.c (update_caller_keys): Fix availability test.
14647 (update_callee_keys): Likewise.
14648 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14649 to follow ELF standard.
14650
107eea2c
JU
146512013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14652
14653 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14654 (mips64r5900el-*-elf*): New configurations.
14655 * config/mips/mips-cpus.def (r5900): New processor.
14656 * config/mips/mips-tables.opt: Regenerate.
14657 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14658 (mips_issue_rate): Handle PROCESSOR_R5900.
14659 (mips_reorg_process_insns): Force reorder mode for the R5900.
14660 * config/mips/mips.h (TARGET_MIPS5900): Define.
14661 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14662 TARGET_MIPS5900.
14663 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14664 TARGET_MIPS5900.
14665 * config/mips/mips.md (processor): Add r5900.
14666 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14667
b7342d25
IB
146682013-06-04 Ian Bolton <ian.bolton@arm.com>
14669
14670 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14671 into function to generate MOVI instruction.
c1e183a9 14672 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
b7342d25
IB
14673 (aarch64_preferred_simd_mode): Turn into wrapper.
14674 (aarch64_output_scalar_simd_mov_immediate): New function.
14675 * config/aarch64/aarch64-protos.h: Add prototype for above.
14676
81c2dfb9
IB
146772013-06-04 Ian Bolton <ian.bolton@arm.com>
14678
14679 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14680 element_char member.
14681 (sizetochar): Return signed char.
14682 (aarch64_simd_valid_immediate): Remove elchar and other
14683 unnecessary variables.
14684 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14685 Calculate element_char as required.
14686 * config/aarch64/aarch64-protos.h: Update and move prototype
14687 for aarch64_output_simd_mov_immediate.
14688 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14689 Update arguments.
14690
48063b9d
IB
146912013-06-04 Ian Bolton <ian.bolton@arm.com>
14692
14693 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14694 information completed by aarch64_simd_valid_immediate.
14695 (aarch64_legitimate_constant_p): Update arguments.
14696 (aarch64_simd_valid_immediate): Work with struct rather than many
14697 pointers.
14698 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14699 (aarch64_simd_make_constant): Update arguments.
14700 (aarch64_output_simd_mov_immediate): Work with struct rather than
14701 many pointers. Output immediate directly rather than as operand.
14702 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14703 Update prototype.
14704 * config/aarch64/constraints.md (Dn): Update arguments.
14705
3ea63f60
IB
147062013-06-04 Ian Bolton <ian.bolton@arm.com>
14707
14708 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14709 longer static.
14710 (aarch64_simd_immediate_valid_for_move): Remove.
14711 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14712 (aarch64_simd_make_constant): Update call.
14713 (aarch64_output_simd_mov_immediate): Update call.
14714 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14715 Add prototype.
14716 * config/aarch64/constraints.md (Dn): Update call.
14717
d8edd899
IB
147182013-06-04 Ian Bolton <ian.bolton@arm.com>
14719
14720 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14721 return type to bool for prototype.
14722 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14723 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14724 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14725
44b20bb8
CM
147262013-06-04 Catherine Moore <clm@codesourcery.com>
14727
14728 * config/mips/mips.opt (meva): New.
c1e183a9 14729 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
44b20bb8
CM
14730 (ASM_SPEC): Handle -meva.
14731 * doc/invoke.texi (meva): Document.
14732
52befbd8
AM
147332013-06-04 Alan Modra <amodra@gmail.com>
14734
14735 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14736 constant output.
14737
aadaf24e
KT
147382013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14739
14740 * rtl.def: Add extra fourth optional field to define_cond_exec.
14741 * gensupport.c (process_one_cond_exec): Process attributes from
14742 define_cond_exec.
14743 * doc/md.texi: Document fourth field in define_cond_exec.
14744
0cd9e9ee
EB
147452013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14746
14747 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14748 out the processing order as in store_bit_field_1.
14749
cb0f665d
JH
147502013-06-04 Jan Hubicka <jh@suse.cz>
14751
14752 PR middle-end/57500
14753 * cgraphunit.c (cgraph_process_same_body_aliases): Create
14754 non-VAR_DECL node if it does not exist yet.
14755
53984b9b
RS
147562013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
14757
14758 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
14759 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
14760 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
14761 target_cpu_default setting.
14762
5d318fd4
TJ
147632013-06-03 Teresa Johnson <tejohnson@google.com>
14764
14765 * dumpfile.c (opt_info_switch_p): Change -fopt-info
14766 default to -fopt-info=optimized instead of all.
14767 * doc/invoke.texi: Ditto.
14768 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
14769 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
14770 (execute_vect_slp): Emit BB vectorization success under
14771 MSG_OPTIMIZED_LOCATIONS.
14772 * tree-vect-slp.c (vect_slp_transform_bb): Change
14773 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
14774 * tree-vect-loop.c (vect_transform_loop): Ditto.
14775
ec3af349
JM
147762013-06-03 Jason Merrill <jason@redhat.com>
14777
14778 PR c++/57415
14779 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
14780 Use TARGET_EXPR for C++.
14781
1baf9159
JJ
147822013-06-03 Jakub Jelinek <jakub@redhat.com>
14783
14784 PR rtl-optimization/57268
0cd9e9ee 14785 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
1baf9159
JJ
14786 if DEBUG_INSN_P (insn).
14787
14788 Reapply
14789 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14790
14791 PR rtl-optimization/57268
14792 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14793 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14794
0f1d3965
YR
147952013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14796
14797 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
14798 (ix86_avoid_lea_for_addr): Likewise.
14799 (exact_dependency_1): Likewise.
14800 (ix86_adjust_cost): Likewise.
14801 (swap_top_of_ready_list): Fix formatting and !reload_completed check
14802 removed.
14803 (do_reorder_for_imul): Fix typo, formatting and
14804 !reload_completed check removed.
14805 (ix86_sched_reorder): Fix typo and formatting.
14806 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
14807 list.
14808
3bc30db4
SN
148092013-06-03 Sofiane Naci <sofiane.naci@arm.com>
14810
14811 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
14812
19cc6fac
EB
148132013-06-03 Eric Botcazou <ebotcazou@adacore.com>
14814
14815 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
14816 <STRING_CST>: Likewise.
14817 <VECTOR_CST>: Likewise.
14818
c68a6e08
JW
148192013-06-01 Janus Weil <janus@gcc.gnu.org>
14820 Mikael Morin <mikael@gcc.gnu.org>
14821
14822 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
14823 * config.in: Regenerated.
14824 * configure: Regenerated.
14825
38e55e5c
JH
148262013-06-01 Jan Hubicka <jh@suse.cz>
14827
14828 PR middle-end/57366
14829 * cgraphunit.c (compile): When weakref is not supported,
14830 set up transparent aliases before final output pass.
14831 * varasm.c (assemble_alias): Do not try to do it here.
14832
eb51d2ff
JH
148332013-06-01 Jan Hubicka <jh@suse.cz>
14834
14835 PR middle-end/57467
14836 * passes.c (for_per_function): Skip unanalyzed functions.
14837
40a7fe1e
JH
148382013-06-01 Jan Hubicka <jh@suse.cz>
14839
14840 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
14841 (lto_symtab_merge_symbols_1): ... this one.
14842 (lto_symtab_merge_cgraph_nodes): Rename to ...
14843 (lto_symtab_merge_symbols): ... this one; simplify.
14844 * cgraph.c (same_body_aliases_done): Rename to ...
14845 (cpp_implicit_aliases_done): ... this one.
14846 (cgraph_create_function_alias): Update.
14847 (cgraph_same_body_alias): Update.
c1e183a9 14848 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
40a7fe1e 14849 (verify_edge_corresponds_to_fndecl): Simplify.
c1e183a9 14850 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
40a7fe1e
JH
14851 (cgraph_node): Remove same_body_alias.
14852 (varpool_node): Remove alias_of and extra_name_alias.
14853 (same_body_aliases_done): Rename to ..
14854 (cpp_implicit_aliases_done): ... this one.
14855 (symtab_alias_ultimate_target): Add default parameter.
14856 (symtab_resolve_alias): New function.
14857 (fixup_same_cpp_alias_visibility): Declare.
14858 (cgraph_function_node): Add default parameter.
14859 (cgraph_node_asm_name): Likewise.
14860 (cgraph_function_or_thunk_node): Add default parameter; do
14861 not ICE when it is NULL.
14862 (varpool_variable_node): Likewise.
14863 * tree-emutls.c (create_emultls_var): Update.
14864 (ipa_lower_emutls): Update.
14865 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
14866 (cgraph_reset_node): Reset alias info.
14867 (cgraph_finalize_function): Update.
14868 (fixup_same_cpp_alias_visibility): Move to symtab.c.
14869 (analyze_function): Simplify.
14870 (cgraph_process_same_body_aliases): Simplify.
14871 (analyze_functions): Fixup same body aliases.
14872 (handle_alias_pairs): Simplify.
14873 (assemble_thunk): Update.
14874 (assemble_thunks_and_aliases): Update.
14875 (output_weakrefs): Rewrite.
14876 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
14877 (lto_output_varpool_node): Likewise.
14878 (compute_ltrans_boundary): Remve assert.
14879 (get_alias_symbol): New functoin.
14880 (input_node): Rewrite alias handling.
14881 (input_varpool_node): Likewise.
14882 * ipa-pure-const.c (propagate_pure_const): Fix formating.
14883 * ipa.c (process_references): Handle weakrefs correctly.
14884 (symtab_remove_unreachable_nodes): Likewise.
14885 * trans-mem.c (get_cg_data): Update.
14886 (ipa_tm_create_version_alias): Update.
14887 (ipa_tm_execute): Update.
14888 * symtab.c (dump_symtab_base): Dump aliases.
14889 (verify_symtab_base): Verify aliases.
14890 (symtab_node_availability): New function.
14891 (symtab_alias_ultimate_target): Simplify.
14892 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
14893 handle all the fixup cases.
14894 (symtab_resolve_alias): New function.
14895 * passes.c (ipa_write_summaries): Handle weakrefs.
14896 * varpool.c (varpool_analyze_node): Simplify.
14897 (assemble_aliases): Update.
14898 (varpool_create_variable_alias): Simplify.
14899 (varpool_extra_name_alias): Simplify.
14900 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
14901 (lto_symtab_merge_symbols): ... this one.
14902
21759881
DT
149032013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
14904
14905 Revert
14906 PR rtl-optimization/57268
14907 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14908 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14909
daf4e940
TB
149102013-06-01 Tobias Burnus <burnus@net-b.de>
14911
14912 Partially reverted:
14913 2013-05-31 Tobias Burnus <burnus@net-b.de>
14914
14915 PR middle-end/57073
14916 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
14917 further up.
14918
c359d8d0
DT
149192013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14920
14921 PR rtl-optimization/57268
14922 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14923 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14924
4172215d
EB
149252013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14926
14927 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
14928 unordered comparison operators when -fno-trapping-math is in effect
14929 on the e500.
14930 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
14931 and implement unordered comparison operators properly on the e500.
14932
a8c50132
EB
149332013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14934
14935 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
14936 for constant scalar integers.
14937 (simplify_relational_operation_1): Likewise.
14938
44626634
SB
149392013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
14940
14941 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
14942 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
14943 Fix comment.
14944
55805e54
YR
149452013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14946 Igor Zamyatin <igor.zamyatin@intel.com>
14947
14948 Silvermont (SLM) architecture performance tuning.
14949 * config/i386/i386.h (enum ix86_tune_indices): Add
14950 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
14951 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
14952
14953 * config/i386/i386.c (initial_ix86_tune_features)
14954 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
14955 (ix86_lea_outperforms): Handle Silvermont tuning.
14956 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
14957 call.
14958 (ix86_use_lea_for_mov): Likewise.
14959 (ix86_avoid_lea_for_addr): Likewise.
14960 (ix86_lea_for_add_ok): Likewise.
14961 (exact_dependency_1): New function.
14962 (exact_store_load_dependency): Likewise.
14963 (ix86_adjust_cost): Handle Silvermont tuning.
14964 (do_reoder_for_imul): Likewise.
14965 (swap_top_of_ready_list): New function.
14966 (ix86_sched_reorder): Changed to handle Silvermont tuning.
14967
14968 * config/i386/i386.md (peepholes that split memory operand in fp
14969 converts): New.
14970
e19c9de2
MS
149712013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14972
14973 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
14974 Remove un-necessary braces.
14975
38e6c9a6
MS
149762013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14977
14978 * config/aarch64/aarch64.c (aarch64_classify_symbol):
14979 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
14980
e3530904
TB
149812013-05-31 Tobias Burnus <burnus@net-b.de>
14982
14983 PR middle-end/57073
c1e183a9 14984 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
e3530904 14985
b6af05a9
KT
149862013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14987
14988 PR target/56315
14989 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
14990 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
14991 * config/arm/neon.md (iordi3_neon): Remove.
14992 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
14993 * config/arm/predicates.md (imm_for_neon_logic_operand):
14994 Move to earlier in the file.
14995 (neon_logic_op2): Likewise.
14996 (arm_iordi_operand_neon): New predicate.
14997
f800c166
RB
149982013-05-31 Richard Biener <rguenther@suse.de>
14999
15000 PR tree-optimization/57478
15001 PR tree-optimization/57453
15002 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
15003 are life as well.
15004
e571fa59
KP
150052013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
15006
15007 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
15008 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
15009
0fa6e0ef
TB
150102013-05-30 Tobias Burnus <burnus@net-b.de>
15011 Thomas Koenig <tkoenig@gcc.gnu.org>
15012
15013 PR middle-end/57073
15014 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
15015 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
15016
4688ddf5
SB
150172013-05-30 Steven Bosscher <steven@gcc.gnu.org>
15018
0fa6e0ef 15019 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
4688ddf5 15020
3b9ceb4b
VM
150212013-05-30 Vladimir Makarov <vmakarov@redhat.com>
15022
15023 * target.def (register_usage_leveling_p): New hook.
15024 * targhooks.c (default_register_usage_leveling_p): New.
15025 * targhooks.h (default_register_usage_leveling_p): New prototype.
15026 * lra-assigns.c (register_usage_leveling_p): Use the hook.
15027 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
15028 * doc/tm.texi: Update.
15029 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
15030
65f9e789
IB
150312013-05-30 Ian Bolton <ian.bolton@arm.com>
15032
15033 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15034 (*insv_reg<mode>): New define_insn.
15035
12211b99 150362013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
ca035367
JR
15037
15038 PR rtl-optimization/57439
15039 * postreload.c (move2add_valid_value_p): Check that we have
15040 a zero subreg_regno_offset when accessing the register in
15041 the requested mode.
15042
0b871ccf
YR
150432013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15044 Igor Zamyatin <igor.zamyatin@intel.com>
15045
15046 Silvermont (SLM) architecture pipeline model, tuning and
15047 insn selection.
15048 * config.gcc: Add slm config options and target.
15049
15050 * config/i386/slm.md: New.
15051
15052 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15053
f43245d1 15054 * config/i386/i386-c.c (ix86_target_macros_internal): New case
0b871ccf
YR
15055 PROCESSOR_SLM.
15056 (ix86_target_macros_internal): Likewise.
15057
f43245d1 15058 * config/i386/i386.c (slm_cost): New cost.
0b871ccf
YR
15059 (m_SLM): New macro flag.
15060 (initial_ix86_tune_features): Set m_SLM.
15061 (x86_accumulate_outgoing_args): Likewise.
15062 (x86_arch_always_fancy_math_387): Likewise.
15063 (processor_target_table): Add slm cost.
15064 (cpu_names): Add slm cpu name.
15065 (x86_option_override_internal): Set SLM ISA.
15066 (ix86_issue_rate): New case PROCESSOR_SLM.
15067 (ia32_multipass_dfa_lookahead): Likewise.
15068 (fold_builtin_cpu): Add slm.
15069
15070 * config/i386/i386.h (TARGET_SLM): New target macro.
15071 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15072 (processor_type): Add PROCESSOR_SLM.
15073
15074 * config/i386/i386.md (cpu): Add new value "slm".
15075 (slm.md): Include slm.md.
15076
24d5b097
XG
150772013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15078 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15079
15080 * config/arm/arm-protos.h: Add and update function protos.
15081 * config/arm/arm.c (use_simple_return_p): New added.
15082 (thumb2_expand_return): Check simple_return flag.
15083 * config/arm/arm.md: Add simple_return and conditional simple_return.
15084 * config/arm/iterators.md: Add iterator for return and simple_return.
15085
c1cccc15
ZC
150862013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15087
15088 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15089 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15090 (arm_emit_vfp_multi_reg_pop): Likewise.
15091 (thumb2_emit_ldrd_pop): Likewise.
15092 (arm_expand_epilogue): Add misc REG_CFA notes.
15093 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15094
3a4f280b
LC
150952013-05-29 Lawrence Crowl <crowl@google.com>
15096
15097 * config/arm/t-arm: Update for below.
15098
15099 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15100 Change type to hash_table. Update dependent calls and types.
15101
15102 * config/i386/t-cygming: Update for below.
15103
15104 * config/i386/t-interix: Update for below.
15105
15106 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15107 Change type to hash_table. Update dependent calls and types.
15108 (i386_find_on_wrapper_list::wrappers): Likewise.
15109
15110 * config/ia64/t-ia64: Update for below.
15111
15112 * config/ia64/ia64.c (bundle_state_table):
15113 Change type to hash_table. Update dependent calls and types.
15114
15115 * config/mips/mips.c (mips_reorg_process_insns::htab):
15116 Change type to hash_table. Update dependent calls and types.
15117
15118 * config/sol2.c (solaris_comdat_htab):
15119 Change type to hash_table. Update dependent calls and types.
15120
15121 * config/t-sol2: Update for above.
15122
1388a0e3
TJ
151232013-05-29 Teresa Johnson <tejohnson@google.com>
15124
15125 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15126 functions are not yet marked as defined.
15127
a5965b52 151282013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
15129 Pat Haugen <pthaugen@us.ibm.com>
15130 Peter Bergner <bergner@vnet.ibm.com>
a5965b52
MM
15131
15132 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15133 instructions.
15134 (VEC_A): Likewise.
15135 (VEC_C): Likewise.
15136 (vrotl<mode>3): Likewise.
15137 (vashl<mode>3): Likewise.
15138 (vlshr<mode>3): Likewise.
15139 (vashr<mode>3): Likewise.
15140
15141 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15142 support for power8 V2DI builtins.
15143
15144 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15145 power8 V2DI builtins.
15146 (vupkhsw): Likewise.
15147 (vupklsw): Likewise.
15148 (vaddudm): Likewise.
15149 (vminsd): Likewise.
15150 (vmaxsd): Likewise.
15151 (vminud): Likewise.
15152 (vmaxud): Likewise.
15153 (vpkudum): Likewise.
15154 (vpksdss): Likewise.
15155 (vpkudus): Likewise.
15156 (vpksdus): Likewise.
15157 (vrld): Likewise.
15158 (vsld): Likewise.
15159 (vsrd): Likewise.
15160 (vsrad): Likewise.
15161 (vsubudm): Likewise.
15162 (vcmpequd): Likewise.
15163 (vcmpgtsd): Likewise.
15164 (vcmpgtud): Likewise.
15165 (vcmpequd_p): Likewise.
15166 (vcmpgtsd_p): Likewise.
15167 (vcmpgtud_p): Likewise.
15168 (vupkhsw): Likewise.
15169 (vupklsw): Likewise.
15170 (vaddudm): Likewise.
15171 (vmaxsd): Likewise.
15172 (vmaxud): Likewise.
15173 (vminsd): Likewise.
15174 (vminud): Likewise.
15175 (vpksdss): Likewise.
15176 (vpksdus): Likewise.
15177 (vpkudum): Likewise.
15178 (vpkudus): Likewise.
15179 (vrld): Likewise.
15180 (vsld): Likewise.
15181 (vsrad): Likewise.
15182 (vsrd): Likewise.
15183 (vsubudm): Likewise.
15184
15185 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15186 support for power8 V2DI instructions.
15187
15188 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15189 power8 V2DI instructions. Combine pack and unpack insns to use an
15190 iterator for each mode. Check whether a particular mode supports
15191 Altivec instructions instead of just checking TARGET_ALTIVEC.
15192 (UNSPEC_VPKUWUM): Likewise.
15193 (UNSPEC_VPKSHSS): Likewise.
15194 (UNSPEC_VPKSWSS): Likewise.
15195 (UNSPEC_VPKUHUS): Likewise.
15196 (UNSPEC_VPKSHUS): Likewise.
15197 (UNSPEC_VPKUWUS): Likewise.
15198 (UNSPEC_VPKSWUS): Likewise.
15199 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15200 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15201 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15202 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15203 (UNSPEC_VUPKHSB): Likewise.
15204 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15205 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15206 (UNSPEC_VUPKHSH): Likewise.
15207 (UNSPEC_VUPKLSB): Likewise.
15208 (UNSPEC_VUPKLSH): Likewise.
15209 (VI2): Likewise.
15210 (VI_char): Likewise.
15211 (VI_scalar): Likewise.
15212 (VI_unit): Likewise.
15213 (VP): Likewise.
15214 (VP_small): Likewise.
15215 (VP_small_lc): Likewise.
15216 (VU_char): Likewise.
15217 (add<mode>3): Likewise.
15218 (altivec_vaddcuw): Likewise.
15219 (altivec_vaddu<VI_char>s): Likewise.
15220 (altivec_vadds<VI_char>s): Likewise.
15221 (sub<mode>3): Likewise.
15222 (altivec_vsubcuw): Likewise.
15223 (altivec_vsubu<VI_char>s): Likewise.
15224 (altivec_vsubs<VI_char>s): Likewise.
15225 (altivec_vavgs<VI_char>): Likewise.
15226 (altivec_vcmpbfp): Likewise.
15227 (altivec_eq<mode>): Likewise.
15228 (altivec_gt<mode>): Likewise.
15229 (altivec_gtu<mode>): Likewise.
15230 (umax<mode>3): Likewise.
15231 (smax<mode>3): Likewise.
15232 (umin<mode>3): Likewise.
15233 (smin<mode>3): Likewise.
15234 (altivec_vpkuhum): Likewise.
15235 (altivec_vpkuwum): Likewise.
15236 (altivec_vpkshss): Likewise.
15237 (altivec_vpkswss): Likewise.
15238 (altivec_vpkuhus): Likewise.
15239 (altivec_vpkshus): Likewise.
15240 (altivec_vpkuwus): Likewise.
15241 (altivec_vpkswus): Likewise.
15242 (altivec_vpks<VI_char>ss): Likewise.
15243 (altivec_vpks<VI_char>us): Likewise.
15244 (altivec_vpku<VI_char>us): Likewise.
15245 (altivec_vpku<VI_char>um): Likewise.
15246 (altivec_vrl<VI_char>): Likewise.
15247 (altivec_vsl<VI_char>): Likewise.
15248 (altivec_vsr<VI_char>): Likewise.
15249 (altivec_vsra<VI_char>): Likewise.
15250 (altivec_vsldoi_<mode>): Likewise.
15251 (altivec_vupkhsb): Likewise.
15252 (altivec_vupkhs<VU_char>): Likewise.
15253 (altivec_vupkls<VU_char>): Likewise.
15254 (altivec_vupkhsh): Likewise.
15255 (altivec_vupklsb): Likewise.
15256 (altivec_vupklsh): Likewise.
15257 (altivec_vcmpequ<VI_char>_p): Likewise.
15258 (altivec_vcmpgts<VI_char>_p): Likewise.
15259 (altivec_vcmpgtu<VI_char>_p): Likewise.
15260 (abs<mode>2): Likewise.
15261 (vec_unpacks_hi_v16qi): Likewise.
15262 (vec_unpacks_hi_v8hi): Likewise.
15263 (vec_unpacks_lo_v16qi): Likewise.
15264 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15265 (vec_unpacks_lo_v8hi): Likewise.
15266 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15267 (vec_pack_trunc_v8h): Likewise.
15268 (vec_pack_trunc_v4si): Likewise.
15269 (vec_pack_trunc_<mode>): Likewise.
15270
15271 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15272 V2DI builtins.
15273 (vec_vmaxsd): Likewise.
15274 (vec_vmaxud): Likewise.
15275 (vec_vminsd): Likewise.
15276 (vec_vminud): Likewise.
15277 (vec_vpksdss): Likewise.
15278 (vec_vpksdus): Likewise.
15279 (vec_vpkudum): Likewise.
15280 (vec_vpkudus): Likewise.
15281 (vec_vrld): Likewise.
15282 (vec_vsld): Likewise.
15283 (vec_vsrad): Likewise.
15284 (vec_vsrd): Likewise.
15285 (vec_vsubudm): Likewise.
15286 (vec_vupkhsw): Likewise.
15287 (vec_vupklsw): Likewise.
15288
e70670cf
JH
152892013-05-29 Jan Hubicka <jh@suse.cz>
15290
15291 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15292 flags; reorder rest of fields in more consistent way.
15293 (varpool_node): Remove analyzed, finalized and alias.
15294 (cgraph_ndoe): Likewise.
15295 (symtab_alias_ultimate_target): New function.
15296 (cgraph_function_node): Move offline.
15297 (cgraph_reset_node): Declare.
15298 (cgraph_comdat_can_be_unshared_p): Remove.
15299 (varpool_remove_initializer): Declare.
15300 (varpool_first_defined_variable, varpool_next_defined_variable
15301 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15302 (cgraph_function_with_gimple_body_p): Update.
15303 (varpool_all_refs_explicit_p): Update.
15304 (symtab_alias_target): New function.
15305 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15306 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
c1e183a9
UB
15307 (cgraph_function_or_thunk_node): Simplify using
15308 symtab_alias_ultimate_target.
e70670cf
JH
15309 (varpool_variable_node): Likewise.
15310 * cgraph.c (cgraph_create_function_alias): Update.
15311 (cgraph_add_thunk): Update.
15312 (cgraph_remove_node): Update.
15313 (dump_cgraph_node): Do not dump removed flags.
15314 (cgraph_function_body_availability): Update.
15315 (cgraph_propagate_frequency): Update.
15316 (verify_cgraph_node): Check sanity of local flag.
15317 (cgraph_function_node): Move here from cgraph.h; revamp for
15318 cgraph_function_or_thunk_node.
15319 * lto-symtab.c (lto_varpool_replace_node): Update.
15320 (lto_symtab_resolve_can_prevail_p): Update.
15321 (lto_symtab_merge_cgraph_nodes): Update.
15322 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15323 propagate_constants_accross_call, devirtualization_time_bonus,
15324 ipcp_propagate_stage): Update.
15325 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
c1e183a9
UB
15326 * ipa-inline-transform.c (clone_inlined_nodes,
15327 preserve_function_body_p): Update.
e70670cf
JH
15328 * ipa-reference.c (propagate): Update.
15329 (write_node_summary_p): Update.
15330 * toplev.c (wrapup_global_declaration_2): Update.
15331 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15332 (analyze_function) ... this one.
15333 (cgraph_process_new_functions): Update.
15334 (cgraph_reset_node): Export.
15335 (cgraph_finalize_function): Update.
15336 (cgraph_add_new_function): Update.
15337 (process_function_and_variable_attributes): Update.
15338 (varpool_finalize_decl): Update.
15339 (symbol_finalized): Remove.
15340 (symbol_finalized_and_needed): Rename to ...
15341 (symbol_defined_and_needed): ... update.
15342 (cgraph_analyze_functions): Update.
15343 (handle_alias_pairs): Update.
15344 (mark_functions_to_output): Update.
15345 (assemble_thunk): Update.
15346 (output_in_order): Update.
15347 (output_weakrefs): Update.
15348 (finalize_compilation_unit): Update.
15349 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15350 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15351 input_node, input_varpool_node): Update.
15352 * dbxout.c (dbxout_expand_expr): Update.
15353 * cgraphclones.c (cgraph_clone_node): Update.
15354 (cgraph_copy_node_for_versioning): Update.
15355 (cgraph_materialize_clone): Update.
15356 (cgraph_materialize_all_clones): Update.
15357 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15358 propagate_pure_const, propagate_nothrow): Update.
15359 * lto-streamer-out.c (lto_output, write_symbol): Update.
15360 * ipa-utils.c (ipa_reverse_postorder): Update.
15361 * ipa-inline.c (can_inline_edge_p): Update.
15362 (update_caller_keys, ipa_inline): Update.
15363 * dwarf2out.c (reference_to_unused,
15364 premark_types_used_by_global_vars_helper): Update.
15365 * tree-eh.c (tree_could_trap_p): Update.
15366 * ipa-split.c (consider_split, execute_split_functions): Update.
15367 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
c1e183a9
UB
15368 has_addr_references_p): Update; move ahead in file for better
15369 readability.
e70670cf
JH
15370 (process_references): Simplify.
15371 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15372 bodies are removed.
15373 (cgraph_comdat_can_be_unshared_p): Make static.
15374 (cgraph_externally_visible_p): Update.
15375 (varpool_externally_visible_p): Update.
15376 (function_and_variable_visibility): Update.
15377 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15378 ipa_tm_mark_force_output_node): Update.
15379 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15380 estimate_edge_devirt_benefit, inline_generate_summary,
15381 inline_write_summary): Update.
15382 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15383 * ipa-prop.c (ipa_compute_jump_functions): Update.
c1e183a9
UB
15384 (ipa_print_node_params, ipa_prop_read_section,
15385 ipa_update_after_lto_read, read_replacements_section): Update.
e70670cf
JH
15386 * varasm.c (mark_decl_referenced): Update.
15387 (assemble_alias, dump_tm_clone_pairs): Update.
15388 * tree-inline.c (copy_bb): Update.
15389 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15390 Update.
15391 * symtab.c (dump_symtab_base): Print new flags.
15392 (verify_symtab_base): Verify new flags.
15393 (symtab_alias_ultimate_target): New function.
15394 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15395 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15396 Update.
c1e183a9
UB
15397 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15398 Update.
e70670cf
JH
15399 * i386.c (ix86_get_function_versions_dispatcher,
15400 ix86_generate_version_dispatcher_body): Update.
15401 (fold_builtin_cpu): Use varpool_add_new_variable.
15402 * varpool.c (varpool_remove_initializer): Break out from ...
15403 (varpool_remove_node): ... this one.
15404 (dump_varpool_node, varpool_node_for_asm,
15405 cgraph_variable_initializer_availability, varpool_analyze_node,
15406 varpool_assemble_decl, varpool_remove_unreferenced_decls,
c1e183a9
UB
15407 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15408 Update.
e70670cf 15409
182802ad
JH
154102013-05-29 Jan Hubicka <jh@suse.cz>
15411
15412 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15413
8190b609
ER
154142013-05-29 Easwaran Raman <eraman@google.com>
15415
15416 PR tree-optimization/57442
15417 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15418 when control exits the main loop.
15419
69f5aa9b
SKS
154202013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15421
15422 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15423 and RX600.
c1e183a9 15424 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
69f5aa9b
SKS
15425 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15426 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15427 option.
15428
4b847da9
BS
154292013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15430
15431 PR tree-optimization/57441
15432 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15433 Don't limit size of incr_vec to number of candidates.
15434
4fc43c39
SE
154352013-05-29 Steve Ellcey <sellcey@imgtec.com>
15436
15437 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15438 and mips16 directories.
c1e183a9 15439 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
4fc43c39
SE
15440 (MULTILIB_DIRNAMES): Ditto.
15441 (MULTILIB_EXCEPTIONS): Add new exceptions.
15442 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15443 (MULTILIB_DIRNAMES): Ditto.
15444 (MULTILIB_EXCEPTIONS): Add new exceptions.
15445
12211b99 154462012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
a5350ddc
CSS
15447 Marcus Shawcroft <marcus.shawcroft@arm.com>
15448
15449 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15450 SYMBOL_TINY_ABSOLUTE.
15451 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15452 SYMBOL_TINY_ABSOLUTE.
15453 (aarch64_expand_mov_immediate): Likewise.
15454 (aarch64_classify_symbol): Likewise.
15455 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15456 Permit SYMBOL_TINY_ABSOLUTE.
15457 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15458
12211b99 154592013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
17f4d4bf
CSS
15460 Marcus Shawcroft <marcus.shawcroft@arm.com>
15461
15462 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15463 Refactor if/switch. Replace gcc_assert with if.
15464
c0186656
GG
154652013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15466
15467 * config/i386/i386.c (initial_ix86_tune_features): Enable
15468 FP Reassociation for AMD bdver1 and bdver2.
15469
d20188f3
MJ
154702013-05-29 Martin Jambor <mjambor@suse.cz>
15471
15472 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15473 and IMAGPART_EXPR do not occur within other handled_components.
15474
292cba13
RB
154752013-05-29 Richard Biener <rguenther@suse.de>
15476
15477 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15478 access on whether the use is in the BB we currently try to
15479 vectorize.
15480 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15481 vectorize to vect_bb_slp_scalar_cost.
15482
6eddf228
RB
154832013-05-29 Richard Biener <rguenther@suse.de>
15484
15485 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15486 computing scalar cost offsetted by stmts that are kept live
15487 by scalar uses.
15488 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15489 for computation of scalar cost.
15490
7df36117
SE
154912013-05-28 Steve Ellcey <sellcey@mips.com>
15492
15493 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15494
36536d79
BI
154952013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15496
15497 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15498 array notation built-in reduction functions.
15499 * doc/passes.texi (Passes): Added documentation about changes done
15500 for Cilk Plus.
15501 * doc/invoke.texi (C Dialect Options): Added documentation about
15502 the -fcilkplus flag.
15503 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15504 (BUILTINS_DEF): Depend on cilkplus.def.
15505 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15506 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15507 * cilkplus.def: New file.
15508
12211b99 155092013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
b49eefa5
JR
15510
15511 PR rtl-optimization/57439
15512 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15513
9daf6dbb
ER
155142013-05-28 Easwaran Raman <eraman@google.com>
15515
15516 PR tree-optimization/57337
15517 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15518 (find_insert_point): Correctly identify the insertion point
15519 when two statements with the same UID is compared.
15520
fbd7e877
RB
155212013-05-28 Richard Biener <rguenther@suse.de>
15522
15523 PR tree-optimization/56787
15524 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15525 from the list of data references.
15526 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15527 clobbers.
15528 (vect_analyze_loop_operations): Likewise.
15529 (vect_transform_loop): Remove clobbers.
15530
bbba1117
MJ
155312013-05-28 Martin Jambor <mjambor@suse.cz>
15532
15533 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15534 and REALPART_EXPRs have scalar type.
15535
bd388c2a
RB
155362013-05-28 Richard Biener <rguenther@suse.de>
15537
15538 PR tree-optimization/57411
15539 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15540 virtual operands.
15541 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15542 virtual operand propagation.
15543
2f56a311
EB
155442013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15545
15546 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15547 destination register for bmasksi_vis.
15548 (vector_init_bshuffle): Likewise.
15549 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15550
5c3eacbb
EB
155512013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15552
15553 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15554 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15555 mode if the instruction isn't available in the original mode.
15556 * config/sparc/sparc.opt (mfix-ut699): New option.
15557 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15558 (divdf3): Turn into expander.
15559 (divdf3_nofix): New insn.
15560 (divdf3_fix): Likewise.
15561 (divsf3): Disable if -mfix-ut699.
15562 (sqrtdf2): Turn into expander.
15563 (sqrtdf2_nofix): New insn.
15564 (sqrtdf2_fix): Likewise.
15565 (sqrtsf2): Disable if -mfix-ut699.
15566
a1756c0a
RB
155672013-05-27 Richard Biener <rguenther@suse.de>
15568
15569 PR middle-end/57412
15570 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15571 block for the new loop.
15572
5a892248
RB
155732013-05-27 Richard Biener <rguenther@suse.de>
15574
15575 PR tree-optimization/57343
15576 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15577 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15578 (number_of_iterations_cond): Do not build the folded tree.
15579
d1de852b
RB
155802013-05-27 Richard Biener <rguenther@suse.de>
15581
15582 Revert
15583 PR middle-end/57381
15584 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15585 OEP_CONSTANT_ADDRESS_OF retained.
15586
15587 PR tree-optimization/57417
15588 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15589 for unchanged base.
15590 (set_ssa_val_to): Compare addresses using
15591 get_addr_base_and_unit_offset.
15592
12211b99 155932013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
7894bc6b
JR
15594
15595 PR rtl-optimization/56833
15596 * postreload.c (move2add_record_mode): New function.
15597 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15598 (move2add_use_add2_insn): Use move2add_record_sym_value.
15599 (move2add_use_add3_insn): Likewise.
15600 (reload_cse_move2add): Use move2add_valid_value_p and
15601 move2add_record_mode. Invalidate call-clobbered and REG_INC
15602 affected regs by setting reg_mode to VOIDmode.
15603 (move2add_note_store): Don't pretend the inside of a SUBREG is
15604 the actual destination. Invalidate single/leading registers by
15605 setting reg_mode to VOIDmode.
15606 Use move2add_record_sym_value, move2add_valid_value_p and
15607 move2add_record_mode.
15608
b03be25f
RB
156092013-05-27 Richard Biener <rguenther@suse.de>
15610
15611 PR tree-optimization/57396
15612 * tree-affine.c (double_int_constant_multiple_p): Properly
15613 return false for val == 0 and div != 0.
15614
44e88db2
RS
156152013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15616
15617 * config/mips/mips.h: Use #elif in preprocessor conditions.
15618
3b859704
RS
156192013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15620
15621 PR target/53916
15622 * config/mips/constraints.md (kl): New constraint.
15623 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15624 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15625 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15626 from LO for MIPS16.
15627 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15628
c979d5f5
RS
156292013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15630
15631 PR target/55777
15632 * config/mips/mips.c (mips_can_inline_p): New function.
15633 (TARGET_CAN_INLINE_P): Define.
15634
8e90de43
SB
156352013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15636
15637 * sched-int.h (ds_t, dw_t): Make unsigned int.
15638 Fix documentation that describes how all the ds_t bits are used.
15639 Reserve the last bit for delayed-branch scheduling.
15640 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15641 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15642 (gen_dep_weak_1): Remove prototype.
15643 * sched-deps.c (get_dep_weak_1): Make static.
15644 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15645 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15646 * doc/tm.texi: Regenerate.
15647 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15648
cb5cb194
SB
156492013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15650
15651 PR debug/56950
15652 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15653
5bd93ff6
NS
156542013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15655 Sandra Loosemore <sandra@codesourcery.com>
15656
15657 * config.gcc (powerpc-*): Allow native for with-cpu.
15658
2343af65
JL
156592013-05-24 Jeff Law <law@redhat.com>
15660
15661 PR tree-optimization/57124
15662 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15663 conversion feeding a condition if the range has an overflow
15664 if -fstrict-overflow. Add warnings for when we do make the
15665 transformation.
15666
3ad695b9
DC
156672013-05-24 Dehao Chen <dehao@google.com>
15668
04960246 15669 * tree-cfg.c (locus_discrim_map): Fix the typo.
3ad695b9
DC
15670 (locus_discrim_hasher): Likewise.
15671 (locus_discrim_hasher::hash): Likewise.
15672 (locus_discrim_hasher::equal): Likewise.
15673
82338059
MJ
156742013-05-24 Martin Jambor <mjambor@suse.cz>
15675
15676 PR tree-optimization/57294
15677 * cgraph.h (ipa_record_stmt_references): Declare.
15678 * cgraphbuild.c (ipa_record_stmt_references): New function.
15679 (build_cgraph_edges): Use ipa_record_stmt_references.
15680 (rebuild_cgraph_edges): Likewise.
15681 (cgraph_rebuild_references): Likewise.
15682 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15683 associated with the old statement and build references from the
15684 newly built statements.
15685 * ipa-ref.c (ipa_remove_stmt_references): New function.
15686 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15687
1ccd4874
VM
156882013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15689
55805e54 15690 * lra-constraints.c (emit_spill_move): Use smaller mode for
1ccd4874 15691 mem-mem moves.
55805e54 15692 (check_and_process_move): Consider mem-reg moves for secondary
1ccd4874 15693 too.
55805e54 15694 (curr_insn_transform): Don't lose insns emitted before for
1ccd4874
VM
15695 secondary memory moves.
15696 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15697 reg set up in the current insn.
15698
25e25c73
DC
156992013-05-24 Dehao Chen <dehao@google.com>
15700
04960246 15701 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
25e25c73
DC
15702 hash function.
15703 (locus_descrim_hasher::equal): Likewise.
04960246 15704 (build_gimple_cfg): New discriminator assignment algorithm.
25e25c73
DC
15705 (make_edges): Likewise.
15706 (next_discriminator_for_locus): Likewise.
15707 (same_line_p): Likewise.
15708 (assign_discriminators): Likewise.
15709 (make_cond_expr_edges): Likewise.
15710 (make_gimple_switch_edges): Likewise.
15711 (make_goto_expr_edges): Likewise.
15712 (make_gimple_asm_edges): Likewise.
15713
50d38551
IB
157142013-05-24 Ian Bolton <ian.bolton@arm.com>
15715
15716 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15717 X format specifier to only display bottom 16 bits.
15718 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15719 immediate to match for operand 2, since it will be masked.
15720
aea0101d
RB
157212013-05-24 Richard Biener <rguenther@suse.de>
15722
15723 PR tree-optimization/57287
15724 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15725 all SSA names that occur in abnormal PHIs.
15726
634e03d3
AI
157272013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15728
15729 PR tree-ssa/57385
15730 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15731 that index is not negative.
15732
b17c024f
EB
157332013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15734
15735 PR rtl-optimization/55177
15736 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15737 (simplify_byte_swapping_operation): New.
15738 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15739 (simplify_relational_operation_1): Deal with BSWAP.
15740
46aeac1b
RH
157412013-05-23 Richard Henderson <rth@redhat.com>
15742
15743 PR target/56742
15744 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15745 (ix86_reorg): Call it.
15746
70cc1536
UB
157472013-05-23 Uros Bizjak <ubizjak@gmail.com>
15748
15749 PR target/57379
15750 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
15751 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
15752 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
15753
b3851501
CB
157542013-05-23 Christian Bruel <christian.bruel@st.com>
15755
15756 PR debug/57351
15757 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
15758
12211b99 157592013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
4e194912
CSS
15760 Marcus Shawcroft <marcus.shawcroft@arm.com>
15761
15762 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
15763 * config/aarch64/constraints.md (Usa): Remove.
15764 * doc/md.texi (AArch64 Usa): Remove.
15765
12211b99 157662013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
83f8c414
CSS
15767 Marcus Shawcroft <marcus.shawcroft@arm.com>
15768
15769 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
15770 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
15771 * config/aarch64/predicates.md (aarch64_const_address): Remove.
15772 (aarch64_mov_operand): Use aarch64_mov_operand_p.
15773
12211b99 157742013-05-23 Vidya Praveen <vidyapraveen@arm.com>
b5574232
VP
15775
15776 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
15777 instruction (AdvSIMD).
15778 * config/aarch64/aarch64-builtins.c
15779 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
15780 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
15781
76c36cb1 157822013-05-23 Martin Jambor <mjambor@suse.cz>
7d2fb524
MJ
15783
15784 PR middle-end/57347
15785 * tree.h (contains_bitfld_component_ref_p): Declare.
15786 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
70cc1536
UB
15787 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
15788 caller.
7d2fb524
MJ
15789 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
15790 not access a bit-field. Assert all final offsets are byte-aligned.
15791
ce521ae6
RB
157922013-05-23 Richard Biener <rguenther@suse.de>
15793
15794 PR tree-optimization/57380
15795 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
15796 least one invariant or re-used load.
15797 * passes.c (init_optimization_passes): Move pass_phiprop before
15798 pass_forwprop.
15799
75c7257f
JG
158002013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
15801
15802 * config/aarch64/aarch64-simd.md
15803 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
15804
af6d44b5
RB
158052013-05-23 Richard Biener <rguenther@suse.de>
15806
15807 PR middle-end/57381
15808 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15809 OEP_CONSTANT_ADDRESS_OF retained.
15810
bd3647bf
JJ
158112013-05-23 Jakub Jelinek <jakub@redhat.com>
15812
15813 PR middle-end/57344
70cc1536
UB
15814 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
15815 don't lower unit. Handle unit not being always BITS_PER_WORD.
bd3647bf 15816
a55757ea
RB
158172013-05-23 Richard Biener <rguenther@suse.de>
15818
15819 PR rtl-optimization/57341
15820 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
15821 instead of true_dependence.
15822
cfb00b41
DM
158232013-05-22 David Malcolm <dmalcolm@redhat.com>
15824
dd1c676f
DM
15825 * bb-reorder.c (branch_threshold): Make const.
15826 (exec_threshold): Ditto.
cfb00b41 15827
f62511da 158282013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
15829 Pat Haugen <pthaugen@us.ibm.com>
15830 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
15831
15832 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
15833 documentation for the power8 crypto builtins.
15834
15835 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
15836
15837 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
15838 macros for defining power8 builtin functions.
15839 (BU_P8V_AV_2): Likewise.
15840 (BU_P8V_AV_P): Likewise.
15841 (BU_P8V_VSX_1): Likewise.
15842 (BU_P8V_OVERLOAD_1): Likewise.
15843 (BU_P8V_OVERLOAD_2): Likewise.
15844 (BU_CRYPTO_1): Likewise.
15845 (BU_CRYPTO_2): Likewise.
15846 (BU_CRYPTO_3): Likewise.
15847 (BU_CRYPTO_OVERLOAD_1): Likewise.
15848 (BU_CRYPTO_OVERLOAD_2): Likewise.
15849 (XSCVSPDP): Fix typo, point to the correct instruction.
15850 (VCIPHER): Add power8 crypto builtins.
15851 (VCIPHERLAST): Likewise.
15852 (VNCIPHER): Likewise.
15853 (VNCIPHERLAST): Likewise.
15854 (VPMSUMB): Likewise.
15855 (VPMSUMH): Likewise.
15856 (VPMSUMW): Likewise.
15857 (VPERMXOR_V2DI): Likewise.
15858 (VPERMXOR_V4SI: Likewise.
15859 (VPERMXOR_V8HI: Likewise.
15860 (VPERMXOR_V16QI: Likewise.
15861 (VSHASIGMAW): Likewise.
15862 (VSHASIGMAD): Likewise.
15863 (VPMSUM): Likewise.
15864 (VPERMXOR): Likewise.
15865 (VSHASIGMA): Likewise.
15866
15867 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15868 __CRYPTO__ if the crypto instructions are available.
15869 (altivec_overloaded_builtins): Add support for overloaded power8
15870 builtins.
15871
15872 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
15873 support for power8 crypto builtins.
15874 (builtin_function_type): Likewise.
15875 (altivec_init_builtins): Add support for builtins that take vector
15876 long long (V2DI) arguments.
15877
15878 * config/rs6000/crypto.md: New file, define power8 crypto
15879 instructions.
15880
158812013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
12211b99
UB
15882 Pat Haugen <pthaugen@us.ibm.com>
15883 Peter Bergner <bergner@vnet.ibm.com>
f62511da
MM
15884
15885 * doc/invoke.texi (Option Summary): Add power8 options.
15886 (RS/6000 and PowerPC Options): Likewise.
15887
15888 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
15889 constraints.md instead of rs6000.h. Reorder w* constraints. Add
15890 wm, wn, wr documentation.
15891
f43245d1 15892 * config/rs6000/constraints.md (wm): New constraint for VSX
f62511da
MM
15893 registers if direct move instructions are enabled.
15894 (wn): New constraint for no registers.
15895 (wq): New constraint for quad word even GPR registers.
15896 (wr): New constraint if 64-bit instructions are enabled.
15897 (wv): New constraint if power8 vector instructions are enabled.
15898 (wQ): New constraint for quad word memory locations.
15899
f43245d1 15900 * config/rs6000/predicates.md (const_0_to_15_operand): New
f62511da
MM
15901 constraint for 0..15 for crypto instructions.
15902 (gpc_reg_operand): If VSX allow registers in VSX registers as well
15903 as GPR and floating point registers.
15904 (int_reg_operand): New predicate to match only GPR registers.
15905 (base_reg_operand): New predicate to match base registers.
15906 (quad_int_reg_operand): New predicate to match even GPR registers
15907 for quad memory operations.
15908 (vsx_reg_or_cint_operand): New predicate to allow vector logical
15909 operations in both GPR and VSX registers.
15910 (quad_memory_operand): New predicate for quad memory operations.
15911 (reg_or_indexed_operand): New predicate for direct move support.
15912
f43245d1 15913 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
f62511da
MM
15914 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
15915 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
15916 (POWERPC_MASKS): Add power8 options.
15917 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
15918 various options.
15919
f43245d1 15920 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
f62511da
MM
15921 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
15922
f43245d1 15923 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
f62511da
MM
15924 (-mpower8-fusion): New power8 options.
15925 (-mpower8-fusion-sign): Likewise.
15926 (-mpower8-vector): Likewise.
15927 (-mcrypto): Likewise.
15928 (-mdirect-move): Likewise.
15929 (-mquad-memory): Likewise.
15930
f43245d1 15931 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
f62511da
MM
15932 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
15933 registers.
70cc1536 15934 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
f62511da
MM
15935 (rs6000_debug_vector_unit): Add p8_vector.
15936 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
15937 definitions. Also print fusion state.
15938 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
15939 (rs6000_builtin_mask_calculate): Add power8 builtin support.
15940 (rs6000_option_override_internal): Add support for power8.
15941 (rs6000_common_init_builtins): Add debugging for skipped builtins
15942 if -mdebug=builtin.
15943 (rs6000_adjust_cost): Add power8 support.
15944 (rs6000_issue_rate): Likewise.
15945 (insn_must_be_first_in_group): Likewise.
15946 (insn_must_be_last_in_group): Likewise.
15947 (force_new_group): Likewise.
15948 (rs6000_register_move_cost): Likewise.
15949 (rs6000_opt_masks): Likewise.
15950
15951 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
15952 power8 capable assembler, default to power7 options.
15953 (TARGET_DIRECT_MOVE): Likewise.
15954 (TARGET_CRYPTO): Likewise.
15955 (TARGET_P8_VECTOR): Likewise.
15956 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
15957 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
15958 (VECTOR_MEM_P8_VECTOR_P): Likewise.
15959 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
15960 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
15961 (TARGET_XSCVDPSPN): Likewise.
15962 (TARGET_XSCVSPDPN): Likewsie.
15963 (TARGET_SYNC_HI_QI): Likewise.
15964 (TARGET_SYNC_TI): Likewise.
15965 (MASK_CRYPTO): Likewise.
15966 (MASK_DIRECT_MOVE): Likewise.
15967 (MASK_P8_FUSION): Likewise.
15968 (MASK_P8_VECTOR): Likewise.
70cc1536
UB
15969 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
15970 temporary used by some of the direct move instructions to get two FP
15971 temporary registers does not force creation of a stack frame.
f62511da
MM
15972 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
15973 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
15974 that any VSX registers are tieable, even if they are also an
15975 Altivec vector mode.
15976 (r6000_reg_class_enum): Add wm, wr, wv constraints.
15977 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
15978 (RS6000_BTM_CRYPTO): Likewise.
15979 (RS6000_BTM_COMMON): Likewise.
15980
15981 * config/rs6000/rs6000.md (cpu attribute): Add power8.
15982 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
15983 (enum rs6000_vector): Add power8 vector support.
15984
73a1a707
RR
159852013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15986
15987 PR target/19599
15988 PR target/57340
15989 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
15990 (any_sibcall_could_use_r3): this and handle indirect calls.
15991 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
15992
d28073d4
BS
159932013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15994
15995 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
15996
bee0b10c
RB
159972013-05-22 Richard Biener <rguenther@suse.de>
15998
15999 PR middle-end/57349
16000 * profile.c (branch_prob): Do not split blocks that are
16001 abnormally receiving from ECF_RETURNS_TWICE functions.
16002
98409b51
RS
160032013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
16004
16005 * recog.c (offsettable_address_addr_space_p): Fix calculation of
16006 address mode. Move pointer mode initialization to the same place.
16007
c0602ab8
MZ
160082013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
16009
16010 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
16011 while it has any effect.
16012
4f7a634e
ER
160132013-05-21 Easwaran Raman <eraman@google.com>
16014
16015 PR tree-optimization/57322
9daf6dbb
ER
16016 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
16017 UID of the statement added to the BB to be 1.
4f7a634e 16018
a4ce1258
JJ
160192013-05-21 Jakub Jelinek <jakub@redhat.com>
16020
16021 PR tree-optimization/57331
70cc1536
UB
16022 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
16023 of conversion from pointer type to integral type with integer.
a4ce1258 16024
1b14621a
MJ
160252013-05-21 Martin Jambor <mjambor@suse.cz>
16026
16027 PR lto/57289
16028 * ipa-prop.c (ipa_read_node_info): Process param_used and
16029 controlled_uses in the same order as when writing.
16030
e60661f0
MG
160312013-05-21 Magnus Granberg <baldrick@free.fr>
16032
16033 PR plugins/56754
ee49aa34 16034 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
e60661f0 16035
4a61be9a
RB
160362013-05-21 Richard Biener <rguenther@suse.de>
16037
16038 PR tree-optimization/57318
16039 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16040 estimate stmts with side-effects as likely eliminated.
16041
c52da5f7
RB
160422013-05-21 Richard Biener <rguenther@suse.de>
16043
16044 PR tree-optimization/57330
16045 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16046 preserve the call stmts fntype.
16047
7ec67e2a
RB
160482013-05-21 Richard Biener <rguenther@suse.de>
16049
16050 PR tree-optimization/57303
16051 * tree-ssa-sink.c (statement_sink_location): Improve killing
16052 stmt detection and properly handle self-assignments.
16053
b112d513
CB
160542013-05-21 Christian Bruel <christian.bruel@st.com>
16055
55805e54
YR
16056 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16057 spanning registers. LEAF_REG_REMAP is supported only for contiguous
b112d513
CB
16058 registers. Set register size out of the PARALLEL loop.
16059
14c2ec26
OE
160602013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16061
16062 PR target/56547
16063 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16064 (*fmasf4, *fmasf4_media): New insns.
16065
da734fa1
RS
160662013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16067
16068 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16069 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16070 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16071 (mips_idiv_insns): Update the comments to say that the returned
16072 instruction counts are in units of BASE_INSN_LENGTH.
16073 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16074 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16075 using 2 rather than 4 as the length of indirect MIPS16 and
16076 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16077 length of a NOP. Don't divide MIPS16 lengths by 2.
16078 (mips16_split_long_branches): Assume a branch is long if the
16079 length is greater than 4 rather than 8.
16080 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16081 rather than multiplying them by 2. Multiply instruction counts
16082 by BASE_INSN_LENGTH rather than 4.
16083 (*jump_mips16, tls_get_tp_mips16_<mode>)
16084 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16085
13719e8b
RS
160862013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16087
16088 * config/mips/mips.md (extended_mips16): Remove branch case.
16089 (length): Remove duplicated extended_mips16 test.
16090
c3850d14
RS
160912013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16092
16093 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16094
8da2e059
RS
160952013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16096
16097 * recog.h (Recog_data): Rename to...
16098 (recog_data_d): ...this.
16099 (recog_data): Update accordingly.
16100 * recog.c (recog_data): Likewise.
16101 * reload.c (save_recog_data): Likewise.
16102 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16103 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16104
2df013f3
JB
161052013-05-17 Julian Brown <julian@codesourcery.com>
16106
16107 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16108 found in a REG_EQUAL note, invalidate it.
16109
371e77e3 161102013-05-17 Easwaran Raman <eraman@google.com>
933f507d
ER
16111
16112 * tree-ssa-reassoc.c (find_insert_point): New function.
16113 (insert_stmt_after): Likewise.
16114 (get_def_stmt): Likewise.
16115 (ensure_ops_are_available): Likewise.
16116 (not_dominated_by): Likewise.
16117 (rewrite_expr_tree): Do not move statements beyond what is
16118 necessary. Remove call to swap_ops_for_binary_stmt...
16119 (reassociate_bb): ... and move it here.
16120 (build_and_add_sum): Assign UIDs for new statements.
16121 (linearize_expr): Likewise.
16122 (do_reassoc): Renumber gimple statement UIDs.
16123
e01c7cca
JH
161242013-05-17 Jan Hubicka <jh@suse.cz>
16125
16126 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16127 weakrefs.
16128 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
e2420687
UB
16129 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16130 weakrefs.
e01c7cca
JH
16131 (output_weakrefs): Update.
16132
c3272a92
PCC
161332013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16134 Martin Jambor <mjambor@suse.cz>
16135
16136 PR middle-end/57276
16137 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16138 value that corresponds to the given aggval is found in values vector.
16139
11c2aa39
UB
161402013-05-17 Uros Bizjak <ubizjak@gmail.com>
16141
16142 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16143 sse, sse2, sse3, ssse3 and sse4a flags to options.
16144
abecc8c6
DM
161452013-05-17 David Malcolm <dmalcolm@redhat.com>
16146
16147 * gengtype-state.c: (s_expr_writer): New class, to handle
16148 prettifying of output layout of s-expressions.
16149 (state_writer): New class, to write out gtype.state.
16150 (state_written_type_count): Move this variable into member data of
16151 state_writer.
16152 (s_expr_writer::s_expr_writer): New code: constructor for new class
16153 (state_writer::state_writer(): ditto
16154 (s_expr_writer::write_new_line): New function
16155 (s_expr_writer::write_any_indent): ditto
16156 (s_expr_writer::begin_s_expr): ditto
16157 (s_expr_writer::end_s_expr): ditto
16158 (write_state_fileloc): convert to method of state_writer...
16159 (state_writer:: write_state_fileloc): ...and use methods of
16160 s_expr_writer to write indentation into the gtype.state output file
16161 to visually represent the hierarchical structure of the list
16162 structures
16163 (write_state_fields): ditto, renaming to...
16164 (state_writer::write_state_fields)
16165 (write_state_a_string): ditto, renaming to...
16166 (state_writer::write_state_a_string)
16167 (write_state_string_option): ditto, renaming to...
16168 (state_writer::write_state_string_option)
16169 (write_state_type_option): ditto, renaming to...
16170 (state_writer::write_state_type_option)
16171 (write_state_nested_option): ditto, renaming to...
16172 (state_writer::write_state_nested_option)
16173 (write_state_option): ditto, renaming to...
16174 (state_writer::write_state_option)
16175 (write_state_options): ditto, renaming to...
16176 (state_writer::write_state_options)
16177 (write_state_lang_bitmap): ditto, renaming to...
16178 (state_writer::write_state_lang_bitmap)
16179 (write_state_version): ditto, renaming to...
16180 (state_writer::write_state_version)
16181 (write_state_scalar_type): ditto, renaming to...
16182 (state_writer::write_state_scalar_type)
16183 (write_state_string_type): ditto, renaming to...
16184 (state_writer::write_state_string_type)
16185 (write_state_undefined_type): ditto, renaming to...
16186 (state_writer::write_state_undefined_type)
16187 (write_state_struct_union_type): ditto, renaming to...
16188 (state_writer::write_state_struct_union_type)
16189 (write_state_struct_type): ditto, renaming to...
16190 (state_writer::write_state_struct_type)
16191 (write_state_user_struct_type): ditto, renaming to...
16192 (state_writer::write_state_user_struct_type)
16193 (write_state_lang_struct_type): ditto, renaming to...
16194 (state_writer::write_state_lang_struct_type)
16195 (write_state_param_struct_type): ditto, renaming to...
16196 (state_writer::write_state_param_struct_type)
16197 (write_state_pointer_type): ditto, renaming to...
16198 (state_writer::write_state_pointer_type)
16199 (write_state_array_type): ditto, renaming to...
16200 (state_writer::write_state_array_type)
16201 (write_state_gc_used): ditto, renaming to...
16202 (state_writer::write_state_gc_used)
16203 (write_state_common_type_content): ditto, renaming to...
16204 (state_writer::write_state_common_type_content)
16205 (write_state_type): ditto, renaming to...
16206 (state_writer::write_state_type)
16207 (write_state_pair_list): ditto, renaming to...
16208 (state_writer::write_state_pair_list)
16209 (write_state_pair): ditto, renaming to...
16210 (state_writer::write_state_pair)
16211 (write_state_typedefs): ditto, renaming to...
16212 (state_writer::write_state_typedefs)
16213 (write_state_structures): ditto, renaming to...
16214 (state_writer::write_state_structures)
16215 (write_state_param_structs): ditto, renaming to...
16216 (state_writer::write_state_param_structs)
16217 (write_state_variables): ditto, renaming to...
16218 (state_writer::write_state_variables)
16219 (write_state_srcdir): ditto, renaming to...
16220 (state_writer::write_state_srcdir)
16221 (write_state_files_list): ditto, renaming to...
16222 (state_writer::write_state_files_list)
16223 (write_state_languages): ditto, renaming to...
16224 (state_writer::write_state_languages)
16225 (write_state): create a state_writer instance and use it when
16226 writing out the state file
16227
d6545f29
MS
162282013-05-17 Mike Stump <mikestump@comcast.net>
16229
816a3f73 16230 PR rtl-optimization/57304
d6545f29
MS
16231 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16232 accessing DF_REF_REAL_LOC.
16233
38c821cf
JJ
162342013-05-17 Jakub Jelinek <jakub@redhat.com>
16235
16236 PR rtl-optimization/57281
16237 PR rtl-optimization/57300
16238 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16239 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16240 what the other splitter did if the registers are dead.
16241
2a293391
RB
162422013-05-17 Richard Biener <rguenther@suse.de>
16243
16244 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16245 MEM_REF offsets.
16246
17042d2b
JJ
162472013-05-17 Jakub Jelinek <jakub@redhat.com>
16248
16249 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16250 linking.
16251
5b115c1f
MP
162522013-05-17 Marek Polacek <polacek@redhat.com>
16253
11c2aa39
UB
16254 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16255 length when doing non-zero store of storing '\0' to '\0'.
5b115c1f 16256
52d84413
JJ
162572013-05-17 Jakub Jelinek <jakub@redhat.com>
16258
68119618
JJ
16259 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16260 vect_external_def oprnd1 with loop_vinfo, try to emit
16261 optional cast, negation and and stmts on the loop preheader
16262 edge instead of into the pattern def seq.
16263
52d84413
JJ
16264 PR tree-optimization/57051
16265 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16266 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16267
17b962bd
NC
162682013-05-16 Nick Clifton <nickc@redhat.com>
16269
16270 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16271 (rl78_is_naked_func): New function.
56aefbf7
UB
16272 (rl78_expand_prologue): Skip prologue generation for naked functions.
16273 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
17b962bd
NC
16274 * doc/extend.texi (naked): Add RL78 to the list of processors
16275 that supports this attribute.
16276
b1a0f84e
JL
162772013-05-16 Jeff Law <law@redhat.com>
16278
16279 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16280
af0e415b
UB
162812013-05-16 Uros Bizjak <ubizjak@gmail.com>
16282
16283 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16284 cache parameters using detect_caches_amd also for CYRIX,
16285 NSC and TM2 signatures.
16286
19db293a
UB
162872013-05-16 Uros Bizjak <ubizjak@gmail.com>
16288 Dzianis Kahanovich <mahatma@eu.by>
16289
16290 PR target/45359
16291 PR target/46396
16292 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16293 VIA/Centaur processors and determine their cache parameters
16294 using detect_caches_amd.
16295
251a41b9
TJ
162962013-05-16 Teresa Johnson <tejohnson@google.com>
16297
16298 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16299 (rtl_verify_edges): New function.
16300 (rtl_verify_bb_insns): Ditto.
16301 (rtl_verify_bb_pointers): Ditto.
16302 (rtl_verify_bb_insn_chain): Ditto.
16303 (rtl_verify_fallthru): Ditto.
16304 (rtl_verify_bb_layout): Ditto.
16305 (rtl_verify_flow_info_1): Outline checks into new functions.
16306 (rtl_verify_flow_info): Ditto.
16307
f14540b6
SE
163082013-05-16 Steve Ellcey <sellcey@imgtec.com>
16309
16310 * cfghooks.c (copy_bbs): Add update_dominance argument.
16311 * cfghooks.h (copy_bbs): Update prototype.
16312 * tree-cfg.c (gimple_duplicate_sese_region):
16313 Add update_dominance argument.
16314 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16315 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16316 gimple_duplicate_sese_region call.
16317 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16318 Update copy_bbs call.
16319 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16320 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16321
7e9a3abb
JJ
163222013-05-16 Jakub Jelinek <jakub@redhat.com>
16323
16324 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16325 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16326 vect_recog_rotate_pattern.
16327 (vect_recog_rotate_pattern): New function.
16328
427b248d
JM
163292013-05-16 Jason Merrill <jason@redhat.com>
16330
16331 * Makefile.in (LLINKER): New variable.
16332 (mostlyclean): Remove link mutex.
16333 * configure.ac: Handle --enable-link-mutex.
16334 * lock-and-run.sh: New script.
16335
b871e3d2
RR
163362013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16337
16338 PR target/19599
16339 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16340 for NULL decl.
16341
ce858126
RO
163422013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16343
16344 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16345
798d3d04
GY
163462013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16347
16348 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16349 * config/arm/arm.c (next_consecutive_mem): New function.
16350 (gen_movmem_ldrd_strd): Likewise.
16351 * config/arm/arm.md (movmemqi): Update condition and code.
16352 (unaligned_loaddi, unaligned_storedi): New patterns.
16353
0baddc45
RO
163542013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16355
16356 * config.gcc: Obsolete *-*-solaris2.9*.
16357 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16358
aa9d5bcf
RB
163592013-05-16 Richard Biener <rguenther@suse.de>
16360
16361 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16362 earlier, after GRAPHITE transforms and IV canonicalization.
16363
5a0f4dd3
JJ
163642013-05-16 Jakub Jelinek <jakub@redhat.com>
16365
16366 * omp-low.c (extract_omp_for_data): For collapsed loops,
16367 if at least one of the loops is known at compile time to
16368 iterate zero times, set count to 0.
16369 (expand_omp_regimplify_p): New function.
16370 (expand_omp_for_generic): For collapsed loops, if at least
16371 one of the loops isn't known to iterate at least once,
16372 add runtime check with setting count to 0.
16373 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16374 For unsigned types if it isn't known at compile time that
16375 the loop will iterate at least once, add runtime check to bypass
16376 the whole loop if initial condition isn't true.
16377
e3753785
NS
163782013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16379
16380 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16381
43bb4dd1
MG
163822013-05-16 Marc Glisse <marc.glisse@inria.fr>
16383
16384 PR middle-end/57286
16385 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16386 transformations to avoid an infinite loop.
16387
3571dde6
MP
163882013-05-16 Marek Polacek <polacek@redhat.com>
16389
16390 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16391
b8b3f0ca
LE
163922013-05-15 Leif Ekblad <leif@rdos.net>
16393
16394 * config/i386/i386.c (ix86_decompose_address): Use
16395 DEFAULT_TLS_SEG_REG to access TLS segment register.
16396 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16397 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16398 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16399
e299a383
RS
164002013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16401
16402 PR target/57260
16403 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16404 sibling calls to functions that would normally be lazily bound,
16405 unless $gp is call-clobbered.
16406
e7208ea3
UB
164072013-05-15 Uros Bizjak <ubizjak@gmail.com>
16408
19db293a 16409 * config/i386/i386.c (ix86_option_override_internal): Update
e7208ea3 16410 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
5d280e2b
UB
16411 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16412 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
e7208ea3
UB
16413 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16414 of TARGET_3DNOW.
16415 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16416
e7413f3d
AS
164172013-05-15 Andreas Schwab <schwab@suse.de>
16418
16419 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16420 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16421 third operand.
16422
ad4db775
TJ
164232013-05-15 Teresa Johnson <tejohnson@google.com>
16424
16425 * loop-unroll.c (report_unroll_peel): Check decision before
16426 emitting unroll/peel message.
16427
af205f67
TJ
164282013-05-15 Teresa Johnson <tejohnson@google.com>
16429
16430 * function.h (has_bb_partition): New rtl_data flag.
16431 (bb_reorder_complete): Ditto.
16432 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16433 instead of flag_reorder_blocks_and_partition.
16434 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16435 with some enhancements.
16436 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16437 * bb-reorder.c (connect_traces): Check for has_bb_partition
16438 instead of flag_reorder_blocks_and_partition.
16439 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16440 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16441 verify_hot_cold_block_grouping.
16442 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16443
9adcfa3c
RR
164442013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16445
16446 PR target/19599
16447 * config/arm/predicates.md (call_insn_operand): New predicate.
16448 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16449 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16450 if insn is not a tail call.
16451 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16452 registers.
16453 * config/arm/arm.h (enum reg_class): New caller save register class.
16454 (REG_CLASS_NAMES): Likewise.
16455 (REG_CLASS_CONTENTS): Likewise.
16456 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16457 without decls.
16458
ccb3ad87
RB
164592013-05-15 Richard Biener <rguenther@suse.de>
16460
16461 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16462 of MSG_OPTIMIZED_LOCATIONS.
16463 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16464 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16465 message.
16466 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16467 of MSG_OPTIMIZED_LOCATIONS.
16468 (execute_vect_slp): Likewise.
16469 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16470 (vect_create_cond_for_alias_checks): Likewise.
16471 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16472 (vect_recog_widen_mult_pattern): Likewise.
16473 (vect_recog_widen_sum_pattern): Likewise.
16474 (vect_recog_over_widening_pattern): Likewise.
16475 (vect_recog_widen_shift_pattern): Likewise.
16476 (vect_recog_vector_vector_shift_pattern): Likewise.
16477 (vect_recog_divmod_pattern): Likewise.
16478 (vect_recog_mixed_size_cond_pattern): Likewise.
16479 (vect_recog_bool_pattern): Likewise.
16480 (vect_pattern_recog_1): Likewise.
16481
48b1474e
MJ
164822013-05-15 Martin Jambor <mjambor@suse.cz>
16483
16484 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16485 non-functions to builtin_unreachable.
16486 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16487 correct when new direct edges were discovered.
16488
9de04252
MJ
164892013-05-15 Martin Jambor <mjambor@suse.cz>
16490
16491 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16492 header, print symbol order instead of node uid, print more information
16493 about indirect edge targets.
16494 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16495 uids.
16496 (ipa_make_edge_direct_to_target): Likewise.
16497 (remove_described_reference): Likewise.
16498 (propagate_controlled_uses): Likewise.
16499 (ipa_print_node_params): Also print symbol order.
16500 (ipcp_transform_function): Print symbol order instead of node uids.
16501 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16502 (cgraph_get_create_real_symbol_node): Likewise.
16503 * ipa-cp.c (print_lattice): Likewise.
16504 (print_all_lattices): Likewise.
16505 (determine_versionability): Likewise.
16506 (initialize_node_lattices): Likewise.
16507 (estimate_local_effects): Likewise.
16508 (update_profiling_info): Likewise.
16509 (create_specialized_node): Likewise.
16510 (perhaps_add_new_callers): Likewise.
16511 (decide_about_value): Likewise.
16512 (decide_whether_version_node): Likewise.
16513 (identify_dead_nodes): Likewise.
16514 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16515 (dump_inline_summary): Likewise.
16516 (estimate_node_size_and_time): Likewise.
16517 (inline_analyze_function): Likewise.
16518 * ipa-inline.c (report_inline_failed_reason): Likewise.
16519 (want_early_inline_function_p): Likewise.
16520 (edge_badness): Likewise.
16521 (update_edge_key): Likewise.
16522 (inline_small_functions): Likewise. Add dumping of order to two other
16523 dumps.
16524 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16525 instead of node uids.
16526 (propagate_pure_const): Likewise.
16527 (propagate_pure_const): Likewise.
16528 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16529 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16530 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16531 of node uids.
16532 * tree-pretty-print.c (dump_function_header): Likewise.
16533 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16534 Print symbol order instead of node uids.
16535
1dd03b91
AK
165362013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16537
16538 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16539 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16540
fdf6a7b9
RB
165412013-05-15 Richard Biener <rguenther@suse.de>
16542
16543 PR tree-optimization/57275
16544 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16545 return value for fail to do runtime alias checks for gather loads.
16546
2d6e4603
JH
165472013-05-15 Jan Hubicka <jh@suse.cz>
16548
16549 PR lto/57038
16550 PR lto/47375
e7208ea3
UB
16551 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16552 weakrefs are not external.
16553 (lto_symtab_merge_decls): Fix thinko when dealing with
16554 non-lto_symtab decls.
2d6e4603
JH
16555 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16556 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16557 * varpool.c (dump_varpool_node): Dump more flags.
16558
83f44b39
GG
165592013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16560
16561 * config/i386/i386.c (processor_alias_table): Add instruction
16562 FSGSBASE for AMD bdver3 architecture.
16563
2e55d062
JJ
165642013-05-14 Jakub Jelinek <jakub@redhat.com>
16565
16566 * tree.c (warn_deprecated_use): Print file:line using locus color.
16567 * diagnostic.c (diagnostic_report_current_module): Print file:line
16568 and file:line:column using locus color.
16569
0bfdb81e
MS
165702013-05-14 Mike Stump <mikestump@comcast.net>
16571
16572 * gdbinit.in: Add __null.
16573
a508ef22
MS
165742013-05-14 Mike Stump <mikestump@comcast.net>
16575
16576 * recog.h: Rename struct recog_data to Recog_data.
16577 * recog.c: Likewise.
16578 * reload.c (can_reload_into): Likewise.
16579 * config/picochip/picochip.c: Likewise.
16580
e7180acb
MS
165812013-05-14 Mike Stump <mikestump@comcast.net>
16582
16583 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16584
61216c26
SB
165852013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16586
fd6beed4
SB
16587 * resource.h (struct resources): Remove unch_memory member.
16588 (CLEAR_RESOURCE): Don't clear unch_memory.
16589 * resource.c (mark_referenced_resources): Don't set it.
16590 (mark_set_resources): Likewise.
16591 (mark_target_live_regs): Don't clear it.
16592 (init_resource_info): Likewise.
16593 * reorg.c (resource_conflicts_p): Don't compare it.
16594 (redundant_insn): Don't set it.
16595
61216c26
SB
16596 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16597 Remove prototypes.
16598 * emit-rtl.c (next_label): Remove unused function.
16599 (skip_consecutive_labels, link_cc0_insns): Move to ...
16600 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16601 only place where these functions are used, and make them static.
16602
418d1b87
MG
166032013-05-14 Marc Glisse <marc.glisse@inria.fr>
16604
16605 * fold-const.c (fold_negate_expr): Handle vectors.
16606 (fold_truth_not_expr): Make it static.
16607 (fold_invert_truthvalue): New static function.
16608 (invert_truthvalue_loc): Handle vectors. Do not call
16609 fold_truth_not_expr directly.
16610 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16611 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16612 (fold_comparison): Handle vector constants.
16613 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16614 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16615 * tree.h (fold_truth_not_expr): Remove declaration.
16616
fc21784d
JG
166172013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16618
16619 * config/aarch64/aarch64-simd.md
16620 (aarch64_vcond_internal<mode>): Rename to...
16621 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16622 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16623 float modes. Clarify all iterator modes.
16624 (vcond<mode><mode>): Use new name for vcond expanders.
16625 (vcond<v_cmp_result><mode>): Likewise.
16626 (vcondu<mode><mode>: Likewise.
16627 * config/aarch64/iterators.md (VDQF_COND): New.
16628
d4c52634
MG
166292013-05-14 Marc Glisse <marc.glisse@inria.fr>
16630
16631 PR bootstrap/57266
16632 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16633 variable for the shift amount. Check that we shift by non-negative
16634 amounts.
16635
2b261262
CLT
166362013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16637
16638 PR target/42017
16639 * config/arm/arm.h (EPILOGUE_USES): Only return true
16640 for LR_REGNUM after epilogue_completed.
16641
12211b99 166422013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
6e022d7b
JR
16643
16644 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
e7208ea3 16645 is error_mark_node.
6e022d7b 16646
56cf7859
RO
166472013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16648
16649 PR target/57261
16650 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16651 and Solaris 11+/x86 with gld.
16652 * configure: Regenerate.
16653
75776c6d
JJ
166542013-05-14 Jakub Jelinek <jakub@redhat.com>
16655
16656 * expmed.c (expand_shift_1): Canonicalize rotates by
16657 constant bitsize / 2 to bitsize - 1.
7f998021 16658 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
75776c6d
JJ
16659 case ROTATERT>: Likewise.
16660
16661 Revert:
16662 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16663
16664 * config/i386/i386.md (rotateinv): New code attr.
16665 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16666 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16667 roll $31, %eax, etc.
16668
df35498a
RB
166692013-05-14 Richard Biener <rguenther@suse.de>
16670
16671 PR middle-end/57235
16672 * tree-eh.c (sink_clobbers): Give up for successors with
16673 multiple predecessors and no virtual uses.
16674
cc6e7ece
EB
166752013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16676
16677 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16678 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16679
f778dd4d
JJ
166802013-05-14 Jakub Jelinek <jakub@redhat.com>
16681
16682 PR middle-end/57251
16683 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16684 the case when both op0 and op1 have VOIDmode.
16685
bad4df9b
KP
166862013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16687
16688 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16689 in multiply-accumulate mode.
16690
56f3e9ac
GW
166912013-05-13 Guozhi Wei <carrot@google.com>
16692
16693 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16694
a3d7ab92
KT
166952013-05-13 Kai Tietz <ktietz@redhat.com>
16696
16697 PR target/56975
16698 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
e7208ea3 16699 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
a3d7ab92 16700 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
e7208ea3 16701 * config/i386/i386.c (ix86_option_override_internal): Likewise.
a3d7ab92
KT
16702 (ix86_expand_prologue): Likewise.
16703 (ix86_expand_split_stack_prologue): Likewise.
16704 (legitimate_pic_address_disp_p): Likewise.
16705 (legitimize_pic_address): Likewise.
16706 (legitimize_tls_address): Likewise.
16707 (legitimize_pe_coff_symbol): Likewise.
16708 (output_pic_addr_const): Likewise.
16709 (construct_plt_address): Likewise.
16710 (ix86_expand_call): Likewise.
16711 (x86_output_mi_thunk): Likewise.
16712 (x86_function_profiler): Likewise.
16713
c59b7e28
SN
167142013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16715
16716 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16717 similar switch cases.
16718 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16719 (aarch64_simd_mov_to_<mode>low): Delete.
16720 (aarch64_simd_mov_to_<mode>high): Delete.
16721 (move_lo_quad_<mode>): Add w<-r alternative.
16722 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16723 (aarch64_simd_mov_from_*): Update type attribute.
16724 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16725 statement.
16726
5f28524a
JH
167272013-05-13 Jan Hubicka <jh@suse.cz>
16728
16729 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16730 * config/i386/i386.c (ix86_compute_frame_layout,
e7208ea3
UB
16731 ix86_expand_epilogue, emit_i387_cw_initialization,
16732 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16733 ix86_local_alignment): Fix use of size/speed predicates.
5f28524a 16734
ae6fa899
JJ
167352013-05-13 Jakub Jelinek <jakub@redhat.com>
16736
16737 PR tree-optimization/45216
16738 PR tree-optimization/57157
16739 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16740 the (-Y) & (B - 1) variant if OP is |.
16741 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16742 return shifted. Use (-op1) & (prec - 1) as other_amount
16743 instead of prec - op1.
16744
4502fe8d
MJ
167452013-05-13 Martin Jambor <mjambor@suse.cz>
16746
16747 PR middle-end/42371
16748 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
16749 (ipa_constant_data): New type.
16750 (ipa_jump_func): Use ipa_constant_data to hold information about
16751 constant jump functions.
16752 (ipa_get_jf_constant): Adjust to jump function type changes.
16753 (ipa_get_jf_constant_rdesc): New function.
16754 (ipa_param_descriptor): New field controlled_uses.
16755 (ipa_get_controlled_uses): New function.
16756 (ipa_set_controlled_uses): Likewise.
16757 * ipa-ref.h (ipa_find_reference): Declare.
16758 * ipa-prop.c (ipa_cst_ref_desc): New type.
16759 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
16760 changes.
16761 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
16762 New parameter cs. Adjust all callers.
16763 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
16764 (remove_described_reference): New function.
16765 (jfunc_rdesc_usable): Likewise.
16766 (try_make_edge_direct_simple_call): Decrement controlled use count,
16767 attempt to remove reference if it hits zero.
16768 (combine_controlled_uses_counters): New function.
16769 (propagate_controlled_uses): Likewise.
16770 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
16771 (ipa_edge_duplication_hook): Duplicate reference descriptions.
16772 (ipa_print_node_params): Print described use counter.
16773 (ipa_write_jump_function): Adjust to jump function type changes.
16774 (ipa_read_jump_function): New parameter CS, pass it to
16775 ipa_set_jf_constant. Adjust caller.
16776 (ipa_write_node_info): Stream controlled use count
16777 (ipa_read_node_info): Likewise.
16778 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
16779 asserting.
16780 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
16781 count. Remove cloning-added reference if it reaches zero.
16782 * ipa-ref.c (ipa_find_reference): New function.
16783
0864bfc2
GG
167842013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16785
e7208ea3 16786 * config/i386/i386.c (processor_target_table): Modified default
0864bfc2
GG
16787 alignment values for AMD BD and BT architectures.
16788
640bfeb2
MG
167892013-05-13 Marc Glisse <marc.glisse@inria.fr>
16790
16791 * tree-vect-generic.c (uniform_vector_p): Move ...
16792 * tree.c (uniform_vector_p): ... here.
16793 * tree.h (uniform_vector_p): Declare it.
16794 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
16795 into a scalar.
16796
3a60f32b
JJ
167972013-05-13 Jakub Jelinek <jakub@redhat.com>
16798
198fe1bf
JJ
16799 PR tree-optimization/57230
16800 * tree-ssa-strlen.c (handle_char_store): Record length for
16801 array store from STRING_CST.
16802
3a60f32b
JJ
16803 PR tree-optimization/57230
16804 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
16805 check.
16806
566be57c
JR
168072013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
16808
16809 * config/epiphany/epiphany.c (epiphany_init): Check size of
16810 NUM_MODES_FOR_MODE_SWITCHING.
16811 (epiphany_expand_prologue):
16812 Remove CONFIG_REGNUM initial value handling code.
16813 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
16814 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
2710a27a 16815 (emit_set_fp_mode, epiphany_mode_after): Likewise.
566be57c
JR
16816 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
16817 Don't return 1 for FP_MODE_NONE.
16818 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
16819 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
16820 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
16821 * config/epiphany/epiphany.md (save_config): New pattern.
16822
0f2c2331
UB
168232013-05-12 Uros Bizjak <ubizjak@gmail.com>
16824
16825 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
16826
5b3f93c7
UB
168272013-05-10 Uros Bizjak <ubizjak@gmail.com>
16828
16829 * config/i386/i386.md (memory): Handle sseishft1.
16830 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
16831 (*vec_extractv2di_1): Ditto.
16832
1f873f0e
VM
168332013-05-10 Vladimir Makarov <vmakarov@redhat.com>
16834
16835 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
16836 saved registers.
16837
341427fa 168382013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
d7de6455
SH
16839
16840 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
16841 Add mthumb/march=armv7-a multilib.
16842 Add mthumb/march=armv7-r multilib.
16843 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
16844
9e69bdde
RC
168452013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
16846
16847 * config/v850/t-rtems: Add more multilibs.
16848
9ff09a22
RB
168492013-05-10 Richard Biener <rguenther@suse.de>
16850
16851 PR tree-optimization/57214
16852 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
16853 not propagate from SSA names that occur in abnormal PHI nodes.
16854
a5e0cd1d
MG
168552013-05-10 Marc Glisse <marc.glisse@inria.fr>
16856
16857 * stor-layout.c (element_precision): New function.
16858 * machmode.h (element_precision): Declare it.
16859 * tree.c (build_minus_one_cst): New function.
16860 (element_precision): Likewise.
16861 * tree.h (build_minus_one_cst): Declare new function.
16862 (element_precision): Likewise.
16863 * fold-const.c (operand_equal_p): Use element_precision.
16864 (fold_binary_loc): Handle vector types.
16865 * convert.c (convert_to_integer): Use element_precision.
16866 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
16867 separately.
16868
cb2558bc
RS
168692013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
16870
16871 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
16872 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
16873 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
16874 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16875 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
16876 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
16877 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
16878 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16879 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
16880 (Uuw8): New constraints.
16881 (Usb4): Move into alphabetical order.
16882 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
16883 (sd8_operand, ub8_operand, uw8_operand): New predicates.
16884 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
16885 previously unnamed patterns.
16886 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
16887 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
16888 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
16889 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
16890 of set_attr_alternative/if_then_else. Use extended_mips16 instead
16891 of specific lengths.
16892
cb3b8d33
JJ
168932013-05-10 Jakub Jelinek <jakub@redhat.com>
16894
6f93c008
JJ
16895 * config/i386/i386.md (rotateinv): New code attr.
16896 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16897 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16898 roll $31, %eax, etc.
16899
cb3b8d33
JJ
16900 PR tree-optimization/45216
16901 PR tree-optimization/57157
16902 * tree-ssa-forwprop.c (simplify_rotate): New function.
16903 (ssa_forward_propagate_and_combine): Call it.
16904
afb119be
RB
169052013-05-10 Richard Biener <rguenther@suse.de>
16906
16907 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
16908 disable peeling when we version for aliasing.
16909 (vector_alignment_reachable_p): Honor explicit user alignment.
16910 (vect_supportable_dr_alignment): Likewise.
16911 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
16912 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
16913 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
16914 then peeling to arrange for the cost-model check to come first.
16915
01ae4861
AM
169162013-05-10 Alan Modra <amodra@gmail.com>
16917
16918 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
16919 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
16920 * configure: Regenerate.
16921
ebc9a431
AM
169222013-05-10 Alan Modra <amodra@gmail.com>
16923
16924 PR target/55033
16925 * varasm.c (default_elf_select_section): Move !DECL_P check..
16926 (get_named_section): ..to here before calling get_section_name.
16927 Adjust assertion.
16928 (default_section_type_flags): Add DECL_P check.
16929 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
16930 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
16931
d4bca93c
JR
169322013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
16933
16934 * config/epiphany/epiphany.c (epiphany_expand_prologue):
16935 When using gen_stack_adjust_str with a register offset, add a
16936 REG_FRAME_RELATED_EXPR note.
16937
60ca9a65
UB
169382013-05-09 Uros Bizjak <ubizjak@gmail.com>
16939
16940 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
16941 (*vec_extractv4si_zext_mem): Ditto.
16942 (*vec_extractv2di): Add 0->x and x->x alternatives.
16943 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
16944 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
16945
a3409c02
JM
169462013-05-09 Jason Merrill <jason@redhat.com>
16947
0138d6b2
JM
16948 N3639 C++1y VLA support
16949 * gimplify.c (gimplify_vla_decl): Don't touch an existing
16950 DECL_VALUE_EXPR.
16951
a3409c02
JM
16952 * tree.c (build_constructor_va): New.
16953 * tree.h: Declare it.
16954
66e6b990
MJ
169552013-05-09 Martin Jambor <mjambor@suse.cz>
16956
16957 PR lto/57084
16958 * gimple-fold.c (canonicalize_constructor_val): Call
16959 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
16960
64cfa6c0
JH
169612013-05-09 Jan Hubicka <jh@suse.cz>
16962 Richard Biener <rguenther@suse.de>
16963
16964 PR lto/54095
16965 * symtab.c (symtab_make_decl_local): Do not add private names.
16966
c3167b00
JH
169672013-05-09 Jan Hubicka <jh@suse.cz>
16968
16969 PR lto/54095
16970 * symtab.c (insert_to_assembler_name_hash): Handle clones.
16971 (unlink_from_assembler_name_hash): Likewise.
16972 (symtab_prevail_in_asm_name_hash, symtab_register_node,
16973 symtab_unregister_node, symtab_initialize_asm_name_hash,
16974 change_decl_assembler_name): Update.
16975
12dc6974
SN
169762013-05-09 Sofiane Naci <sofiane.naci@arm.com>
16977
16978 * config/aarch64/aarch64.md: New movtf split.
16979 (*movtf_aarch64): Update.
16980 (aarch64_movdi_tilow): Handle TF modes and rename to
16981 aarch64_movdi_<mode>low.
16982 (aarch64_movdi_tihigh): Handle TF modes and rename to
16983 aarch64_movdi_<mode>high
16984 (aarch64_movtihigh_di): Handle TF modes and rename to
16985 aarch64_mov<mode>high_di
16986 (aarch64_movtilow_di): Handle TF modes and rename to
16987 aarch64_mov<mode>low_di
16988 (aarch64_movtilow_tilow): Remove spurious whitespace.
16989 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
16990 splits.
16991 (aarch64_print_operand): Update.
16992
227eb343
AM
169932013-05-09 Alan Modra <amodra@gmail.com>
16994
16995 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
16996 powerpc64le.
16997 * configure: Regenerate.
16998
0b013847
UB
169992013-05-08 Uros Bizjak <ubizjak@gmail.com>
17000
17001 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
17002 splitter preparation statements.
17003 * config/i386/sse.md (*vec_extract* splitters): Ditto.
17004 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
17005 adjust_address_nv.
17006
1dc3d6e9
BS
170072013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17008
17009 * gimple-ssa-strength-reduction.c (count_candidates): Change
17010 return value to int.
17011 (analyze_candidates_and_replace): Change type of length to int.
17012
e61e7d28
UB
170132013-05-08 Uros Bizjak <ubizjak@gmail.com>
17014
17015 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
17016 (*vec_extract<mode>): Use VI12_128 mode iterator.
17017 (*vec_extract<mode>_mem): Ditto.
17018 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
17019 attribute.
17020
4fbfcf44
DN
170212013-05-08 Diego Novillo <dnovillo@google.com>
17022
17023 PR bootstrap/54659
17024
17025 Revert:
4fbfcf44
DN
17026 2012-08-17 Diego Novillo <dnovillo@google.com>
17027
e61e7d28
UB
17028 PR bootstrap/54281
17029 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
17030 * config.in: Regenerate.
17031 * configure: Regenerate.
17032 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
4fbfcf44 17033
702d8703
JH
170342013-05-08 Jan Hubicka <jh@suse.cz>
17035
17036 PR lto/54095
17037 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17038 * cgraph.h (symtab_node_base): Add unique_name.
17039 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17040 input_overwrite_node, input_varpool_node): Stream unique_name.
17041 * cgraphclones.c (cgraph_create_virtual_clone,
17042 cgraph_function_versioning): Set unique_name.
17043 * ipa.c (function_and_variable_visibility): Set unique_name.
17044
8b28cf47
BS
170452013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17046
17047 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17048 (alloc_cand_and_find_basis): Restrict conditional candidate
17049 processing to CAND_MULTs.
17050
e86074fd
JH
170512013-05-08 Jan Hubicka <jh@suse.cz>
17052
17053 PR lto/54095
17054 lto-symtab.c (lto_symtab_symbol_p): New function.
17055 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17056 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17057 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17058 Skip static symbols.
17059
44398cbe
PC
170602013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17061
17062 PR tree-optimization/57200
17063 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17064 Only call inform if the preceding warning_at returns true.
17065
f6bc1c4a
HS
170662013-05-07 Han Shen <shenhan@google.com>
17067
17068 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17069 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17070 * common.opt (fstack-protector-strong): New option.
17071 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17072 * doc/invoke.texi (Optimization Options): Document
17073 "-fstack-protector-strong".
17074 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17075
4ffecb1f
SB
170762013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17077
17078 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17079
5a107a0f
VM
170802013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17081
17082 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17083 it is not lhs of the insn.
17084 (setup_reg_equiv): Remove insn having equiv memory which it is not
17085 lhs of the insn.
17086 * lra-constraints.c (process_address): Try to improve generation
17087 code for address base + disp.
17088 (lra_constraints): Make correct the code for checking insn setting
17089 up backward equivalence. Remove insn only if it is in the init
17090 insn list.
17091 * lra-eliminations.c (update_reg_eliminate): Change return value.
17092 (lra_eliminate): Use the result.
17093
3f5783ea
UB
170942013-05-07 Uros Bizjak <ubizjak@gmail.com>
17095
17096 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17097 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17098 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17099 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17100 register target operands.
17101 (*vec_extractv8hi_sse2): New pattern.
17102 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17103 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17104 (*vec_extract<mode>_mem): New insn and split pattern.
17105
8a5800b8
CL
171062013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17107
17108 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17109 (TARGET_ASAN_SHADOW_OFFSET): Define.
17110 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17111 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17112
7bf55a70
BS
171132013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17114
17115 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17116 (incr_vec_index): Return -1 if increment not found.
17117 (create_add_on_incoming_edge): Assert if increment not found.
17118 (record_increment): Limit number of increments recorded.
17119 (all_phi_incrs_profitable): Return false if an increment not found.
17120 (replace_profitable_candidates): Don't process increments that were
17121 not recorded.
17122 (analyze_candidates_and_replace): Limit size of incr_vec.
17123
3f8825c0
RB
171242013-05-07 Richard Biener <rguenther@suse.de>
17125
17126 * calls.c (special_function_p): setjmp-like functions are leaf.
17127 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17128 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17129
fd4842cd
SN
171302013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17131
17132 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17133 (aarch64_simd_mov<mode>): New expander.
17134 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17135 (aarch64_simd_mov_to_<mode>high): Likewise.
17136 (aarch64_simd_mov_from_<mode>low): Likewise.
17137 (aarch64_simd_mov_from_<mode>high): Likewise.
17138 (aarch64_dup_lane<mode>): Update.
17139 (aarch64_dup_lanedi): New instruction pattern.
17140 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17141 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17142
a7a7d10e
BS
171432013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17144
17145 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17146 (replace_mult_candidate): Remove unnecessary argument; remove
17147 unnecessary parameter from call to introduce_cast_before_cand.
17148 (replace_unconditional_candidate): Remove unnecessary parameter
17149 from call to replace_mult_candidate.
17150 (replace_conditional_candidate): Likewise.
17151 (insert_initializers): Use make_temp_ssa_name.
17152 (introduce_cast_before_cand): Remove unnecessary argument; use
17153 make_temp_ssa_name.
17154 (replace_one_candidate): Remove unnecessary argument; remove
17155 unnecessary parameter from calls to introduce_cast_before_cand.
17156 (replace_profitable_candidates): Remove unnecessary parameters
17157 from calls to replace_one_candidate.
17158
29105868
BS
171592013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17160
17161 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17162 phi def as possibly hiding a basis for a CAND_ADD whose operands
17163 have been commuted in the analysis.
17164 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17165
4095f9fa
N
171662013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17167
17168 * config/aarch64/aarch64.md
17169 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17170 shift value between 0-4.
17171
39e843e8
RB
171722013-05-07 Richard Biener <rguenther@suse.de>
17173
17174 * double-int.h (rshift): New overload.
17175 * double-int.c (rshift): New function.
17176 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17177 (create_reference_ops_from_ref): Remove.
17178 (vn_reference_insert): Use shared ops for constructing the
17179 reference and copy it.
17180
0a1a83cb
RB
171812013-05-07 Richard Biener <rguenther@suse.de>
17182
17183 PR middle-end/57190
17184 * tree-eh.c (sink_clobbers): Properly propagate
17185 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17186
43303d6f
JJ
171872013-05-07 Jakub Jelinek <jakub@redhat.com>
17188
ba7e83f8
JJ
17189 PR tree-optimization/57149
17190 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17191 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17192 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17193 uninit_undefined_value_p instead of ssa_undefined_value_p.
17194
43303d6f
JJ
17195 PR debug/57184
17196 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17197 for modifier == EXPAND_INITIALIZER.
17198
14523c25
AB
171992013-05-07 Anton Blanchard <anton@samba.org>
17200
17201 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17202 for powerpc64 little endian.
17203 * configure: Regenerate.
17204
cb7c8be9
GS
172052013-05-06 Graham Stott <grahams@btinternet.com>
17206
17207 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17208 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17209 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17210 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17211
172122013-05-06 Graham Stott <grahams@btinternet.com>
a8fd4225
GS
17213
17214 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17215 codes which allow non-lvalues.
17216
9a0ee7b0
MG
172172013-05-06 Marc Glisse <marc.glisse@inria.fr>
17218
17219 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17220 components are all 1s.
17221 (integer_minus_onep): New function.
17222 * tree.h (integer_minus_onep): Declare it.
17223 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17224 integer_minus_onep instead of integer_all_onesp.
17225
f2c17ea9
OE
172262013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17227
17228 PR target/52933
17229 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17230 variations of these patterns.
17231
f75e6a51
UB
172322013-05-06 Uros Bizjak <ubizjak@gmail.com>
17233
17234 * config/i386/i386.md (isa): Add x64_sse4 member.
17235 (enabled): Handle x64_sse4.
17236 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17237 instruction for 64bit SSE4_1 targets. Update insn attributes.
17238 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17239 instruction for SSE4_1 targets. Update insn attributes.
17240 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17241 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17242 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17243 const_1 selector.
17244 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17245 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17246 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17247
a986d468
OE
172482013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17249
17250 PR target/57108
17251 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17252
382522cb
MK
172532013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17254
17255 * final.c (do_assembler_dialects): Don't handle curly braces and
17256 vertical bar escaped by % as dialect delimiters.
17257 (output_asm_insn): Print curly braces and vertical bar if escaped
17258 by % and ASSEMBLER_DIALECT defined.
17259 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17260 * doc/tm.texi: Regenerated.
17261
fb0d5c60
SB
172622013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17263
fb0d5c60
SB
17264 * config/mips/mips.c: Include tree-pass.h.
17265 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17266 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17267 (pass_mips_machine_reorg2): New machine specific pass.
17268 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17269 (mips_option_override): Register the new pass.
17270 * rtl.h (cleanup_barriers): Remove prototype.
17271 (dbr_schedule): Likewise.
17272 * jump.c (cleanup_barriers): Make static.
17273 * reorg.c (dbr_schedule): Likewise.
17274
aa06a978
RB
172752013-05-06 Richard Biener <rguenther@suse.de>
17276
17277 PR tree-optimization/57185
17278 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17279 SSA names properly.
17280
0b953bec
UB
172812013-05-06 Uros Bizjak <ubizjak@gmail.com>
17282
17283 PR target/57106
17284 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17285 between "first_arg" and "insn", not "last" and "insn".
17286
28708525
WS
172872013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17288
17289 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17290 (find_candidates_in_block): Re-enable slsr_process_phi.
17291 (create_phi_basis): Fix double counting of candidate adjustment.
17292
0107dca2
RB
172932013-05-06 Richard Biener <rguenther@suse.de>
17294
17295 PR middle-end/57147
17296 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17297 the edge is also fallthru, preserve it and just clear the
17298 abnormal flag.
17299 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17300 also complex, preserve that and just clear the fallthru flag.
17301 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17302 update virtual operands.
17303
470d4d13
AM
173042013-05-06 Alan Modra <amodra@gmail.com>
17305
17306 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17307 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17308 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17309 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17310 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17311 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17312
8f1dbf8d
AM
173132013-05-06 Alan Modra <amodra@gmail.com>
17314
17315 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17316 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17317 (DEFAULT_ASM_ENDIAN): Define.
17318 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17319 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17320 Update -K PIC clause from sysv4.h.
17321 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17322 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17323
54890767
AM
173242013-05-06 Alan Modra <amodra@gmail.com>
17325
17326 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17327 twice for little-endian.
17328 (ashrdi3_no_power, ashrdi3): Support little-endian.
17329
2353515d
OE
173302013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17331
17332 PR target/55303
17333 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17334 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17335 related expanders.
17336 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17337 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17338 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17339 New predicates.
17340
33e67557
SB
173412013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17342 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17343
17344 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17345 * config/pa/pa.opt: Make mbig-switch a no-op.
17346 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17347 (CASE_VECTOR_MODE): Always return SImode.
17348 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17349 for the !TARGET_BIG_SWITCH case.
17350 * config/pa/pa-linux.h: Likewise.
17351 * config/pa/pa-openbsd.h: Likewise.
17352 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17353 * config/pa/pa.md (short_jump): Remove define_insn.
17354 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17355 (casesi0): Remove define_insn.
17356 (type): Remove btable_branch.
17357 (pa_combine_type): Likewise.
17358 (in_nullified_branch_delay): Likewise.
17359 (in_call_delay): Likewise.
17360 (define_delay): Likewise.
17361 (define_insn_reservation "Z3"): Likewise.
17362 (define_insn_reservation "Z4"): Likewise.
17363 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17364 (pa_adjust_insn_length): Remove adjustment for btable branches.
17365 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17366 and mno-big-switch
17367
3095685e
UB
173682013-05-05 Uros Bizjak <ubizjak@gmail.com>
17369
17370 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17371 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17372 Add m->r,x alternatives.
17373 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17374 splitters using SWI48x mode iterator.
17375 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17376 TARGET_64BIT. Add m->x alternative.
17377 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17378 Add o->x alternative. Enable for TARGET_SSE.
17379 (sse_storeq): Remove expander.
17380 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17381 with memory input operand.
17382 (*vec_extractv2di_1 splitter): New.
17383 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17384 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17385
4b36ae28
SB
173862013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17387
17388 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17389 (INT_LOWPART): Delete.
17390 (extract_MB): Adjust.
17391 (extract_ME): Adjust.
17392 (print_operand): Adjust.
17393
da226db2
SB
173942013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17395
17396 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17397 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17398 (reg_or_logical_cint_operand, easy_fp_constant,
17399 logical_const_operand): Delete "CONST_DOUBLE" case.
17400 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17401 "HOST_BITS_PER_WIDE_INT == 64" test.
3095685e 17402 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
da226db2
SB
17403 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17404 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17405 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17406 test.
17407 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17408 CONST_DOUBLE DImode/VOIDmode case.
17409 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17410 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17411 CONST_DOUBLE VOIDmode case.
17412 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17413 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17414 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17415 Delete CONST_DOUBLE case.
17416 (splitters for mov FMOVE64 const_double): Delete
17417 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17418 "HOST_BITS_PER_WIDE_INT >= 64" test.
17419 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17420 case.
17421 (mov DI const_double): Delete.
17422
40de22d6
JJ
174232013-05-04 Jakub Jelinek <jakub@redhat.com>
17424
17425 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17426 on op shows all bits zero in mode of a lowpart subreg, return zero.
17427
5ec6aff2
MM
174282013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17429
17430 PR target/57150
17431 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17432 to save TFmode registers and DImode to save TImode registers for
17433 caller save operations.
17434 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17435 mark being partially clobbered since they only use the first
17436 double word.
17437
17438 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17439 and TDmode only use the upper 64-bits of each VSX register.
17440
2cefad90
BS
174412013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17442
17443 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17444 (find_candidates_in_block): Disable slsr_process_phi.
17445
d6d7eee1
GW
174462013-05-03 Guozhi Wei <carrot@google.com>
17447
17448 * coverage.c (coverage_obj_init): Move the construction of gcov
3095685e
UB
17449 constructor to ...
17450 (build_init_ctor): ... here.
d6d7eee1 17451
9b92d12b
BS
174522013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17453
17454 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17455 (slsr_cand_d): Redefine def_phi.
17456 (stride_status, phi_adjust_status, count_phis_status): New enums.
17457 (find_phi_def): New.
17458 (find_basis_for_base_expr): New.
17459 (find_basis_for_candidate): Handle hidden bases.
17460 (alloc_cand_and_find_basis): Handle phi candidates.
17461 (slsr_process_phi): New.
17462 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17463 (create_mul_imm_cand): Likewise.
17464 (create_add_ssa_cand): Exclude phi base candidates.
17465 (create_add_imm_cand): Likewise.
17466 (slsr_process_cast): Likewise.
17467 (slsr_process_copy): Likewise.
17468 (find_candidates_in_block): Handle phi candidates.
17469 (dump_candidate): Likewise.
17470 (unconditional_cands): Delete.
17471 (unconditional_cands_with_known_stride_p): Delete.
17472 (phi_dependent_cand_p): New.
17473 (cand_increment): Handle phi-dependent candidates.
17474 (replace_dependent): Delete.
17475 (replace_mult_candidate): New.
17476 (replace_unconditional_candidate): New.
17477 (incr_vec_index): Move to avoid forward reference.
17478 (create_add_on_incoming_edge): New.
17479 (create_phi_basis): New.
17480 (replace_dependents): Delete.
17481 (replace_conditional_candidate): New.
17482 (phi_add_costs): New.
17483 (replace_uncond_cands_and_profitable_phis): New.
17484 (record_increment): Handle phi adjustments.
17485 (record_phi_increments): New.
17486 (record_increments): Handle phi adjustments.
17487 (phi_incr_cost): New.
17488 (lowest_cost_path): Handle phis.
17489 (total_savings): Likewise.
17490 (analyze_increments): Likewise.
17491 (ncd_with_phi): New.
17492 (ncd_of_cand_and_phis): New.
17493 (nearest_common_dominator_for_cands): Handle phi increments.
17494 (all_phi_incrs_profitable): New.
17495 (replace_profitable_candidates): Handle phi-dependent candidates.
17496 (analyze_candidates_and_replace): Likewise.
17497
68f073d4
TJ
174982013-05-03 Teresa Johnson <tejohnson@google.com>
17499
17500 PR bootstrap/57154
17501 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17502 do not exceed REG_BR_PROB_BASE.
17503
a4ee7cb9
JL
175042013-05-03 Jeff Law <law@redhat.com>
17505
ade67f70 17506 PR tree-optimization/57144
a4ee7cb9
JL
17507 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17508 operand of the condition will bit into the new type when eliminating
17509 a cast feeding a condition.
17510
47954c4e
JJ
175112013-05-03 Jakub Jelinek <jakub@redhat.com>
17512
17513 PR rtl-optimization/57130
3095685e
UB
17514 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17515 of COMPARE as in_code to the recursive call if needed.
47954c4e 17516
3c21604f
UB
175172013-05-03 Uros Bizjak <ubizjak@gmail.com>
17518
17519 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17520 (enabled): Handle new members.
17521 * config/i386/sse.md (*vec_concatv2si): Merge from
17522 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17523 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17524
12211b99 175252013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
95c03b36
JR
17526
17527 PR tree-optimization/57027
17528 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17529 for fnms opportunity, check we got the prerequisite kind
17530 of tree / gimple before using accessor functions.
17531
07bfc9ec
RB
175322013-05-03 Richard Biener <rguenther@suse.de>
17533
17534 * double-int.h (lshift): New overload without precision
17535 and arith argument.
17536 (operator *=, operator +=, operator -=): Move ...
17537 * double-int.c (operator *=, operator +=, operator -=): ... here
17538 and implement more efficiently.
17539 (mul_double_with_sign): Remove.
17540 (lshift_double): Adjust to take unsinged shift argument, push
17541 dispatching code to callers.
17542 (mul_double_wide_with_sign): Add early out for callers that
17543 are not interested in high parts or overflow.
17544 (lshift): New function.
17545 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17546 dispatch code here.
17547 (lrotate, rrotate): Use logical shifts.
17548 * expr.c (get_inner_reference): Use lshift.
17549 * fixed-value.c (do_fixed_divide): Likewise.
17550 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17551 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17552 (indirect_refs_may_alias_p): Likewise.
17553 (stmt_kills_ref_p_1): Likewise.
17554
7769bb64
VP
175552013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17556
17557 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17558
1d0c8e5c
VP
175592013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17560
17561 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17562 scalar form of FABD instruction.
17563
f15feaf9
VM
175642013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17565
17566 * lra-constraints.c (process_alt_operands): Add checking alt
17567 number to choose the best alternative.
17568
d90e76d4
RB
175692013-05-02 Richard Biener <rguenther@suse.de>
17570
17571 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17572 bitmap and its handling.
17573 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17574
9f8e7a96
RB
175752013-05-02 Richard Biener <rguenther@suse.de>
17576
17577 PR middle-end/57140
17578 * tree-inline.c (copy_loops): Properly handle removed loops.
17579 (copy_cfg_body): Mark destination loops for fixup if source
17580 loops needed fixup.
17581
f3a81b39
GY
175822013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17583
17584 PR target/56732
17585 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17586 generating simple_return for naked functions.
17587
7b920a9a
MJ
175882013-05-02 Martin Jambor <mjambor@suse.cz>
17589
17590 PR middle-end/56988
17591 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17592 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17593 flags match.
17594 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17595 ipa_agg_replacement_value structures.
17596 (known_aggs_to_agg_replacement_list): Likewise.
17597 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17598 (read_agg_replacement_chain): Likewise.
17599 (ipcp_transform_function): Also check that by_ref flags match.
17600
2c41c19d
RB
176012013-05-02 Richard Biener <rguenther@suse.de>
17602
17603 * graphds.h (struct graph): Add obstack member.
17604 * graphds.c (new_graph): Initialize obstack and allocate
17605 vertices from it.
17606 (add_edge): Allocate edge from the obstack.
3c21604f 17607 (free_graph): Free the obstack instead of all edges and vertices.
2c41c19d 17608
8b47039c
TJ
176092013-05-02 Teresa Johnson <tejohnson@google.com>
17610
17611 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17612 divides.
17613 * cfg.c (update_bb_profile_for_threading): Ditto.
17614 * tree-inline.c (copy_bb): Ditto.
17615 (copy_edges_for_bb): Ditto.
17616 (initialize_cfun): Ditto.
17617 (copy_cfg_body): Ditto.
17618 (expand_call_inline): Ditto.
17619 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17620 (estimate_node_size_and_time): Ditto.
17621 (inline_merge_summary): Ditto.
17622 * cgraphclones.c (cgraph_clone_edge): Ditto.
17623 (cgraph_clone_node): Ditto.
17624 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17625 (compute_trg_info): Ditto.
17626
da65928c
IB
176272013-05-02 Ian Bolton <ian.bolton@arm.com>
17628
17629 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17630 S reg when fp attribute set.
17631 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17632
add3c965
IB
176332013-05-02 Ian Bolton <ian.bolton@arm.com>
17634
17635 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17636 New pattern.
17637 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17638 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17639 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17640
fdd43ac4
RB
176412013-05-02 Richard Biener <rguenther@suse.de>
17642
17643 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17644 (struct instantiate_cache_entry): New type.
17645 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17646 (struct instantiate_cache_type): New type.
17647 (set_instantiated_value, get_instantiated_value): Remove.
17648 (get_instantiated_value_entry): New function.
17649 (instantiate_scev_name): Use the new cache and adjust.
17650 (instantiate_scev_poly): Adjust.
17651 (instantiate_scev_binary): Likewise.
17652 (instantiate_array_ref): Likewise.
17653 (instantiate_scev_convert): Likewise.
17654 (instantiate_scev_not): Likewise.
17655 (instantiate_scev_3): Likewise.
17656 (instantiate_scev_2): Likewise.
17657 (instantiate_scev_r): Likewise.
17658 (instantiate_scev): Likewise.
17659 (resolve_mixers): Likewise.
17660
36ff9dfb
VM
176612013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17662
17663 PR target/57091
17664 * lra-constraints.c (best_small_class_operands_num): Remove.
17665 (process_alt_operands): Remove small_class_operands_num. Take
17666 small classes operands into losers and only if the operand is not
17667 matched. Modify debugging output.
17668 (curr_insn_transform): Remove best_small_class_operands_num.
17669 Print insn name.
17670
36054fab
JG
176712013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17672
17673 * config/aarch64/aarch64-builtins.c
17674 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17675 * config/aarch64/aarch64-simd-builtins.def
17676 (reduc_splus_): Add new modes.
17677 (reduc_uplus_): New.
17678 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17679 (reduc_uplus_v4sf): Likewise.
17680 (reduc_splus_v4sf): Likewise.
17681 (aarch64_addv<mode>): Likewise.
17682 (reduc_uplus_<mode>): Likewise.
17683 (reduc_splus_<mode>): Likewise.
17684 (aarch64_addvv2di): Likewise.
17685 (reduc_uplus_v2di): Likewise.
17686 (reduc_splus_v2di): Likewise.
17687 (aarch64_addvv2si): Likewise.
17688 (reduc_uplus_v2si): Likewise.
17689 (reduc_splus_v2si): Likewise.
17690 (reduc_<sur>plus_<mode>): New.
17691 (reduc_<sur>plus_v2di): Likewise.
17692 (reduc_<sur>plus_v2si): Likewise.
17693 (reduc_<sur>plus_v4sf): Likewise.
17694 (aarch64_addpv4sf): Likewise.
17695 * config/aarch64/arm_neon.h
17696 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17697 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17698 add UNSPEC_SADDV, UNSPEC_UADDV.
17699 (SUADDV): New.
17700 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17701
6dce23a8
JG
177022013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17703
17704 * config/aarch64/arm_neon.h
17705 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17706
1598945b
JG
177072013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17708
17709 * config/aarch64/aarch64-builtins
17710 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17711
998eaf97
JG
177122013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17713
17714 * config/aarch64/aarch64-simd-builtins.def
17715 (reduc_smax_): New.
17716 (reduc_smin_): Likewise.
17717 (reduc_umax_): Likewise.
17718 (reduc_umin_): Likewise.
17719 (reduc_smax_nan_): Likewise.
17720 (reduc_smin_nan_): Likewise.
17721 (fmax): Remove.
17722 (fmin): Likewise.
17723 (smax): Update for V2SF, V4SF and V2DF modes.
17724 (smin): Likewise.
17725 (smax_nan): New.
17726 (smin_nan): Likewise.
17727 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17728 (<su><maxmin><mode>3): ...This, refactor.
17729 (s<maxmin><mode>3): New.
17730 (<maxmin_uns><mode>3): Likewise.
17731 (reduc_<maxmin_uns>_<mode>): Refactor.
17732 (reduc_<maxmin_uns>_v4sf): Likewise.
17733 (reduc_<maxmin_uns>_v2si): Likewise.
17734 (aarch64_<fmaxmin><mode>: Remove.
17735 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17736 new builtin names.
17737 (vmin<q>_f<32,64>): Likewise.
17738 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17739 (FMAXMIN): New.
17740 (su): Add mappings for smax, smin, umax, umin.
17741 (maxmin): New.
17742 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17743 (FMAXMIN): Rename as...
17744 (FMAXMIN_UNS): ...This.
17745 (maxminv): Remove.
17746 (fmaxminv): Likewise.
17747 (fmaxmin): Likewise.
17748 (maxmin_uns): New.
17749 (maxmin_uns_op): Likewise.
17750
bd11644e
JG
177512013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17752
17753 * config/aarch64/arm_neon.h
17754 (vac<ge, gt><sd>_f<32, 64>): Rename to...
17755 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
17756 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
17757
75dd5ace
JG
177582013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17759
17760 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
17761 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
17762
7c19979f
JG
177632013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17764
17765 * config/aarch64/aarch64-simd.md
17766 (vcond<mode>_internal): Handle special cases for constant masks.
17767 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
17768 (vcondu<mode><mode>): Likewise.
17769 (vcond<v_cmp_result><mode>): New.
17770
bb60efd9
JG
177712013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17772
17773 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
17774 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
17775 * config/aarch64/aarch64-simd-builtins.def
17776 (cmeq): Update to BUILTIN_VALLDI.
17777 (cmgt): Likewise.
17778 (cmge): Likewise.
17779 (cmle): Likewise.
17780 (cmlt): Likewise.
17781 * config/aarch64/arm_neon.h
17782 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
17783 to builtins or C as appropriate.
17784
889b9412
JG
177852013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17786
17787 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
17788 (cmgeu): ...This.
17789 (cmhi): Rename to...
17790 (cmgtu): ...This.
17791 * config/aarch64/aarch64-simd.md
17792 (simd_mode): Add SF.
17793 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
17794 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
17795 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
17796 (cstore<mode>_neg): ...This.
17797 * config/aarch64/iterators.md
17798 (VALLF): new.
17799 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
17800 (COMPARISONS): New.
17801 (UCOMPARISONS): Likewise.
17802 (optab): Add missing comparisons.
17803 (n_optab): New.
17804 (cmp_1): Likewise.
17805 (cmp_2): Likewise.
17806 (CMP): Likewise.
17807 (cmp): Remove.
17808 (VCMP_S): Likewise.
17809 (VCMP_U): Likewise.
17810 (V_cmp_result): Add DF, SF modes.
17811 (v_cmp_result): Likewise.
17812 (v): Likewise.
17813 (vmtype): Likewise.
17814 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
17815
0a7dbb76
GY
178162013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
17817
17818 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
17819 define_insn to define_insn_and_split.
17820 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
17821 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
17822 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
17823 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
17824 (thumb2_negscc): Likewise.
17825
fb614ca6
GY
178262013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
17827
17828 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
17829
9e64a0bf
GY
178302013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
17831
17832 * config/arm/thumb2.md: Remove trailing whitespaces.
17833
d6b28156
RS
178342013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17835
17836 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
17837 Use gen_int_mode rather than GEN_INT.
17838
f91674c3
L
178392013-04-30 H.J. Lu <hongjiu.lu@intel.com>
17840
b0dec607 17841 * value-prof.c (stream_in_histogram_value): Remove the strayed
f91674c3
L
17842 debug_gimple_stmt.
17843
3551257c
RB
178442013-04-30 Richard Biener <rguenther@suse.de>
17845
17846 PR middle-end/57122
3c21604f 17847 * cfghooks.c (split_edge): Properly check for the loop latch edge.
3551257c 17848
4c1aff1c
RB
178492013-04-30 Richard Biener <rguenther@suse.de>
17850
17851 PR middle-end/57107
17852 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
17853
54b8379a
AB
178542013-04-30 Andrey Belevantsev <abel@ispras.ru>
17855
17856 PR rtl-optimization/56957
17857 PR rtl-optimization/57105
54b8379a
AB
17858 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
17859 variable. Use just INSN_UID for determining whether an insn
17860 should be only disconnected from the insn stream.
17861 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
17862
abc27962
JJ
178632013-04-30 Jakub Jelinek <jakub@redhat.com>
17864
17865 PR tree-optimization/57104
17866 * tsan.c (instrument_expr): Don't instrument accesses to
17867 DECL_HARD_REGISTER VAR_DECLs.
17868
0fc822d0
RB
178692013-04-30 Richard Biener <rguenther@suse.de>
17870
17871 * function.h (loops_for_fn): New inline function.
17872 (set_loops_for_fn): Likewise.
17873 * cfgloop.h (place_new_loop): Add struct function parameter.
17874 (get_loop): Likewise.
17875 (get_loops): Likewise.
17876 (number_of_loops): Likewise.
17877 (fel_next): Adjust.
17878 (fel_init): Likewise.
17879 * cfg.c (get_loop_copy): Adjust.
17880 * cfgloop.c (flow_loops_dump): Likewise.
17881 (record_loop_exits): Likewise.
17882 (verify_loop_structure): Likewise.
17883 * cfgloopanal.c (mark_irreducible_loops): Likewise.
17884 (estimate_reg_pressure_cost): Likewise.
17885 (mark_loop_exit_edges): Likewise.
17886 * cfgloopmanip.c (place_new_loop): Likewise.
17887 (add_loop): Likewise.
17888 (duplicate_loop): Likewise.
17889 * graph.c (draw_cfg_nodes): Likewise.
17890 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
17891 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
17892 (extract_affine_chrec): Likewise.
17893 (build_scop_iteration_domain): Likewise.
17894 * graphite.c (graphite_initialize): Likewise.
17895 * ira-build.c (create_loop_tree_nodes): Likewise.
17896 (more_one_region_p): Likewise.
17897 (rebuild_regno_allocno_maps): Likewise.
17898 (mark_loops_for_removal): Likewise.
17899 (mark_all_loops_for_removal): Likewise.
17900 (remove_unnecessary_regions): Likewise.
17901 (ira_build): Likewise.
17902 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17903 * loop-init.c (fix_loop_structure): Likewise.
17904 (gate_rtl_move_loop_invariants): Likewise.
17905 (gate_rtl_unswitch): Likewise.
17906 (gate_rtl_unroll_and_peel_loops): Likewise.
17907 (rtl_doloop): Likewise.
17908 * lto-streamer-in.c (input_cfg): Likewise.
17909 * lto-streamer-out.c (output_cfg): Likewise.
17910 * modulo-sched.c (sms_schedule): Likewise.
17911 * predict.c (tree_estimate_probability): Likewise.
17912 (tree_estimate_probability_driver): Likewise.
17913 (estimate_loops): Likewise.
17914 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
17915 (move_sese_region_to_fn): Likewise.
17916 (debug_loop_num): Likewise.
17917 * tree-chrec.c (chrec_evaluate): Likewise.
17918 (hide_evolution_in_other_loops_than_loop): Likewise.
17919 (chrec_component_in_loop_num): Likewise.
17920 (reset_evolution_in_loop): Likewise.
17921 (evolution_function_is_invariant_rec_p): Likewise.
17922 * tree-if-conv.c (main_tree_if_conversion): Likewise.
17923 * tree-inline.c (copy_loops): Likewise.
17924 (copy_cfg_body): Likewise.
17925 (tree_function_versioning): Likewise.
17926 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
17927 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
17928 Likewise.
17929 (add_to_evolution_1): Likewise.
17930 (scev_const_prop): Likewise.
17931 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
17932 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
17933 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
17934 (tree_ssa_lim_initialize): Likewise.
17935 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
17936 (verify_loop_closed_ssa): Likewise.
17937 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
17938 (tree_ssa_loop_im): Likewise.
17939 (tree_ssa_loop_unswitch): Likewise.
17940 (tree_vectorize): Likewise.
17941 (check_data_deps): Likewise.
17942 (tree_ssa_loop_ivcanon): Likewise.
17943 (tree_ssa_loop_bounds): Likewise.
17944 (tree_complete_unroll): Likewise.
17945 (tree_complete_unroll_inner): Likewise.
17946 (tree_parallelize_loops): Likewise.
17947 (tree_ssa_loop_prefetch): Likewise.
17948 (tree_ssa_loop_ivopts): Likewise.
17949 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
17950 * tree-vectorizer.c (vectorize_loops): Likewise.
17951
37953bd3
MF
179522013-04-29 Mike Frysinger <vapier@gentoo.org>
17953
17954 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
17955 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
17956 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
17957 with EABI_LINK_SPEC.
17958
f9ed28db
UB
179592013-04-29 Uros Bizjak <ubizjak@gmail.com>
17960
17961 PR target/44578
17962 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
17963 alternative.
17964
deca73f5
VM
179652013-04-29 Vladimir Makarov <vmakarov@redhat.com>
17966
17967 PR target/57097
37953bd3 17968 * lra-constraints.c (process_alt_operands): Discourage a bit more
deca73f5
VM
17969 using memory for pseudos. Print cost dump for alternatives.
17970 Modify cost values for conflicts with early clobbers.
17971 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
17972
fdca7d03
UB
179732013-04-29 Uros Bizjak <ubizjak@gmail.com>
17974
17975 PR target/57098
17976 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
17977
0b064172
IB
179782013-04-29 Ian Bolton <ian.bolton@arm.com>
17979
17980 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
17981 from/to S register.
17982 (movdi_aarch64): Support LDR/STR from/to D register.
17983
473cec55
IB
179842013-04-29 Ian Bolton <ian.bolton@arm.com>
17985
17986 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
17987 or higher by default.
17988
a6f30e66
RB
179892013-04-29 Richard Biener <rguenther@suse.de>
17990
17991 PR middle-end/57075
17992 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
17993 even if not adding abnormal edges for calls that can make
17994 abnormal gotos.
17995
0c2b2040
RB
179962013-04-29 Richard Biener <rguenther@suse.de>
17997
17998 PR middle-end/57103
17999 * tree-cfg.c (move_stmt_op): Fix condition under which to update
18000 TREE_BLOCK.
18001 (move_stmt_r): Remove redundant checking.
18002
f41f80f9
TJ
180032013-04-29 Teresa Johnson <tejohnson@google.com>
18004
18005 PR bootstrap/57077
18006 * basic-block.h (apply_scale): New function.
18007 (apply_probability): Use apply_scale.
18008 * gimple-streamer-in.c (input_bb): Ditto.
18009 * lto-streamer-in.c (input_cfg): Ditto.
18010 * lto-cgraph.c (merge_profile_summaries): Ditto.
18011 * tree-optimize.c (execute_fixup_cfg): Ditto.
fdca7d03 18012 * tree-inline.c (copy_bb): Update comment to use apply_scale.
f41f80f9
TJ
18013 (copy_edges_for_bb): Ditto.
18014 (copy_cfg_body): Ditto.
18015
315bbd2e
TV
180162013-04-29 Tom de Vries <tom@codesourcery.com>
18017
18018 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
18019 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
18020 (tail_merge_optimize): Handle current_loops == NULL.
18021
ebbd90d8
JL
180222013-04-26 Jeff Law <law@redhat.com>
18023
18024 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
18025 (simplify_cond_using_ranges): Generalize code to simplify
18026 COND_EXPRs where one argument is a constant and the other
18027 is an SSA_NAME created by an integral type conversion.
18028
8b9b57eb
KT
180292013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18030
18031 * config/arm/arm.md (store_minmaxsi): Use only when
18032 optimize_insn_for_size_p.
18033
9498e5dc
CB
180342013-04-29 Christian Bruel <christian.bruel@st.com>
18035
18036 PR target/57108
18037 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18038
6093bc06
RB
180392013-04-29 Richard Biener <rguenther@suse.de>
18040
18041 PR middle-end/57089
fdca7d03
UB
18042 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18043 loop tree make sure to schedule a fixup for the child as well.
6093bc06
RB
18044 (expand_omp_for_generic): Properly add loops.
18045 (expand_omp_for_static_nochunk): Likewise.
18046 (expand_omp_for_static_chunk): Likewise.
18047 (expand_omp_for): For the degenerate case fixup loops.
18048 (expand_omp_sections): Fix default bb placement in loops.
18049 (expand_omp_atomic_pipeline): Properly add loops.
18050
84aacbfd
KT
180512013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18052
18053 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18054
8dee4479
TV
180552013-04-29 Tom de Vries <tom@codesourcery.com>
18056
18057 * tree-ssa-tail-merge.c: Update header comment.
18058
47934dc4
JG
180592013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18060
18061 * config/aarch64/arm_neon.h
18062 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18063 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18064 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18065 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18066 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18067 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18068 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18069 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18070
384be29f
JG
180712013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18072
18073 * config/aarch64/aarch64-simd.md
18074 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18075 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18076 fix_trunc, fixuns_trunc.
18077 (ftrunc<VDQF:mode>2): New.
18078 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18079 (fix_trunc_optab): New.
18080
0386b123
JG
180812013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18082
18083 * config/aarch64/aarch64-builtins.c
18084 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18085 iceilf, lround, iroundf.
18086
00fcb892
UB
180872013-04-29 Uros Bizjak <ubizjak@gmail.com>
18088
18089 PR target/54349
18090 * config/i386/i386.h (enum ix86_tune_indices)
18091 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18092 New, split from X86_TUNE_INTER_UNIT_MOVES.
18093 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18094 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18095 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18096 (TARGET_INTER_UNIT_MOVES): Remove.
18097 * config/i386/i386.c (initial_ix86_tune_features): Update.
18098 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18099 (ix86_expand_convert_uns_didf_sse): Use
18100 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18101 (ix86_expand_vector_init_one_nonzero): Ditto.
18102 (ix86_expand_vector_init_interleave): Ditto.
18103 (inline_secondary_memory_needed): Return true for moves from SSE class
18104 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18105 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18106 * config/i386/constraints.md (Yi, Ym): Depend on
18107 TARGET_INTER_UNIT_MOVES_TO_VEC.
18108 (Yj, Yn): New constraints.
18109 * config/i386/i386.md (*movdi_internal): Change constraints of
18110 operand 1 from Yi to Yj and from Ym to Yn.
18111 (*movsi_internal): Ditto.
18112 (*movdf_internal): Ditto.
18113 (*movsf_internal): Ditto.
18114 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18115 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18116 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18117 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18118 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18119 * config/i386/sse.md (movdi_to_sse): Ditto.
18120 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18121 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18122 TARGET_INTER_UNIT_MOVES.
18123 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18124 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18125 instead of TARGET_INTER_UNIT_MOVES.
18126 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18127 operand 1 from Yi to Yj and from Ym to Yn.
18128
4c871069
JG
181292013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18130
18131 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18132 (float_truncate_hi_): Likewise.
18133 (float_extend_lo_): Likewise.
18134 (float_truncate_lo_): Likewise.
18135 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18136 (aarch64_float_extend_lo_v2df): Likewise.
18137 (vec_unpacks_hi_v4sf): Likewise.
18138 (aarch64_float_truncate_lo_v2sf): Likewise.
18139 (aarch64_float_truncate_hi_v4sf): Likewise.
18140 (vec_pack_trunc_v2df): Likewise.
18141 (vec_pack_trunc_df): Likewise.
18142
1709ff9b
JG
181432013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18144
18145 * config/aarch64/aarch64-builtins.c
18146 (aarch64_fold_builtin): Fold float conversions.
18147 * config/aarch64/aarch64-simd-builtins.def
18148 (floatv2si, floatv4si, floatv2di): New.
18149 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18150 * config/aarch64/aarch64-simd.md
18151 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18152 * config/aarch64/iterators.md (FLOATUORS): New.
18153 (optab): Add float, floatuns.
18154 (su_optab): Likewise.
18155
ce966824
JG
181562013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18157
18158 * config/aarch64/aarch64-builtins.c
18159 (aarch64_builtin_vectorized_function): Use new names for
18160 fcvt builtins.
18161 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18162 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18163 (fcvtzu): Split as...
18164 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18165 (fcvtas): Split as...
18166 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18167 (fcvtau): Split as...
18168 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18169 (fcvtps): Split as...
18170 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18171 (fcvtpu): Split as...
18172 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18173 (fcvtms): Split as...
18174 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18175 (fcvtmu): Split as...
18176 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18177 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18178 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18179 (lfrintnusf, lfrintnudf): Likewise.
18180 * config/aarch64/aarch64-simd.md
18181 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18182 define_insn.
18183 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18184 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18185 (fcvt_pattern): Likewise.
18186
b9de24fe
JG
181872013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18188
18189 * config/aarch64/aarch64-simd.md
18190 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18191 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18192
77a205be
JG
181932013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18194
18195 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18196 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18197 (vrnd<a,m,n,p>_f32): Implement using builtins.
18198 (vrnd<i,x><q>_f<32, 64>): New.
18199
0659ce6f
JG
182002013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18201
18202 * config/aarch64/aarch64-builtins.c
18203 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18204 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18205 (frintz): Rename to...
18206 (btrunc): ...this.
18207 (frintp): Rename to...
18208 (ceil): ...this.
18209 (frintm): Rename to...
18210 (floor): ...this.
18211 (frinti): Rename to...
18212 (nearbyint): ...this.
18213 (frintx): Rename to...
18214 (rint): ...this.
18215 (frinta): Rename to...
18216 (round): ...this.
18217 * config/aarch64/aarch64-simd.md
18218 (aarch64_frint<frint_suffix><mode>): Delete.
18219 (<frint_pattern><mode>2): Convert to insn.
18220 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18221 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18222 (frint_pattern): Likewise.
18223 (frint_suffix): Likewise.
18224
ea78906a
RB
182252013-04-29 Richard Biener <rguenther@suse.de>
18226
18227 PR tree-optimization/57081
18228 * loop-init.c: Include tree-flow.h.
18229 (loop_optimizer_finalize): Free number of iteration estimates.
18230 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18231
baee1763
JJ
182322013-04-29 Jakub Jelinek <jakub@redhat.com>
18233
94dc5332
JJ
18234 PR tree-optimization/57083
18235 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18236 non-singleton shift count range, zero extend low_bound for uns case.
18237
baee1763
JJ
18238 * config/i386/predicates.md (general_vector_operand): New predicate.
18239 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18240 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18241 if they aren't nonimmediate operands. If their original values
18242 satisfy const_vector_equal_evenodd_p, don't shift them.
18243 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18244 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18245 if not nonimmediate_operand.
18246 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18247 instead of register_operand.
18248 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18249
a9073727 182502013-04-28 Eric Botcazou <ebotcazou@adacore.com>
22ad5e0a
EB
18251
18252 * stor-layout.c (finalize_size_functions): Allocate a structure and
18253 reset cfun before dumping the functions.
18254
ba8011e6
JJ
182552013-04-27 Jakub Jelinek <jakub@redhat.com>
18256
d6fde69e
JJ
18257 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18258
ba8011e6
JJ
18259 PR target/56866
18260 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18261 use xop_pmacsdqh if uns_p.
18262 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18263 the immediate rotate count.
18264
2c62cbaa
VM
182652013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18266
18267 * rtl.h (struct rtx_def): Add comment for field jump.
18268 (LRA_SUBREG_P): New macro.
18269 * recog.c (register_operand): Check LRA_SUBREG_P.
18270 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18271 stack frame.
18272 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18273 (lra_final_code_change): Skip subreg change for operators.
18274 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18275 if there are no operand changes.
18276 * lra-constraints.c (curr_insn_set): New.
18277 (match_reload): Set LRA_SUBREG_P.
18278 (emit_spill_move): Ditto.
18279 (check_and_process_move): Use curr_insn_set. Process only single
18280 set insns. Don't initialize sec_mem_p and change_p.
18281 (simplify_operand_subreg): Use LRA_SUBREG_P.
18282 (reg_in_class_p): New function.
18283 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18284 of #ifdef. Add code to remove cycling.
18285 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18286 non-null disp. Reload inner instead of disp when base and index
18287 are null. Try to put lo_sum into register.
18288 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18289 (check_and_process_move): Move code for move cost check to
18290 simple_move_p. Remove equiv_substitution.
18291 (simple_move_p): New function.
18292 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18293 curr_insn_set. Call check_and_process_move only for single set
18294 insns. Use the new function. Move call of check_and_process_move
18295 after operand equiv substitution and address process.
18296
e7d764f3
JJ
182972013-04-26 Jakub Jelinek <jakub@redhat.com>
18298
18299 PR go/57045
18300 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18301 with nonlocal goto receivers or returns twice calls, ignore
18302 unininitialized values from abnormal edges to nl goto receiver
18303 or returns twice call.
18304
41e10689
JJ
183052013-04-26 Jakub Jelinek <jakub@redhat.com>
18306
18307 PR tree-optimization/57051
18308 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18309 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18310 bitsize.
18311
d7ed20db
RB
183122013-04-26 Richard Biener <rguenther@suse.de>
18313
18314 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18315 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18316 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18317 (fixup_loop_arrays_after_move): New function.
18318 (move_sese_region_to_fn): Properly outline the loop tree parts
18319 of the SESE region.
18320
df93505e
UB
183212013-04-26 Uros Bizjak <ubizjak@gmail.com>
18322
18323 * config/i386/i386.md (type, unit): Fix long lines.
18324
dd366ec3
RB
183252013-04-26 Richard Biener <rguenther@suse.de>
18326
18327 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18328 (lto-streamer-out.o): Likewise.
18329 * cfgloop.c (init_loops_structure): Export, add struct function
18330 argument and adjust.
18331 (flow_loops_find): Adjust.
18332 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18333 (init_loops_structure): Declare.
18334 * lto-streamer-in.c: Include cfgloop.h.
18335 (input_cfg): Input the loop tree.
18336 * lto-streamer-out.c: Include cfgloop.h.
18337 (output_cfg): Output the loop tree.
18338 (output_struct_function_base): Do not drop PROP_loops.
18339
a9e0d843
RB
183402013-03-26 Richard Biener <rguenther@suse.de>
18341
18342 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18343 (pass_build_cfg): Provide PROP_loops.
18344 (move_sese_region_to_fn): Remove loops that are outlined into fn
18345 for now.
18346 * tree-inline.c: Include cfgloop.h.
18347 (initialize_cfun): Do not drop PROP_loops.
18348 (copy_loops): New function.
18349 (copy_cfg_body): Copy loop structure.
18350 (tree_function_versioning): Initialize destination loop tree.
18351 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18352 (pass_parallelize_loops): Do IL verification.
18353 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18354 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18355 the CFG make sure we fixup loops as well.
18356 * tree-ssa-tail-merge.c: Include cfgloop.h.
18357 (replace_block_by): When merging loop latches mark loops for fixup.
18358 * lto-streamer-out.c (output_struct_function_base): Drop
18359 PROP_loops for now.
18360 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18361 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18362 * ipa-split.c: Include cfgloop.h.
18363 (split_function): Add the new return block to the loop tree root.
18364 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18365 whether we have removed the forwarder block.
18366 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18367 * cfgloop.h (place_new_loop): Declare.
18368 * cfgloopmanip.c (place_new_loop): Export.
18369 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18370 (tree-switch-conversion.o): Likewise.
18371 (tree-complex.o): Likewise.
18372 (tree-inline.o): Likewise.
18373 (tree-ssa-tailmerge.o): Likewise.
18374 (ipa-split.o): Likewise.
18375 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18376 (tree-ssa-copy.o): Likewise.
18377 * tree-switch-conversion.c: Include cfgloop.h
18378 (process_switch): If we emit a bit-test cascade, schedule loops
18379 for fixup.
18380 * tree-complex.c: Include cfgloop.h.
18381 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18382 * asan.c: Include cfgloop.h.
18383 (create_cond_insert_point): Properly add new basic-blocks to
18384 loops, schedule loop fixup.
18385 * cfgloop.c (verify_loop_structure): Check that looks are not
18386 marked for fixup.
18387 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18388 to loops.
18389 (expand_omp_for_generic): Likewise.
18390 (expand_omp_sections): Likewise.
18391 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18392 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18393 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18394 is initialized, not when loops are present.
18395 * tree-parloops.c (parallelize_loops): Remove checking here.
18396 * passes.c (init_optimization_passes): Schedule a copy-propagation
18397 pass before complete unrolling of inner loops.
18398
e78e8a0b
JJ
183992013-04-26 Jakub Jelinek <jakub@redhat.com>
18400
a2e836b2
JJ
18401 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18402 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18403 (colorize_init): Add argument to _WIN32 version.
18404 * toplev.c: Include diagnostic-color.h.
18405 (process_options): Default to -fdiagnostics-color=auto if
18406 GCC_COLORS env var is in the environment.
18407 * common.opt (fdiagnostics-color=): Add Var and Init.
df93505e
UB
18408 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18409 env var is in the environment, the default is auto rather than never.
a2e836b2 18410
e78e8a0b
JJ
18411 * diagnostic.h (file_name_as_prefix): Add context argument.
18412 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18413 the string as locus.
18414 * langhooks.c (lhd_print_error_function): Adjust caller.
18415
013e5ef9
LC
184162013-04-25 Lawrence Crowl <crowl@google.com>
18417
18418 * var-tracking.c (shared_hash_def::htab):
18419 Change type to hash_table. Update dependent calls and types.
18420
4a8fb1a1
LC
184212013-04-25 Lawrence Crowl <crowl@google.com>
18422
18423 * Makefile.in: Update as needed below.
18424
18425 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18426 Move declaration to after the type's method definitons.
18427
18428 * attribs.c (htab_t scoped_attributes::attribute_hash):
18429 Change type to hash_table. Update dependent calls and types.
18430
18431 * bitmap.c (htab_t bitmap_desc_hash):
18432 Change type to hash_table. Update dependent calls and types.
18433
18434 * cselib.c (htab_t cselib_hash_table):
18435 Change type to hash_table. Update dependent calls and types.
18436
18437 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18438 (hash_string_slot_node): Move implementation into lto-streamer.h
18439 struct string_slot_hasher.
18440 (eq_string_slot_node): Likewise.
18441
18442 * data-streamer-out.c: Update output_block::string_hash_table
18443 dependent calls and types.
18444
18445 * dwarf2cfi.c (htab_t trace_index):
18446 Change type to hash_table. Update dependent calls and types.
18447
18448 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18449 Change type to hash_table. Update dependent calls and types.
18450 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18451 (htab_t optimize_external_refs::map): Likewise.
18452 (htab_t output_comp_unit::extern_map): Likewise.
18453 (htab_t output_comdat_type_unit::extern_map): Likewise.
18454 (htab_t output_macinfo::macinfo_htab): Likewise.
18455 (htab_t optimize_location_lists::htab): Likewise.
18456 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18457
18458 * except.c (htab_t ehspec_hash_type):
18459 Change type to hash_table. Update dependent calls and types.
18460 (assign_filter_values::ttypes): Likewise.
18461 (assign_filter_values::ehspec): Likewise.
18462 (sjlj_assign_call_site_values::ar_hash): Likewise.
18463 (convert_to_eh_region_ranges::ar_hash): Likewise.
18464
18465 * gcse.c (htab_t pre_ldst_table):
18466 Change type to hash_table. Update dependent calls and types.
18467
18468 * ggc-common.c (htab_t saving_htab):
18469 Change type to hash_table. Update dependent calls and types.
18470 (htab_t loc_hash): Likewise.
18471 (htab_t ptr_hash): Likewise.
18472 (call_count): Rename ggc_call_count.
18473 (call_alloc): Rename ggc_call_alloc.
18474 (loc_descriptor): Rename make_loc_descriptor.
18475 (add_statistics): Rename ggc_add_statistics.
18476
18477 * ggc-common.c (saving_htab):
18478 Change type to hash_table. Update dependent calls and types.
18479
18480 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18481 (push_gimplify_context): Likewise.
18482 (pop_gimplify_context): Likewise.
18483 (struct gimple_temp_hash_elt): Added.
18484 (struct gimplify_hasher): Likewise.
18485 (struct gimplify_ctx.temp_htab):
18486 Change type to hash_table. Update dependent calls and types.
18487
18488 * gimple-fold.c: Include gimplify-ctx.h.
18489
18490 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18491 Change type to hash_table. Update dependent calls and types.
18492 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18493 avoid potential global name collision.
18494
18495 * gimplify.c: Include gimplify-ctx.h.
18496 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18497 (htab_t gimplify_ctx::temp_htab):
18498 Update dependent calls and types for new type hash_table.
18499 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18500 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18501
18502 * gimplify-ctx.h: New.
18503 (struct gimple_temp_hash_elt): Move from gimplify.c.
18504 (class gimplify_hasher): New.
18505 (struct gimplify_ctx): Move from gimple.h.
18506 (htab_t gimplify_ctx::temp_htab):
18507 Change type to hash_table. Update dependent calls and types.
18508
18509 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18510 (htab_t ivs_params::newivs_index):
18511 Change type to hash_table. Update dependent calls and types.
18512 (htab_t ivs_params::params_index): Likewise.
18513 (htab_t print_generated_program::params_index): Likewise.
18514 (htab_t gloog::newivs_index): Likewise.
18515 (htab_t gloog::params_index): Likewise.
18516
18517 * graphite.c: Include graphite-htab.h.
18518 4htab_t graphite_transform_loops::bb_pbb_mapping):
18519 Change type to hash_table. Update dependent calls and types.
18520
18521 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18522 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18523 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18524
18525 * graphite-dependences.c: Include graphite-htab.h.
18526 (loop_is_parallel_p): Change hash table type of parameter.
18527
18528 * graphite-htab.h: New.
18529 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18530 (extern find_pbb_via_hash): Move from graphite-poly.h.
18531 (extern loop_is_parallel_p): Move from graphite-poly.h.
18532 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18533
18534 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18535 (extern loop_is_parallel_p): Move to graphite-htab.h.
18536 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18537
18538 * haifa-sched.c (htab_t delay_htab):
18539 Change type to hash_table. Update dependent calls and types.
18540 (htab_t delay_htab_i2): Likewise.
18541
18542 * ira-color.c (htab_t allocno_hard_regs_htab):
18543 Change type to hash_table. Update dependent calls and types.
18544
18545 * ira-costs.c (htab_t cost_classes_htab):
18546 Change type to hash_table. Update dependent calls and types.
18547
18548 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18549 Change type to hash_table. Update dependent calls and types.
18550
18551 * loop-iv.c (htab_t bivs):
18552 Change type to hash_table. Update dependent calls and types.
18553
18554 * loop-unroll.c (htab_t opt_info::insns_to_split):
18555 Change type to hash_table. Update dependent calls and types.
18556 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18557
18558 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18559 (struct string_slot_hasher): New.
18560 (htab_t output_block::string_hash_table):
18561 Change type to hash_table. Update dependent calls and types.
18562
18563 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18564 (htab_t file_name_hash_table):
18565 Change type to hash_table. Update dependent calls and types.
18566
18567 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18568 calls and types.
18569
18570 * lto-streamer.c (htab_t tree_htab):
18571 Change type to hash_table. Update dependent calls and types.
18572
18573 * omp-low.c: Include gimplify-ctx.h.
18574
18575 * passes.c (htab_t name_to_pass_map):
18576 Change type to hash_table. Update dependent calls and types.
18577 (pass_traverse): Rename to passes_pass_traverse.
18578
18579 * plugin.c (htab_t event_tab):
18580 Change type to hash_table. Update dependent calls and types.
18581
18582 * postreload-gcse.c (htab_t expr_table):
18583 Change type to hash_table. Update dependent calls and types.
18584 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18585
18586 * sese.c (debug_rename_map_1): Make extern.
18587 (htab_t copy_bb_and_scalar_dependences::rename_map):
18588 Change type to hash_table. Update dependent calls and types.
18589
18590 * sese.h (extern debug_rename_map): Move to .c file.
18591
18592 * store-motion.c (htab_t store_motion_mems_table):
18593 Change type to hash_table. Update dependent calls and types.
18594
18595 * trans-mem.c (htab_t tm_new_mem_hash):
18596 Change type to hash_table. Update dependent calls and types.
18597
18598 * tree-browser.c (htab_t TB_up_ht):
18599 Change type to hash_table. Update dependent calls and types.
18600
18601 * tree-cfg.c (htab_t discriminator_per_locus):
18602 Change type to hash_table. Update dependent calls and types.
18603
18604 * tree-complex.c: Include tree-hasher.h
18605 (htab_t complex_variable_components):
18606 Change type to hash_table. Update dependent calls and types.
18607
18608 * tree-eh.c (htab_t finally_tree):
18609 Change type to hash_table. Update dependent calls and types.
18610
18611 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18612 struct int_tree_hasher.
18613 (extern int_tree_map_eq): Likewise.
18614 (uid_decl_map_hash): Removed.
18615 (extern decl_tree_map_eq): Likewise.
18616
18617 * tree-hasher.h: New.
18618 (struct int_tree_hasher): New.
18619 (typedef int_tree_htab_type): New.
18620
18621 * tree-inline.c: Include gimplify-ctx.h.
18622
18623 * tree-mudflap.c: Include gimplify-ctx.h.
18624
18625 * tree-parloops.c: Include tree-hasher.h.
18626 (htab_t eliminate_local_variables_stmt::decl_address):
18627 Change type to hash_table. Update dependent calls and types.
18628 (htab_t separate_decls_in_region::decl_copies): Likewise.
18629
18630 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18631 Change type to hash_table. Update dependent calls and types.
18632
18633 * tree-sra.c (candidates):
18634 Change type to hash_table. Update dependent calls and types.
18635
18636 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18637 in tree-flow.h.
18638 (int_tree_map_hash): Likewise.
18639
18640 * tree-ssa-dom.c (htab_t avail_exprs):
18641 Change type to hash_table. Update dependent calls and types.
18642
18643 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18644 Change type to hash_table. Update dependent calls and types.
18645
18646 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18647 Change type to hash_table. Update dependent calls and types.
18648
18649 * tree-ssa-phiopt.c (seen_ssa_names):
18650 Change type to hash_table. Update dependent calls and types.
18651
18652 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18653 Change type to hash_table. Update dependent calls and types.
18654
18655 * tree-ssa-uncprop.c (equiv):
18656 Change type to hash_table. Update dependent calls and types.
18657
c5a44004
JJ
186582013-04-25 Jakub Jelinek <jakub@redhat.com>
18659
18660 PR rtl-optimization/57003
18661 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18662 call note_stores with kill_clobbered_value callback again after
18663 killing regs_invalidated_by_call.
18664
09962a4a
JG
186652013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18666
18667 * config/aarch64/aarch64-simd.md
18668 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18669 (aarch64_simd_bsl<mode>): Likewise.
18670 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18671
ea28bb0b
MP
186722013-04-25 Marek Polacek <polacek@redhat.com>
18673
18674 PR tree-optimization/57066
3c21604f 18675 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
ea28bb0b 18676
96659611
JG
186772013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18678
18679 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18680
9697e620
JG
186812013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18682
18683 * config/aarch64/aarch64-builtins.c
18684 (aarch64_fold_builtin): New.
18685 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18686 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18687 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18688 * config/aarch64/arm_neon.h
18689 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18690
0ac198d3
JG
186912013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18692 Tejas Belagod <tejas.belagod@arm.com>
18693
18694 * config/aarch64/aarch64-builtins.c
18695 (aarch64_gimple_fold_builtin): New.
18696 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18697 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18698 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18699 (addvv4sf): Update.
18700 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18701
58cff58c
N
187022013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18703
df93505e 18704 * config/aarch64/aarch64.md
58cff58c
N
18705 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18706
7e0228bf
N
187072013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18708
18709 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18710 (*ngcsi_uxtw): New pattern.
18711
5819f96f 187122013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3c21604f 18713 Julian Brown <julian@codesourcery.com>
5819f96f
KT
18714
18715 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18716 (TB_DREG): Add T_V4HF.
18717 (v4hf_UP): New macro.
18718 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
df93505e 18719 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
18720 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18721 built-ins.
df93505e 18722 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
5819f96f
KT
18723 (arm_vector_mode_supported_p): Handle V4HF.
18724 (arm_mangle_map): Handle V4HFmode.
18725 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18726 * config/arm/arm_neon_builtins.def: Add entries for
18727 vcvtv4hfv4sf, vcvtv4sfv4hf.
18728 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18729 (neon_vcvtv4hfv4sf): Likewise.
18730 * config/arm/neon-gen.ml: Handle half-precision floating point
18731 features.
18732 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18733 * config/arm/arm_neon.h: Regenerate.
18734 * config/arm/neon.ml (type elts): Add F16.
18735 (type vectype): Add T_float16x4, T_floatHF.
18736 (type vecmode): Add V4HF.
18737 (type features): Add Requires_FP_bit feature.
18738 (elt_width): Handle F16.
18739 (elt_class): Likewise.
18740 (elt_of_class_width): Likewise.
18741 (mode_of_elt): Refactor.
18742 (type_for_elt): Handle F16, fix error messages.
18743 (vectype_size): Handle T_float16x4.
18744 (vcvt_sh): New function.
18745 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18746 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18747 (string_of_mode): Handle V4HF.
18748 * doc/arm-neon-intrinsics.texi: Regenerate.
18749
1ef395e4
JG
187502013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18751
18752 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
18753 format specifier in 'X' case.
18754
41c34e94
AM
187552013-04-25 Alan Modra <amodra@gmail.com>
18756
18757 PR target/57052
18758 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
18759 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
18760 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
18761 Repeat for many other rotate/shift and mask patterns using subregs.
18762 Name lshiftrt insns.
18763 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
18764 on WORDS_BIG_ENDIAN.
18765
b9a7eb5d
AM
187662013-04-25 Alan Modra <amodra@gmail.com>
18767
18768 * config.gcc: Support little-endian powerpc-linux targets.
18769 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
18770 (LINK_OS_LINUX_SPEC): Define.
18771 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
18772 Preserve MASK_LITTLE_ENDIAN.
18773 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
18774 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
18775 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
18776 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
18777 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
18778 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
18779 Correct fp word order for little-endian. Don't shift toc entries
18780 smaller than a word for little-endian.
18781 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
18782 (bswapdi2 splits): Correct low-part subreg for little-endian.
18783 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
18784 low/high where such is correct only for be.
18785 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
18786 little-endian for -mcall-aixdesc.
18787
87f73374
AM
187882013-04-25 Alan Modra <amodra@gmail.com>
18789
18790 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
18791 replace_equiv_address_nv.
18792
cabf91cd
AM
187932013-04-25 Alan Modra <amodra@gmail.com>
18794
18795 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
18796
0ae24cc8
VM
187972013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18798
18799 Revert:
18800 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18801 * rtl.h (struct rtx_def): ...
cabf91cd 18802
77bce07c
VM
188032013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18804
18805 PR rtl-optimizations/57046
18806 * lra-constraints (split_reg): Set up lra_risky_transformations_p
18807 for multi-reg splits.
18808
0db63e7f
L
188092013-04-24 H.J. Lu <hongjiu.lu@intel.com>
18810
18811 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
18812
3362b6b6
SA
188132013-04-24 Sterling Augustine <saugustine@google.com>
18814
18815 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
18816 (comp_dir_string, debug_str_dwo_section): New.
18817 (DEBUG_STR_DWO_SECTION): Rename to ...
18818 (DEBUG_DWO_STR_SECTION): ... this.
18819 (DEBUG_NORM_STR_SECTION): Delete.
18820 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
18821 (DEBUG_STR_DWO_SECTION_FLAGS): New.
18822 (find_AT_string): Move most logic to ...
18823 (find_AT_string_in_table): ... here. New.
18824 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
18825 add_skeleton_AT_string. Delete logic.
18826 (output_skeleton_debug_sections): Remove call to
18827 add_top_level_skeleton_die_attrs.
18828 (add_comp_dir_attribute): Move logic to comp_dir_string.
18829 (dwarf2out_init): Initialize debug_str_dwo_section.
18830 (output_indirect_string): Call find_string_form.
18831 (output_indirect_strings): Rewrite.
18832 (prune_unused_types): Empty skeleton_debug_str_hash.
18833 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
18834 (dwarf2out_finish): Call output_indirect_strings.
18835
e93e18e9
PC
188362013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
18837
18838 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
18839
f6ce35ac
VM
188402013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18841
cabf91cd 18842 * rtl.h (struct rtx_def): Add comment for field jump.
f6ce35ac
VM
18843 (LRA_SUBREG_P): New macro.
18844 * recog.c (register_operand): Check LRA_SUBREG_P.
18845 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18846 stack frame.
18847 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18848 (lra_final_code_change): Skip subreg change for operators.
18849 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18850 if there are no operand changes.
18851 * lra-constraints.c (curr_insn_set): New.
18852 (match_reload): Set LRA_SUBREG_P.
18853 (emit_spill_move): Ditto.
18854 (check_and_process_move): Use curr_insn_set. Process only single
18855 set insns. Don't initialize sec_mem_p and change_p.
18856 (simplify_operand_subreg): Use LRA_SUBREG_P.
18857 (reg_in_class_p): New function.
18858 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18859 of #ifdef. Add code to remove cycling.
18860 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18861 non-null disp. Reload inner instead of disp when base and index
18862 are null. Try to put lo_sum into register.
18863 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
cabf91cd 18864 (check_and_process_move): Move code for move cost check to
f6ce35ac 18865 simple_move_p. Remove equiv_substitution.
cabf91cd 18866 (simple_move_p): New function.
f6ce35ac
VM
18867 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18868 curr_insn_set. Call check_and_process_move only for single set
18869 insns. Use the new function. Move call of check_and_process_move
18870 after operand equiv substitution and address process.
18871
38047d90
JG
188722013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
18873
18874 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
18875 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
18876 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
18877
13f39b2e
PC
188782013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
18879
18880 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
18881
97c116dc
MP
188822013-04-24 Marek Polacek <polacek@redhat.com>
18883
18884 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
18885 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
18886 (select_loops_exit_conditions): Likewise.
18887 (number_of_iterations_for_all_loops): Likewise.
18888 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
18889 (scev_analysis): Likewise.
18890
83082391 188912013-04-02 Catherine Moore <clm@codesourcery.com>
12211b99 18892 Chao-ying Fu <fu@mips.com>
83082391 18893
cabf91cd
AM
18894 * config/mips/micromips.md (jraddiusp): New pattern.
18895 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
18896 instruction if possible.
83082391 18897
19e34aa2
AM
188982013-04-24 Alan Modra <amodra@gmail.com>
18899
18900 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
18901
fdb6603c
JB
189022013-04-24 Julian Brown <julian@codesourcery.com>
18903 Chung-Lin Tang <cltang@codesourcery.com>
18904
18905 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
df93505e
UB
18906 dependency behavior in enumeration type DIE generation. Add TODO note
18907 to comments about future DW_FORM_sdata/udata re-work of related code.
fdb6603c 18908
bf190e8d
LC
189092013-04-23 Lawrence Crowl <crowl@google.com>
18910
18911 * Makefile.in: Update as needed below.
18912
18913 * hash-table.h (class hash_table):
18914 Correct many methods with parameter types compare_type to the correct
18915 value_type. (Correct code was unlikely to notice the change.)
18916 (hash_table::elements_with_deleted) New.
18917 (class hashtable::iterator): New.
18918 (hashtable::begin()): New.
18919 (hashtable::end()): New.
18920 (FOR_EACH_HASH_TABLE_ELEMENT): New.
18921
18922 * statistics.c (statistics_hashes):
18923 Change type to hash_table. Update dependent calls and types.
18924
18925 * tree-into-ssa.c (var_infos):
18926 Change type to hash_table. Update dependent calls and types.
18927
18928 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
18929 Change type to hash_table. Update dependent calls and types.
18930
18931 * tree-ssa-loop-im.c (struct mem_ref.refs):
18932 Change type to hash_table. Update dependent calls and types.
18933
18934 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
18935 Change type to hash_table. Update dependent calls and types.
18936
18937 * tree-ssa-sccvn.c (vn_tables_s::nary):
18938 Change type to hash_table. Update dependent calls and types.
18939 (vn_tables_s::phis): Likewise.
18940 (vn_tables_s::references): Likewise.
18941
18942 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
18943 (vn_reference_eq): Update parameter and return types.
18944
18945 * tree-ssa-structalias.c (pointer_equiv_class_table):
18946 Change type to hash_table. Update dependent calls and types.
18947 (location_equiv_class_table): Likewise.
18948
18949 * tree-vect-data-refs.c: Consequential changes for making
18950 peeling a hash_table.
18951
18952 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
18953 (destroy_loop_vec_info): Dependent hash_table update.
18954
18955 * tree-vectorizer.h (peeling_htab):
18956 Change type to hash_table. Update dependent calls and types.
18957
d70a81dd
SC
189582013-04-23 Shiva Chen <shiva0217@gmail.com>
18959
cabf91cd
AM
18960 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
18961 to check the register content is equal or not.
18962 * lra-constraints.c (match_reload): Use lra_assign_reg_val
18963 to assign register content record.
18964 * lra-eliminations.c (update_reg_eliminate): Use
d70a81dd 18965 lra_update_reg_val_offset to update register content offset.
cabf91cd
AM
18966 * lra-int.h (struct lra_reg): Add offset member.
18967 (lra_reg_val_equal_p): New static inline function.
18968 (lra_update_reg_val_offset): New static inline function.
18969 (lra_assign_reg_val): New static inline function.
18970 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
18971 to assign register content record.
18972 (initialize_lra_reg_info_element): Initial offset to zero.
d70a81dd 18973
b894a1f3
CM
189742013-04-23 Catherine Moore <clm@codesourcery.com>
18975
18976 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
18977 operands. Record compression.
18978
ecd14de9
XDL
189792013-04-23 Xinliang David Li <davidxl@google.com>
18980
18981 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
18982
92e776e9
RB
189832013-04-23 Richard Biener <rguenther@suse.de>
18984
18985 PR middle-end/57036
18986 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
18987 parameter, only add abnormal goto edges from the copied body
18988 if the call could perform abnormal gotos.
18989 (copy_cfg_body): Adjust.
18990
a15ee567
SN
189912013-04-23 Sofiane Naci <sofiane.naci@arm.com>
18992
18993 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
18994
08c52234
AS
189952013-04-23 Andreas Schwab <schwab@linux-m68k.org>
18996
18997 * coretypes.h (gimple_stmt_iterator): Add struct to make
18998 compatible with C.
18999
999c1171
RB
190002013-04-23 Richard Biener <rguenther@suse.de>
19001
19002 PR tree-optimization/57026
19003 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
19004 from SSA names occuring in abnormal PHI nodes.
19005
53669259
AK
190062013-04-22 Andi Kleen <ak@linux.intel.com>
19007
19008 * lto/lto.c (print_lto_report_1): Fix LTO report names.
19009
1a0ad150
AK
190102013-04-22 Andi Kleen <ak@linux.intel.com>
19011
19012 * lto/lto.c (print_lto_report_1): Declare early.
19013 (read_cgraph_and_symbols): Call print_lto_report_1 early.
19014
057f8f20
AK
190152013-04-22 Andi Kleen <ak@linux.intel.com>
19016
19017 * common.opt (-flto-report-wpa): Add.
19018 * doc/invoke.texi (-flto-report-wpa): Add.
19019 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
19020 (lto_main): dito.
19021
473b1e05
XDL
190222013-04-22 Xinliang David Li <davidxl@google.com>
19023
19024 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
19025 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
19026 * Makefile.in: New dependency
19027
cabf91cd 19028 David Daney <ddaney.cavm@gmail.com>
b1485a33 19029
cabf91cd
AM
19030 * configure.ac (gcc_cv_as_micromips_support): Use the
19031 --fatal-warnings option.
19032 * configure: Regenerate.
b1485a33 19033
829d0168
MP
190342013-04-22 Marek Polacek <polacek@redhat.com>
19035
19036 PR sanitizer/56990
19037 * tsan.c (instrument_expr): Don't instrument expression
19038 in case its size is zero.
19039
6d9b7208
UB
190402013-04-22 Uros Bizjak <ubizjak@gmail.com>
19041
19042 PR target/57032
19043 Revert:
19044 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19045
19046 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19047
ea679d55
JG
190482013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19049
19050 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19051 (gimple_stmt_iterator): New typedef.
19052 * gimple.h (gimple_stmt_iterator): Rename to...
19053 (gimple_stmt_iterator_d): ... This.
19054 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19055 trees be valid for GIMPLE and GENERIC.
19056 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19057 * gimple-fold.c (gimple_fold_call): Call target hook
19058 gimple_fold_builtin.
19059 * hooks.c (hook_bool_gsiptr_false): New.
19060 * hooks.h (hook_bool_gsiptr_false): New.
19061 * target.def (fold_stmt): New.
19062 * doc/tm.texi: Regenerate.
19063
88a581da
VM
190642013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19065
19066 PR target/57018
19067 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19068 a set sp if no stack realignment.
19069
92be22dc
NC
190702013-04-22 Nick Clifton <nickc@redhat.com>
19071
19072 * config.gcc (tilegx-linux): Extend extra_objs rather than
19073 overwriting it.
19074 (tilepro-linux): Likewise.
19075
0ddec79f
JG
190762013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19077
19078 * config/aarch64/aarch64-builtins.c
19079 (CF): Remove.
19080 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19081 (VAR<1-12>): Add MAP parameter.
19082 (BUILTIN_*): Likewise.
19083 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19084 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19085 (aarch64_ushl_n<mode>): Likewise.
19086 (aarch64_sshr_n<mode>): Likewise.
19087 (aarch64_ushr_n<mode>): Likewise.
19088 (aarch64_<maxmin><mode>): Likewise.
19089 (aarch64_sqrt<mode>): Likewise.
19090 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19091 (vshr<q>_n_*): Likewise.
19092
0050faf8
JG
190932013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19094
19095 * config/aarch64/aarch64-builtins.c
19096 (aarch64_simd_builtin_type_mode): Handle SF types.
19097 (sf_UP): Define.
19098 (BUILTIN_GPF): Define.
19099 (aarch64_init_simd_builtins): Handle SF types.
19100 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19101 (frecps): Likewise.
19102 (frecpx): Likewise.
19103 * config/aarch64/aarch64-simd.md
19104 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19105 (aarch64_frecpe<mode>): New.
19106 (aarch64_frecps<mode>): Likewise.
19107 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19108 (v8type): Add frecp<esx>.
19109 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19110 (aarch64_frecps<mode>): Likewise.
19111 * config/aarch64/iterators.md (FRECP): New.
19112 (frecp_suffix): Likewise.
19113 * config/aarch64/arm_neon.h
19114 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19115
0fad3dbc 191162013-04-22 Christian Bruel <christian.bruel@st.com>
1818d01c
CB
19117
19118 PR target/56995
19119 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19120 (REG_CLASS_NAMES): Idem.
19121 (REG_CLASS_CONTENTS): Idem.
19122 (REGCLASS_HAS_FP_REG): Idem.
19123 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19124 (sh_conditional_register_usage): Idem.
19125
3e8a33f9
JL
191262013-04-21 Jeff Law <law@redhat.com>
19127
19128 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19129 (ssa_forward_propagate_and_combine): Use it.
19130
f38e1b0a
VM
191312013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19132
19133 * lra.c: Update the flow chart diagram.
19134
682303da
VM
191352013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19136
19137 PR rtl-optimization/56847
19138 * lra-constraints.c (process_alt_operands): Discourage alternative
19139 with non-matche doffsettable memory constraint fro memory with
19140 known offset.
19141
f6b64c35
RB
191422013-04-19 Richard Biener <rguenther@suse.de>
19143
19144 PR tree-optimization/56982
19145 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19146 function.
19147 * gimplify.c (gimplify_call_expr): Notice special calls.
19148 (gimplify_modify_expr): Likewise.
19149 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19150 abnormal control flow receivers.
19151 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19152 in the same way as cfun->has_nonlocal_labels.
19153 (gimple_purge_dead_abnormal_call_edges): Likewise.
19154 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19155 receivers start a basic-block.
19156
01d8bf07
RB
191572013-04-19 Richard Biener <rguenther@suse.de>
19158
19159 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19160 member ...
19161 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19162 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19163 (SLP_TREE_LOAD_PERMUTATION): Add.
19164 (vect_transform_slp_perm_load): Adjust prototype.
19165 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19166 (vect_free_slp_instance): Likewise.
19167 (vect_create_new_slp_node): Likewise.
19168 (vect_supported_slp_permutation_p): Remove.
19169 (vect_slp_rearrange_stmts): Adjust.
19170 (vect_supported_load_permutation_p): Likewise. Inline
19171 vect_supported_slp_permutation_p here.
19172 (vect_analyze_slp_instance): Compute load permutations per
19173 slp node instead of per instance.
19174 (vect_get_slp_defs): Adjust.
19175 (vect_transform_slp_perm_load): Likewise.
19176 (vect_schedule_slp_instance): Remove redundant code.
19177 (vect_schedule_slp): Remove hack for PR56270, add it ...
19178 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19179 CSE loads for SLP. Adjust.
19180
ede22fc3
GY
191812013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19182
19183 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19184 spelling in two comments.
19185
67bc84fb
GY
191862013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19187
19188 PR target/56797
19189 * config/arm/arm.c (load_multiple_sequence): Require SP
19190 as base register for loads if SP is in the register list.
19191
e248d83f
MJ
191922013-04-19 Martin Jambor <mjambor@suse.cz>
19193
19194 PR tree-optimization/56718
19195 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19196 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19197 and made public. Adjusted all callers.
19198 (ipa_intraprocedural_devirtualization): New function.
19199 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19200 (ipa_intraprocedural_devirtualization): Likewise.
19201 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19202
4891e8f8
RB
192032013-04-19 Richard Biener <rguenther@suse.de>
19204
19205 PR tree-optimization/57000
19206 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19207
dad89f7c
TG
192082013-04-19 Terry Guo <terry.guo@arm.com>
19209
19210 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19211 Replace with ...
19212 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19213 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19214 (cortex_m4_fmacs): Use new reservations.
19215 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19216
72ea0d47
VM
192172013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19218
f1e6512c 19219 PR rtl-optimization/56999
72ea0d47
VM
19220 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19221 related code.
19222 (lra_coalesce): Remove split_origin_bitmap and related code.
19223 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19224 ranges if necessary.
19225
780a5b71
UB
192262013-04-18 Uros Bizjak <ubizjak@gmail.com>
19227
19228 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19229 New array.
19230 (ix86_expand_call): Remove clobbered_registers array and use
19231 x86_64_ms_sysv_extra_clobbered_registers instead.
19232 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19233 Declare here.
19234 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19235 predicate.
19236 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19237 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19238 (*call_value_rex64_ms_sysv): Ditto.
19239
6f5a366a
CC
192402013-04-18 Cary Coutant <ccoutant@google.com>
19241
19242 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19243 parent instead of die_mark.
19244
475b8f37
DN
192452013-04-18 Diego Novillo <dnovillo@google.com>
19246
19247 * gimple.c (create_gimple_tmp): New.
19248 (get_expr_type): New.
19249 (build_assign): New.
19250 (build_type_cast): New.
19251 * gimple.h (enum ssa_mode): Define.
19252 (gimple_seq_set_location): New.
19253 * asan.c (build_check_stmt): Change some gimple_build_* calls
780a5b71 19254 to use build_assign and build_type_cast.
475b8f37 19255
08940f33
RB
192562013-04-18 Richard Biener <rguenther@suse.de>
19257
19258 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19259 handle negative step. Remove redundant checks.
19260 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19261 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19262 for negative step and grouped loads fail to vectorize.
19263
0e0f87d4
SB
192642013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19265
19266 * emit-rtl.c (reset_insn_used_flags): New function.
19267 (reset_all_used_flags): Use it.
19268 (verify_insn_sharing): New function.
19269 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19270
4c445590
JJ
192712013-04-18 Jakub Jelinek <jakub@redhat.com>
19272
19273 PR tree-optimization/56984
19274 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19275 and (x >> M) >= N don't register any assertion if N << M is the
19276 minimum value.
19277
6873ecab
SB
192782013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19279
19280 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19281 do not delete_insn insns that have not yet been emitted, only
19282 unlink them with remove_insn.
19283 * df-scan.c (df_insn_delete): Revert r197492.
19284
3ccb989e
SB
192852013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19286
19287 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19288 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19289
192902013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
0480fff0
GY
19291
19292 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19293 define_insn_and_split.
19294 (and_scc,ior_scc,negscc): Likewise.
19295 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19296
3ccb989e 192972013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
367075fe
GY
19298
19299 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19300 can benefit from using a sequence of LDRD instructions in epilogue
19301 instead of a single LDM instruction.
19302
6d10a203
MLI
193032013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19304
19305 PR 45688
19306 * doc/extend.texi: Fix typo.
19307
6983e6b5
RB
193082013-04-17 Richard Biener <rguenther@suse.de>
19309
19310 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19311 (vect_build_slp_tree): ... here.
19312 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19313 match. Remove special-casing of mismatched complex loads.
19314 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19315 re-try the match with swapped commutative operands.
19316 (vect_supported_load_permutation_p): Remove special-casing of
19317 mismatched complex loads.
19318 (vect_analyze_slp_instance): Adjust.
19319
ef23e6a2
RB
193202013-04-17 Richard Biener <rguenther@suse.de>
19321
19322 PR rtl-optimization/56921
19323 * cfgloop.h (struct loop): Add simple_loop_desc member.
19324 (struct niter_desc): Mark with GTY(()).
19325 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19326 * loop-iv.c (get_simple_loop_desc): Likewise.
19327 (free_simple_loop_desc): Likewise.
19328
19329 Revert
19330 2013-04-16 Richard Biener <rguenther@suse.de>
19331
19332 PR rtl-optimization/56921
19333 * loop-init.c (pass_rtl_move_loop_invariants): Add
19334 TODO_do_not_ggc_collect to todo_flags_finish.
19335 (pass_rtl_unswitch): Same.
19336 (pass_rtl_unroll_and_peel_loops): Same.
19337 (pass_rtl_doloop): Same.
19338
fc6f94f5
EB
193392013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19340
19341 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19342 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19343 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19344 references.
19345 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19346 * tree-streamer.c (record_common_node): Adjust reference in comment.
19347
10a88311
TG
193482013-04-17 Terry Guo <terry.guo@arm.com>
19349
19350 * config/arm/cortex-m4.md: Add a new bypass.
19351
6d9b7208 193522013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f3d096b0
N
19353
19354 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19355 New pattern.
19356 (*subs_<optab><mode>_multp2): New pattern.
19357 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19358 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19359
6d9b7208 193602013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
860ad33f
N
19361
19362 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19363 (*subs_mul_imm_<mode>): New pattern.
19364
18a6701e
DE
193652013-04-16 David Edelsohn <dje.gcc@gmail.com>
19366
19367 PR target/56948
19368 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19369 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19370 (vsx_movti_32bit): Change j->wa to O->wa.
19371
07c37b2f
RB
193722013-04-16 Richard Biener <rguenther@suse.de>
19373
19374 PR rtl-optimization/56921
19375 * loop-init.c (pass_rtl_move_loop_invariants): Add
19376 TODO_do_not_ggc_collect to todo_flags_finish.
19377 (pass_rtl_unswitch): Same.
19378 (pass_rtl_unroll_and_peel_loops): Same.
19379 (pass_rtl_doloop): Same.
19380
0e0f87d4 193812013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
974ba22d
GY
19382
19383 * config/arm/arm.c (emit_multi_reg_push): New declaration
19384 for an existing function.
19385 (arm_emit_strd_push): New function.
19386 (arm_expand_prologue): Used here.
19387 (arm_emit_ldrd_pop): New function.
19388 (arm_expand_epilogue): Used here.
19389 (arm_get_frame_offsets): Update condition.
19390 (arm_emit_multi_reg_pop): Add a special case for load of a single
19391 register with writeback.
19392
5e8e2af4
UB
193932013-04-16 Uros Bizjak <ubizjak@gmail.com>
19394
19395 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19396 description.
19397
9fd9ccf7
RB
193982013-04-16 Richard Biener <rguenther@suse.de>
19399
19400 PR tree-optimization/56756
19401 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19402 (first_mem_ref_loc): New.
19403 (execute_sm): Place the load temporarily before a previous
19404 access instead of in the latch edge to ensure its SSA dependencies
19405 are defined at points dominating the load.
19406
96fba521
SB
194072013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19408
4c8af858
SB
19409 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19410 correct fix by moving header and footer insn to the footer of
19411 the merged basic block. Clear BB_END of the merged-away block.
19412
96fba521
SB
19413 PR middle-end/43631
19414 * emit-rtl.c (make_note_raw): New function.
19415 (link_insn_into_chain): New static inline function.
19416 (add_insn): Use it.
19417 (add_insn_before, add_insn_after): Factor insn chain linking code...
19418 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19419 using link_insn_into_chain.
19420 (note_outside_basic_block_p): New helper function for emit_note_after
19421 and emit_note_before.
19422 (emit_note_after): Use nobb variant of add_insn_after if the note
19423 should not be contained in a basic block.
19424 (emit_note_before): Use nobb variant of add_insn_before if the note
19425 should not be contained in a basic block.
19426 (emit_note_copy): Use make_note_raw.
19427 (emit_note): Likewise.
19428 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19429 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19430 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19431 the moved barrier the tail of the basic block it follows.
19432 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19433
7b8265ba
JJ
194342013-04-15 Jakub Jelinek <jakub@redhat.com>
19435
19436 PR tree-optimization/56962
19437 * gimple-ssa-strength-reduction.c (record_increment): Only set
19438 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19439 either rhs1 or rhs2 is equal to c->base_expr.
19440
5185d248
RB
194412013-04-15 Richard Biener <rguenther@suse.de>
19442
19443 PR tree-optimization/56933
19444 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19445 member.
19446 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19447 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19448 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19449 dependence check ...
19450 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19451 ... here.
19452 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19453 GROUP_READ_WRITE_DEPENDENCE.
19454
a24243a0
AK
194552013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19456
19457 * emit-rtl.c (reset_all_used_flags): New function.
19458 (verify_rtl_sharing): Call reset_all_used_flags before and after
19459 performing the checks.
19460
1c50eada
KT
194612013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19462
19463 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19464 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19465 * config/arm/constraints.md (De): New constraint.
19466 * config/arm/neon.md (anddi3_neon): Delete.
19467 (neon_vand<mode>): Expand to standard anddi3 pattern.
19468 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19469 Move earlier in the file.
19470 (neon_inv_logic_op2): Likewise.
19471 (arm_anddi_operand_neon): New predicate.
19472
e927b6ad
RO
194732013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19474
19475 * configure.ac (gcc_cv_ld_as_needed): Set
19476 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19477 Use -z ignore, -z record on *-*-solaris2*.
19478 (HAVE_LD_AS_NEEDED): Update comment.
19479 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19480 * configure: Regenerate.
19481 * config.in: Regenerate.
19482 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19483 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19484 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19485 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19486 equivalents. Fix markup.
19487 * doc/tm.texi: Regenerate.
19488
e0ea8797
AH
194892013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19490
19491 * config/i386/i386.opt: New option mstack-protector-guard=.
19492 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19493 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19494 TARGET_SSP_TLS_GUARD.
19495 * config/i386/i386.c (ix86_option_override_internal): Set
19496 ix86_stack_protector_guard.
19497 * config/i386/i386.md (stack_protect_set): Enable for
19498 TARGET_SSP_TLS_GUARD only.
19499 (stack_protect_set_<mode>): Ditto.
19500 (stack_protect_test): Ditto.
19501 (stack_protect_test_<mode>): Ditto.
19502 * doc/invoke.texi (i386 Option): Document.
19503
811b72f9
EB
195042013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19505
19506 PR target/56890
19507 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19508 (S_MODES): Set H_MODE bit.
19509 (SF_MODES): Set only S_MODE and SF_MODE bits.
19510 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19511 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19512 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19513 <MODE_FLOAT>: Likewise.
19514
5529fdd6
JY
195152013-04-15 Joey Ye <joey.ye@arm.com>
19516
19517 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19518
517b1da2
JY
195192013-04-15 Joey Ye <joey.ye@arm.com>
19520
19521 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19522 for real far jump.
19523 (thumb_far_jump_used_p): Count instruction size and set
19524 far_jump_used.
19525
01007ae0
EB
195262013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19527
19528 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19529 * resource.c (mark_target_live_regs): Reformat conditional block.
19530
c46f6580
SB
195312013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19532
19533 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19534 notes, they are emitted only just before final.
19535 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19536
1f397f45
SB
195372013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19538
19539 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19540 * cfgrtl.c (delete_insn): Call it here instead.
19541 * lra-spills.c (lra_final_code_change): Use delete_insn.
19542 * haifa-sched.c (sched_remove_insn): Likewise.
19543 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19544 returning to the nop pool.
19545 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19546 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19547
58a51369
SB
195482013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19549
19550 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19551 * doc/tm.texi: Regenerated.
19552
33159866
UB
195532013-04-12 Uros Bizjak <ubizjak@gmail.com>
19554
19555 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19556 QImode checks.
19557
226e378f
SB
195582013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19559
19560 * df-core.c (df_find_def): Compare register numbers.
19561 (df_find_use): Likewise.
19562
fafb9b18
VM
195632013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19564
19565 PR target/56903
19566 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19567 lra_in_progress for return.
19568
9a946fd6
GY
195692013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19570
19571 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19572 define_insn into define_insn_and_split and emit movsicc patterns.
19573
41b83758
GY
195742013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19575
19576 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19577
d3afd9aa
RB
195782013-04-12 Richard Biener <rguenther@suse.de>
19579
19580 * tree-pass.h (TODO_do_not_ggc_collect): New.
19581 * passes.c (execute_one_ipa_transform_pass): Honor
19582 TODO_do_not_ggc_collect.
19583 (execute_one_pass): Likewise.
19584
19585 Revert
19586 2013-04-10 Richard Biener <rguenther@suse.de>
19587
19588 * passes.c (init_optimization_passes): Remove reload pass.
19589 * ira.c (do_reload): Merge into ...
19590 (ira): ... this.
19591 (rest_of_handle_reload): Remove.
19592 (pass_reload): Likewise.
19593 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19594 of reload for vzeroupper pass placement.
19595
06f9b387
JJ
195962013-04-12 Jakub Jelinek <jakub@redhat.com>
19597
19598 PR tree-optimization/56918
19599 PR tree-optimization/56920
19600 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19601 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19602 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19603 use wide_mul_with_sign method.
19604
953094d2
RB
196052013-04-12 Richard Biener <rguenther@suse.de>
19606
19607 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19608 not be considered a gimple constant.
19609
bb506982
MG
196102013-04-12 Marc Glisse <marc.glisse@inria.fr>
19611
19612 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19613 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19614 operations.
19615
4b84d650
JJ
196162013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19617 Jakub Jelinek <jakub@redhat.com>
19618
19619 * opts.c: Include diagnostic-color.h.
19620 (common_handle_option): Handle OPT_fdiagnostics_color_.
19621 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19622 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19623 (diagnostic-color.o): New.
19624 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19625 (diagnostic_color_rule): New enum.
19626 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19627 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19628 the location string.
19629 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19630 either NULL, or color kind.
19631 * diagnostic-color.c: New file.
19632 * diagnostic-color.h: New file.
19633 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19634 arguments.
19635 * doc/invoke.texi (-fdiagnostics-color): Document.
19636 * pretty-print.h (pp_show_color): Define.
19637 (struct pretty_print_info): Add show_color field.
19638 * diagnostic.c: Include diagnostic-color.h.
19639 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19640 macros. Colorize error:, warning: etc. strings and also the location
19641 string.
19642 (diagnostic_show_locus): Colorize the caret line.
19643 * pretty-print.c: Include diagnostic-color.h.
19644 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19645 inside of %< %> quotes or quoted through q format modifier.
19646
067a1e71
AK
196472013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19648
33159866 19649 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
067a1e71 19650
33159866 196512013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
b056c910
N
19652
19653 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19654 code in CC_NZ mode.
19655 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19656 pattern.
19657
7b55f98f
MP
196582013-04-11 Marek Polacek <polacek@redhat.com>
19659
19660 PR tree-optimization/48184
33159866 19661 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
7b55f98f 19662
966b587e
EB
196632013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19664
19665 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19666 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19667 (skip_simple_arithmetic): Tidy up.
19668 * tree.h (skip_simple_constant_arithmetic): Declare.
19669
33159866 196702013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
f8bf91ab
N
19671
19672 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19673
1d42c1ec
RB
196742013-04-11 Richard Biener <rguenther@suse.de>
19675
19676 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19677 generate vector constants.
19678
4ba5ea11
RB
196792013-04-11 Richard Biener <rguenther@suse.de>
19680
19681 PR tree-optimization/56878
19682 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19683 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19684 New function.
19685 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19686 Prefer to align the DR with the most invariant base address.
19687
f0defe58
SKS
196882013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19689
19690 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19691 comment.
19692
d07458be
JG
196932013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19694
19695 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19696 floating-point vector comparisons against 0.
19697
146b8692
JJ
196982013-04-11 Jakub Jelinek <jakub@redhat.com>
19699
19700 PR tree-optimization/56899
19701 * fold-const.c (extract_muldiv_1): Apply distributive law
19702 only if TYPE_OVERFLOW_WRAPS (ctype).
19703
b8578ff7
BC
197042013-04-11 Bin Cheng <bin.cheng@arm.com>
19705
19706 PR target/56124
19707 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19708 loading has side effect.
19709
0ea8a6f9
SB
197102013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19711
19712 * config/sparc/sparc.c: Include tree-pass.h.
19713 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19714 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19715 head of file. Change return type. Split off gate function.
19716 (sparc_gate_work_around_errata): New function.
19717 (pass_work_around_errata): New pass definition.
19718 (insert_pass_work_around_errata) New pass insert definition to
19719 insert pass_work_around_errata just after delayed-branch scheduling.
19720 (sparc_option_override): Insert the pass.
19721 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19722
42e37616
DM
197232013-04-10 David S. Miller <davem@davemloft.net>
19724
89deeb3b
DM
19725 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19726 or -mcpu=hypersparc.
19727
42e37616
DM
19728 * target.def (cstore_mode): New hook.
19729 * target.h: Include insn-codes.h
19730 * targhooks.c: Likewise.
19731 (default_cstore_mode): New function.
19732 * targhooks.h: Declare it.
19733 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19734 * doc/tm.texi: Rebuild.
19735 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19736 target hook, rather than inspecting the insn_data.
19737 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19738 (TARGET_CSTORE_MODE): Redefine.
19739 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19740 result patterns.
19741 * config/sparc/predicates.md (cstore_result_operand): New special
19742 predicate.
19743 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19744 Use it for operand 0.
19745 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19746 (*snesi_special): Likewise.
19747 (*snesi_zero): Likewise.
19748 (*seqsi_zero): Likewise.
19749 (*sltu_insn): Likewise.
19750 (*sgeu_insn): Likewise.
19751 (*seqdi_special): Make operand 0 and comparison operation be of
19752 DImode.
19753 (*snedi_special): Likewise.
19754 (*snedi_special_vis3): Likewise.
19755 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
19756 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
19757 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
19758 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
19759 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
19760 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
19761 (*sltu_extend_sp64): Likewise.
19762 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
19763 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
19764 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
19765 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
19766 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
19767
95ca411e
YZ
197682013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
19769
19770 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
19771 (aarch64_start_file): Use the new function.
19772
6782438d 197732013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
33159866 19774 Jason Merrill <jason@redhat.com>
6782438d
SKS
19775
19776 * common.opt: Add -gdwarf.
19777 * opts.c (common_handle_option): Handle it.
19778 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
19779
bb313b93
RB
197802013-04-10 Richard Biener <rguenther@suse.de>
19781
19782 * passes.c (execute_todo): Do not call ggc_collect conditional here.
19783 (execute_one_ipa_transform_pass): But unconditionally here.
19784 (execute_one_pass): And here.
19785 (init_optimization_passes): Remove reload pass.
19786 * tree-pass.h (TODO_ggc_collect): Remove.
19787 (pass_reload): Likewise.
19788 * ira.c (do_reload): Merge into ...
19789 (ira): ... this.
19790 (rest_of_handle_reload): Remove.
19791 (pass_reload): Likewise.
19792 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19793 of reload for vzeroupper pass placement.
19794 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
19795 and todo_flags_finish of all passes.
19796
793d9a16
RB
197972013-04-10 Richard Biener <rguenther@suse.de>
19798
19799 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
19800 first_const_oprnd field, rename first_def_type to first_op_type.
19801 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
19802 (vect_get_and_check_slp_defs): Always use the type of the
19803 operand. Allow mixed vect_external_def, vect_constant_def types.
19804 (vect_get_constant_vectors): Handle mixed vect_external_def,
19805 vect_constant_def types.
19806
12211b99 198072013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
ee8a9b7b
JR
19808
19809 PR tree-optimization/55524
19810 * tree-ssa-math-opts.c
19811 (convert_mult_to_fma): Don't use an fms construct
19812 when we don't have an fms operation, but fnma, and it looks
19813 likely that we'll be able to use the latter.
19814
12211b99 198152013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
6957a6f6
ZZ
19816
19817 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
19818 function.
19819 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
19820 inline fail caused by overwritable functions.
19821
34ab4a5b
CJW
198222013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
19823
19824 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
19825 unnecessary bits in the constant power of two case.
19826
abf9bfbc
RB
198272013-04-10 Richard Biener <rguenther@suse.de>
19828
19829 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
19830 broken code swapping operands.
19831 (vect_build_slp_tree): Do not compute load permutations here.
19832 (vect_analyze_slp_instance): Compute load permutations here,
19833 after building the SLP tree.
19834
f408477e
CB
198352013-04-09 Christian Bruel <christian.bruel@st.com>
19836
19837 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
19838 of next/prev_real_insn.
19839
5ac42672
JH
198402013-04-09 Jan Hubicka <jh@suse.cz>
19841
abf9bfbc
RB
19842 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19843 Drop aliased parameter.
5ac42672
JH
19844 (function_and_variable_visibility): Do not handle alias pairs.
19845 * cgraph.c (varpool_externally_visible_p): Update prototype.
19846 * varpool.c (varpool_add_new_variable): Update.
19847
5017f1d2
KT
198482013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19849
19850 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
19851
48eecbee
SB
198522013-04-09 Steven Bosscher <steven@gcc.gnu.org>
19853
a949cf1c
SB
19854 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
19855
48eecbee
SB
19856 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
19857
75ef8e3d
MP
198582013-04-09 Marek Polacek <polacek@redhat.com>
19859
19860 PR tree-optimization/48762
33159866 19861 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
75ef8e3d 19862
23847df4
RB
198632013-04-09 Richard Biener <rguenther@suse.de>
19864
19865 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
19866 dealing with cost.
19867 (vect_build_slp_tree): Likewise.
19868 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
19869 calculating the cost of a SLP instance.
19870 (vect_analyze_slp_instance): Use it from here, after building
19871 the SLP tree.
19872
ea3a0fde
JJ
198732013-04-09 Jakub Jelinek <jakub@redhat.com>
19874
19875 PR middle-end/56883
19876 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
19877 expand_omp_for_static_chunk): Use simple_p = true in
19878 force_gimple_operand_gsi calls when assigning to addressable decls.
19879
a32dfe9d
JL
198802013-04-09 Jeff Law <law@redhat.com>
19881
19882 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
19883 when the boolean was created by converting a wider object which
19884 had a boolean range.
19885
d755c7ef
RB
198862013-04-09 Richard Biener <rguenther@suse.de>
19887
19888 * tree-vectorizer.h (slp_void_p): Remove.
19889 (slp_tree): Typedef before _slp_tree declaration.
19890 (struct _slp_tree): Use a vector of slp_tree as children.
19891 (vect_get_place_in_interleaving_chain): Remove.
19892 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
19893 Move ...
19894 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
19895 and make static.
19896 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
19897 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
19898 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
19899 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
19900 Use slp_node instead of slp_void_p and adjust.
19901
3d741091
RB
199022013-04-09 Richard Biener <rguenther@suse.de>
19903
19904 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
19905 work that is not necessary.
19906
39307ba7
JJ
199072013-04-09 Jakub Jelinek <jakub@redhat.com>
19908
19909 PR tree-optimization/56854
19910 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
19911 forward into clobber stmts if it would change MEM_REF lhs into
19912 non-MEM_REF.
19913
343881fd
MK
199142013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
19915
19916 * tree.c (type_hash_lookup, type_hash_add): Make static.
19917 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
19918
3922658a
RB
199192013-04-09 Richard Biener <rguenther@suse.de>
19920
19921 * tree.h (unsave_expr_now): Remove.
19922 * tree-inline.c (mark_local_for_remap_r): Remove.
19923 (unsave_expr_1): Likewise.
19924 (unsave_r): Likewise.
19925 (unsave_expr_now): Likewise.
19926 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
19927 (propagate_tree_value): Likewise.
19928
9fb6b620
SB
199292013-04-08 Steven Bosscher <steven@gcc.gnu.org>
19930
19931 * doc/rtl.texi (sequence): Rewrite documentation to match the
19932 current use of SEQUENCE rtl objects.
19933 * rtl.def (SEQUENCE): Likewise.
19934
19935 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
19936 Update documentation.
19937 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
19938 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
19939
19940 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
19941
8ddb5a29
TJ
199422013-04-08 Teresa Johnson <tejohnson@google.com>
19943
19944 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
19945 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
33159866 19946 methods.
8ddb5a29
TJ
19947 (estimate_edge_size_and_time): Add comment to suggest using rounding
19948 methods.
19949 (estimate_node_size_and_time): Ditto.
19950 (remap_edge_change_prob): Use helper rounding divide methods.
19951 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
19952 (gimple_mod_pow2_value_transform): Ditto.
19953 (gimple_mod_subtract_transform): Ditto.
19954 (gimple_ic_transform): Ditto.
19955 (gimple_stringops_transform): Ditto.
19956 * stmt.c (conditional_probability): Ditto.
19957 (emit_case_dispatch_table): Ditto.
19958 * lto-cgraph.c (merge_profile_summaries): Ditto.
19959 * tree-optimize.c (execute_fixup_cfg): Ditto.
19960 * cfgcleanup.c (try_forward_edges): Ditto.
19961 * cfgloopmanip.c (scale_loop_profile): Ditto.
19962 (loopify): Ditto.
19963 (duplicate_loop_to_header_edge): Ditto.
19964 (lv_adjust_loop_entry_edge): Ditto.
19965 * tree-vect-loop.c (vect_transform_loop): Ditto.
19966 * profile.c (compute_branch_probabilities): Ditto.
19967 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
19968 * lto-streamer-in.c (input_cfg): Ditto.
19969 * gimple-streamer-in.c (input_bb): Ditto.
19970 * ipa-cp.c (update_profiling_info): Ditto.
19971 (update_specialized_profile): Ditto.
19972 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
19973 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
33159866 19974 rounding methods.
8ddb5a29
TJ
19975 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19976 (compute_trg_info): Ditto.
19977 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
19978 (purge_dead_edges): Ditto.
19979 * loop-unswitch.c (unswitch_loop): Ditto.
19980 * cgraphclones.c (cgraph_clone_edge): Ditto.
19981 (cgraph_clone_node): Ditto.
19982 * tree-inline.c (copy_bb): Ditto.
19983 (copy_edges_for_bb): Ditto.
19984 (initialize_cfun): Ditto.
19985 (copy_cfg_body): Ditto.
19986 (expand_call_inline): Ditto.
19987
661e6bd7
KT
199882013-04-08 Kai Tietz <ktietz@redhat.com>
19989
19990 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
19991 TARGET_CYGWIN64 by TARGET_64BIT.
19992
105766f3
JR
199932013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19994
19995 * config/epiphany/epiphany.md (GPR_1): New constant.
19996 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
19997 * config/epiphany/epiphany.c (gen_compare_reg):
19998 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
19999 is already in place.
20000 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
20001 Don't require being called during rtl expansion; If y operlaps r0,
20002 return 0.
20003 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
20004 (epiphany_expand_epilogue): Likewise.
20005
8afab237
JR
20006 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
20007 Don't use CC_FPmode for ORDERED / UNORDERED.
33159866 20008 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
8afab237 20009
fa7707d6
JR
20010 * config/epiphany/constraints.md (CnL): New constraint.
20011 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
20012 * config/epiphany/predicates.md (add_operand): Allow 1024.
20013
5138e18d
JR
20014 * config/epiphany/epiphany.md (logical_op): New code iterator.
20015 (op_mnc): New code attribute.
20016 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
20017 (mov_f+1, mov_f+2): New peephole2 patterns.
20018
2ccc703d
JR
20019 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
20020 (cstoresi4): Also allow re-use of zero result when doing a NE
20021 comparison to a non-zero operand.
aefb0819 20022 Use (clobber (scratch)) for first insn if the gpr output is not needed.
2ccc703d 20023
093ac0a5
JR
20024 * config/epiphany/epiphany.md (<insn_opname>v2si3):
20025 Use gen_addsi3_i / gen_subsi3_i.
20026
f223bb13
JJ
200272013-04-08 Jakub Jelinek <jakub@redhat.com>
20028
20029 PR c++/34949
20030 PR c++/50243
20031 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
20032 contain anything but clobbers, at most one __builtin_stack_restore,
20033 optionally debug stmts and final resx, and if it has at least one
20034 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20035 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20036 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20037 which isn't defaut definition, remove them.
20038 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20039 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20040 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20041 with MEM_REF LHS with SSA_NAME address.
20042
4481581f
JL
200432013-04-08 Jeff Law <law@redhat.com>
20044
20045 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
a32dfe9d 20046
451dabda
RB
200472013-04-08 Richard Biener <rguenther@suse.de>
20048
20049 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20050 extra newline.
20051 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20052 determined vector type.
20053 (vect_analyze_data_refs): Likewise.
20054 (vect_get_new_vect_var): Adjust.
20055 (vect_create_destination_var): Preserve SSA name versions.
20056 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20057 not dump anything here.
20058
3b088b47
JR
200592013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20060
20061 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20062 Add member lr_slot_known.
20063 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20064 if necessary.
20065 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20066 Remove code that sets lr_slot_offset according to what a previous
20067 version of epiphany_emit_save_restore used to do.
20068 (epiphany_emit_save_restore): When doing an lr save or restore,
20069 set/verify lr_slot_known and lr_slot_offset.
20070
d8484d41
XQ
200712013-04-08 Xinyu Qi <xyqi@marvell.com>
20072
33159866
UB
20073 PR target/54338
20074 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
d8484d41
XQ
20075 in ALL_REGS.
20076
481be1c4
RB
200772013-04-08 Richard Biener <rguenther@suse.de>
20078
20079 * alias.c (find_base_term): Fix thinko in previous change.
20080
401f3a81
JJ
200812013-04-08 Jakub Jelinek <jakub@redhat.com>
20082
20083 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20084 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20085 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20086 if possible to compute val.
20087 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20088 For QImode integers don't require anything about precision. Use
20089 const_with_all_bytes_same to find out if the constant doesn't have
20090 repeated bytes in it.
20091
03ed99a8
AK
200922013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20093
20094 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20095 within mode size.
20096
781b2e62
MP
200972013-04-08 Marek Polacek <polacek@redhat.com>
20098
20099 PR rtl-optimization/48182
20100 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20101 value to 1.
20102
27e430a2
JDA
201032013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20104
20105 PR target/55487
20106 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20107 nuses, make sure we have a label.
20108
4902aa64
BS
201092013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20110
20111 PR target/56843
20112 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20113 (rs6000_emit_swdiv_low_precision): Remove.
20114 (rs6000_emit_swdiv): Rewrite to handle between one and four
20115 iterations of Newton-Raphson generally; modify required number of
20116 iterations for some cases.
20117 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20118
7bca81dc
SB
201192013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20120
20121 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20122 set-but-unused variable.
20123
20124 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20125 basic blocks of released function bodies garbage-collectable.
20126
20127 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20128 (struct rtl_opt_pass): Add TODO_df_finish.
20129
20130 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20131
4542a38a
GY
201322013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20133
20134 * config/arm/constraints.md (q): New constraint.
20135 * config/arm/ldrdstrd.md: New file.
20136 * config/arm/arm.md (ldrdstrd.md) New include.
20137 (arm_movdi): Use "q" instead of "r" constraint
20138 for double-word memory access.
20139 (movdf_soft_insn): Likewise.
20140 * config/arm/vfp.md (movdi_vfp): Likewise.
20141 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
33159866 20142 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
4542a38a
GY
20143 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20144 (mem_ok_for_ldrd_strd): Likewise.
20145 (output_move_double): Update assertion.
20146
2385b218
GY
201472013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20148
20149 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20150
75fe1cb5
GY
201512013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20152
20153 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20154 define_insn_and_split.
20155 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20156
dd660e8e
GY
201572013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20158
20159 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20160 define_insn_and_split.
33159866 20161 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
dd660e8e
GY
20162 (shiftsi3_compare): New pattern.
20163 (rrx): New pattern.
20164 * config/arm/unspecs.md (UNSPEC_RRX): New.
20165
045e472c
GY
201662013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20167
20168 * config/arm/arm.md (negdi_extendsidi): New pattern.
20169 (negdi_zero_extendsidi): Likewise.
20170
3f3bf1a8
GY
201712013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20172
20173 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20174 define_insn_and_split.
20175 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20176 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20177
b0b49556
GY
201782013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20179
20180 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20181 define_insn_and_split.
20182 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20183 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20184
d633dd84
GY
201852013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20186
20187 * config/arm/arm.md (subsi3_carryin): New pattern.
20188 (subsi3_carryin_const): Likewise.
20189 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20190 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20191
f4499066
GY
201922013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20193
20194 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20195
ceef6fd9
GY
201962013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20197
20198 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
d633dd84 20199 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
ceef6fd9 20200
ddbdd8a7
KT
202012013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20202
20203 * config/arm/arm.c (arm_expand_builtin): Change fcode
20204 type to unsigned int.
20205
8456d78a
RR
202062013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20207
20208 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20209
526c230f
ILT
202102013-04-04 Ian Lance Taylor <iant@google.com>
20211
33159866
UB
20212 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20213 language standard.
526c230f 20214
3cfbe04d
SB
202152013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20216
20217 PR middle-end/56729
20218 * df-scan.c (df_insn_delete): Disable failing assert.
20219
dfa3f8d0
KT
202202013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20221
20222 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20223 New function prototype.
20224 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20225 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20226 (arm_builtin_vectorized_function): New function.
20227
f7837758
KT
202282013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20229
20230 * config/arm/arm_neon_builtins.def: New file.
20231 * config/arm/arm.c (neon_builtin_data): Move contents to
20232 arm_neon_builtins.def.
20233 (enum arm_builtins): Include neon builtin definitions.
20234 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
33159866 20235 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
f7837758 20236
39c1b6db
MP
202372013-04-04 Marek Polacek <polacek@redhat.com>
20238
20239 PR tree-optimization/48186
20240 * predict.c (maybe_hot_frequency_p): Return false if
20241 HOT_BB_FREQUENCY_FRACTION is 0.
20242 (cgraph_maybe_hot_edge_p): Likewise.
20243
314f64eb
RB
202442013-04-04 Richard Biener <rguenther@suse.de>
20245
20246 PR tree-optimization/56826
20247 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20248 more accurately.
20249
51a905b2
RB
202502013-04-04 Richard Biener <rguenther@suse.de>
20251
20252 PR tree-optimization/56213
20253 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
33159866 20254 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
51a905b2 20255
f1bf4f3a
RB
202562013-04-04 Richard Biener <rguenther@suse.de>
20257
20258 PR tree-optimization/56837
20259 * tree-loop-distribution.c (classify_partition): For non-zero
20260 values require that the value has the same precision as its
20261 mode to be useful as memset value.
20262
0bca7ded
NC
202632013-04-03 Nick Clifton <nickc@redhat.com>
20264
33159866 20265 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
0bca7ded
NC
20266 (fmssf4): Use fmsf.s on E3V5 architectures.
20267 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20268 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20269
b4019227
JL
202702013-04-03 Jeff Law <law@redhat.com>
20271
20272 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20273 (lra-eliminations.o): Likewise.
20274
f57ddb5b
TJ
202752013-04-03 Teresa Johnson <tejohnson@google.com>
20276
20277 * gcov-io.c (compute_working_sets): Moved most of body of old
33159866 20278 compute_working_sets here from profile.c.
f57ddb5b 20279 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
33159866
UB
20280 (gcov_working_set_t): Moved typedef here from basic-block.h
20281 (compute_working_set): Declare.
f57ddb5b
TJ
20282 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20283 (get_working_sets): Renamed from compute_working_set,
33159866 20284 replace most of body with call to new compute_working_sets.
f57ddb5b 20285 (get_exec_counts): Replace call to compute_working_sets
33159866
UB
20286 to get_working_sets.
20287 * profile.h (get_working_sets): Renamed from compute_working_set.
f57ddb5b 20288 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
33159866 20289 to get_working_sets.
f57ddb5b
TJ
20290 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20291 * gcov-dump.c (dump_working_sets): New function.
20292
12211b99 202932013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
5e0919f1
KZ
20294
20295 * hwint.c (sext_hwi, zext_hwi): New functions.
20296 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20297 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20298 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20299 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20300 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20301 (sext_hwi, zext_hwi): New functions.
0bca7ded 20302
be672e08
JL
203032013-04-03 Jeff Law <law@redhat.com>
20304
20305 PR tree-optimization/56799
33159866
UB
20306 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20307 back test for widening conversion erroneously dropped in prior change.
be672e08 20308
9d821fa5
KT
203092013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20310
20311 PR target/56809
20312 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20313 instead of next_real_insn.
20314
71c581e7
MP
203152013-04-03 Marek Polacek <polacek@redhat.com>
20316
20317 PR sanitizer/55702
33159866 20318 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
71c581e7 20319
4a32ef80
KT
203202013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20321
20322 PR target/56809
20323 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20324 next_real_insn.
20325 (thumb1_output_casesi): Likewise.
20326 (thumb2_output_casesi): Likewise.
20327
1b2253d4
RB
203282013-04-03 Richard Biener <rguenther@suse.de>
20329
20330 PR tree-optimization/56817
20331 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20332 Split out ...
20333 (tree_unroll_loops_completely_1): ... new function to manually
20334 walk the loop tree, properly defering outer loops of unrolled
20335 loops to later iterations.
20336
38000232
MG
203372013-04-03 Marc Glisse <marc.glisse@inria.fr>
20338
20339 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20340 (vectorizable_load): Likewise.
20341 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20342 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20343
3fa35298
MG
203442013-04-03 Marc Glisse <marc.glisse@inria.fr>
20345
20346 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20347 BIT_FIELD_REF.
20348
b3d45ff0
UW
203492013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20350
20351 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20352
ec9202a8
BC
203532013-04-03 Bin Cheng <bin.cheng@arm.com>
20354
20355 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20356
6805bd36
MG
203572013-04-03 Marc Glisse <marc.glisse@inria.fr>
20358
20359 PR tree-optimization/56790
33159866
UB
20360 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20361 folding.
6805bd36 20362
da694a77
MG
203632013-04-03 Marc Glisse <marc.glisse@inria.fr>
20364
20365 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20366 Handle VEC_MERGE.
20367 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20368 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20369 equal arguments.
20370
4e7d7b3d
JJ
203712013-04-03 Jakub Jelinek <jakub@redhat.com>
20372
20373 PR c/19449
20374 * tree.h (force_folding_builtin_constant_p): New decl.
20375 * builtins.c (force_folding_builtin_constant_p): New variable.
20376 (fold_builtin_constant_p): Fold immediately also if
20377 force_folding_builtin_constant_p.
20378
e6c9d234
RB
203792013-04-03 Richard Biener <rguenther@suse.de>
20380
20381 PR tree-optimization/56812
20382 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20383 DRs of the same interleaving chain are independent.
20384
984d07dd
JM
203852013-04-02 Jason Merrill <jason@redhat.com>
20386
20387 * gdbinit.in (pbb): Use debug fn.
20388
622849c9
LC
203892013-04-02 Lawrence Crowl <crowl@google.com>
20390
20391 * sese.h (struct ivtype_map_elt_s): Remove unused.
20392 (extern debug_ivtype_map): Remove unused.
20393 (extern eq_ivtype_map_elts): Remove unused.
20394 * sese.c (debug_ivtype_map): Removed unused.
20395 (debug_ivtype_map_1): Removed unused.
20396 (debug_ivtype_elt): Remove unused.
20397 (eq_ivtype_map_elts): Remove unused.
20398
20399
82c0e1a0
KT
204002013-04-02 Kai Tietz <ktietz@redhat.com>
20401
20402 PR target/52790
20403 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20404 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
33159866
UB
20405 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20406 function.
82c0e1a0
KT
20407 (legitimize_pe_coff_symbol): Likewise.
20408 (is_imported_p): New helper-function.
20409 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20410 for Windows x64 targets.
20411 (ix86_expand_prologue): Optimize for pe-coff targets.
20412 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20413 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20414 medium/large code-model.
20415 (legitimize_pic_address): Likewise.
20416 (legitimize_tls_address): Likewise.
20417 (ix86_expand_call): Likewise.
20418 (x86_output_mi_thunk): Likewise.
20419 (get_dllimport_decl): Add new beimport argument.
20420 (construct_plt_address): Don't assert for x64 pe-coff targets.
20421 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20422 targets.
20423 (SYMBOL_FLAG_STUBVAR): New macro.
20424 (SYMBOL_REF_STUBVAR_P): Likewise.
20425 * config/i386/winnt.c (stub_list): New structure.
20426 (stub_head): New local variable.
20427 (i386_pe_record_stub): New function.
20428 (i386_pe_file_end): Emit refptr-stubs.
20429
5d751b0c
JJ
204302013-04-02 Jakub Jelinek <jakub@redhat.com>
20431
09bb4c99
JJ
20432 PR rtl-optimization/56745
20433 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20434 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20435
a1d8947a
JJ
20436 PR c++/34949
20437 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20438 and both of them are MEM_REFs, just compare first argument for
20439 equality and attempt to deal even with differing offsets.
20440
5d751b0c
JJ
20441 PR c++/34949
20442 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20443 of gimple_clobber_p to be MEM_REF.
20444 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20445 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20446 after gimplification.
20447 * asan.c (get_mem_ref_of_assignment): Don't instrument
20448 gimple_clobber_p stmts.
20449 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20450 gimple_clobber_p stmt if they have MEM_REF lhs and
20451 are dead because of another gimple_clobber_p stmt.
20452 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20453 gimple_clobber_p stmts like debug stmts.
20454 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20455 that refer to unused VAR_DECLs or uninitialized values.
20456 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20457 gimple_clobber_p stmts if they refer to removed parameters.
20458 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20459 formatting.
20460
e594716a
UB
204612013-04-02 Uros Bizjak <ubizjak@gmail.com>
20462
20463 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20464 using SWI48 mode attribute.
20465
7a80735b
WM
204662013-04-02 Wei Mi <wmi@google.com>
20467
20468 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20469 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20470 *<rotate_insn><mode>3_mask in i386.md.
20471
f423a9e4
AI
204722013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20473
20474 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20475
90eb75f2
RB
204762013-04-02 Richard Biener <rguenther@suse.de>
20477
20478 PR tree-optimization/56778
20479 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20480 Runtime alias tests are not supported for gather loads.
20481 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20482 stmts referenced from SSA operands before updating SSA form.
20483
d8c69a92
IC
204842013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20485 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20486
20487 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
e594716a
UB
20488 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20489 * config/arm/cortex-a53.md: New file.
20490 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20491 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20492 * config/arm/arm.c (arm_issue_rate): Likewise.
20493 * config/arm/arm-tune.md: Regenerate
20494 * config/arm/arm-tables.opt: Regenerate.
20495 * config/arm/arm-cores.def: Add cortex-a53.
d8c69a92 20496
239eb04c
ZC
204972013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20498
20499 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20500 non-static link.
20501
c902d3c8
SN
205022013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20503
d8c69a92
IC
20504 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20505 scalar load/store operations using B/H registers.
c902d3c8
SN
20506 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20507
051d0e2f
SN
205082013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20509
20510 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20511 scalar move.
20512 * config/aarch64/aarch64.c
20513 (aarch64_simd_scalar_immediate_valid_for_move): New.
20514 * config/aarch64/aarch64-protos.h
20515 (aarch64_simd_scalar_immediate_valid_for_move): New.
20516 * config/aarch64/constraints.md (Dh, Dq): New.
20517 * config/aarch64/iterators.md (hq): New.
20518
0ee1e3d9
EB
205192013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20520
20521 * reorg.c (get_branch_condition): Deal with conditional returns.
20522 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20523
136fb3f7
WM
205242013-04-01 Wei Mi <wmi@google.com>
20525
20526 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20527 Truncate operand 2 using %b asm operand modifier.
20528 (*<shift_insn><mode>3_mask): Ditto.
20529 (*<rotate_insn><mode>3_mask): Ditto.
20530
6388c738
SB
205312013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20532
20533 PR middle-end/56798
20534 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20535
f7a4d826
KK
205362013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20537
20538 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20539 of next_real_insn.
20540 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20541
4b943a49
LC
205422013-03-30 Lawrence Crowl <crowl@google.com>
20543
20544 * dse.c (clear_alias_sets): Remove never set.
20545 (disqualified_clear_alias_sets): Remove never set.
20546 (clear_alias_mode_pool): Remove never set.
20547 (dse_step0): Remove condition that is never true.
20548 (canon_address): Remove condition that is never true.
20549 (dse_step7): Remove condition that is never true.
20550 (rest_of_handle_dse): Remove condition that is never true.
20551 (rest_of_handle_dse::did_global): Remove never read from above.
20552 (dse_step2_spill): Remove never called from above.
20553 (dse_step5_spill): Remove never called from above.
20554
39718607
SB
205552013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20556
da5c6bde
SB
20557 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20558 JUMP_TABLE_DATA changes.
20559 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20560 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20561 (Insns) <jump_table_data>: New entry.
20562 * doc/tm.texi: Regenerate.
20563
39718607
SB
20564 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20565
20566 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20567 for table jump at the end of a basic block using tablejump_p.
20568 * targhooks.c (default_invalid_within_doloop): Likewise.
20569 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20570 target hook implementation that is identical to the default hook.
20571 (rs6000_invalid_within_doloop): Remove.
20572
20573 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20574 unused variable from tablejump_p call.
20575
20576 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20577 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20578 (INSN_DELETED_P): Likewise.
20579 (emit_jump_table_data): New prototype.
20580 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20581 after 4th as unused.
20582 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20583 * sched-vis.c (print_insn): Likewise.
20584 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20585 insn for compatibility with back ends that use next_active_insn to
20586 identify jump table data.
20587 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20588 (remove_insn): Likewise.
20589 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20590 to be emitted.
20591 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20592 (emit_jump_table_data): New function.
20593
20594 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20595 basic block, a JUMP_TABLE_DATA never is.
20596 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20597 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20598 off from code handling real insns.
20599 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20600 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20601 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20602 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20603 is not a NONDEBUG_INSN_P.
20604 * ira-costs.c (scan_one_insn): Likewise.
20605 * jump.c (mark_all_labels): Likewise.
20606 (mark_jump_label_1): Likewise.
20607 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20608 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20609 a basic block.
20610 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20611 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20612 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20613 JUMP_TABLE_DATA_P insns.
20614 (calculate_elim_costs_all_insns): Likewise.
20615 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20616 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20617 (delete_output_reload): Code style fixups.
20618 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20619 insn flags on this non-insn.
20620 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20621 as scheduling barriers, for pre-change compatibility.
20622 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20623 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20624
20625 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20626 redundant JUMP_TABLE_DATA_P test.
20627 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20628 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20629 (frv_for_each_packet): Likewise.
20630 * config/i386/i386.c (min_insn_size): Likewise.
20631 (ix86_avoid_jump_mispredicts): Likewise.
20632 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20633 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20634 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20635 (mips16_insn_length): Robustify.
20636 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20637 (mips16_split_long_branches): Likewise.
20638 * config/pa/pa.c (pa_combine_instructions): Likewise.
20639 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20640 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20641 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20642 as contributing to pool range lengths.
20643 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20644 Remove redundant JUMP_TABLE_DATA_P test.
20645 (sh_loop_align): Likewise.
20646 (split_branches): Likewise.
20647 (sh_insn_length_adjustment): Likewise.
20648 * config/spu/spu.c (get_branch_target): Likewise.
20649
0208f7da
JH
206502013-03-29 Jan Hubicka <jh@suse.cz>
20651
20652 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20653 gcov streaming; stream hot bb threshold to ltrans.
20654 * predict.c (get_hot_bb_threshold): Break out from ....
20655 (maybe_hot_count_p): ... here.
20656 (set_hot_bb_threshold): New function.
20657 * lto-section-in.c (lto_section_name): Add profile.
20658 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20659 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20660 and data-streamer.h
20661 (histogram_entry): New structure.
20662 (histogram, histogram_pool): New global vars.
20663 (histogram_hash): New structure.
20664 (histogram_hash::hash): New method.
20665 (histogram_hash::equal): Likewise.
20666 (account_time_size): New function.
20667 (cmp_counts): New function.
20668 (dump_histogram): New function.
20669 (ipa_profile_generate_summary): New function.
20670 (ipa_profile_write_summary): New function.
20671 (ipa_profile_read_summary): New function.
20672 (ipa_profile): Decide on threshold.
e594716a
UB
20673 (pass_ipa_profile): Add ipa_profile_write_summary and
20674 ipa_profile_read_summary.
0208f7da
JH
20675 * Makefile.in (ipa.o): Update dependencies.
20676 * lto-streamer.h (LTO_section_ipa_profile): New section.
20677
5a6ccc94
GDR
206782013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20679
20680 * tree.h (VAR_P): New.
20681
39385fa6
PC
206822013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20683
20684 PR lto/56777
20685 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20686
34f0d87a
SB
206872013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20688
20689 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20690 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20691 (control_flow_insn_p): Likewise.
20692 * cfgrtl.c (duplicate_insn_chain): Likewise.
20693 * final.c (get_attr_length_1): Likewise.
20694 (shorten_branches): Likewise.
20695 (final_scan_insn): Likewise.
20696 * function.c (instantiate_virtual_regs): Likewise.
20697 * gcse.c (insert_insn_end_basic_block): Likewise.
20698 * ira-costs.c (scan_one_insn): Likewise.
20699 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20700 * lra.c (check_rtl): Likewise.
20701 * reload1.c (elimination_costs_in_insn): Likewise.
20702 * reorg.c (follow_jumps): Likewise.
20703
20704 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20705 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20706 (thumb_far_jump_used_p): Likewise.
20707 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20708 (workaround_speculation): Likewise.
20709 (add_sched_insns_for_speculation): Likewise.
20710 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20711 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20712 (frv_for_each_packet): Likewise.
20713 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20714 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20715 (final_emit_insn_group_barriers): Likewise.
20716 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20717 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20718 (mips16_insn_length): Likewise.
20719 * config/pa/pa.c (pa_reorg): Likewise.
20720 (pa_combine_instructions): Likewise.
20721 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20722 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20723 (sh_reorg): Likewise.
20724 (split_branches): Likewise.
20725 * config/spu/spu.c (get_branch_target): Likewise.
20726
20727 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20728 JUMP_TABLE_DATA_P.
20729
4ac761b0
KY
207302013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20731
39385fa6 20732 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
4ac761b0
KY
20733 Fix declaration name.
20734
58a49189
LC
207352013-03-28 Lawrence Crowl <crowl@google.com>
20736
20737 * graphds.h (struct graph.indicies): Remove unused.
20738 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20739 (SCOP_ORIGINAL_PDDRS): Remove unused.
20740 * sese.h (extern insert_loop_close_phis): Removed unused.
20741 (extern insert_guard_phis): Removed unused.
20742 (extern ivtype_map_elt_info): Removed unused.
20743 (new_ivtype_map_elt): Removed unused.
20744 * sese.c (ivtype_map_elt_info): Removed unused.
20745
7b3b6ae4
LC
207462013-03-28 Lawrence Crowl <crowl@google.com>
20747
20748 * Makefile.in: Add several missing include dependences.
20749 (DUMPFILE_H): New.
20750 (test-dump.o): New. This object is not added to any executable,
20751 but is present for ad-hoc testing.
20752 * bitmap.c
20753 (debug (const bitmap_head_def &)): New.
20754 (debug (const bitmap_head_def *)): New.
20755 * bitmap.h
20756 (extern debug (const bitmap_head_def &)): New.
20757 (extern debug (const bitmap_head_def *)): New.
20758 * cfg.c
20759 (debug (edge_def &)): New.
20760 (debug (edge_def *)): New.
20761 * cfghooks.c
20762 (debug (basic_block_def &)): New.
20763 (debug (basic_block_def *)): New.
20764 * dumpfile.h
20765 (dump_node (const_tree, int, FILE *)): Correct source file.
20766 * dwarf2out.c
20767 (debug (die_struct &)): New.
20768 (debug (die_struct *)): New.
20769 * dwarf2out.h
20770 (extern debug (die_struct &)): New.
20771 (extern debug (die_struct *)): New.
20772 * gimple-pretty-print.c
20773 (debug (gimple_statement_d &)): New.
20774 (debug (gimple_statement_d *)): New.
20775 * gimple-pretty-print.h
20776 (extern debug (gimple_statement_d &)): New.
20777 (extern debug (gimple_statement_d *)): New.
20778 * ira-build.c
20779 (debug (ira_allocno_copy &)): New.
20780 (debug (ira_allocno_copy *)): New.
20781 (debug (ira_allocno &)): New.
20782 (debug (ira_allocno *)): New.
20783 * ira-int.h
20784 (extern debug (ira_allocno_copy &)): New.
20785 (extern debug (ira_allocno_copy *)): New.
20786 (extern debug (ira_allocno &)): New.
20787 (extern debug (ira_allocno *)): New.
20788 * ira-lives.c
20789 (debug (live_range &)): New.
20790 (debug (live_range *)): New.
20791 * lra-int.h
20792 (debug (lra_live_range &)): New.
20793 (debug (lra_live_range *)): New.
20794 * lra-lives.c
20795 (debug (lra_live_range &)): New.
20796 (debug (lra_live_range *)): New.
20797 * omega.c
20798 (debug (omega_pb_d &)): New.
20799 (debug (omega_pb_d *)): New.
20800 * omega.h
20801 (extern debug (omega_pb_d &)): New.
20802 (extern debug (omega_pb_d *)): New.
20803 * print-rtl.c
20804 (debug (const rtx_def &)): New.
20805 (debug (const rtx_def *)): New.
20806 * print-tree.c
20807 (debug_tree (tree): Move within file.
20808 (debug_raw (const tree_node &)): New.
20809 (debug_raw (const tree_node *)): New.
20810 (dump_tree_via_hooks (const tree_node *, int)): New.
20811 (debug (const tree_node &)): New.
20812 (debug (const tree_node *)): New.
20813 (debug_verbose (const tree_node &)): New.
20814 (debug_verbose (const tree_node *)): New.
20815 (debug_head (const tree_node &)): New.
20816 (debug_head (const tree_node *)): New.
20817 (debug_body (const tree_node &)): New.
20818 (debug_body (const tree_node *)): New.
20819 (debug_vec_tree (tree): Move and reimplement in terms of dump.
20820 (debug (vec<tree, va_gc> &)): New.
20821 (debug (vec<tree, va_gc> *)): New.
20822 * rtl.h
20823 (extern debug (const rtx_def &)): New.
20824 (extern debug (const rtx_def *)): New.
20825 * sbitmap.c
20826 (debug_raw (simple_bitmap_def &)): New.
20827 (debug_raw (simple_bitmap_def *)): New.
20828 (debug (simple_bitmap_def &)): New.
20829 (debug (simple_bitmap_def *)): New.
20830 * sbitmap.h
20831 (extern debug (simple_bitmap_def &)): New.
20832 (extern debug (simple_bitmap_def *)): New.
20833 (extern debug_raw (simple_bitmap_def &)): New.
20834 (extern debug_raw (simple_bitmap_def *)): New.
20835 * sel-sched-dump.c
20836 (debug (vinsn_def &)): New.
20837 (debug (vinsn_def *)): New.
20838 (debug_verbose (vinsn_def &)): New.
20839 (debug_verbose (vinsn_def *)): New.
20840 (debug (expr_def &)): New.
20841 (debug (expr_def *)): New.
20842 (debug_verbose (expr_def &)): New.
20843 (debug_verbose (expr_def *)): New.
20844 (debug (vec<rtx> &)): New.
20845 (debug (vec<rtx> *)): New.
20846 * sel-sched-dump.h
20847 (extern debug (vinsn_def &)): New.
20848 (extern debug (vinsn_def *)): New.
20849 (extern debug_verbose (vinsn_def &)): New.
20850 (extern debug_verbose (vinsn_def *)): New.
20851 (extern debug (expr_def &)): New.
20852 (extern debug (expr_def *)): New.
20853 (extern debug_verbose (expr_def &)): New.
20854 (extern debug_verbose (expr_def *)): New.
20855 (extern debug (vec<rtx> &)): New.
20856 (extern debug (vec<rtx> *)): New.
20857 * sel-sched-ir.h
20858 (_list_iter_cond_expr): Make inline instead of static.
20859 * sreal.c
20860 (debug (sreal &)): New.
20861 (debug (sreal *)): New.
20862 * sreal.h
20863 (extern debug (sreal &)): New.
20864 (extern debug (sreal *)): New.
20865 * tree.h
20866 (extern debug_raw (const tree_node &)): New.
20867 (extern debug_raw (const tree_node *)): New.
20868 (extern debug (const tree_node &)): New.
20869 (extern debug (const tree_node *)): New.
20870 (extern debug_verbose (const tree_node &)): New.
20871 (extern debug_verbose (const tree_node *)): New.
20872 (extern debug_head (const tree_node &)): New.
20873 (extern debug_head (const tree_node *)): New.
20874 (extern debug_body (const tree_node &)): New.
20875 (extern debug_body (const tree_node *)): New.
20876 (extern debug (vec<tree, va_gc> &)): New.
20877 (extern debug (vec<tree, va_gc> *)): New.
20878 * tree-cfg.c
20879 (debug (struct loop &)): New.
20880 (debug (struct loop *)): New.
20881 (debug_verbose (struct loop &)): New.
20882 (debug_verbose (struct loop *)): New.
20883 * tree-dump.c: Add header dependence.
20884 * tree-flow.h
20885 (extern debug (struct loop &)): New.
20886 (extern debug (struct loop *)): New.
20887 (extern debug_verbose (struct loop &)): New.
20888 (extern debug_verbose (struct loop *)): New.
20889 * tree-data-ref.c
20890 (debug (data_reference &)): New.
20891 (debug (data_reference *)): New.
20892 (debug (vec<data_reference_p> &)): New.
20893 (debug (vec<data_reference_p> *)): New.
20894 (debug (vec<ddr_p> &)): New.
20895 (debug (vec<ddr_p> *)): New.
20896 * tree-data-ref.h
20897 (extern debug (data_reference &)): New.
20898 (extern debug (data_reference *)): New.
20899 (extern debug (vec<data_reference_p> &)): New.
20900 (extern debug (vec<data_reference_p> *)): New.
20901 (extern debug (vec<ddr_p> &)): New.
20902 (extern debug (vec<ddr_p> *)): New.
20903 * tree-ssa-alias.c
20904 (debug (pt_solution &)): New.
20905 (debug (pt_solution *)): New.
20906 * tree-ssa-alias.h
20907 (extern debug (pt_solution &)): New.
20908 (extern debug (pt_solution *)): New.
20909 * tree-ssa-alias.c
20910 (debug (_var_map &)): New.
20911 (debug (_var_map *)): New.
20912 (debug (tree_live_info_d &)): New.
20913 (debug (tree_live_info_d *)): New.
20914 * tree-ssa-alias.h
20915 (extern debug (_var_map &)): New.
20916 (extern debug (_var_map *)): New.
20917 (extern debug (tree_live_info_d &)): New.
20918 (extern debug (tree_live_info_d *)): New.
20919
be77e1e5
JH
209202013-03-28 Jan Hubicka <jh@suse.cz>
20921
20922 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
20923
777e6976
IB
209242013-03-28 Ian Bolton <ian.bolton@arm.com>
20925
20926 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
20927 record only when desired or required.
20928
9cd347ae
UB
209292013-03-28 Uros Bizjak <ubizjak@gmail.com>
20930
20931 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
20932 *vec_extractv2di_1_rex64. Use x64 isa attribute.
20933
33159866 209342013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
2a1d1991
N
20935
20936 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
20937 (*andsi3_compare0_uxtw): New pattern.
20938 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
20939 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
20940
89ab31c1
JH
209412013-03-28 Jan Hubicka <jh@suse.cz>
20942
20943 * data-streamer-in.c (streamer_read_gcov_count): New function.
20944 * gimple-streamer-out.c: Include value-prof.h.
20945 (output_gimple_stmt): Output histogram.
20946 (output_bb): Use streamer_write_gcov_count.
20947 * value-prof.c: Include data-streamer.h
20948 (dump_histogram_value): Add HIST_TYPE_MAX.
20949 (stream_out_histogram_value): New function.
20950 (stream_in_histogram_value): New function.
20951 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
20952 (stream_out_histogram_value, stream_in_histogram_value): Declare.
20953 * data-streamer-out.c (streamer_write_gcov_count): New function.
20954 (streamer_write_gcov_count_stream): New function.
20955 * lto-cgraph.c (lto_output_edge): Update counter streaming.
20956 (lto_output_node): Likewise.
20957 (input_node, input_edge): Likewise.
20958 * lto-streamer-out.c (output_cfg): Update streaming.
20959 * lto-streamer-in.c (input_cfg): Likewise.
20960 * data-streamer.h (streamer_write_gcov_count,
20961 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
20962 * gimple-streamer-in.c: Include value-prof.h
20963 (input_gimple_stmt): Input histograms.
20964 (input_bb): Update profile streaming.
20965
e594716a 209662013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
b4c73eed 20967
33159866 20968 * genmodes.c (emit_max_int): New function.
8697be17 20969 (emit_insn_modes_h): Added call to emit_max_function.
39385fa6
PC
20970 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
20971 Added doc.
8697be17 20972 * machmode.def: Fixed comment.
89ab31c1 20973
e594716a 209742013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
8697be17
KZ
20975
20976 * combine.c (try_combine): Removed useless assert.
b4c73eed
KZ
20977 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
20978
74946978
MP
209792013-03-28 Marek Polacek <polacek@redhat.com>
20980 Richard Biener <rguenther@suse.de>
20981
20982 PR tree-optimization/56695
20983 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
20984 build signed result of a vector comparison.
20985 * tree-cfg.c (verify_gimple_comparison): Check that a result
20986 of a vector comparison has signed type.
20987
a64b9c26
RB
209882013-03-28 Richard Biener <rguenther@suse.de>
20989
20990 PR tree-optimization/37021
20991 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
20992 do not restrict gaps between groups.
20993 * tree-vect-stmts.c (vectorizable_load): Properly account for
20994 a gap between groups.
20995
a9dc2a2f
EB
209962013-03-28 Eric Botcazou <ebotcazou@adacore.com>
20997
20998 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
20999 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
21000 is not enabled.
21001
53cb97f9
GP
210022013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
21003
21004 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
21005 * doc/extend.texi (Named Address Spaces): Ditto.
21006 (Variable Attributes): Ditto.
21007
b802ae5c
KT
210082013-03-27 Kai Tietz <ktietz@redhat.com>
21009
eddae10a
KT
21010 * config.build: Add support for cygwin x64 target.
21011 * config.gcc: Likewise.
21012 * config.host: Likewise.
21013 * configure.ac: Likewise
21014 * configure: Regenerated.
b802ae5c 21015
371e77e3 210162013-03-27 Kai Tietz <ktietz@redhat.com>
8269de54
KT
21017
21018 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
21019 * config/i386/t-cygwin-w64: New file.
21020 * config/i386/cygwin-w64.h: New file.
21021 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
21022 and add support for x64-cygwin target.
21023 (CPP_SPEC): Likewise.
21024 (CXX_WRAP_SPEC_LIST): Undefine before define.
21025 (LIBGCJ_SONAME): Use 15 as version.
21026
f49b33cb
RB
210272013-03-27 Richard Biener <rguenther@suse.de>
21028
21029 PR tree-optimization/56716
21030 * tree-ssa-structalias.c (perform_var_substitution): Adjust
21031 dumping for ref nodes.
21032
b37a6ce5
MJ
210332013-03-27 Martin Jambor <mjambor@suse.cz>
21034
21035 PR tree-optimization/55334
21036 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21037 restricted pointers to arrays.
21038
9469b9b2
GDR
210392013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21040
21041 * Makefile.in (.SUFFIXES): Add .cc.
21042 (.c.o): Apply same recipe for implicit rule .cc.o.
21043
7d24f650
RB
210442013-03-27 Richard Biener <rguenther@suse.de>
21045
21046 PR tree-optimization/37021
21047 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21048 REALPART/IMAGPART_EXPRs around the supported refs.
21049 * tree-ssa-structalias.c (find_func_aliases): Assume that
21050 floating-point values are not used to transfer pointers.
21051
2f251a05
AI
210522013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21053
f43245d1
UB
21054 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21055 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21056 * doc/tm.texi: Regenerate.
21057 * targhooks.h (default_has_ifunc_p): New.
21058 * targhooks.c (default_has_ifunc_p): Ditto.
21059 * config/linux-protos.h: New file.
21060 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21061 hook for linux which disables support of indirect functions in android.
21062 * config/linux-android.c: New file.
21063 * config/t-linux-android.c: Ditto.
21064 * config.gcc: Added new object file linux-android.o.
21065 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
2f251a05 21066 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
f43245d1 21067 * varasm.c (do_assemble_alias): Likewise.
2f251a05
AI
21068 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21069 doesn't support indirect functions.
21070 * configure: Regenerate.
21071
78b4e425
BC
210722013-03-27 Bin Cheng <bin.cheng@arm.com>
21073
21074 PR target/56102
21075 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21076 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21077 mult-word mode.
21078
a5ba7b92
AK
210792013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21080
21081 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21082
be624986
TG
210832013-03-27 Terry Guo <terry.guo@arm.com>
21084
21085 * config/arm/arm-cores.def: Added core cortex-r7.
21086 * config/arm/arm-tune.md: Regenerated.
21087 * config/arm/arm-tables.opt: Regenerated.
21088 * doc/invoke.texi: Added entry for core cortex-r7.
21089
0a514f47
WL
210902013-03-27 Walter Lee <walt@tilera.com>
21091
21092 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21093 double-decrement of next_scratch_regno.
21094
9332b0d2
WL
210952013-03-27 Walter Lee <walt@tilera.com>
21096
801d9b2a 21097 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
9332b0d2
WL
21098 input operands.
21099 (insn_v1mulus): Ditto.
21100 (insn_v2muls): Ditto.
21101
f54ea5dd
WL
211022013-03-27 Walter Lee <walt@tilera.com>
21103
e594716a 21104 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
f54ea5dd
WL
21105 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21106
bf60f4ca
WL
211072013-03-27 Walter Lee <walt@tilera.com>
21108
e594716a 21109 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
bf60f4ca
WL
21110 (*sibcall_value): Ditto.
21111
e3b51eeb
WL
211122013-03-27 Walter Lee <walt@tilera.com>
21113
21114 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21115 (insn_mnz_v8qi): ... this ...
21116 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21117 vector equivalent.
21118 (insn_v<n>mnz): Replaced by ...
21119 (insn_v1mnz): ... this ...
21120 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21121 equivalent.
21122 (insn_mz_<mode>): Replaced by ...
21123 (insn_mz_v8qi): ... this ...
21124 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21125 vector equivalent.
21126 (insn_v<n>mz): Replaced by ...
21127 (insn_v1mz): ... this ...
21128 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21129 equivalent.
21130
4fc7b145
EB
211312013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21132
21133 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21134
6f33817e
RM
211352013-03-26 Roland McGrath <mcgrathr@google.com>
21136
21137 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21138 than fprintf with a non-constant, non-format string.
21139
93a4145b
UB
211402013-03-26 Uros Bizjak <ubizjak@gmail.com>
21141
21142 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21143 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21144 operand 0 predicate.
21145 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21146 attribute. Use general_x64nomem_operand as operand 1 predicate.
21147 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21148 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21149 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21150 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21151 (mov<mode>_insv_1): Remove expander. Merge insn with
21152 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21153 Use general_x64nomem_operand as operand 1 predicate.
21154 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21155 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21156 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21157 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21158 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21159 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21160 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21161 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21162 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21163 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21164 (general_x64nomem_operand): Ditto.
21165
c6a9ed5a 211662013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
df599020
SH
21167
21168 * config/rtems.opt: Add -pthread option.
21169
31b0a960
RB
211702013-03-26 Richard Biener <rguenther@suse.de>
21171
93a4145b
UB
21172 * alias.c (find_base_term): Avoid redundant and not used recursion.
21173 (base_alias_check): Get the initial base term from the caller.
31b0a960
RB
21174 (true_dependence_1): Compute and pass base terms to base_alias_check.
21175 (write_dependence_p): Likewise.
21176 (may_alias_p): Likewise.
21177
79517551
SN
211782013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21179
21180 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21181 PC-relative load in SI modes and above only.
21182
a76213b9
XQ
211832013-03-26 Xinyu Qi <xyqi@marvell.com>
21184
21185 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
33159866
UB
21186 * config/arm/iwmmxt.md (WCGR0): Update.
21187 (WCGR1, WCGR2, WCGR3): Likewise.
a76213b9 21188
37ff9355
UB
211892013-03-26 Uros Bizjak <ubizjak@gmail.com>
21190
21191 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21192 Use x64 and nox64 isa attributes.
21193
30b0317c
RB
211942013-03-26 Richard Biener <rguenther@suse.de>
21195
21196 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21197 alignment computations and rely on get_object_alignment_1
21198 for the !TYPE_P case.
21199 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21200
cbcdb140
WL
212012013-03-26 Walter Lee <walt@tilera.com>
21202
21203 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21204 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21205
dd884525 212062013-03-25 Jeff Law <law@redhat.com>
be672e08 21207
33159866
UB
21208 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21209 check for INTEGRAL_TYPE_P that was missing due to checking in
dd884525
JL
21210 wrong version of prior patch.
21211
0051d3ec
WL
212122013-03-25 Walter Lee <walt@tilera.com>
21213
21214 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21215 TILEGX_INSN_SHUFFLEBYTES1.
21216 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21217 shufflebytes1.
21218 (tilegx_builtins): Ditto.
21219 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21220
065a3605
WL
212212013-03-25 Walter Lee <walt@tilera.com>
21222
21223 * config/tilegx/tilegx.md (floatsisf2): New pattern.
0051d3ec
WL
21224 (floatunssisf2): New pattern.
21225 (floatsidf2): New pattern.
21226 (floatunssidf2): New pattern.
065a3605 21227
5b2a3c82
WL
212282013-03-25 Walter Lee <walt@tilera.com>
21229
21230 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21231 tests for constraint J, K, N, P.
21232
192ea533
WL
212332013-03-25 Walter Lee <walt@tilera.com>
21234
21235 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21236 Use indirect/pcrel encoding.
21237 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21238 Ditto.
21239
855e0d0b
SE
212402013-03-25 Steve Ellcey <sellcey@mips.com>
21241
21242 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21243 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21244 * config/mips/mips.c (mips_option_override): Set IMADD default.
21245 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21246 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21247 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21248 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21249 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21250
39e45653
JL
212512013-03-25 Jeff Law <law@redhat.com>
21252
21253 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21254 slightly to avoid creating and folding useless trees. Simplify
21255 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21256
7482c470
UB
212572013-03-25 Uros Bizjak <ubizjak@gmail.com>
21258
21259 * config/i386/i386.md (*zero_extendsidi2): Merge with
21260 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21261 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21262 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21263 targets. Clarify comment.
21264
4a53743e
MJ
212652013-03-25 Martin Jambor <mjambor@suse.cz>
21266
21267 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21268 pass-through jump functions differently.
21269 (ipa_read_jump_function): Likewise. Also use setter functions to set
21270 up jump functions.
21271
162712de
MJ
212722013-03-25 Martin Jambor <mjambor@suse.cz>
21273
21274 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21275 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21276 process it.
21277 (ipa_get_indirect_edge_target): New function.
21278 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21279 ipa_get_indirect_edge_target. Update all callers.
21280 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21281 ipa_get_indirect_edge_target_1 instead of calling
21282 ipa_get_indirect_edge_target.
21283 (create_specialized_node): Pass aggvlas to
21284 ipcp_discover_new_direct_edges.
21285
a5a4c20a
KT
212862013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21287
21288 * config/arm/arm.md (f_sels, f_seld): New types.
21289 (*cmov<mode>): New pattern.
21290 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21291 predicate.
21292
f992b9fc
KT
212932013-03-25 Kai Tietz <ktietz@redhat.com>
21294
21295 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21296 POSIX-printf for mingw-hosted builds.
21297
c16fd676
RB
212982013-03-25 Richard Biener <rguenther@suse.de>
21299
21300 PR middle-end/56694
21301 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21302 must-not-throw stmt location.
21303
5ad29f12
KT
213042013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21305
21306 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21307 Emit load-acquire versions when acq is true.
21308 (arm_emit_store_exclusive): Add rel parameter.
21309 Emit store-release versions when rel is true.
21310 (arm_split_compare_and_swap): Use acquire-release instructions
21311 instead.
21312 of barriers when appropriate.
21313 (arm_split_atomic_op): Likewise.
21314 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21315 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21316 (VUNSPEC_SLX): Likewise.
21317 (VUNSPEC_LDA): Likewise.
21318 (VUNSPEC_STL): Likewise.
21319 * config/arm/sync.md (atomic_load<mode>): New pattern.
21320 (atomic_store<mode>): Likewise.
21321 (arm_load_acquire_exclusive<mode>): Likewise.
21322 (arm_load_acquire_exclusivesi): Likewise.
21323 (arm_load_acquire_exclusivedi): Likewise.
21324 (arm_store_release_exclusive<mode>): Likewise.
21325
03a7dddb
CM
213262013-03-25 Catherine Moore <clm@codesourcery.com>
21327
21328 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21329 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21330 * config/mip/predicates.md (lwsp_swsp_operand,
21331 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21332 sb16_operand, db4_operand, db7_operand, ib3_operand,
21333 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21334 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21335 andi16_operand): New predicates.
21336 * config/mips/mips.md (compression): New attribute.
21337 (enabled): New attribute.
21338 (length): Consider compression in computing length.
21339 (shift_compression): New code attribute.
21340 (*add<mode>3): New operands. Record compression.
21341 (sub<mode>3): Likewise.
21342 (one_cmpl<mode>2): Likewise.
21343 (*and<mode>3): Likewise.
21344 (*ior<mode>3): Likewise.
21345 (unnamed pattern for xor): Likewise.
21346 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21347 (*<optab><mode>3): Likewise.
21348 (*mov<mode>_internal: Likewise.
21349 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21350 (mips_unsigned_immediate_p): New.
21351 (umips_lwsp_swsp_address_p): New.
21352 (m16_based_address_p): New.
21353 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21354 (mips_unsigned_immediate_p): New prototype.
21355 (lwsp_swsp_address_p): New prototype.
21356 (m16_based_address_p): New prototype.
21357 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21358 (mips_signed_immediate_p): New function.
21359 (m16_based_address_p): New function.
21360 (lwsp_swsp_address_p): New function.
21361 (mips_print_operand_punctuation): Recognize short delay slot insns
21362 for microMIPS.add<mode>3"
21363
f35c297f 213642013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
a5784ce6 21365
f35c297f
KT
21366 PR target/56720
21367 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21368 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21369
051b9446
RB
213702013-03-25 Richard Biener <rguenther@suse.de>
21371
21372 PR tree-optimization/56689
21373 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21374 any edge.
21375
374001cb
RB
213762013-03-25 Richard Biener <rguenther@suse.de>
21377
21378 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21379 of bitmap.
21380 (memory_references): Likewise.
21381 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21382 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21383 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21384 (gather_mem_refs_in_loops): Fold into ...
21385 (analyze_memory_references): ... this. Move initialization
21386 to tree_ssa_lim_initialize.
21387 (fill_always_executed_in): Rename to ...
21388 (fill_always_executed_in_1): ... this.
21389 (fill_always_executed_in): Move contains_call computation to
21390 this new function from ...
21391 (tree_ssa_lim_initialize): ... here.
21392 (tree_ssa_lim): Call fill_always_executed_in.
21393
57895947
EB
213942013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21395
21396 * postreload.c (reload_combine): Fix code detecting returns.
21397
1e1b18c1
EB
213982013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21399
21400 * function.c (emit_use_return_register_into_block): On cc0 targets,
21401 do not emit the sequence between cc0 setter and user.
21402
9216baf1
KT
214032013-03-25 Kai Tietz <ktietz@redhat.com>
21404
1e1b18c1
EB
21405 * config/i386/predicates.md (local_symbolic_operand): Interpret
21406 dll-imported symbols as none-local.
9216baf1 21407
76421b44
RB
214082013-03-25 Richard Biener <rguenther@suse.de>
21409
21410 * tree-ssa-loop-im.c (struct depend): Remove.
21411 (struct lim_aux_data): Make depends a vec of gimples.
21412 (free_lim_aux_data): Adjust.
21413 (add_dependency): Likewise.
21414 (set_level): Likewise.
21415
d154bfa2
RB
214162013-03-25 Richard Biener <rguenther@suse.de>
21417
21418 PR middle-end/56434
21419 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21420 the pointer returned by calls with ECF_MALLOC set.
21421
2cd9804e
UB
214222013-03-24 Uros Bizjak <ubizjak@gmail.com>
21423
acef5fe0
UB
21424 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21425
214262013-03-24 Uros Bizjak <ubizjak@gmail.com>
21427
21428 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
2cd9804e
UB
21429 using MMXMODE mode iterator.
21430 (*move<mode>_internal): Merge with *movv2sf_internal and
21431 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21432
225ccc68
SB
214332013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21434
85c0f02d
SB
21435 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21436 (record_last_mem_set_info): Likewise.
21437
225ccc68
SB
21438 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21439 of XNEWVEC followed by memset.
21440 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21441
b64925dc
SB
214422013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21443
21444 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21445 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21446 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21447 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21448 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21449 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21450 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21451 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21452 BARRIER_P instead of GET_CODE.
21453
b9aaf52e
EB
214542013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21455
21456 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21457 inaccuracy in the probing code.
21458
21459 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21460 (ctrapdi4): Likewise.
21461
66de4d7c
EB
214622013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21463
21464 * calls.c (expand_call): Add missing guard to code handling return
21465 of non-BLKmode structures in MSB.
21466 * function.c (expand_function_end): Likewise.
21467
4f25c8fb
EB
214682013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21469
21470 * combine.c (try_combine): Adjust comment. Do not add the set of
21471 insn #0 if the destination indirectly is set or dies in insn #2.
21472 Tidy up code to distribute a new note.
21473
23b7850d
UB
214742013-03-22 Uros Bizjak <ubizjak@gmail.com>
21475
21476 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21477 also for alternatives 16 and 17.
21478
bed852cf
UB
214792013-03-22 Uros Bizjak <ubizjak@gmail.com>
21480
21481 * config/i386/sse.md (*mov<mode>_internal): Merge with
21482 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21483 Emit insn template depending on type attribute. Use
21484 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21485 movd instead of movq mnemonic for interunit moves. Rewrite mode
21486 attribute calculation. Remove unit attribute calculation.
21487 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21488 Set prefix_data16 attribute for DImode ssemov types.
21489 Use Ym instead of y for SSE-MMX conversion alternatives.
21490 Reorder operand constraints.
21491
1e8a7937
SB
214922013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21493
21494 * df.h (df_insn_delete): Adjust prototype.
21495 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21496 and let it decide whether mark the basic block dirty.
21497 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21498 * df-scan.c (df_insn_info_delete): New helper function, split
21499 off from df_insn_delete.
21500 (df_scan_free_bb_info): Use it.
21501 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21502 Likewise.
21503 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21504 that the insn is actually an insn and it has a non-NULL basic block.
21505 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21506
e14d094c
RB
215072013-03-22 Richard Biener <rguenther@suse.de>
21508
21509 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21510 dep_ref members.
21511 (mem_ref_alloc): Do not allocate them.
21512 (refs_independent_p): Do not query or maintain a cache.
21513
e9cf7316
RB
215142013-03-22 Richard Biener <rguenther@suse.de>
21515
21516 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21517 (gather_mem_refs_in_loops): Do not compute it.
21518 (analyze_memory_references): Do not allocate it.
21519 (tree_ssa_lim_finalize): Do not free it.
21520 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21521
f046e81b
RB
215222013-03-22 Richard Biener <rguenther@suse.de>
21523
21524 * is-a.h (as_a): Use gcc_checking_assert.
bed852cf 21525
a05c0ddf
IB
215262013-03-22 Ian Bolton <ian.bolton@arm.com>
21527
21528 * config/aarch64/aarch64.c (aarch64_print_operand): New
21529 format specifier for printing a constant in hex.
21530 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21531 format specifier for printing second operand.
21532
c00217fc
RB
215332013-03-22 Richard Biener <rguenther@suse.de>
21534
21535 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21536 bitmaps.
21537 (gather_mem_refs_in_loops): Perform store accumulation here.
21538 (create_vop_ref_mapping_loop): Remove.
21539 (create_vop_ref_mapping): Likewise.
21540 (analyze_memory_references): Initialize refs_stored_in_loop.
bed852cf 21541 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
c00217fc
RB
21542 (record_indep_loop): Remove.
21543 (record_dep_loop): New function.
21544 (ref_indep_loop_p_1): Adjust to only walk over references
21545 in the loop, not its subloops.
21546 (ref_indep_loop_p): Rename to ...
21547 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21548 maintaining a more fine-grained cache.
21549 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21550 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21551
15d19bf8
RB
215522013-03-22 Richard Biener <rguenther@suse.de>
21553
21554 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21555 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21556 aggregate mem_ref_loc.
21557 (free_mem_ref_locs): Inline into ...
21558 (memref_free): ... this and adjust.
21559 (mem_ref_alloc): Adjust.
21560 (mem_ref_locs_alloc): Remove.
21561 (record_mem_ref_loc): Adjust.
21562 (get_all_locs_in_loop): Rewrite into ...
21563 (for_all_locs_in_loop): ... this iterator.
21564 (rewrite_mem_ref_loc): New functor.
21565 (rewrite_mem_refs): Use for_all_locs_in_loop.
21566 (sm_set_flag_if_changed): New functor.
21567 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21568 (ref_always_accessed): New functor.
21569 (ref_always_accessed_p): Use for_all_locs_in_loop.
21570
6f37411d
MG
215712013-03-21 Marc Glisse <marc.glisse@inria.fr>
21572
21573 * tree-pass.h (PROP_gimple_lvec): New.
21574 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21575 (init_optimization_passes): Move pass_lower_vector.
21576 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21577 PROP_gimple_lvec.
21578 (pass_lower_vector): Provide PROP_gimple_lvec.
21579 (pass_lower_vector_ssa): Likewise.
21580 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21581
511ef689
MW
215822013-03-21 Mark Wielaard <mjw@redhat.com>
21583
21584 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21585
cc1df30b
UB
215862013-03-21 Uros Bizjak <ubizjak@gmail.com>
21587
21588 * config/i386/i386.md (*movdi_internal): Disparage slightly
21589 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21590 conversion alternatives.
21591
c845cfe1
JJ
215922013-03-21 Jakub Jelinek <jakub@redhat.com>
21593
37e99116
JJ
21594 PR middle-end/48087
21595 * diagnostic.def (DK_WERROR): New kind.
21596 * diagnostic.h (werrorcount): Define.
21597 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21598 promoted to DK_ERROR, increment DK_WERROR counter instead of
21599 DK_ERROR counter.
21600 * toplev.c (toplev_main): Call print_ignored_options even if
21601 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21602 even if just werrorcount is non-zero.
21603
c845cfe1
JJ
21604 PR debug/55608
21605 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21606 on failure.
21607 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21608 (string_cst_pool_decl): New function.
21609 (optimize_one_addr_into_implicit_ptr): New function.
21610 (resolve_addr_in_expr): Optimize DWARF location expression
21611 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21612 which doesn't live in memory, but has DW_AT_location or
21613 DW_AT_const_value, or refers to a string literal, into
21614 DW_OP_GNU_implicit_pointer.
21615 (optimize_location_into_implicit_ptr): New function.
21616 (resolve_addr): If removing DW_AT_location of a variable because
21617 it was DW_OP_addr of address of the variable, but the variable doesn't
21618 live in memory, try to emit const value attribute for the initializer.
21619
08e0cda6
MG
216202013-03-21 Marc Glisse <marc.glisse@inria.fr>
21621
21622 * tree.h (VECTOR_TYPE_P): New macro.
21623 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21624 TYPE_MODE): Use it.
21625 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21626 VEC_COND_EXPR cannot be lvalues.
21627 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21628
d08633b4
MG
216292013-03-21 Marc Glisse <marc.glisse@inria.fr>
21630
21631 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21632 Restrict the transformation to equal modes.
21633
e6647190
RB
216342013-03-21 Richard Biener <rguenther@suse.de>
21635
21636 PR tree-optimization/39326
21637 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21638 (MEM_ANALYZABLE): Adjust.
21639 (record_mem_ref_loc): Move bitmap ops ...
21640 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21641 unanalyzable refs, do not record locations for it.
21642 (analyze_memory_references): Allocate ref zero as shared
21643 unanalyzable ref.
21644 (refs_independent_p): Do not test for unanalyzed mems here.
21645 (ref_indep_loop_p_1): Special-case disambiguation against
21646 the unanalyzed ref.
cc1df30b 21647 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
e6647190 21648
65074f54
CL
216492013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21650
21651 * config/arm/arm-protos.h (tune_params): Add
21652 prefer_neon_for_64bits field.
21653 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21654 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21655 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21656 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21657 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21658 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21659 (arm_option_override): Handle -mneon-for-64bits new option.
21660 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21661 (prefer_neon_for_64bits): Declare new variable.
21662 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
cc1df30b 21663 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
65074f54
CL
21664 (arch_enabled): Handle new arch types. Remove support for onlya8
21665 and nota8.
21666 (one_cmpldi2): Use new arch names.
bc5faa5b 21667 (zero_extend<mode>di2, extend<mode>di2): Ditto.
65074f54
CL
21668 * config/arm/arm.opt (mneon-for-64bits): Add option.
21669 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21670 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21671 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21672 of onlya8.
21673 * doc/invoke.texi (-mneon-for-64bits): Document.
21674
5a2d2a79
RB
216752013-03-21 Richard Biener <rguenther@suse.de>
21676
21677 PR tree-optimization/39326
21678 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21679 (sort_bbs_in_loop_postorder_cmp): New function.
cc1df30b 21680 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
5a2d2a79 21681
5abe1e05
RB
216822013-03-21 Richard Biener <rguenther@suse.de>
21683
21684 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21685 (vect_insert_into_interleaving_chain): Likewise.
21686 (vect_drs_dependent_in_basic_block): Inline ...
21687 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21688 split out from ...
21689 (vect_analyze_data_ref_dependence): ... here. Simplify.
21690 (vect_check_interleaving): Simplify.
21691 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21692 (vect_slp_analyze_data_ref_dependences): ... this new function.
21693 (dr_group_sort_cmp): New function.
21694 (vect_analyze_data_ref_accesses): Compute data-reference groups
21695 here instead of in vect_analyze_data_ref_dependence. Use
21696 a more efficient algorithm.
21697 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21698 vect_slp_analyze_data_ref_dependences. Call
21699 vect_analyze_data_ref_accesses earlier.
21700 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21701 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21702 (vect_slp_analyze_data_ref_dependences): New prototype.
21703
cad1735b
RB
217042013-03-21 Richard Biener <rguenther@suse.de>
21705
21706 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21707 ref is stored in the loop.
21708 (find_refs_for_sm): Walk only over all stores.
21709 (store_motion_loop): Allocate from lim_bitmap_obstack.
21710 (store_motion): Likewise.
21711
141310ef
RB
217122013-03-21 Richard Biener <rguenther@suse.de>
21713
21714 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21715 Update virtual SSA form.
21716
5022315a
RO
217172013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21718
21719 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21720 * configure: Regenerate.
21721 * config.in: Regenerate.
21722 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21723 if !HAVE_LD_EH_FRAME_CIEV3.
21724
50d4421c
RB
217252013-03-21 Richard Biener <rguenther@suse.de>
21726
21727 * tree-cfg.c (verify_expr_no_block): New function.
21728 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21729 nor DECL_VALUE_EXPR have locations with associated blocks.
21730 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21731 (clear_unused_block_pointer): Remove code dealing with
21732 blocks in DECL_DEBUG_EXPR locations.
21733
839b422f
RB
217342013-03-21 Richard Biener <rguenther@suse.de>
21735
21736 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21737 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21738 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21739 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21740 instead of DECL_DEBUG_EXPR_IS_FROM.
21741 * gimplify.c (gimplify_modify_expr): Likewise.
21742 * tree-cfg.c (verify_expr_location_1): Likewise.
21743 * tree-complex.c (create_one_component_var): Likewise.
21744 * tree-sra.c (create_access_replacement): Likewise.
21745 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21746 (clear_unused_block_pointer): Likewise.
21747 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21748 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
21749 * var-tracking.c (var_debug_decl): Likewise.
21750 (track_expr_p): Likewise.
21751 * tree-inline.c (add_local_variables): Likewise. Set
21752 DECL_HAS_DEBUG_EXPR_P after copying it.
21753 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
21754 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
21755
fe04878d
UB
217562013-03-21 Uros Bizjak <ubizjak@gmail.com>
21757
21758 PR bootstrap/56656
21759 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
21760 * configure: Regenerate.
21761 * config.in: Regenerate.
21762 * config/i386/i386.md (*movdf_internal): Use
21763 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21764 movd instead of movq mnemonic for interunit moves.
21765 (*movdi_internal): Ditto.
21766
bd059b26 217672013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
fa2e9a58
N
21768
21769 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
21770 (abd<mode>_3): New pattern.
21771 (aba<mode>_3): New pattern.
21772 (fabd<mode>_3): New pattern.
21773
bd059b26 217742013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
50ce6f88
N
21775
21776 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
21777 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
21778 occurrence of REGISTER_PREFIX as its empty string.
21779
4f1f78b9
JL
217802013-03-20 Jeff Law <law@redhat.com>
21781
21782 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
21783 addititional equivalences for equality comparisons between an SSA_NAME
21784 and a constant where the SSA_NAME was set from a widening conversion.
21785
327a1118
WL
217862013-03-20 Walter Lee <walt@tilera.com>
21787
21788 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
21789
813e0036
UB
217902013-03-20 Uros Bizjak <ubizjak@gmail.com>
21791
21792 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
21793 depending on type attribute.
21794 (*movti_internal): Ditto.
21795 (*movtf_internal): Ditto.
21796 (*movxf_internal): Ditto.
21797 (*movdf_internal): Ditto.
21798 (*movsf_internal): Ditto.
21799
7cf34aae
UB
218002013-03-20 Uros Bizjak <ubizjak@gmail.com>
21801
21802 * config/i386/i386.md (*movti_internal): Set prefix attribute to
21803 maybe_vex for sselog1 and ssemov types.
21804 (*movdi_internal): Reorder operand constraints.
21805 (*movsi_internal): Ditto. Set prefix attribute to
21806 maybe_vex for sselog1 and ssemov types.
21807 (*movtf_internal): Set prefix attribute to maybe_vex
21808 for sselog1 and ssemov types.
21809 (*movdf_internal): Ditto. Set prefix_data16 attribute for
21810 DImode ssemov types. Reorder operand constraints.
21811 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
21812 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
21813 attribute for SImode ssemov types. Reorder operand constraints.
21814
19321415
MJ
218152013-03-20 Martin Jambor <mjambor@suse.cz>
21816
21817 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
21818 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
21819
12211b99 218202013-03-20 Pat Haugen <pthaugen@us.ibm.com>
c6d5ff83
MM
21821
21822 * config/rs6000/predicates.md (indexed_address, update_address_mem
21823 update_indexed_address_mem): New predicates.
21824 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
21825 attribute for load/store instructions.
21826 * config/rs6000/dfp.md (movsd_store): Likewise.
21827 (movsd_load): Likewise.
21828 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
21829 (unnamed HI->DI extend define_insn): Likewise.
21830 (unnamed SI->DI extend define_insn): Likewise.
21831 (unnamed QI->SI extend define_insn): Likewise.
21832 (unnamed QI->HI extend define_insn): Likewise.
21833 (unnamed HI->SI extend define_insn): Likewise.
21834 (unnamed HI->SI extend define_insn): Likewise.
21835 (extendsfdf2_fpr): Likewise.
21836 (movsi_internal1): Likewise.
21837 (movsi_internal1_single): Likewise.
21838 (movhi_internal): Likewise.
21839 (movqi_internal): Likewise.
21840 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
21841 attribute for load/store instructions.
21842 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
21843 instructions.
21844 (mov<mode>_softfloat): Likewise.
21845 (mov<mode>_hardfloat32): Likewise.
21846 (mov<mode>_hardfloat64): Likewise.
21847 (mov<mode>_softfloat64): Likewise.
21848 (movdi_internal32): Likewise.
21849 (movdi_internal64): Likewise.
21850 (probe_stack_<mode>): Likewise.
21851
218522013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
21853
21854 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
21855 floating point, and decimal floating point to reload iterator.
21856
21857 * config/rs6000/constraints.md (wl constraint): New constraints to
21858 return FLOAT_REGS if certain options are used to reduce the number
21859 of separate patterns that exist in the file.
21860 (wx constraint): Likewise.
21861 (wz constraint): Likewise.
21862
21863 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21864 -mdebug=reg, print wg, wl, wx, and wz constraints.
21865 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21866 Initialize the reload functions for 64-bit binary/decimal floating
21867 point types.
21868 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21869 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21870 create the buffer on the stack to overcome not having a 32-bit
21871 load and store.
21872 (rs6000_emit_move): Likewise.
21873 (rs6000_secondary_memory_needed_rtx): Likewise.
21874 (rs6000_alloc_sdmode_stack_slot): Likewise.
21875 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21876 via xxlxor, just like DFmode 0.0.
21877
21878 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
21879 define as 1 if we are running on a power7 or newer.
21880 (enum r6000_reg_class_enum): Add new constraints.
21881
21882 * config/rs6000/dfp.md (movsd): Delete, combine with binary
21883 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21884 with other moves by using conditional constraits (wg). Use LFIWZX
bd059b26 21885 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
c6d5ff83
MM
21886 (movsd splitter): Likewise.
21887 (movsd_hardfloat): Likewise.
21888 (movsd_softfloat): Likewise.
21889
21890 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
21891 binary and decimal floating point moves.
21892 (fmove_ok): New attributes to combine binary and decimal floating
21893 point moves, and to combine power6x (mfpgpr) moves along normal
21894 floating moves.
21895 (real_value_to_target): Likewise.
21896 (f32_lr): Likewise.
21897 (f32_lm): Likewise.
21898 (f32_li): Likewise.
21899 (f32_sr): Likewise.
21900 (f32_sm): Likewise.
21901 (f32_si): Likewise.
21902 (movsf): Combine binary and decimal floating point moves. Combine
21903 power6x (mfpgpr) moves with other moves by using conditional
bd059b26 21904 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
c6d5ff83
MM
21905 (mov<mode> for SFmode/SDmode); Likewise.
21906 (SFmode/SDmode splitters): Likewise.
21907 (movsf_hardfloat): Likewise.
21908 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
21909 (movsf_softfloat): Likewise.
21910 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
21911
21912 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
21913 wx and wz constraints.
21914
21915 * config/rs6000/constraints.md (wg constraint): New constraint to
21916 return FLOAT_REGS if -mmfpgpr (power6x) was used.
21917
21918 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
21919 constraint.
21920
21921 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21922 -mdebug=reg, print wg, wl, wx, and wz constraints.
21923 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21924 Initialize the reload functions for 64-bit binary/decimal floating
21925 point types.
21926 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21927 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21928 create the buffer on the stack to overcome not having a 32-bit
21929 load and store.
21930 (rs6000_emit_move): Likewise.
21931 (rs6000_secondary_memory_needed_rtx): Likewise.
21932 (rs6000_alloc_sdmode_stack_slot): Likewise.
21933 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21934 via xxlxor, just like DFmode 0.0.
21935
c6d5ff83
MM
21936 * config/rs6000/dfp.md (movdd): Delete, combine with binary
21937 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21938 with other moves by using conditional constraits (wg). Use LFIWZX
21939 and STFIWX for loading SDmode on power7.
21940 (movdd splitters): Likewise.
21941 (movdd_hardfloat32): Likewise.
21942 (movdd_softfloat32): Likewise.
21943 (movdd_hardfloat64_mfpgpr): Likewise.
21944 (movdd_hardfloat64): Likewise.
21945 (movdd_softfloat64): Likewise.
21946
21947 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
21948 64-bit binary and decimal floating point moves.
21949 (FMOVE64X): Likewise.
21950 (movdf): Combine 64-bit binary and decimal floating point moves.
21951 Combine power6x (mfpgpr) moves with other moves by using
21952 conditional constraits (wg).
21953 (mov<mode> for DFmode/DDmode): Likewise.
21954 (DFmode/DDmode splitters): Likewise.
21955 (movdf_hardfloat32): Likewise.
21956 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
21957 (movdf_softfloat32): Likewise.
21958 (movdf_hardfloat64_mfpgpr): Likewise.
21959 (movdf_hardfloat64): Likewise.
21960 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
21961 (movdf_softfloat64): Likewise.
21962 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
21963 (reload_<mode>_load): Move to later in the file so they aren't in
21964 the middle of the floating point move insns.
21965 (reload_<mode>_store): Likewise.
21966
21967 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
21968 constraint.
21969
21970 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
21971 constraint if -mdebug=reg.
bd059b26
UB
21972 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
21973 Enable using dd reload support if needed.
c6d5ff83
MM
21974
21975 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
21976 binary and decimal floating point moves in rs6000.md.
21977 (movtd_internal): Likewise.
21978
21979 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
21980 decimal floating point moves.
21981 (movtf): Likewise.
21982 (movtf_internal): Likewise.
21983 (mov<mode>_internal, TDmode/TFmode): Likewise.
21984 (movtf_softfloat): Likewise.
21985 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
21986
21987 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
21988 movdi_internal64, using wg constraint for move direct operations.
21989 (movdi_internal64): Likewise.
21990
21991 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
21992 MODES_TIEABLE_P for selected modes. Print the numerical value of
21993 the various virtual registers. Use GPR/FPR first/last values,
21994 instead of hard coding the register numbers. Print which modes
21995 have reload functions registered.
bd059b26
UB
21996 (rs6000_option_override_internal): If -mdebug=reg, trace the options
21997 settings before/after setting cpu, target and subtarget settings.
21998 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
21999 and for secondary reload failures in rs6000_secondary_reload_inner.
c6d5ff83
MM
22000 (rs6000_secondary_reload_fail): Likewise.
22001 (rs6000_secondary_reload_inner): Likewise.
22002
22003 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
22004 macros for first/last GPR and FPR registers.
22005 (LAST_GPR_REGNO): Likewise.
22006 (FIRST_FPR_REGNO): Likewise.
22007 (LAST_FPR_REGNO): Likewise.
22008
22009 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
22010 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
22011 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
22012 (vcond<mode><mode>): Likewise.
22013 (vcondu<mode><mode>): Likewise.
22014 (vector_gtu<mode>): Likewise.
22015 (vector_gte<mode>): Likewise.
22016 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
bd059b26 22017 to prevent the compiler from converting DImode operations to TImode.
c6d5ff83
MM
22018 (ior<mode>3): Likewise.
22019 (and<mode>3): Likewise.
22020 (one_cmpl<mode>2): Likewise.
22021 (nor<mode>3): Likewise.
22022 (andc<mode>3): Likewise.
22023
22024 * config/rs6000/constraints.md (wt constraint): New constraint
22025 that returns VSX_REGS if TImode is allowed in VSX registers.
22026
22027 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
22028 constant under VSX.
22029
22030 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
22031 similar to TImode, but it is restricted to being in the GPRs.
22032
22033 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22034 TImode to occupy a single VSX register.
22035
22036 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22037 -mvsx-timode for power7/power8.
22038 (power7 cpu): Likewise.
22039 (power8 cpu): Likewise.
22040
22041 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22042 sure that TFmode/TDmode take up two registers if they are ever
22043 allowed in the upper VSX registers.
22044 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22045 registers.
22046 (rs6000_init_hard_regno_mode_ok): Likewise.
22047 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22048 constraint. Print if LRA is turned on.
22049 (rs6000_option_override_internal): Give an error if -mvsx-timode
22050 and VSX is not enabled.
22051 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22052 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22053 to reg+offset addressing. Use PTImode when checking offset
22054 addresses for validity.
22055 (reg_offset_addressing_ok_p): Likewise.
22056 (rs6000_legitimate_offset_address_p): Likewise.
22057 (rs6000_legitimize_address): Likewise.
22058 (rs6000_legitimize_reload_address): Likewise.
22059 (rs6000_legitimate_address_p): Likewise.
22060 (rs6000_eliminate_indexed_memrefs): Likewise.
22061 (rs6000_emit_move): Likewise.
22062 (rs6000_secondary_reload): Likewise.
22063 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22064 reloads to fpr registers to continue to use reg+offset addressing,
22065 but 64-bit reloads to altivec registers need reg+reg addressing.
22066 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22067 it. Treat LO_SUM like a PLUS operation.
22068 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
bd059b26 22069 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
c6d5ff83
MM
22070 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22071 registers to share a register with a smaller sized type, since VSX
22072 puts scalars in the upper 64-bits.
22073 (print_operand): Add support for PTImode.
22074 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22075 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22076 registers, but don't have arithmetic support.
22077 (rs6000_memory_move_cost): Add test for VSX.
22078 (rs6000_opt_masks): Add -mvsx-timode.
22079
22080 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22081 for TImode.
22082 (VSs): Likewise.
22083 (VSr): Use wt constraint for TImode.
22084 (VSv): Drop TImode support.
22085 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22086 (vsx_movti_64bit): Likewise.
22087 (vsx_movti_32bit): Likewise.
22088 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22089 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22090 one '?' on the appropriate output constraint. Do not allow TImode
22091 logical operations on 32-bit systems.
22092 (vsx_ior<mode>3): Likewise.
22093 (vsx_xor<mode>3): Likewise.
22094 (vsx_one_cmpl<mode>2): Likewise.
22095 (vsx_nor<mode>3): Likewise.
22096 (vsx_andc<mode>3): Likewise.
22097 (vsx_concat_<mode>): Likewise.
22098 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22099
22100 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22101 OPTION_MASK_VSX_TIMODE.
22102 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22103 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22104
22105 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22106 (TI2 iterator): New iterator for TImode, PTImode.
22107 (wd mode attribute): Add values for vector types.
bd059b26
UB
22108 (movti_string): Replace TI move operations with operations for TImode
22109 and PTImode. Add support for TImode being allowed in VSX registers.
c6d5ff83
MM
22110 (mov<mode>_string, TImode/PTImode): Likewise.
22111 (movti_ppc64): Likewise.
22112 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22113 (TI mode splitters): Likewise.
22114
22115 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22116 constraint.
22117
1fc5eced
MG
221182013-03-20 Marc Glisse <marc.glisse@inria.fr>
22119
22120 PR tree-optimization/56355
22121 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22122 Also handle integers with undefined overflow.
22123
22c4c869
CM
221242013-03-20 Catherine Moore <clm@codesourcery.com>
22125 Maciej W. Rozycki <macro@codesourcery.com>
22126 Tom de Vries <tom@codesourcery.com>
12211b99 22127 Nathan Sidwell <nathan@codesourcery.com>
22c4c869
CM
22128 Iain Sandoe <iain@codesourcery.com>
22129 Nathan Froyd <froydnj@codesourcery.com>
12211b99 22130 Chao-ying Fu <fu@mips.com>
22c4c869
CM
22131
22132 * doc/extend.texi: (micromips, nomicromips, nocompression):
7cf34aae 22133 Document new function attributes.
22c4c869
CM
22134 * doc/invoke.texi (minterlink-compressed, mmicromips,
22135 m14k, m14ke, m14kec): Document new options.
22136 (minterlink-mips16): Update documentation.
22137 * doc/md.texi (ZC, ZD): Document new constraints.
22138 * configure.ac (gcc_cv_as_micromips): Check if linker
22139 supports the .set micromips directive.
22140 * configure: Regenerate.
22141 * config.in: Regenerate.
22142 * config/mips/mips-tables.opt: Regenerate.
22143 * config/mips/micromips.md: New file.
22144 * constraints.md (ZC, ZD): New constraints.
22145 * config/mips/predicates.md (movep_src_register): New predicate.
22146 (movep_src_operand): New predicate.
22147 (non_volatile_mem_operand): New predicate.
22148 * config/mips/mips.md (multimem): New type.
22149 (length): Differentiate between 17-bit and 18-bit branch offsets.
22150 (MOVEP1, MOVEP2): New mode iterator.
33159866 22151 (mov_<load>l): Use ZC constraint.
22c4c869
CM
22152 (mov_<load>r): Likewise.
22153 (mov_<store>l): Likewise.
22154 (mov_<store>r): Likewise.
22155 (*branch_equality<mode>_inverted): Add microMIPS support.
22156 (*branch_equality<mode>): Likewise.
22157 (*jump_absolute): Likewise.
22158 (indirect_jump_<mode>): Likewise.
22159 (tablejump_<mode>): Likewise.
22160 (<optab>_internal): Likewise.
22161 (sibcall_internal): Likewise.
22162 (sibcall_value_internal): Likewise.
22163 (prefetch): Use constraint ZD.
22164 * config/mips/mips.opt (minterlink-compressed): New option.
22165 (minterlink-mips16): Now an alias for minterlink-compressed.
22166 (mmicromips): New option.
22167 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22168 (compare_and_swap_12): Likewise.
22169 (sync_add<mode>): Likewise.
22170 (sync_<optab>_12): Likewise.
22171 (sync_old_<optab>_12): Likewise.
22172 (sync_new_<optab>_12): Likewise.
22173 (sync_nand_12): Likewise.
22174 (sync_old_nand_12): Likewise.
22175 (sync_new_nand_12): Likewise.
22176 (sync_sub<mode>): Likewise.
22177 (sync_old_add<mode>): Likewise.
22178 (sync_old_sub<mode>): Likewise.
22179 (sync_new_add<mode>): Likewise.
22180 (sync_new_sub<mode>): Likewise.
22181 (sync_<optab><mode>): Likewise.
22182 (sync_old_<optab><mode>): Likewise.
22183 (sync_new_<optab><mode>): Likewise.
22184 (sync_nand<mode>): Likewise.
22185 (sync_old_nand<mode>): Likewise.
22186 (sync_new_nand<mode>): Likewise.
22187 (sync_lock_test_and_set<mode>): Likewise.
22188 (test_and_set_12): Likewise.
22189 (atomic_compare_and_swap<mode>): Likewise.
22190 (atomic_exchange<mode>_llsc): Likewise.
22191 (atomic_fetch_add<mode>_llsc): Likewise.
22192 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22193 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22194 (umips_save_restore_pattern_p): Likewise.
22195 (umips_load_store_pair_p): Likewise.
22196 (umips_output_load_store_pair): Likewise.
22197 (umips_movep_target_p): Likewise.
22198 (umips_12bit_offset_address_p): Likewise.
22199 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22200 (mips_base_mips16): Rename this...
22201 (mips_base_compression_flags): ...to this. Update all uses.
22202 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22203 (mips_mips16_decl_p): Delete.
22204 (mips_nomips16_decl_p): Delete.
33159866 22205 (mips_get_compress_on_flags): New function.
22c4c869
CM
22206 (mips_get_compress_off_flags): New function.
22207 (mips_get_compress_mode): New function.
22208 (mips_get_compress_on_name): New function.
22209 (mips_get_compress_off_name): New function.
22210 (mips_insert_attributes): Support multiple compression types.
22211 (mips_merge_decl_attributes): Likewise.
22212 (umips_12bit_offset_address_p): New function.
22213 (mips_start_function_definition): Emit .set micromips directive.
22214 (mips_call_may_need_jalx_p): New function.
22215 (mips_function_ok_for_sibcall): Add microMIPS support.
22216 (mips_print_operand_punctuation): Support short delay slots and
22217 compact jumps.
22218 (umips_swm_mask, umips_swm_encoding): New.
22219 (umips_build_save_restore): New function.
22220 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22221 (was_mips16_p): Remove.
22222 (old_compression_mode): New.
22223 (mips_set_compression_mode): New function.
22224 (mips_set_current_function): Add microMIPS support.
22225 (mips_option_override): Likewise.
22226 (umips_save_restore_pattern_p): New function.
22227 (umips_output_save_restore): New function.
22228 (umips_load_store_pair_p_1): New function.
22229 (umips_load_store_pair_p): New function.
22230 (umips_output_load_store_pair_1): New function.
22231 (umips_output_load_store_pair): New function.
22232 (umips_movep_target_p) New function.
22233 (mips_prepare_pch_save): Add microMIPS support.
22234 * config/mips/mips.h (TARGET_COMPRESSION): New.
22235 (TARGET_CPU_CPP_BUILTINS): Update macro
22236 to use new compression flags and to support microMIPS.
22237 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22238 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22239 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22240 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22241 (ASM_SPEC): Support mmicromips and mno-micromips.
22242 (M16STORE_REG_P): New macro.
22243 (MIPS_CALL): Support TARGET_MICROMIPS.
22244 (MICROMIPS_J): New macro.
22245 (mips_base_mips16): Rename this...
22246 (mips_base_compression_flags): ...to this.
22247 (UMIPS_12BIT_OFFSET_P): New macro.
22248 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22249 (MULTILIB_DIRNAMES): Likewise.
92a8d7a7
RB
222502013-03-20 Richard Biener <rguenther@suse.de>
22251
22252 PR tree-optimization/56661
22253 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22254 the result does not have to be distinct.
22255
54714c68
RB
222562013-03-20 Richard Biener <rguenther@suse.de>
22257
22258 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22259 remap_gimple_op_r.
22260
cca1130d 222612013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
12211b99 22262 Steven Bosscher <steven@gcc.gnu.org>
cca1130d
BS
22263
22264 PR rtl-optimization/56605
22265 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22266
2a930377
UB
222672013-03-20 Uros Bizjak <ubizjak@gmail.com>
22268
22269 PR bootstrap/56656
22270 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22271 that require movd instead of movq.
22272
d6d305fe
RB
222732013-03-20 Richard Biener <rguenther@suse.de>
22274
22275 * tree-ssa-structalias.c (struct variable_info): Add pointer
22276 to the first field of an aggregate with sub-vars. Make
22277 this and the pointer to the next subfield its ID.
22278 (vi_next): New function.
22279 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22280 storedanything_id, integer_id): Increment by one.
22281 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22282 get_call_clobber_vi): Adjust.
22283 (solution_set_expand): Simplify and speedup.
22284 (solution_set_add): Inline into ...
22285 (set_union_with_increment): ... this. Adjust accordingly.
22286 (do_sd_constraint): Likewise.
22287 (do_ds_constraint): Likewise.
22288 (do_complex_constraint): Simplify.
22289 (build_pred_graph): Adjust.
22290 (solve_graph): Likewise. Simplify and speedup.
22291 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22292 get_constraint_for_component_ref, get_constraint_for_1,
22293 first_vi_for_offset, first_or_preceding_vi_for_offset,
22294 create_function_info_for, create_variable_info_for_1,
22295 create_variable_info_for, intra_create_variable_infos): Adjust.
22296 (init_base_vars): Push NULL for ID zero.
22297 (compute_points_to_sets): Adjust.
22298
a271b42d
RB
222992013-03-20 Richard Biener <rguenther@suse.de>
22300
22301 * cfgloop.c (verify_loop_structure): Streamline and avoid
22302 ICEing on corrupt loop tree.
22303 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22304 loop tree.
22305
ebd65954
RB
223062013-03-20 Richard Biener <rguenther@suse.de>
22307
22308 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22309 check whether an SSA update is needed.
22310
4547b7ee
RS
223112013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22312
22313 * config/mips/constraints.md (T): Rename to...
22314 (Yf): ...this.
22315 (U): Rename to...
22316 (Yd): ...this.
22317 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22318 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22319
0b8cdc58
IB
223202013-03-19 Ian Bolton <ian.bolton@arm.com>
22321
22322 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22323 (*subsi3_carryin_uxtw): Likewise.
22324
5977a10d
IB
223252013-03-19 Ian Bolton <ian.bolton@arm.com>
22326
22327 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22328 (*rorsi3_insn_uxtw): Likewise.
22329
bd83ff2c
IB
223302013-03-19 Ian Bolton <ian.bolton@arm.com>
22331
22332 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22333 (*extrsi5_insn_uxtw): Likewise.
22334
ba65123c
RB
223352013-03-19 Richard Biener <rguenther@suse.de>
22336
22337 PR tree-optimization/56273
22338 * passes.c (init_optimization_passes): Move second VRP after DOM.
22339
2eac0476
UB
223402013-03-19 Uros Bizjak <ubizjak@gmail.com>
22341
22342 * config/i386/i386.md (*movti_internal): Merge from
22343 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22344 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22345 nox64 isa attributes.
22346
6e55eda7
RB
223472013-03-18 Richard Biener <rguenther@suse.de>
22348
22349 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22350 (unite): Likewise.
22351 (merge_node_constraints): Likewise.
22352 (build_succ_graph): Likewise.
22353 (valid_graph_edge): Inline into single caller.
22354 (unify_nodes): Likewise. Use bitmap_set_bit return value
22355 and cache varinfo.
22356 (scc_visit): Fix formatting and variable use.
22357 (do_sd_constraint): Use gcc_checking_assert.
22358 (do_ds_constraint): Likewise.
22359 (do_complex_constraint): Likewise.
22360 (condense_visit): Likewise. Cleanup.
22361 (dump_pred_graph): New function.
22362 (perform_var_substitution): Dump the pred-graph before
22363 variable substitution.
22364 (find_equivalent_node): Use gcc_checking_assert.
22365 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22366
4bdd44c4
RB
223672013-03-18 Richard Biener <rguenther@suse.de>
22368
22369 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22370 Remove cond_expr_stmt_list argument and do not gimplify the
22371 built expression.
22372 (vect_loop_versioning): Adjust.
22373 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22374 Cleanup to use less temporaries.
22375 (vect_create_data_ref_ptr): Cleanup.
22376
38c56a5b
JJ
223772013-03-18 Jakub Jelinek <jakub@redhat.com>
22378
22379 PR tree-optimization/56635
22380 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22381 require types_compatible_p types.
22382
20b2e6a0
NC
223832013-03-18 Nick Clifton <nickc@redhat.com>
22384
a6178a25
NC
22385 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22386 spurious backslash.
22387
20b2e6a0
NC
22388 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22389 Add missing line to comment describing function.
22390
92608d0e
RB
223912013-03-18 Richard Biener <rguenther@suse.de>
22392
22393 PR tree-optimization/56210
22394 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22395 Handle string / character search functions.
22396 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22397
31348d52
RB
223982013-03-18 Richard Biener <rguenther@suse.de>
22399
22400 PR middle-end/56483
22401 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22402 and implement properly.
22403 * gimple.h (gimple_cond_single_var_p): Remove.
22404
fcac74a1
RB
224052013-03-18 Richard Biener <rguenther@suse.de>
22406
22407 * tree-data-ref.h (find_data_references_in_loop): Declare.
22408 * tree-data-ref.c (get_references_in_stmt): Use a stack
22409 vector pre-allocated in the callers.
22410 (find_data_references_in_stmt): Adjust.
22411 (graphite_find_data_references_in_stmt): Likewise.
22412 (create_rdg_vertices): Likewise.
22413 (find_data_references_in_loop): Export.
22414 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22415 Compute dependences here...
22416 (vect_analyze_data_refs): ...not here. When we encounter
22417 a non-vectorizable data reference in basic-block vectorization
22418 truncate the data reference vector. Do not bother to
22419 fixup data-dependence information for gather loads.
22420 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22421 of data references, as reported.
22422
0d5a1b56
RB
224232013-03-18 Richard Biener <rguenther@suse.de>
22424
22425 PR tree-optimization/3713
22426 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22427 has_constants and expr.
22428 (stmt_has_constants): Properly valueize SSA names when deciding
22429 whether the stmt has constants.
22430
789c34e3
RB
224312013-03-18 Richard Biener <rguenther@suse.de>
22432
22433 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22434 whole function when there is nothing to do.
22435 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22436 * tree-vectorizer.c (vectorize_loops): Update virtual and
22437 loop-closed SSA once.
22438 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22439
076b4605
RB
224402013-03-18 Richard Biener <rguenther@suse.de>
22441
22442 PR middle-end/56113
22443 * domwalk.c (bb_postorder): New global static.
22444 (cmp_bb_postorder): New function.
22445 (walk_dominator_tree): Replace scheme imposing an order for
22446 visiting dominator sons by one sorting them at the time they
22447 are pushed on the stack.
22448
bdb01696
RB
224492013-03-18 Richard Biener <rguenther@suse.de>
22450
22451 PR tree-optimization/39326
22452 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22453 (struct mem_ref): Replace mem member with ao_ref typed member.
22454 (MEM_ANALYZABLE): Adjust.
22455 (memref_eq): Likewise.
22456 (mem_ref_alloc): Likewise.
22457 (gather_mem_refs_stmt): Likewise.
22458 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22459 (execute_sm_if_changed_flag_set): Adjust.
22460 (execute_sm): Likewise.
22461 (ref_always_accessed_p): Likewise.
22462 (refs_independent_p): Likewise.
22463 (can_sm_ref_p): Likewise.
22464
12d2dc5e
JJ
224652013-03-18 Jakub Jelinek <jakub@redhat.com>
22466
22467 PR c/56566
22468 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22469 return 1 even for !unsignedp.
22470
286fb677
UB
224712013-03-17 Uros Bizjak <ubizjak@gmail.com>
22472
22473 * config/i386/i386.md (isa): Add x64 and nox64.
22474 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22475 (*pushtf): Enable *roF alternative for x64 isa only.
22476 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22477 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22478 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22479 (*movtf_internal): Merge from *movtf_internal_rex64 and
22480 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22481 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22482 nox64 isa attributes.
22483 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22484 nox64 isa attributes.
22485 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22486
88b97037
UB
224872013-03-17 Uros Bizjak <ubizjak@gmail.com>
22488
22489 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22490
9f4f1735
JJ
224912013-03-17 Jakub Jelinek <jakub@redhat.com>
22492
22493 PR target/56640
22494 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22495 class names. Remove trailing comma after "ALL_REGS".
22496
48f4a6fa
JH
224972013-03-16 Jan Hubicka <jh@suse.cz>
22498
22499 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22500 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22501 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22502 of cgraph_get_create_node.
22503 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22504
98e81123
JM
225052013-03-16 Jason Merrill <jason@redhat.com>
22506
22507 PR debug/49090
22508 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22509 with DW_AT_default_value.
22510
8a298c04
JJ
225112013-03-16 Jakub Jelinek <jakub@redhat.com>
22512
22513 * BASE-VER: Set to 4.9.0.
22514
4323afa0
AK
225152013-03-14 Andi Kleen <ak@linux.intel.com>
22516
22517 PR target/56619
22518 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22519 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22520 Document _x* TSX intrinsics.
22521
b3c0d469
JJ
225222013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22523 David Holsgrove <david.holsgrove@xilinx.com>
8cc9a5a5
EI
22524
22525 * configure.ac: Add MicroBlaze TLS support detection.
22526 * configure: Regenerate.
88b97037
UB
22527 * config/microblaze/microblaze-protos.h
22528 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22529 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22530 * config/microblaze/microblaze.c (microblaze_address_type): Add
22531 ADDRESS_TLS and tls_reloc address types.
8cc9a5a5
EI
22532 (microblaze_address_info): Add tls_reloc.
22533 (TARGET_HAVE_TLS): Define.
22534 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22535 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
88b97037
UB
22536 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22537 load_tls_operand, microblaze_call_tls_get_addr,
22538 microblaze_legitimize_tls_address): New functions.
8cc9a5a5
EI
22539 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22540 (get_base_reg): Use microblaze_tls_symbol_p.
22541 (microblaze_classify_address): Handle TLS.
88b97037
UB
22542 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22543 label_mentioned_p and microblaze_tls_referenced_p.
8cc9a5a5
EI
22544 (microblaze_legitimize_address): Handle TLS.
22545 (microblaze_address_insns): Handle ADDRESS_TLS.
22546 (pic_address_needs_scratch): Handle TLS.
22547 (print_operand_address): Handle TLS.
22548 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22549 (microblaze_expand_move): Handle TLS.
88b97037
UB
22550 (microblaze_legitimate_constant_p): Check
22551 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
8cc9a5a5
EI
22552 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22553 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22554 (PIC_OFFSET_TABLE_REGNUM): Set.
22555 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22556 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22557 (addsi3, movsi_internal2, movdf_internal): Update constraints
22558 * config/microblaze/predicates.md (arith_plus_operand): Define
88b97037
UB
22559 (move_operand): Redefine as move_src_operand,
22560 check microblaze_tls_referenced_p.
8cc9a5a5 22561
d803a491
IB
225622013-03-14 Ian Bolton <ian.bolton@arm.com>
22563
22564 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22565 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22566
a8504f22
IB
225672013-03-14 Ian Bolton <ian.bolton@arm.com>
22568
22569 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22570 CC mode for AND.
22571
df2dfaea
JJ
225722013-03-14 Jakub Jelinek <jakub@redhat.com>
22573
fbd28bc3
JJ
22574 PR tree-optimization/53265
22575 * common.opt (Waggressive-loop-optimizations): New option.
22576 * tree-ssa-loop-niter.c: Include tree-pass.h.
22577 (do_warn_aggressive_loop_optimizations): New function.
22578 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22579 if number_of_latch_executions returned constant.
22580 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22581 early. If number_of_latch_executions returned constant, set
22582 nb_iterations_upper_bound back to it.
22583 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22584 field.
22585 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22586 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22587
df2dfaea
JJ
22588 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22589 (MULTILIB_OSDIRNAMES): Set.
22590 * genmultilib: If defaultosdirname doesn't start with :: , set
22591 defaultosdirname2 instead, clear it and emit two . multilib_raw
22592 entries instead of just one.
22593
ee0d2b68
KK
225942013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22595
22596 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22597 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22598 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22599 (SUBTARGET_OVERRIDE_OPTIONS): New.
22600
decc676e
OE
226012013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22602
22603 PR target/49880
22604 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22605 (musermode): Convert to Var(TARGET_USERMODE).
22606 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22607 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22608 * config/sh/sh.c (sh_option_override): Use
22609 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22610 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22611 condition.
22612 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22613 TARGET_SH4.
22614 (udivsi3_i4_single, divsi3_i4_single): Use
22615 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22616
f4b719c7
DK
226172013-03-13 Dave Korn <dave.korn.cygwin@....>
22618
22619 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22620 default setting.
22621
c40eced0
RB
226222013-03-13 Richard Biener <rguenther@suse.de>
22623
22624 PR tree-optimization/56608
22625 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22626 calls when vectorizing basic-blocks.
22627
1bfa5973
JJ
226282013-03-13 Jakub Jelinek <jakub@redhat.com>
22629
22630 PR plugins/45078
22631 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22632 tm_file.
22633
6fcf5434
JJ
226342013-03-12 Jakub Jelinek <jakub@redhat.com>
22635
22636 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22637
4f38fa8c
JH
226382013-03-11 Jan Hubicka <jh@suse.cz>
22639
22640 PR lto/56557
22641 * lto-streamer-out.c (output_symbol_p): Skip references from
22642 constructors of external variables.
22643
c5c5ba89
JH
226442013-03-11 Jan Hubicka <jh@suse.cz>
22645
22646 PR middle-end/56571
22647 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22648 from pseudos.
22649 * emit-rtl.c (verify_rtx_sharing): Likewise.
22650 (copy_insn_1): Likewise.
22651 * rtl.c (copy_rtx): Likewise.
22652
c2a939b1
GJL
226532013-03-11 Georg-Johann Lay <avr@gjlay.de>
22654
22655 PR target/56591
22656 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22657 output_operand_lossage message.
88b97037 22658
3c3279fb
RE
226592013-03-11 Richard Earnshaw <rearnsha@arm.com>
22660
22661 PR target/56470
22662 * arm.c (shift_op): Validate RTL pattern on the fly.
22663 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22664 the RTL.
22665
aef5ef9d
JDA
226662013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22667
22668 PR target/56347
22669 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22670 new call patterns that clobber %fr12.
22671 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22672 split and postreload patterns.
22673 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22674 registers %fr12 and %fr12R as call used.
22675
456610d3
SB
226762013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22677
22678 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22679 (canon_address, record_store, replace_read, check_mem_read_rtx,
22680 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22681 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22682 rest_of_handle_dse): Likewise.
22683
4b1baac8
RS
226842013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22685
22686 PR middle-end/56524
22687 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22688 Add base_optabs.
22689 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22690 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22691 (save_optabs_if_changed): Replace with...
22692 (init_tree_optimization_optabs): ...this.
22693 * optabs.c (save_optabs_if_changed): Rename to...
22694 (init_tree_optimization_optabs): ...this. Take the optimization node
22695 as argument. Do nothing if the base optabs are already correct.
22696 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22697 to recompute optabs.
22698 * function.h (function): Remove optabs field.
22699 * function.c (invoke_set_current_function_hook): Call
22700 init_tree_optimization_optabs. Use the result to initialize
22701 this_fn_optabs.
22702
b7a78683
AH
227032013-02-27 Aldy Hernandez <aldyh@redhat.com>
22704
22705 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22706 if GTMA_HAS_NO_INSTRUMENTATION.
22707 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22708 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22709 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22710 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22711 GTMA_HAS_NO_INSTRUMENTATION.
22712
6384c29b
JJ
227132013-03-08 Jakub Jelinek <jakub@redhat.com>
22714
22715 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22716 libasan_preinit.o.
22717
ecd4f20a
MP
227182013-03-08 Marek Polacek <polacek@redhat.com>
22719 Jakub Jelinek <jakub@redhat.com>
22720
22721 PR tree-optimization/56478
22722 * predict.c (is_comparison_with_loop_invariant_p): Change the
22723 type of loop_step to tree.
22724 (predict_loops): Adjust.
22725 (predict_iv_comparison): Perform the computations on double_ints.
22726
64366d35
RB
227272013-03-08 Richard Biener <rguenther@suse.de>
22728
22729 PR tree-optimization/56570
22730 * tree-cfg.c (verify_expr_location_1): Verify locations for
22731 DECL_DEBUG_EXPR.
22732 * tree-sra.c (create_access_replacement): Strip locations
22733 from DECL_DEBUG_EXPRs.
22734
a9d5a059
RB
227352013-03-08 Richard Biener <rguenther@suse.de>
22736
22737 * tree-inline.c (expand_call_inline): Do not associate
22738 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22739 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22740
b7aa4e9a
RB
227412013-03-08 Richard Biener <rguenther@suse.de>
22742
22743 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22744 or block changes with -Og. Fix for location / block encoding
22745 changes and PHI arguments with locations.
22746
c4c2f9fa
SB
227472013-03-07 Steven Bosscher <steven@gcc.gnu.org>
22748
22749 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
22750 for all counters.
22751 (struct output_info): Likewise.
22752 (register_overhead): Remove bad gcc_assert.
22753 (bitmap_find_bit): If there is only a single bitmap element, do not
22754 count a miss as a search.
22755 (print_statistics): Update for counter type changes.
22756 (dump_bitmap_statistics): Likewise. Print headers such that they
22757 are properly lined up with the printed counters.
22758
5bf6606a
JJ
227592013-03-07 Jakub Jelinek <jakub@redhat.com>
22760
22761 PR tree-optimization/56559
22762 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
22763 check that it has only a single use.
22764
2c653d46
RB
227652013-03-07 Richard Biener <rguenther@suse.de>
22766
22767 * doc/invoke.texi (fwhole-program): Discourage use in combination
22768 with -flto.
22769
a72d8780
JJ
227702013-03-06 Jakub Jelinek <jakub@redhat.com>
22771
01a454df
JJ
22772 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
22773
c1781047
JJ
22774 PR tree-optimization/56539
22775 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
22776 instead of GSI_CONTINUE_LINKING as last argument to
22777 force_gimple_operand_gsi. Adjust function comment.
22778
9772c47a
JJ
22779 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
22780 aarch64-cores.def.
22781
a72d8780
JJ
22782 PR middle-end/56548
22783 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
22784 promoted mode, convert the result back to the original mode.
22785
fa5556de
RB
227862013-03-06 Richard Biener <rguenther@suse.de>
22787
22788 PR middle-end/56294
22789 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
22790 (insert_updated_phi_nodes_compare_uids): New function.
22791 (update_ssa): Sort symbols_to_rename after UID before
22792 traversing it to insert PHI nodes.
22793
010403d1
RB
227942013-03-06 Richard Biener <rguenther@suse.de>
22795
22796 PR middle-end/50494
22797 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
22798 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
22799
22800 Revert
22801 2013-02-13 Richard Biener <rguenther@suse.de>
22802
22803 PR lto/50494
22804 * varasm.c (output_constant_def_1): Get the decl representing
22805 the constant as argument.
22806 (output_constant_def): Wrap output_constant_def_1.
22807 (make_decl_rtl): Use output_constant_def_1 with the decl
22808 representing the constant.
22809 (build_constant_desc): Optionally re-use a decl already
22810 representing the constant.
22811 (tree_output_constant_def): Adjust.
22812
3c27ce4c
JY
228132013-03-06 Joey Ye <joey.ye@arm.com>
22814
22815 PR lto/50293
22816 * gcc.c (convert_white_space): New function.
22817 (main): Handles white space in function name.
22818
8f6d1c86
OE
228192013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
22820
22821 PR target/56529
22822 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
22823 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
22824 to SH_DIV_CALL_TABLE for TARGET_SH2.
22825 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
22826 list.
22827 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
22828 call-table options.
22829
798a209f
SA
228302013-03-05 Sterling Augustine <saugustine@google.com>
22831 Cary Coutant <ccoutant@google.com>
22832
22833 PR debug/55364
22834 * dwarf2out.c (resolve_addr): Don't call
22835 remove_loc_list_addr_table_entries a second time for the same
22836 expression.
22837
6cfa417f 228382013-03-05 Jakub Jelinek <jakub@redhat.com>
ee61ea38 22839
6cfa417f
JJ
22840 PR debug/56510
22841 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
22842 (avoid_complex_debug_insns): New function.
22843 (expand_debug_locations): Call it.
22844
22845 PR rtl-optimization/56484
22846 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
22847 lifetimes of hard registers on small register class machines.
22848
228492013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
22850
22851 * config/microblaze/microblaze-protos.h: Rename
ee61ea38 22852 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
6cfa417f 22853 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
ee61ea38
DH
22854 fast_interrupt.
22855 (microblaze_fast_interrupt_function_p): New function.
22856 (microblaze_is_interrupt_handler): Rename to
22857 microblaze_is_interrupt_variant and add fast_interrupt check.
22858 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
22859 (save_restore_insns): Likewise.
22860 (compute_frame_size): Likewise.
22861 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
22862 (microblaze_globalize_label): Likewise.
6cfa417f
JJ
22863 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
22864 * config/microblaze/microblaze.md: Use wrapper
ee61ea38
DH
22865 microblaze_is_interrupt_variant.
22866
731aefac
KT
228672013-03-05 Kai Tietz <ktietz@redhat.com>
22868
22869 * sdbout.c (sdbout_one_type): Switch to current function's section
22870 supporting cold/hot.
22871
a72d8780 228722013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
e34fd5a7
DH
22873
22874 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
22875 -mxl-reorder.
22876
334e71e8
JJ
228772013-03-05 Jakub Jelinek <jakub@redhat.com>
22878
0b50e654
JJ
22879 PR middle-end/56461
22880 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
22881 if VALGRIND_GET_VBITS is defined, temporarily make object
22882 memory all defined, and restore previous valgrind addressability
22883 and definability afterwards. Free this_object at the end.
22884
4ccf8f43
JJ
22885 PR middle-end/56461
22886 * lra.c (lra): Call lra_clear_live_ranges if live_p,
22887 right before calling lra_create_live_ranges, also call it
22888 when clearing live_p. Only call lra_clear_live_ranges
22889 at the end if live_p.
22890
334e71e8
JJ
22891 PR middle-end/56461
22892 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
22893
9ca966ca
RB
228942013-03-05 Richard Biener <rguenther@suse.de>
22895
22896 PR tree-optimization/56521
22897 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
22898 value-id.
22899
d273b176
SB
229002013-03-05 Steven Bosscher <steven@gcc.gnu.org>
22901
22902 PR c++/55135
22903 * except.h (remove_unreachable_eh_regions): New prototype.
22904 * except.c (remove_eh_handler_splicer): New function, split out
22905 of remove_eh_handler.
22906 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
22907 warning about running it on many EH regions one at a time.
22908 (remove_unreachable_eh_regions_worker): New function, walk the
22909 EH tree in depth-first order and remove non-marked regions.
22910 (remove_unreachable_eh_regions): New function.
22911 * tree-eh.c (mark_reachable_handlers): New function, split out
22912 from remove_unreachable_handlers.
22913 (remove_unreachable_handlers): Use mark_reachable_handlers and
22914 remove_unreachable_eh_regions.
22915 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
22916 and remove_unreachable_eh_regions.
22917
5e7f1aef
RB
229182013-03-05 Richard Biener <rguenther@suse.de>
22919
22920 PR middle-end/56525
22921 * loop-init.c (fix_loop_structure): Remove loops in two stages,
22922 not freeing them until the end.
22923
f276b762
AK
229242013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22925
22926 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
22927
12a43ab8
RB
229282013-03-05 Richard Biener <rguenther@suse.de>
22929
22930 PR tree-optimization/56270
22931 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
22932 of loads after scheduling an SLP instance.
22933
93675444
JJ
229342013-03-05 Jakub Jelinek <jakub@redhat.com>
22935
db4138e3
JJ
22936 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
22937 tic6x.exp.
22938 (check_gcc_parallelize): Run guality.exp as a separate job from
22939 vect.exp with unsorted.exp and $(dg_target_exps) separately from
22940 struct-layout-1.exp with stackalign.exp.
22941
dd3d1ec0
JJ
22942 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
22943
b8d381a3
JJ
22944 PR middle-end/56461
22945 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
22946 load_index sbitmap even if some bit in it isn't set.
22947
b4f9786b
JJ
22948 PR middle-end/56461
22949 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
22950 (discover_iteration_bound_by_body_walk): Change queues to
22951 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
22952 spelling in comment. Call safe_push on queues[bound_index] directly.
22953 Release queues[queue_index] in every iteration unconditionally.
22954 Release bounds vector.
22955
93675444
JJ
22956 PR middle-end/56461
22957 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
22958 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
22959 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
22960 inner_phis vector.
22961
3e492e9c
RB
229622013-03-05 Richard Biener <rguenther@suse.de>
22963
22964 PR lto/56515
22965 * tree-inline.c (remap_blocks_to_null): New function.
22966 (expand_call_inline): When expanding a call stmt without
22967 an associated block inline remap all callee blocks to NULL.
22968
a5d1569a
JJ
229692013-03-05 Jakub Jelinek <jakub@redhat.com>
22970
86efb5cd
JJ
22971 PR rtl-optimization/56494
22972 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
22973 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
22974 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
22975
85f5dbea
JJ
22976 PR middle-end/56461
22977 * sel-sched-ir.c (free_sched_pools): Release
22978 succs_info_pool.stack[succs_info_pool.max_top] vectors too
22979 if succs_info_pool.max_top isn't -1.
22980
a5d1569a
JJ
22981 PR bootstrap/56509
22982 * opts.c (opts_obstack, opts_concat): Moved to...
22983 * opts-common.c (opts_obstack, opts_concat): ... here.
22984
4432aa6c
JJ
229852013-03-04 Jakub Jelinek <jakub@redhat.com>
22986
22987 PR middle-end/56461
22988 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
22989
f3a1fb91
MJ
229902013-03-04 Martin Jambor <mjambor@suse.cz>
22991
22992 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
22993 all appropriate places.
22994
22110e6c
EB
229952013-01-04 Eric Botcazou <ebotcazou@adacore.com>
22996
22997 PR tree-optimization/56424
22998 * ipa-split.c (split_function): Do not set the RSO flag if result is
22999 not by reference and its type is a register type.
23000
a72d8780 230012013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
91c4e421 23002
88b97037 23003 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
91c4e421 23004 (microblaze_legitimate_pic_operand): Likewise
88b97037 23005 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
91c4e421
DH
23006 new function microblaze_legitimate_pic_operand
23007 * config/microblaze/microblaze-protos.h
23008 (microblaze_legitimate_pic_operand): Declare.
23009
a72d8780 230102013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
82f8f53a 23011
a72d8780 23012 * config/microblaze/predicates.md (call_insn_simple_operand):
82f8f53a 23013 New predicate for supported rtx code types.
a72d8780 23014 * config/microblaze/microblaze.md (call_internal1): Use
82f8f53a
EI
23015 call_insn_simple_operand predicate.
23016
541d9ac8
JJ
230172013-03-04 Jakub Jelinek <jakub@redhat.com>
23018
5eb010bc
JJ
23019 PR middle-end/56461
23020 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
23021 partitions.ordered_remove.
23022
30862efc
JJ
23023 PR middle-end/56461
23024 * tree-vect-stmts.c (vectorizable_conversion): Don't call
23025 vec_oprnds0.create (1) for modifier == NONE.
23026
8930f723
JJ
23027 PR middle-end/56461
23028 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
23029 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
23030 vec_oprnds1 right before pushing anything to it for
23031 scalar_shift_arg.
23032
541d9ac8
JJ
23033 PR middle-end/56461
23034 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23035 set nbbs to 0 instead of having separate code path.
23036 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23037 instead of false as last argument if returning NULL.
23038
7aa7f2e3
SL
230392013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23040
23041 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23042 the attribute is now called "target" instead of "option".
23043 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23044 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23045 attribute/pragma name for TARGET_OPTION_VALID_P and
23046 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23047 * doc/tm.texi: Regenerated.
23048
8930f723 230492013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
cb8a1637 23050
8930f723 23051 * config/microblaze/microblaze.c:
cb8a1637 23052 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
8930f723
JJ
23053 * config/microblaze/microblaze.h: Add -mxl-reorder to
23054 DRIVER_SELF_SPECS.
23055 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23056 instructions emitted if TARGET_REORDER.
88b97037
UB
23057 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23058 or 0 for -m/-mno case, but initialises as 2 to detect default use case
8930f723 23059 separately.
cb8a1637 23060
4dc1d68c
XDL
230612013-03-01 Xinliang David Li <davidxl@google.com>
23062
23063 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23064 walk length.
23065
689eaba3
JJ
230662013-03-01 Jakub Jelinek <jakub@redhat.com>
23067
9d676bf6
JJ
23068 PR middle-end/56461
23069 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23070 vector even when returning true. Fix up function comment formatting.
23071
caff7edf
JJ
23072 PR middle-end/56461
23073 * ira-build.c (ira_loop_nodes_count): New variable.
23074 (create_loop_tree_nodes): Initialize it.
23075 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23076
b6b9227d
JJ
23077 PR middle-end/56461
23078 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23079 method on dr_chain and result_chain.
23080 * tree-vect-stmts.c (vectorizable_store): Only call
23081 result_chain.create if j == 0.
23082
689eaba3
JJ
23083 PR middle-end/56461
23084 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23085 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23086 before overwriting it.
23087
40bf31ed
TB
230882013-03-01 Tobias Burnus <burnus@net-b.de>
23089
23090 * doc/extended.texi (C Extensions): Change order in @menu
23091 to match @node.
23092 (Other MIPS Built-in Functions): Move last MIPS entry before
23093 "picoChip Built-in Functions".
23094 (SH Built-in Functions): Move after RX Built-in Functions.
23095 * doc/gcc.texi (Introduction): Change order in @menu
23096 to match @node.
23097 * doc/md.texi (Constraints): Ditto.
23098 * gty.texi (Type Information): Ditto.
23099 (User-provided marking routines for template types): Make
23100 subsection.
23101 * doc/invoke.texi (AArch64 Options): Move before
23102 "Adapteva Epiphany Options".
23103
e664c61c
KS
231042013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23105 Jakub Jelinek <jakub@redhat.com>
23106
23107 PR sanitizer/56454
23108 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23109 no_address_safety_analysis attribute.
23110 * doc/extend.texi (no_address_safety_attribute): Rename to
23111 no_sanitize_address attribute, mention no_address_safety_analysis
23112 attribute as deprecated alias.
23113
37b5ec8f
JJ
231142013-02-28 Jakub Jelinek <jakub@redhat.com>
23115
23116 PR middle-end/56461
23117 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23118 type to vec<vec<tree> > *.
23119 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23120 to be vec<tree> instead of vec<tree> *, set vec_defs
23121 to vNULL and call vec_defs.create (number_of_vects), adjust other
23122 uses of vec_defs.
23123 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23124 vectorizable_condition): Adjust vect_get_slp_defs callers.
23125
ba96cdfb
JG
231262013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23127
23128 * config/aarch64/aarch64.c
23129 (aarch64_float_const_representable): Remove unused variable.
23130
6f549691
JG
231312013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23132
23133 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23134
af55e82d
JG
231352013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23136
23137 * config/aarch64/aarch64-builtins.c
23138 (aarch64_init_simd_builtins): Make static.
23139
1df3f464
JG
231402013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23141
23142 * config/aarch64/aarch64.c
23143 (aarch64_simd_make_constant): Make static.
23144
f8f42513
MJ
231452013-02-28 Martin Jambor <mjambor@suse.cz>
23146
23147 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23148 with no initialization to the RHS of debug statements.
23149
b48b3fc4
MJ
231502013-02-28 Martin Jambor <mjambor@suse.cz>
23151
23152 PR tree-optimization/56294
23153 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23154 Adjust dumping.
23155 (get_access_replacement): Do not call create_access_replacement.
23156 Assert a replacement exists.
23157 (get_repl_default_def_ssa_name): Create the replacement declaration
23158 itself.
23159
c3ae224c
RR
231602013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23161
23162 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23163 final_end_function.
23164
45fa0eef
MP
231652013-02-28 Marek Polacek <polacek@redhat.com>
23166
23167 PR rtl-optimization/56466
23168 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23169 if we're changing a loop.
23170 (peel_loops_completely): Likewise.
23171
502c067d
PC
231722013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23173
23174 PR c++/55813
23175 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23176
ff24afc8
GJL
231772013-02-28 Georg-Johann Lay <avr@gjlay.de>
23178
23179 PR target/56445
23180 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23181 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23182 INTX_FTYPE_FX, FX_FTYPE_INTX.
23183 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23184
c7b1fc1b
GJL
231852013-02-28 Georg-Johann Lay <avr@gjlay.de>
23186
23187 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23188 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23189 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23190 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23191 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23192 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23193 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23194 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23195 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23196 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23197 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23198 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23199 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23200 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23201 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23202 (avrxmega6): Increase max flash segments from 5 to 6.
23203 * config/avr/t-multilib: Regenerate.
23204 * config/avr/avr-tables.opt: Regenerate.
23205 * doc/avr-mmcu.texi: Regenerate.
23206
4a0e3cfe
GJL
232072013-02-28 Georg-Johann Lay <avr@gjlay.de>
23208
23209 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23210 (avr_device_to_arch): Rename to avr_device_to_ld.
23211 (avr_device_to_as): New prototype.
23212 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23213 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23214 * config/avr/driver-avr.c (avr_device_to_as): New.
23215 (avr_device_to_arch): Rename to avr_device_to_ld.
23216
97785e52
JJ
232172013-02-27 Jakub Jelinek <jakub@redhat.com>
23218
3f292312
JJ
23219 PR middle-end/56461
23220 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23221 method on dr_chain and result_chain.
23222
a344216b
JJ
23223 PR middle-end/56461
23224 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23225 pointer_set_destroy on not_executed_last_iteration.
23226
f121ad02 23227 PR middle-end/56461
88b97037 23228 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
f121ad02 23229
307f83a3
JJ
23230 PR middle-end/56461
23231 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23232 FOR_EACH_DEFINED_FUNCTION when freeing state.
23233
e19624ee
JJ
23234 PR middle-end/56461
23235 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23236 pool_free.
23237 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23238 overwriting it.
23239
90e709fd
JJ
23240 PR middle-end/56461
23241 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23242 known_aggs[i].items and release known_aggs vector.
23243
97785e52
JJ
23244 PR middle-end/56461
23245 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23246
232472013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
bbb3a9e2 23248
97785e52
JJ
23249 * config/microblaze/microblaze.c (microblaze_emit_compare):
23250 Use xor for EQ/NE comparisions.
23251 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23252 (cbranchsf4): Adjust operator to comparison_operator.
bbb3a9e2 23253
6fa5e0ed
JJ
232542013-02-27 Jakub Jelinek <jakub@redhat.com>
23255
23256 PR middle-end/56461
23257 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23258 vector.
23259 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23260 vec_safe_push, always update *slot.
23261 (redirect_edge_var_map_clear): Use vec_free.
23262 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23263 (free_var_map_entry): Use vec_free.
23264 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23265 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23266
436a956a
AB
232672013-02-27 Andrey Belevantsev <abel@ispras.ru>
23268
23269 PR middle-end/45472
436a956a
AB
23270 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23271 when the may_trap_p bit of the exprs being merged differs.
436a956a
AB
23272 Reorder tests for speculativeness in the logical and operator.
23273
f5c2caca
JJ
232742013-02-27 Jakub Jelinek <jakub@redhat.com>
23275
0fcb564b
JJ
23276 * incpath.c (add_standard_paths): Use reconcat instead of concat
23277 where appropriate and avoid leaking memory.
23278
dc357798
JJ
23279 * opts.h: Include obstack.h.
23280 (opts_concat): New prototype.
23281 (opts_obstack): New declaration.
23282 * opts.c (opts_concat): New function.
23283 (opts_obstack): New variable.
23284 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23285 (finish_options): Use opts_concat instead of concat
23286 and XOBNEWVEC instead of XNEWVEC.
23287 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23288 generate_option): Likewise.
23289 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23290 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23291
f5c2caca
JJ
23292 PR target/56455
23293 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23294 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23295
d0163673
JJ
232962013-02-26 Jakub Jelinek <jakub@redhat.com>
23297
23298 PR middle-end/56461
23299 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23300
b5ad2b8e
JR
233012013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23302
23303 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23304 (arm_block_move_unaligned_straight): Likewise.
23305 (arm_adjust_block_mem): Likewise.
23306
9b639e2c
JR
233072013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23308
23309 PR target/48901
23310 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23311 temp, cond and label.
23312 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23313
e13a0ccb
JR
23314 PR target/52500
23315 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23316 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23317
fbe4f171
JR
23318 PR target/52501
23319 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23320 of prologue/epilogue functions.
23321
ae006543
JR
23322 PR target/52550
23323 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23324 Remove unused variable cfa_offset.
23325 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23326
c81369fa
JR
23327 PR target/54639
23328 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23329 type promotion to unsigned.
23330
f8a8fea7
JR
23331 PR target/54640
23332 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23333 for HOST_WIDE_INT of 32 bit / same size as int.
23334 (arm_block_move_unaligned_straight): Likewise.
23335 (arm_adjust_block_mem): Likewise.
23336
f8be5169
JR
23337 PR target/54662
23338 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23339 ALL_CFLAGS.
23340
f1ad3354
MP
233412013-02-26 Marek Polacek <polacek@redhat.com>
23342
23343 PR tree-optimization/56426
88b97037 23344 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
f1ad3354 23345
a6af0f42
RB
233462013-02-26 Richard Biener <rguenther@suse.de>
23347
23348 PR target/56444
23349 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23350 unused variable loops.
23351
aca43c6c
JJ
233522013-02-26 Jakub Jelinek <jakub@redhat.com>
23353
cecbe5d9
JJ
23354 PR tree-optimization/56448
23355 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23356 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23357 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23358 later operands of the references, or even first operand for
23359 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23360
aca43c6c
JJ
23361 PR tree-optimization/56443
23362 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23363 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23364 to type_for_mode langhook.
23365
53e2e141
MT
233662013-02-25 Matt Turner <mattst88@gmail.com>
23367
23368 * doc/invoke.texi: Document r4700.
23369
259ee451
RB
233702013-02-25 Richard Biener <rguenther@suse.de>
23371
23372 PR tree-optimization/56175
23373 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23374 split out from ...
23375 (simplify_bitwise_binary): ... here. Also guard the conversion
23376 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23377
6d65e8f1
CM
233782013-02-25 Catherine Moore <clm@codesourcery.com>
23379
23380 Revert:
23381 2013-02-24 Catherine Moore <clm@codesourcery.com>
23382 Maciej W. Rozycki <macro@codesourcery.com>
23383 Tom de Vries <tom@codesourcery.com>
a72d8780 23384 Nathan Sidwell <nathan@codesourcery.com>
6d65e8f1
CM
23385 Iain Sandoe <iain@codesourcery.com>
23386 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23387 Chao-ying Fu <fu@mips.com>
88b97037 23388
6d65e8f1 23389 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23390 Document new function attributes.
6d65e8f1
CM
23391 * doc/invoke.texi (minterlink-compressed, mmicromips,
23392 m14k, m14ke, m14kec): Document new options.
23393 (minterlink-mips16): Update documentation.
23394 * doc/md.texi (ZC, ZD): Document new constraints.
23395 * configure.ac (gcc_cv_as_micromips): Check if linker
23396 supports the .set micromips directive.
23397 * configure: Regenerate.
23398 * config.in: Regenerate.
23399 * config/mips/mips-tables.opt: Regenerate.
23400 * config/mips/micromips.md: New file.
23401 * constraints.md (ZC, AD): New constraints.
23402 * config/mips/predicates.md (movep_src_register): New predicate.
23403 (movep_src_operand): New predicate.
23404 (non_volatile_mem_operand): New predicate.
23405 * config/mips/mips.md (multimem): New type.
23406 (length): Differentiate between 17-bit and 18-bit branch offsets.
23407 (MOVEP1, MOVEP2): New mode iterator.
23408 (mov_<load>l): Use ZC constraint.
23409 (mov_<load>r): Likewise.
23410 (mov_<store>l): Likewise.
23411 (mov_<store>r): Likewise.
23412 (*branch_equality<mode>_inverted): Add microMIPS support.
23413 (*branch_equality<mode>): Likewise.
23414 (*jump_absolute): Likewise.
23415 (indirect_jump_<mode>): Likewise.
23416 (tablejump_<mode>): Likewise.
23417 (<optab>_internal): Likewise.
23418 (sibcall_internal): Likewise.
23419 (sibcall_value_internal): Likewise.
23420 (prefetch): Use constraint ZD.
23421 * config/mips/mips.opt (minterlink-compressed): New option.
23422 (minterlink-mips16): Now an alias for minterlink-compressed.
23423 (mmicromips): New option.
23424 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23425 (compare_and_swap_12): Likewise.
23426 (sync_add<mode>): Likewise.
23427 (sync_<optab>_12): Likewise.
23428 (sync_old_<optab>_12): Likewise.
23429 (sync_new_<optab>_12): Likewise.
23430 (sync_nand_12): Likewise.
23431 (sync_old_nand_12): Likewise.
23432 (sync_new_nand_12): Likewise.
23433 (sync_sub<mode>): Likewise.
23434 (sync_old_add<mode>): Likewise.
23435 (sync_old_sub<mode>): Likewise.
23436 (sync_new_add<mode>): Likewise.
23437 (sync_new_sub<mode>): Likewise.
23438 (sync_<optab><mode>): Likewise.
23439 (sync_old_<optab><mode>): Likewise.
23440 (sync_new_<optab><mode>): Likewise.
23441 (sync_nand<mode>): Likewise.
23442 (sync_old_nand<mode>): Likewise.
23443 (sync_new_nand<mode>): Likewise.
23444 (sync_lock_test_and_set<mode>): Likewise.
23445 (test_and_set_12): Likewise.
23446 (atomic_compare_and_swap<mode>): Likewise.
23447 (atomic_exchange<mode>_llsc): Likewise.
23448 (atomic_fetch_add<mode>_llsc): Likewise.
23449 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23450 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23451 (umips_save_restore_pattern_p): Likewise.
23452 (umips_load_store_pair_p): Likewise.
23453 (umips_output_load_store_pair): Likewise.
23454 (umips_movep_target_p): Likewise.
23455 (umips_12bit_offset_address_p): Likewise.
23456 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23457 (mips_base_mips16): Rename this...
23458 (mips_base_compression_flags): ...to this. Update all uses.
23459 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23460 (mips_mips16_decl_p): Delete.
23461 (mips_nomips16_decl_p): Delete.
23462 (mips_get_compress_on_flags): New function.
23463 (mips_get_compress_off_flags): New function.
23464 (mips_get_compress_mode): New function.
23465 (mips_get_compress_on_name): New function.
23466 (mips_get_compress_off_name): New function.
23467 (mips_insert_attributes): Support multiple compression types.
23468 (mips_merge_decl_attributes): Likewise.
23469 (umips_12bit_offset_address_p): New function.
23470 (mips_start_function_definition): Emit .set micromips directive.
23471 (mips_call_may_need_jalx_p): New function.
23472 (mips_function_ok_for_sibcall): Add microMIPS support.
23473 (mips_print_operand_punctuation): Support short delay slots and
23474 compact jumps.
23475 (umips_swm_mask, umips_swm_encoding): New.
23476 (umips_build_save_restore): New function.
23477 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23478 (was_mips16_p): Remove.
23479 (old_compression_mode): New.
23480 (mips_set_compression_mode): New function.
23481 (mips_set_current_function): Add microMIPS support.
23482 (mips_option_override): Likewise.
23483 (umips_save_restore_pattern_p): New function.
23484 (umips_output_save_restore): New function.
23485 (umips_load_store_pair_p_1): New function.
23486 (umips_load_store_pair_p): New function.
23487 (umips_output_load_store_pair_1): New function.
23488 (umips_output_load_store_pair): New function.
23489 (umips_movep_target_p) New function.
23490 (mips_prepare_pch_save): Add microMIPS support.
23491 * config/mips/mips.h (TARGET_COMPRESSION): New.
23492 (TARGET_CPU_CPP_BUILTINS): Update macro
23493 to use new compression flags and to support microMIPS.
23494 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23495 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23496 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23497 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23498 (ASM_SPEC): Support mmicromips and mno-micromips.
23499 (M16STORE_REG_P): New macro.
23500 (MIPS_CALL): Support TARGET_MICROMIPS.
23501 (MICROMIPS_J): New macro.
23502 (mips_base_mips16): Rename this...
23503 (mips_base_compression_flags): ...to this.
23504 (UMIPS_12BIT_OFFSET_P): New macro.
23505 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23506 (MULTILIB_DIRNAMES): Likewise.
23507
5e5df392
TV
235082013-02-25 Tom de Vries <tom@codesourcery.com>
23509
23510 PR rtl-optimization/56131
23511 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23512 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23513 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23514
3fdb53c1
TB
235152013-02-25 Tobias Burnus <burnus@net-b.de>
23516
23517 * doc/invoke.texi (-fsanitize=): Move from optimization
23518 to debugging options.
23519
ed358aea
AB
235202013-02-25 Andrey Belevantsev <abel@ispras.ru>
23521
23522 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23523
f45e9053
AB
235242013-02-25 Andrey Belevantsev <abel@ispras.ru>
23525 Alexander Monakov <amonakov@ispras.ru>
23526
23527 PR middle-end/56077
23528 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23529 flush pending lists also on non-jumps. Adjust comment.
23530
6941b508
CM
235312013-02-24 Catherine Moore <clm@codesourcery.com>
23532 Maciej W. Rozycki <macro@codesourcery.com>
23533 Tom de Vries <tom@codesourcery.com>
a72d8780 23534 Nathan Sidwell <nathan@codesourcery.com>
6941b508
CM
23535 Iain Sandoe <iain@codesourcery.com>
23536 Nathan Froyd <froydnj@codesourcery.com>
a72d8780 23537 Chao-ying Fu <fu@mips.com>
6941b508
CM
23538
23539 * doc/extend.texi: (micromips, nomicromips, nocompression):
88b97037 23540 Document new function attributes.
6941b508
CM
23541 * doc/invoke.texi (minterlink-compressed, mmicromips,
23542 m14k, m14ke, m14kec): Document new options.
23543 (minterlink-mips16): Update documentation.
23544 * doc/md.texi (ZC, ZD): Document new constraints.
23545 * configure.ac (gcc_cv_as_micromips): Check if linker
23546 supports the .set micromips directive.
23547 * configure: Regenerate.
23548 * config.in: Regenerate.
23549 * config/mips/mips-tables.opt: Regenerate.
23550 * config/mips/micromips.md: New file.
23551 * constraints.md (ZC, AD): New constraints.
23552 * config/mips/predicates.md (movep_src_register): New predicate.
23553 (movep_src_operand): New predicate.
23554 (non_volatile_mem_operand): New predicate.
23555 * config/mips/mips.md (multimem): New type.
23556 (length): Differentiate between 17-bit and 18-bit branch offsets.
23557 (MOVEP1, MOVEP2): New mode iterator.
3fdb53c1 23558 (mov_<load>l): Use ZC constraint.
6941b508
CM
23559 (mov_<load>r): Likewise.
23560 (mov_<store>l): Likewise.
23561 (mov_<store>r): Likewise.
23562 (*branch_equality<mode>_inverted): Add microMIPS support.
23563 (*branch_equality<mode>): Likewise.
23564 (*jump_absolute): Likewise.
23565 (indirect_jump_<mode>): Likewise.
23566 (tablejump_<mode>): Likewise.
23567 (<optab>_internal): Likewise.
23568 (sibcall_internal): Likewise.
23569 (sibcall_value_internal): Likewise.
23570 (prefetch): Use constraint ZD.
23571 * config/mips/mips.opt (minterlink-compressed): New option.
23572 (minterlink-mips16): Now an alias for minterlink-compressed.
23573 (mmicromips): New option.
23574 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23575 (compare_and_swap_12): Likewise.
23576 (sync_add<mode>): Likewise.
23577 (sync_<optab>_12): Likewise.
23578 (sync_old_<optab>_12): Likewise.
23579 (sync_new_<optab>_12): Likewise.
23580 (sync_nand_12): Likewise.
23581 (sync_old_nand_12): Likewise.
23582 (sync_new_nand_12): Likewise.
23583 (sync_sub<mode>): Likewise.
23584 (sync_old_add<mode>): Likewise.
23585 (sync_old_sub<mode>): Likewise.
23586 (sync_new_add<mode>): Likewise.
23587 (sync_new_sub<mode>): Likewise.
23588 (sync_<optab><mode>): Likewise.
23589 (sync_old_<optab><mode>): Likewise.
23590 (sync_new_<optab><mode>): Likewise.
23591 (sync_nand<mode>): Likewise.
23592 (sync_old_nand<mode>): Likewise.
23593 (sync_new_nand<mode>): Likewise.
23594 (sync_lock_test_and_set<mode>): Likewise.
23595 (test_and_set_12): Likewise.
23596 (atomic_compare_and_swap<mode>): Likewise.
23597 (atomic_exchange<mode>_llsc): Likewise.
23598 (atomic_fetch_add<mode>_llsc): Likewise.
23599 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23600 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23601 (umips_save_restore_pattern_p): Likewise.
23602 (umips_load_store_pair_p): Likewise.
23603 (umips_output_load_store_pair): Likewise.
23604 (umips_movep_target_p): Likewise.
23605 (umips_12bit_offset_address_p): Likewise.
23606 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23607 (mips_base_mips16): Rename this...
23608 (mips_base_compression_flags): ...to this. Update all uses.
23609 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23610 (mips_mips16_decl_p): Delete.
23611 (mips_nomips16_decl_p): Delete.
3fdb53c1 23612 (mips_get_compress_on_flags): New function.
6941b508
CM
23613 (mips_get_compress_off_flags): New function.
23614 (mips_get_compress_mode): New function.
23615 (mips_get_compress_on_name): New function.
23616 (mips_get_compress_off_name): New function.
23617 (mips_insert_attributes): Support multiple compression types.
23618 (mips_merge_decl_attributes): Likewise.
23619 (umips_12bit_offset_address_p): New function.
23620 (mips_start_function_definition): Emit .set micromips directive.
23621 (mips_call_may_need_jalx_p): New function.
23622 (mips_function_ok_for_sibcall): Add microMIPS support.
23623 (mips_print_operand_punctuation): Support short delay slots and
23624 compact jumps.
23625 (umips_swm_mask, umips_swm_encoding): New.
23626 (umips_build_save_restore): New function.
23627 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23628 (was_mips16_p): Remove.
23629 (old_compression_mode): New.
23630 (mips_set_compression_mode): New function.
23631 (mips_set_current_function): Add microMIPS support.
23632 (mips_option_override): Likewise.
23633 (umips_save_restore_pattern_p): New function.
23634 (umips_output_save_restore): New function.
23635 (umips_load_store_pair_p_1): New function.
23636 (umips_load_store_pair_p): New function.
23637 (umips_output_load_store_pair_1): New function.
23638 (umips_output_load_store_pair): New function.
23639 (umips_movep_target_p) New function.
23640 (mips_prepare_pch_save): Add microMIPS support.
23641 * config/mips/mips.h (TARGET_COMPRESSION): New.
23642 (TARGET_CPU_CPP_BUILTINS): Update macro
23643 to use new compression flags and to support microMIPS.
23644 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23645 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23646 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23647 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23648 (ASM_SPEC): Support mmicromips and mno-micromips.
23649 (M16STORE_REG_P): New macro.
23650 (MIPS_CALL): Support TARGET_MICROMIPS.
23651 (MICROMIPS_J): New macro.
23652 (mips_base_mips16): Rename this...
23653 (mips_base_compression_flags): ...to this.
23654 (UMIPS_12BIT_OFFSET_P): New macro.
23655 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23656 (MULTILIB_DIRNAMES): Likewise.
23657
cdebbc6d
JJ
236582013-02-24 Jakub Jelinek <jakub@redhat.com>
23659
23660 PR target/52555
23661 * target-globals.c (save_target_globals): For init_reg_sets and
23662 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23663
18c63565
JG
236642013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23665
23666 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23667 * config/aarch64/t-aarch64
23668 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23669
2194f7a2
VM
236702013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23671
23672 PR inline-asm/56148
23673 * lra-constraints.c (process_alt_operands): Reload operand
23674 conflicting with earlier clobber only if no more other conflicting
23675 operands.
23676
7d613735
JJ
236772013-02-22 Jakub Jelinek <jakub@redhat.com>
23678
23679 PR sanitizer/56393
23680 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23681 if not linking a shared library.
23682
ac8d93a7
SL
236832013-02-22 Seth LaForge <sethml@google.com>
23684
23685 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23686
e0237780
GY
236872013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23688
3fdb53c1
TB
23689 * config/arm/arm.md (split for extendsidi): Update condition.
23690 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23691 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23692 (qhs_zextenddi_cstr): Likewise.
e0237780 23693
d7fde18c
JJ
236942013-02-21 Jakub Jelinek <jakub@redhat.com>
23695
be63b77d
JJ
23696 PR middle-end/56420
23697 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23698 avoid signed wrapping.
23699 (expand_mult): Handle properly multiplication by
23700 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23701 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23702 in the compiler if coeff is HOST_WIDE_INT_MIN.
23703 (expand_divmod): Don't make ext_op1 static, change it's type to
23704 uhwi. Avoid undefined behavior in -INTVAL (op1).
23705
d7fde18c
JJ
23706 PR rtl-optimization/50339
23707 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23708 field.
23709 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23710 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23711 into splitting_ashiftrt field.
23712 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23713 ASHIFTRT.
23714 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23715 choices.
23716
6aad4455
AH
237172013-02-20 Aldy Hernandez <aldyh@redhat.com>
23718
23719 PR middle-end/56108
23720 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23721 are sure to go irrevocable.
23722
38fe784d
HPN
237232013-02-21 Hans-Peter Nilsson <hp@axis.com>
23724
23725 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23726 scalars are valid operands.
23727
0fd44da3
MJ
237282013-02-21 Martin Jambor <mjambor@suse.cz>
23729
23730 PR tree-optimization/56310
23731 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23732 only matching indices and non-negative final offsets.
23733 (intersect_aggregates_with_edge): Pass src_idx to
23734 agg_replacements_to_vector. Pass src_idx insstead of index to
23735 intersect_with_agg_replacements.
23736
7a92038b
MJ
237372013-02-21 Martin Jambor <mjambor@suse.cz>
23738
23739 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23740 instead of hard-wired defaults.
23741
c0da9c37
MR
237422013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23743
23744 * doc/invoke.texi (MIPS Options): Update documentation of the
23745 floating-point multiply-accumulate instruction restrictions.
23746
d247ea0c 237472013-02-21 Kostya Serebryany <kcc@google.com>
7df59255
KS
23748
23749 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
23750 asan_shadow_offset on x86_64 linux.
23751
22deefcb
RB
237522013-02-21 Richard Biener <rguenther@suse.de>
23753
23754 PR tree-optimization/56415
23755 Revert
23756 2013-02-11 Richard Biener <rguenther@suse.de>
23757
23758 PR tree-optimization/56273
23759 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23760 first VRP run.
23761
7bcc6e75
JJ
237622013-02-21 Jakub Jelinek <jakub@redhat.com>
23763
a0ad148f
JJ
23764 PR bootstrap/56258
23765 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
23766 instead of @itemx.
23767
7bcc6e75
JJ
23768 PR inline-asm/56405
23769 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
23770 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
23771
a0a7b611
JH
237722013-02-20 Jan Hubicka <jh@suse.cz>
23773
23774 PR tree-optimization/56265
88b97037
UB
23775 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
23776 when target is referenced for first time.
a0a7b611 23777
c0e50f72
RB
237782013-02-20 Richard Biener <rguenther@suse.de>
23779
23780 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
23781 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
23782 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
23783 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
23784 not return anything.
23785 (rename_ssa_copies): Do not remove unused locals.
23786 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
88b97037 23787 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
c0e50f72
RB
23788 * passes.c (execute_function_todo): Do not schedule unused locals
23789 removal if cleanup_tree_cfg did something.
23790 * tree-ssa-live.c (remove_unused_locals): Dump statistics
23791 about the number of removed locals.
23792
a52ca739
RB
237932013-02-20 Richard Biener <rguenther@suse.de>
23794
23795 PR tree-optimization/56398
88b97037 23796 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
a52ca739 23797
ec9c9d1b
MJ
237982013-02-20 Martin Jambor <mjambor@suse.cz>
23799
23800 PR tree-optimization/55334
23801 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
23802 restricted pointers to arrays.
23803
e91c8ed6 238042013-02-20 Richard Biener <rguenther@suse.de>
88b97037 23805 Jakub Jelinek <jakub@redhat.com>
e91c8ed6
RB
23806
23807 PR tree-optimization/56396
23808 * tree-ssa-ccp.c (n_const_val): New static variable.
23809 (get_value): Return NULL for SSA names we don't have a lattice
23810 entry for.
23811 (ccp_initialize): Initialize n_const_val.
23812 * tree-ssa-copy.c (n_copy_of): New static variable.
23813 (init_copy_prop): Initialize n_copy_of.
23814 (get_value): Return NULL_TREE for SSA names we don't have a
23815 lattice entry for.
23816
3d916479
MJ
238172013-02-20 Martin Jambor <mjambor@suse.cz>
23818
23819 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
23820
71a86758
RB
238212013-02-20 Richard Biener <rguenther@suse.de>
23822
23823 * genpreds.c (write_lookup_constraint): Do not compare first
23824 letter of the constraint again.
23825
79836a12
RB
238262013-02-20 Richard Biener <rguenther@suse.de>
23827
23828 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
23829 and ceil_log2.
23830 (get_use_iv_cost): Terminate hashtable walk when coming across
23831 an empty entry.
23832
bbe4fb2c
IZ
238332013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
23834
23835 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
23836 reassociation for avx2 targets.
23837
a72d8780 238382012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
79836a12 23839
c77f83d5
EI
23840 * config/microblaze/microblaze.c: microblaze_has_clz = 0
23841 Add version check for v8.10.a to enable microblaze_has_clz
88b97037 23842 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
c77f83d5
EI
23843 version and TARGET_PATTERN_COMPARE check
23844 * config/microblaze/microblaze.md: New clzsi2 instruction
3fdb53c1 23845
a72d8780 238462012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
959434e6 23847
a72d8780 23848 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
959434e6
EI
23849 function before branching.
23850
242387fa
AB
238512012-02-19 Andrey Belevantsev <abel@ispras.ru>
23852
23853 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
23854 DUMP_INSN_RTX_UID.
23855 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
23856
b4979ab9
AB
238572012-02-19 Andrey Belevantsev <abel@ispras.ru>
23858
23859 PR middle-end/55889
b4979ab9
AB
23860 * sel-sched.c: Include ira.h.
23861 (implicit_clobber_conflict_p): New function.
23862 (moveup_expr): Use it.
88b97037 23863 * Makefile.in (sel-sched.o): Depend on ira.h.
b4979ab9 23864
24d63016
RB
238652013-02-19 Richard Biener <rguenther@suse.de>
23866
23867 PR tree-optimization/56384
23868 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
23869 (vn_hash_type): Split out from ...
23870 (vn_hash_constant_with_type): ... here.
23871 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
23872 (vn_phi_eq): Compare types from vn_phi_s structure.
23873 (vn_phi_lookup): Populate vn_phi_s type.
23874 (vn_phi_insert): Likewise.
23875
a475fd3d
JJ
238762013-02-19 Jakub Jelinek <jakub@redhat.com>
23877
47cc28f5
JJ
23878 PR tree-optimization/56350
23879 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
23880 if haven't found reduction or nested cycle operand, rather than
23881 asserting we must find it.
23882
a475fd3d
JJ
23883 PR tree-optimization/56381
23884 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
23885 to fold_build3.
23886
238872013-02-18 Aldy Hernandez <aldyh@redhat.com>
135204dd
AH
23888 Jakub Jelinek <jakub@redhat.com>
23889
23890 PR target/52555
23891 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
23892 (swap_optab_enable): Same.
23893 (init_all_optabs): Use argument instead of global.
88b97037 23894 * tree.h (struct tree_optimization_option): New field target_optabs.
135204dd
AH
23895 * expr.h (init_all_optabs): Add argument to prototype.
23896 (TREE_OPTIMIZATION_OPTABS): New.
23897 (save_optabs_if_changed): Protoize.
23898 * optabs.h: Declare this_fn_optabs.
23899 * optabs.c (save_optabs_if_changed): New.
23900 Declare this_fn_optabs.
23901 (init_optabs): Add argument to init_all_optabs() call.
23902 * function.c (invoke_set_current_function_hook): Handle per
23903 function optabs.
23904 * function.h (struct function): New field optabs.
23905 * config/mips/mips.c (mips_set_mips16_mode): Handle when
23906 optimization_current_node has changed.
23907 * target-globals.h (save_target_globals_default_opts): Protoize.
23908 * target-globals.c (save_target_globals_default_opts): New.
23909
3f587ca3
JDA
239102013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23911
23912 PR target/56347
23913 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
23914 registers %fr12 and %fr12R as call used.
ceaca33e
JDA
23915
23916 PR target/56214
23917 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
88b97037
UB
23918 and HImode, require all displacements to be an integer multiple of
23919 their mode size.
ceaca33e
JDA
23920 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
23921 only allow QImode and HImode when reload is in progress and strict is
23922 true. Likewise for symbolic addresses. Use base14_operand to check
23923 displacements in REG+BASE addresses.
23924
fe0b4796
RB
239252013-02-18 Richard Biener <rguenther@suse.de>
23926
23927 PR tree-optimization/56366
23928 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23929 handle sign-conversion of outer-loop initial induction value.
23930
6aaf596b
RB
239312013-02-18 Richard Biener <rguenther@suse.de>
23932
73db8ff1 23933 PR middle-end/56349
6aaf596b
RB
23934 * cfghooks.c (merge_blocks): If we merge a latch into another
23935 block adjust references to it.
23936 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
23937 (verify_loop_structure): Verify that a recorded latch is in fact
23938 a latch.
23939
5e97dfb6
RB
239402013-02-18 Richard Biener <rguenther@suse.de>
23941
23942 PR tree-optimization/56321
23943 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
23944 order SSA name release and virtual operand unlinking.
23945
825527e8
EI
239462013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23947
23948 * config/microblaze/microblaze.md (save_stack_block): Define.
23949 (restore_stack_block): Likewise.
23950
debd11d9
EI
239512013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23952
23953 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
23954 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
23955 * config/microblaze/microblaze.c (microblaze_option_override):
23956 Bail out early for PIC modes when target does not support PIC.
23957
8ec77be0
EI
239582013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23959
88b97037 23960 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
8ec77be0
EI
23961 Replace with a microblaze version.
23962 (microblaze_trampoline_init): Adapt for microblaze.
a752b80f 23963 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
8ec77be0
EI
23964 microblaze.
23965
b41288b3
JJ
239662013-02-16 Jakub Jelinek <jakub@redhat.com>
23967 Dodji Seketeli <dodji@redhat.com>
23968
23969 PR asan/56330
88b97037 23970 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
b41288b3
JJ
23971 (instrument_mem_region_access): Do not forget to always put
23972 instrumentation of the of 'base' and 'base + len' in a "if (len !=
23973 0) statement, even for cases where either 'base' or 'base + len'
23974 are not instrumented -- because they have been previously
23975 instrumented. Simplify the logic by putting all the statements
23976 instrument 'base + len' inside a sequence, and then insert that
23977 sequence right before the current insertion point. Then, to
23978 instrument 'base + len', just get an iterator on that statement.
23979 And do not forget to update the pointer to iterator the function
23980 received as argument.
23981
47918951
VM
239822013-02-15 Vladimir Makarov <vmakarov@redhat.com>
23983
23984 PR rtl-optimization/56348
23985 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
23986
3fb7c699
SB
239872013-02-15 Steven Bosscher <steven@gcc.gnu.org>
23988
23989 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
23990 (clean_graph_dump_file): Pass base to start_graph_dump.
23991
576fe41a
RH
239922013-02-14 Richard Henderson <rth@redhat.com>
23993
23994 PR target/55941
23995 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
23996
bb3accfa
SB
239972013-02-14 Steven Bosscher <steven@gcc.gnu.org>
23998
23999 * collect2-aix.h: Define F_LOADONLY.
24000
03143140
RB
240012013-02-14 Richard Biener <rguenther@suse.de>
24002
24003 PR lto/50494
24004 * varasm.c (output_constant_def_1): Get the decl representing
24005 the constant as argument.
24006 (output_constant_def): Wrap output_constant_def_1.
24007 (make_decl_rtl): Use output_constant_def_1 with the decl
24008 representing the constant.
24009 (build_constant_desc): Optionally re-use a decl already
24010 representing the constant.
24011 (tree_output_constant_def): Adjust.
24012
8fb06726
DS
240132013-02-14 Dodji Seketeli <dodji@redhat.com>
24014
24015 Fix an asan crash
24016 * asan.c (instrument_builtin_call): Really put the length of the
24017 second source argument into src1_len.
24018
c4bfe8bf
JJ
240192013-02-13 Jakub Jelinek <jakub@redhat.com>
24020
24021 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
24022 argument. If it is false, don't create edge from then_bb to
24023 fallthru_bb.
24024 (insert_if_then_before_iter): Pass true to it.
24025 (build_check_stmt): Pass false to it.
24026 (transform_statements): Flush hash table only on extended basic
24027 block boundaries, rather than at the beginning of every bb.
24028 Don't flush hash table on nonfreeing_call_p calls.
24029 * tree-flow.h (nonfreeing_call_p): New prototype.
24030 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
24031
7afe2801
DM
240322013-02-13 David S. Miller <davem@davemloft.net>
24033
24034 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24035
70cc3288
VM
240362013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24037
24038 PR target/56184
24039 * ira.c (max_regno_before_ira): Move from ...
24040 (ira): ... here.
24041 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24042 vec_safe_length.
24043
6422242b
JJ
240442013-02-13 Jakub Jelinek <jakub@redhat.com>
24045
24046 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24047
976a81ee
RB
240482013-02-13 Richard Biener <rguenther@suse.de>
24049
24050 PR lto/56295
24051 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24052 globals in MEM_REFs.
24053
c1874a87
RB
240542013-02-13 Richard Biener <rguenther@suse.de>
24055
24056 * loop-init.c (loop_optimizer_init): Clear loop state when
24057 re-initializing preserved loops.
24058 * loop-unswitch.c (unswitch_single_loop): Return whether
24059 we unswitched the loop. Do not verify loop state here.
88b97037 24060 (unswitch_loops): When we unswitched a loop discover new loops.
c1874a87 24061
b4ab7d34
KS
240622013-02-13 Kostya Serebryany <kcc@google.com>
24063
88b97037
UB
24064 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24065 on x86_64 linux.
b4ab7d34
KS
24066 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24067
bdcbe80c
DS
240682013-02-12 Dodji Seketeli <dodji@redhat.com>
24069
24070 Avoid instrumenting duplicated memory access in the same basic block
24071 * Makefile.in (asan.o): Add new dependency on hash-table.h
24072 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24073 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24074 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24075 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24076 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24077 (get_mem_ref_of_assignment): New functions.
24078 (get_mem_refs_of_builtin_call): Extract from
24079 instrument_builtin_call and tweak a little bit to make it fit with
24080 the new signature.
24081 (instrument_builtin_call): Use the new
24082 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24083 of is_gimple_builtin_call.
24084 (instrument_derefs, instrument_mem_region_access): Insert the
24085 instrumented memory reference into the hash table.
24086 (maybe_instrument_assignment): Renamed instrument_assignment into
24087 this, and change it to advance the iterator when instrumentation
24088 actually happened and return true in that case. This makes it
24089 homogeneous with maybe_instrument_assignment, and thus give a
24090 chance to callers to be more 'regular'.
24091 (transform_statements): Clear the memory reference hash table
24092 whenever we enter a new BB, when we cross a function call, or when
24093 we are done transforming statements. Use
24094 maybe_instrument_assignment instead of instrumentation. No more
24095 need to special case maybe_instrument_assignment and advance the
24096 iterator after calling it; it's now handled just like
24097 maybe_instrument_call. Update comment.
24098
4861a1f7
RB
240992013-02-13 Richard Biener <rguenther@suse.de>
24100
24101 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24102 Fix loop discovery code.
24103
1c86bd80
VM
241042013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24105
24106 PR inline-asm/56148
24107 * lra-constraints.c (process_alt_operands): Match early clobber
70cc3288 24108 operand with itself. Check conflicts with earlyclobber only if
1c86bd80
VM
24109 the operand is not reloaded. Prefer to reload conflicting operand
24110 if earlyclobber and matching operands are the same.
24111
7cbda518
RB
241122013-02-12 Richard Biener <rguenther@suse.de>
24113
24114 PR lto/56297
24115 * lto-streamer-out.c (write_symbol): Do not output symbols
24116 for hard register variables.
24117
e68a4ef6
GJL
241182013-02-12 Georg-Johann Lay <avr@gjlay.de>
24119
24120 PR target/54222
24121 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24122 (umulsidi3_insn, mulsidi3_insn): New insns.
24123
a72d8780 241242013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
2597da22
CL
24125
24126 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24127 (struct tune_params): Add vec_costs field.
24128 * config/arm/arm.c (arm_builtin_vectorization_cost)
24129 (arm_add_stmt_cost): New functions.
24130 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24131 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24132 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24133 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24134 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24135 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24136 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24137
43320568
RB
241382013-02-12 Richard Biener <rguenther@suse.de>
24139
24140 PR lto/56295
24141 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24142 decls again if possible.
24143
a011aa39
RB
241442013-02-12 Richard Biener <rguenther@suse.de>
24145
24146 PR middle-end/56288
24147 * tree-ssa.c (verify_ssa_name): Fix check, move
24148 SSA_NAME_IN_FREE_LIST check up.
24149
6da26889
JJ
241502013-02-12 Jakub Jelinek <jakub@redhat.com>
24151 Steven Bosscher <steven@gcc.gnu.org>
24152
24153 PR rtl-optimization/56151
24154 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24155 equal to op0 or op1, and last_insn pattern is CODE operation
24156 with MEM dest and one of the operands matches that MEM.
24157
f80e0faf
ST
241582013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24159
24160 * doc/extend.texi: Document Function Multiversioning and "default"
24161 parameter string to target attribute.
24162 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24163 target attribute parameter is "default".
24164 (ix86_compare_version_priority): Remove checks for target attribute.
24165 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24166 Remove check for target attribute equal to NULL. Add assert.
24167 (ix86_generate_version_dispatcher_body): Change error to sorry.
24168
6c59ffd1
IS
241692013-02-11 Iain Sandoe <iain@codesourcery.com>
24170 Jack Howarth <howarth@bromo.med.uc.edu>
24171 Patrick Marlier <patrick.marlier@gmail.com>
24172
24173 PR libitm/55693
24174 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24175 define ENDFILE_SPEC as TM_DESTRUCTOR.
24176 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24177
a72d8780 241782013-02-11 Alexander Potapenko <glider@google.com>
6a52ccb3
AP
24179 Jack Howarth <howarth@bromo.med.uc.edu>
24180 Jakub Jelinek <jakub@redhat.com>
24181
24182 PR sanitizer/55617
24183 * config/darwin.c (cdtor_record): Rename ctor_record.
24184 (sort_cdtor_records): Rename sort_ctor_records.
24185 (finalize_dtors): New routine to sort destructors by
24186 priority before use in assemble_integer.
24187 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24188
b63fe007
UB
241892013-02-11 Uros Bizjak <ubizjak@gmail.com>
24190
24191 PR rtl-optimization/56275
24192 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24193 offset is non-negative and less than cmode size before
24194 calling simplify_subreg.
24195
8e89b5b5
RB
241962013-02-11 Richard Biener <rguenther@suse.de>
24197
24198 PR tree-optimization/56264
24199 * cfgloop.h (fix_loop_structure): Adjust prototype.
24200 * loop-init.c (fix_loop_structure): Return the number of
24201 newly discovered loops.
24202 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24203 are discovered, do a full loop-closed SSA rewrite.
24204
b4a4b56d
RB
242052013-02-11 Richard Biener <rguenther@suse.de>
24206
24207 PR tree-optimization/56273
24208 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24209 first VRP run.
24210 (check_array_ref): Fix missing newline in dumps.
24211 (search_for_addr_array): Likewise.
24212
0c885229
DE
242132013-02-09 David Edelsohn <dje.gcc@gmail.com>
24214
24215 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24216
59ac9a55
JJ
242172013-02-09 Jakub Jelinek <jakub@redhat.com>
24218
24219 PR target/56256
24220 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24221
25bb0bb5
VM
242222013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24223
24224 PR rtl-optimization/56246
0c885229 24225 * lra-constraints.c (simplify_operand_subreg): Try to reuse
25bb0bb5
VM
24226 reload pseudo.
24227 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24228 constraints are satisfied.
24229
a698cc03
JL
242302013-02-08 Jeff Law <law@redhat.com>
24231
24232 PR debug/53948
24233 * emit-rtl.c (reg_is_parm_p): New function.
24234 * regs.h (reg_is_parm_p): New prototype.
24235 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24236 callee-clobbered registers.
24237
e1122ddd
MM
242382013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24239
24240 PR target/56043
24241 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24242 If there is no implicit builtin declaration, just return NULL.
24243
19c5f6e6
UB
242442013-02-08 Uros Bizjak <ubizjak@gmail.com>
24245
24246 * config/i386/sse.md (FMAMODEM): New mode iterator.
24247 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24248 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24249
2480f2ca 242502013-02-08 Uros Bizjak <ubizjak@gmail.com>
76ef61fb 24251
2480f2ca
UB
24252 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24253 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24254 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24255
242562013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24257
24258 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24259 (microblaze*-*-elf): Likewise.
24260 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24261 LINK_SPEC.
24262 * config/microblaze/microblaze-c.c: Add builtin defines for
24263 _LITTLE_ENDIAN and _BIG_ENDIAN.
24264 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24265 add to TARGET_DEFAULT flags.
76ef61fb 24266 Expand ASM_SPEC and LINK_SPEC.
2480f2ca
UB
24267 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24268 * config/microblaze/microblaze.md: Update extendsidi2 and
24269 movdi_internal instructions to use low-order / high-order reg
24270 print_operands.
24271 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24272 options and inversemask / mask of LITTLE_ENDIAN.
24273 * config/microblaze/t-microblaze: Expand multilib options to
24274 include mlittle-endian (le) and update exceptions patterns.
24275
600a5961
JJ
242762013-02-08 Jakub Jelinek <jakub@redhat.com>
24277
5df81313
JJ
24278 PR rtl-optimization/56195
24279 * lra-constraints.c (get_reload_reg): Don't reuse regs
24280 if they have smaller mode than requested, if they have
24281 wider mode than requested, try to return a SUBREG.
24282
600a5961
JJ
24283 PR tree-optimization/56250
24284 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24285 if type is unsigned and code isn't MULT_EXPR.
24286
ff544649
GJL
242872013-02-08 Georg-Johann Lay <avr@gjlay.de>
24288
24289 PR tree-optimization/56064
24290 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24291 bits according to mode.
24292 * fixed-value.h (fixed_from_double_int)
24293 (const_fixed_from_double_int): Adjust comments.
24294
e45cde98
RB
242952013-02-08 Richard Biener <rguenther@suse.de>
24296
24297 PR lto/56231
24298 * lto-streamer.h (struct data_in): Remove current_file, current_line
24299 and current_col members.
24300 * lto-streamer-out.c (lto_output_location): Stream changed bits
24301 en-block for efficiency.
24302 * lto-streamer-in.c (clear_line_info): Remove.
24303 (lto_input_location): Cache current file, line and column
24304 globally via local statics. Read changed bits en-block.
24305 (input_function): Do not call clear_line_info.
24306 (lto_read_body): Likewise.
24307 (lto_input_toplevel_asms): Likewise.
24308
c1ca73d8
MM
243092013-02-08 Michael Matz <matz@suse.de>
24310
24311 PR tree-optimization/52448
24312 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24313 (nt_call_phase): New static.
24314 (add_or_mark_expr): Only mark accesses with newer phase than any
24315 call seen.
24316 (nonfreeing_call_p): New.
24317 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24318 (nt_fini_block): Keep blocks marked as visited.
24319 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24320
57548aa2
RB
243212013-02-08 Richard Biener <rguenther@suse.de>
24322
24323 * ira.c (ira): Free broken dominator information.
24324
8e10366f
UB
243252013-02-08 Uros Bizjak <ubizjak@gmail.com>
24326
24327 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24328
f64fb0fa
MP
243292013-02-08 Marek Polacek <polacek@redhat.com>
24330
8e10366f 24331 * cfgloop.c (verify_loop_structure): Add more checking of headers.
f64fb0fa 24332
0375167b
RB
243332013-02-08 Richard Biener <rguenther@suse.de>
24334
24335 PR middle-end/56181
24336 * cfgloop.h (flow_loops_find): Adjust.
24337 (bb_loop_header_p): Declare.
24338 * cfgloop.c (bb_loop_header_p): New function split out from ...
24339 (flow_loops_find): ... here. Adjust function signature,
24340 support incremental loop structure update.
24341 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24342 * cfgloopmanip.c (fix_loop_structure): Move ...
24343 * loop-init.c (fix_loop_structure): ... here.
24344 (apply_loop_flags): Split out from ...
24345 (loop_optimizer_init): ... here.
24346 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24347 in incremental mode, only remove dead loops here.
24348
85d768f3
GJL
243492013-02-08 Georg-Johann Lay <avr@gjlay.de>
24350
24351 PR target/54222
24352 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24353 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24354 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24355 (*round<mode>3.libgcc): New insns for fixed-modes.
24356 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24357 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24358 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24359 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24360 implementations. Define to __builtin_avr_absFX,
24361 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24362 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24363 __builtin_avr_countlsFX, respectively.
24364 * config/avr/avr-c.c (target.h): Include it.
24365 (enum avr_builtin_id): New enum.
24366 (avr_resolve_overloaded_builtin): New static function.
24367 (avr_register_target_pragmas): Use it to set
24368 targetm.resolve_overloaded_builtin.
24369 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24370 tree nodes used by DEF_BUILTIN.
24371 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24372 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24373 <AVR_BUILTIN_xxBITS>: Same.
24374
661bc682
RB
243752013-02-08 Richard Biener <rguenther@suse.de>
24376
24377 * cfgloop.c (verify_loop_structure): Properly handle
24378 a loop exiting to another loop header.
24379 * ira-int.h (ira_loops): Remove.
24380 * ira.c (ira_loops): Remove.
24381 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24382 (do_reload): Use loop_optimizer_finalize.
24383 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24384 number_of_loops to access the loop tree.
24385 (more_one_region_p): Likewise.
24386 (finish_loop_tree_nodes): Likewise.
24387 (rebuild_regno_allocno_maps): Likewise.
24388 (mark_loops_for_removal): Likewise.
24389 (mark_all_loops_for_removal): Likewise.
24390 (remove_unnecessary_regions): Likewise.
24391 (ira_build): Likewise.
24392 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24393
0d5049b2
RB
243942013-02-08 Richard Biener <rguenther@suse.de>
24395
24396 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24397 * ipa-pure-const.c (analyze_function): Avoid calling
24398 mark_irreducible_loops twice.
8e10366f 24399 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
0d5049b2 24400
499f32e8
DM
244012013-02-07 David S. Miller <davem@davemloft.net>
24402
24403 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24404 on 'reg'.
24405 * var-tracking.c (vt_add_function_parameter): Test the presence of
24406 HAVE_window_save properly and do not remap argument registers when
24407 we have a leaf function.
24408
6edc3e32
UB
244092013-02-07 Uros Bizjak <ubizjak@gmail.com>
24410
24411 PR bootstrap/56227
24412 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24413 instead of "ll".
24414 * config/i386/i386.c (ix86_print_operand): Ditto.
24415
5306401f
VM
244162013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24417
6edc3e32 24418 * lra-constraints.c (process_alt_operands): Fix recently added comment.
5306401f 24419
027ece11
VM
244202013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24421
24422 PR rtl-optimization/56225
24423 * lra-constraints.c (process_alt_operands): Check that reload hard
24424 reg can hold value for strict_low_part.
24425
f980dfdb
JJ
244262013-02-07 Jakub Jelinek <jakub@redhat.com>
24427
24428 PR debug/56154
24429 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24430 dwarf2out_end_function.
24431 (in_first_function_p, maybe_at_text_label_p,
24432 first_loclabel_num_not_at_text_label): New variables.
24433 (dwarf2out_var_location): In the first function find out
6edc3e32 24434 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
f980dfdb
JJ
24435 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24436 functions.
24437
73dd3123
EB
244382013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24439
24440 PR rtl-optimization/56178
24441 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24442 SUBREG of a register. Tidy up related block of code.
24443 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24444 note if the source is a register or a SUBREG of a register.
24445
e3936f47
JJ
244462013-02-07 Jakub Jelinek <jakub@redhat.com>
24447
24448 PR target/56228
24449 * config/rs6000/rs6000.md (ptrm): New mode attr.
24450 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24451 call_value_indirect_aix<pttrsize>,
24452 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24453 m in constraints.
24454
d96d674b
MH
244552013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24456
24457 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24458 if -bnortl. Convert to strcmp and strncmp.
24459
921f2dee
AM
244602013-02-07 Alan Modra <amodra@gmail.com>
24461
24462 PR target/54009
24463 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24464 addresses won't wrap when offsetting.
24465 (rs6000_secondary_reload): Provide secondary reloads needed for
24466 wrapping LO_SUM addresses.
24467
d09c7dba
TS
244682013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24469
24470 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24471 MACH, just __MACH__.
24472
a44bbd48
RB
244732013-02-06 Richard Biener <rguenther@suse.de>
24474
24475 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24476 instead of calling fix_loop_structure.
24477
6d840d99
JJ
244782013-02-06 Jakub Jelinek <jakub@redhat.com>
24479
24480 PR middle-end/56217
24481 * omp-low.c (use_pointer_for_field): Return false if
24482 lower_send_shared_vars doesn't generate any copy-out code.
24483
0f33baa9
TV
244842013-02-06 Tom de Vries <tom@codesourcery.com>
24485
24486 PR rtl-optimization/56131
24487 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24488 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24489 of the label is NULL. Add comment.
24490
11f1e3ab
JJ
244912013-02-05 Jakub Jelinek <jakub@redhat.com>
24492
31502f9f
JJ
24493 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24494
a74db9bd
JJ
24495 PR sanitizer/55374
24496 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24497 (STATIC_LIBTSAN_LIBS): Likewise.
24498 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24499 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24500 is defined, don't add anything else beyond that.
24501 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24502 (LINK_COMMAND_SPEC): Use them.
24503
11f1e3ab
JJ
24504 PR tree-optimization/56205
24505 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24506 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24507 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24508
14ac6aa2
RB
245092013-02-05 Richard Biener <rguenther@suse.de>
24510
24511 PR tree-optimization/53342
24512 PR tree-optimization/53185
24513 * tree-vectorizer.h (vect_check_strided_load): Remove.
24514 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24515 not disallow peeling for vectorized strided loads.
24516 (vect_check_strided_load): Make static and simplify.
24517 (vect_analyze_data_refs): Adjust.
24518 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24519 correctly when vectorizing strided loads.
24520
6f22445a
RB
245212013-02-05 Richard Biener <rguenther@suse.de>
24522
24523 * doc/install.texi: Refer to ISL, not PPL.
24524
39f9719e
JH
245252013-02-05 Jan Hubicka <jh@suse.cz>
24526
ec4224ac
JH
24527 PR tree-optimization/55789
24528 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24529
245302013-02-05 Jan Hubicka <jh@suse.cz>
24531
24532 PR tree-optimization/55789
39f9719e
JH
24533 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24534 the dead call anyway.
24535
956d3b33
EB
245362013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24537
24538 PR sanitizer/55374
24539 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24540
245412013-02-04 Alexander Potapenko <glider@google.com>
24542 Jack Howarth <howarth@bromo.med.uc.edu>
ae526fe5
AP
24543 Jakub Jelinek <jakub@redhat.com>
24544
24545 PR sanitizer/55617
24546 * config/darwin.c (sort_ctor_records): Stabilized qsort
24547 on constructor priority by using original position.
24548 (finalize_ctors): New routine to sort constructors by
24549 priority before use in assemble_integer.
24550 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24551
7ac3af38
JJ
245522013-02-04 Jakub Jelinek <jakub@redhat.com>
24553
24554 PR libstdc++/54314
24555 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24556 about visibility on artificial decls.
24557 * config/sol2.c (solaris_assemble_visibility): Likewise.
24558
152689dc
KT
245592013-02-04 Kai Tietz <ktietz@redhat.com>
24560
24561 PR target/56186
24562 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24563 argument and improve checking of return-argument types for 16-byte
24564 modes.
24565 (ix86_function_value_1): Add additional valtype argument on call
24566 of function_value_64.
24567 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24568 handling infunction_value_64 function.
24569
2b5987b5
MGD
245702013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24571
7ac3af38 24572 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
2b5987b5 24573
14597080
RB
245742013-02-04 Richard Biener <rguenther@suse.de>
24575
24576 PR tree-optimization/56188
24577 * tree-ssa-structalias.c (label_visit): Consider case with
24578 initially non-empty points-to set.
24579 (perform_var_substitution): Dump node mapping and clean up.
24580
ed73881e
RG
245812013-02-04 Richard Guenther <rguenther@suse.de>
24582
24583 PR lto/56168
24584 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24585 node prevail as last resort.
24586 (lto_symtab_merge_decls): Remove guard on LTRANS here.
6edc3e32 24587 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
ed73881e 24588
139a0707
RB
245892013-02-04 Richard Biener <rguenther@suse.de>
24590
24591 PR tree-optimization/56113
24592 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24593 Merge into ...
24594 (equiv_class_lookup_or_add): ... this.
24595 (label_visit): Adjust and fix error in previous patch.
24596 (perform_var_substitution): Adjust.
24597
50fe8924
OE
245982013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24599
24600 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24601 * config/sh/sh4-300.md: Likewise.
24602 * config/sh/sh4a.md: Likewise.
24603 * config/sh/constraints.md: Likewise.
24604 * config/sh/sh.md: Likewise.
24605 * config/sh/netbsd-elf.h: Likewise.
24606 * config/sh/predicates.md: Likewise.
24607 * config/sh/sh-protos.h: Likewise.
24608 * config/sh/ushmedia.h: Likewise.
24609 * config/sh/linux.h: Likewise.
24610 * config/sh/sh.c: Likewise.
24611 * config/sh/superh.h: Likewise.
24612 * config/sh/elf.h: Likewise.
24613 * config/sh/sh4.md: Likewise.
24614 * config/sh/sh.h: Likewise.
24615
1a04ac2b
JDA
246162013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24617
24618 * config/pa/constraints.md: Adjust unused letters. Change "T"
24619 constraint to match_test floating_point_store_memory_operand().
24620 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24621 (base14_operand): New.
24622 (floating_point_store_memory_operand): New.
24623 (integer_store_memory_operand): Revise to use base14_operand and
24624 reg_plus_base_memory_operand.
24625 (move_dest_operand): Allow symbolic_memory_operands.
24626 (symbolic_memory_operand): Check for LO_SOM.
24627 (symbolic_operand): Change default case to break.
24628 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24629 CONST_DOUBLE values to be reloaded by putting them into memory when
24630 the destination is a floating point register.
24631 (movdf): Remove code to handle CONST_DOUBLE.
24632 (movsf): Likewise.
24633 (reload_indf_r1): New.
24634 (reload_insf_r1): New.
24635 Consistently use "Q" and "T" constraints with integer and floating
24636 point move instructions, respectively.
24637 (movdi): Remove FAIL.
24638 Change predicate for source operand unamed DImode move from
24639 general_operand to move_src_operand.
24640 (umulsidi3): Change predicate for destination operand to
24641 register_operand.
24642 Likewise for similar unamed patterns.
24643 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24644 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24645 (hppa_legitimize_address): Simplify mask calculation.
24646 (pa_emit_move_sequence): Revised handling of secondary reloads from
24647 REG+D addresses for floating point loads and stores. Directly handle
24648 loading CONST0_RTX (mode) to a floating point register.
24649 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24650 to floating point registers. Don't restrict secondary reloads to
24651 floating point registers to integer modes. Revise some comments and
24652 cleanup some code.
24653 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24654 (pa_legitimate_address_p): New.
24655 (pa_legitimize_reload_address): New.
24656 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24657 (STRICT_REG_OK_FOR_BASE_P): New.
24658 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24659 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24660
3bf65591
DE
246612013-02-03 David Edelsohn <dje.gcc@gmail.com>
24662 Andrew Dixie <andrewd@gentrack.com>
24663
24664 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24665 flag set.
24666
c0a8a3e6
RS
246672013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24668
24669 * expmed.c (extract_bit_field_1): Pass the full width of the
24670 structure to get_best_reg_extraction_insn.
24671
99113dff
DE
246722013-02-01 David Edelsohn <dje.gcc@gmail.com>
24673
24674 PR target/54601
24675 * configure.ac (use_cxa_atexit): Add AIX.
24676 * configure: Regenerate.
24677
24678 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24679
ddd84654
JJ
246802013-02-01 Jakub Jelinek <jakub@redhat.com>
24681
24682 PR debug/54793
24683 * final.c (need_profile_function): New variable.
24684 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24685 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24686 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24687 notes, targetm.asm_out.function_prologue doesn't emit anything,
24688 HAVE_prologue and profiler should be emitted before prologue,
24689 set need_profile_function instead of emitting it.
24690 (final_scan_insn): If need_profile_function, emit
24691 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24692 NOTE_INSN_FUNCTION_BEG note.
24693
246942013-02-01 Richard Henderson <rth@redhat.com>
749af8ee
RH
24695
24696 * config/rs6000/rs6000.md (smulditi3): New.
24697 (umulditi3): New.
24698
ff2a9d88
RH
24699 * config/alpha/alpha.md (umulditi3): New.
24700
14d52b90
DE
247012013-02-01 David Edelsohn <dje.gcc@gmail.com>
24702
24703 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24704 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24705
8c7ca45c
RB
247062013-02-01 Richard Biener <rguenther@suse.de>
24707
24708 PR tree-optimization/56113
24709 * tree-ssa-structalias.c (label_visit): Reduce work for
24710 single-predecessor nodes.
24711
9f419393
EB
247122013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24713
24714 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24715 range isn't testing for zero.
24716
c5f4be84
SB
247172013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24718
24719 PR middle-end/56113
24720 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24721
dbdbd982
NC
247222013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24723 Nick Clifton <nickc@redhat.com>
24724
24725 * config/v850/constraints.md (Q): Define as a memory constraint.
24726 * config/v850/predicates.md (label_ref_operand): New predicate.
24727 (e3v5_shift_operand): New predicate.
24728 (ior_operator): New predicate.
24729 * config/v850/t-v850: Add e3v5 multilib.
24730 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24731 (v850_gen_movdi): Prototype.
24732 * config/v850/v850.c: Add support for e3v5 architecture.
24733 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24734 TARGET_V850E_UP.
24735 (construct_save_jarl): Add e3v5 long JARL support.
24736 (v850_adjust_insn_length): New function. Adjust length of call
24737 insns when using e3v5 instructions.
24738 (v850_gen_movdi): New function: Generate instructions to move a
24739 DImode value.
24740 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24741 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24742 (TARGET_USE_FPU): Enable for e3v5.
24743 (CONST_OK_FOR_W): New macro.
24744 (ADJUST_INSN_LENGTH): Define.
24745 * config/v850/v850.md (UNSPEC_LOOP): Define.
24746 (attr cpu): Add v850e3v5.
24747 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24748 (movdi): New pattern.
24749 (movdi_internal): New pattern.
24750 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
24751 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
24752 (cstoresf4): Likewise.
24753 (cstoredf4): Likewise.
24754 (insv): New pattern.
24755 (rotlso3_a): New pattern.
24756 (rotlsi3_b): New pattern
24757 (rotlsi3_v850e3v5): New pattern.
24758 (doloop_begin): New pattern.
24759 (fix_loop_counter): New pattern.
24760 (doloop_end): New pattern.
24761 (branch_normal): Add e3v5 long branch support.
24762 (branch_invert): Likewise.
24763 (branch_z_normal): Likewise.
24764 (branch_z_invert): Likewise.
24765 (branch_nz_normal): Likewise.
24766 (branch_nz_invert): Likewise.
24767 (call_internal_short): Add e3v5 register-indirect JARL support.
24768 (call_internal_long): Likewise.
24769 (call_value_internal_short): Likewise.
24770 (call_value_internal_long): Likewise.
24771 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
24772 (mloop): New option.
24773 * config.gcc: Add support for configuring v840e3v5 target.
6edc3e32 24774 * doc/invoke.texi: Document new v850 specific command line options.
dbdbd982 24775
73861a41
PK
247762013-01-31 Paul Koning <ni1d@arrl.net>
24777
24778 PR debug/55059
24779 PR debug/54508
24780 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
24781 children if parent is a class.
24782 (prune_unused_types_prune): Don't add DW_AT_declaration.
24783
e44978dc
RB
247842013-01-31 Richard Biener <rguenther@suse.de>
24785
24786 PR tree-optimization/56157
24787 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
24788 match up operand with SLP child.
24789
90b10dec
JM
247902013-01-31 Jason Merrill <jason@redhat.com>
24791
04d2dadd 24792 PR debug/54410
90b10dec
JM
24793 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
24794 parameters the first time.
24795 (gen_scheduled_generic_parms_dies): Check completeness here.
24796
6e616110
RB
247972013-01-31 Richard Biener <rguenther@suse.de>
24798
24799 PR middle-end/53073
24800 * common.opt (faggressive-loop-optimizations): New flag,
24801 enabled by default.
24802 * doc/invoke.texi (faggressive-loop-optimizations): Document.
24803 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
24804 infer_loop_bounds_from_undefined by it.
24805
636f59cf
RB
248062013-01-31 Richard Biener <rguenther@suse.de>
24807
24808 PR tree-optimization/56150
24809 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
24810 visit virtual operands.
24811 (find_uses_to_rename_bb): Likewise.
24812
248132013-01-31 Richard Biener <rguenther@suse.de>
fcfa87ac
RB
24814
24815 PR tree-optimization/56150
24816 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
24817 mixed store non-store stmts.
24818
32887460
JJ
248192013-01-30 Jakub Jelinek <jakub@redhat.com>
24820
e60e09a0
JJ
24821 PR sanitizer/55374
24822 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
24823 LIBASAN_EARLY_SPEC is defined.
24824 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
24825 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
24826 before %o.
24827 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
24828
32887460
JJ
24829 PR c++/55742
24830 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
24831 invalid args instead of ICEing on it.
24832 (ix86_valid_target_attribute_tree): Return error_mark_node if
24833 ix86_valid_target_attribute_inner_p failed.
24834 (ix86_valid_target_attribute_p): Return false only if
24835 ix86_valid_target_attribute_tree returned error_mark_node. Allow
24836 target("default") attribute.
24837 (sorted_attr_string): Change argument from const char * to tree,
24838 merge in all target attribute arguments rather than just one.
24839 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
24840 instead of free. Avoid using strcat.
24841 (ix86_mangle_function_version_assembler_name): Mangle
24842 target("default") as if no target attribute is present. Adjust
24843 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
24844 instead of xmalloc and XDELETEVEC instead of free.
24845 (ix86_function_versions): Don't return true if one of the decls
24846 doesn't have target attribute. If they don't and one of the decls
24847 is DECL_FUNCTION_VERSIONED, report an error. Adjust
24848 sorted_attr_string caller. Use XDELETEVEC instead of free.
24849 (ix86_supports_function_versions): Remove.
24850 (make_name): Fix up formatting.
24851 (make_dispatcher_decl): Remove resolver_name and its initialization.
24852 Avoid leaking memory.
24853 (is_function_default_version): Return true if there is
24854 target("default") attribute rather than no target attribute at all.
24855 (make_resolver_func): Avoid leaking memory.
24856 (ix86_generate_version_dispatcher_body): Likewise.
24857 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
24858 * target.def (supports_function_versions): Remove.
24859 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
24860 * doc/tm.texi: Regenerated.
24861
73cca0cc
VM
248622013-01-30 Vladimir Makarov <vmakarov@redhat.com>
24863
24864 PR rtl-optimization/56144
24865 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
24866 for values with side effects.
24867
11452e7b
RB
248682013-01-30 Richard Biener <rguenther@suse.de>
24869
24870 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
24871 (sparseset_pop): Likewise.
24872 * cfganal.c (compute_idf): Likewise. Increase work-stack size
24873 to be able to use quick_push in the worker loop.
24874
01cb1ef5
MP
248752013-01-30 Marek Polacek <polacek@redhat.com>
24876
24877 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
24878
e892936e
RB
248792013-01-30 Richard Biener <rguenther@suse.de>
24880
24881 PR lto/56147
6edc3e32 24882 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
e892936e 24883
cc06c01d
GJL
248842013-01-30 Georg-Johann Lay <avr@gjlay.de>
24885
24886 PR tree-optimization/56064
24887 * fixed-value.c (fixed_from_double_int): New function.
24888 * fixed-value.h (fixed_from_double_int): New prototype.
24889 (const_fixed_from_double_int): New static inline function.
24890 * fold-const.c (native_interpret_fixed): New static function.
24891 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
24892 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
24893 (native_encode_fixed): New static function.
24894 (native_encode_expr) <FIXED_CST>: Use it.
24895 (native_interpret_int): Move double_int worker code to...
24896 * double-int.c (double_int::from_buffer): ...this new static method.
24897 * double-int.h (double_int::from_buffer): Prototype it.
24898
d394a308
RB
248992013-01-30 Richard Biener <rguenther@suse.de>
24900
24901 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
24902 New pointer-map and obstack.
24903 (init_alias_vars): Allocate pointer-map and obstack.
24904 (delete_points_to_sets): Free them.
24905 (find_what_var_points_to): Cache result.
24906 (find_what_p_points_to): Adjust for changed interface of
24907 find_what_var_points_to.
24908 (compute_points_to_sets): Likewise.
24909 (ipa_pta_execute): Likewise.
24910
20804d96
RO
249112013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24912
24913 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
24914 * configure: Regenerate.
24915 * config.in: Regenerate.
24916 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
24917 #nobits/#progbits if supported.
24918
a7ad88a2
OE
249192013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
24920
24921 PR target/56121
24922 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
24923 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
24924 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
24925
7af79f92
GY
249262013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24927
91bfca59
OE
24928 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
24929 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
7af79f92 24930
753bcf7b
GY
249312013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24932
91bfca59
OE
24933 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
24934 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
753bcf7b 24935
8cbc2ea8
GY
249362013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24937
91bfca59
OE
24938 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
24939 declaration.
24940 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
24941 * config/arm/cortex-a7.md: New bypasses using
24942 arm_mac_accumulator_is_result.
8cbc2ea8 24943
697a3325
GY
249442013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24945
91bfca59 24946 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
697a3325 24947 (cortex_a7_neon_mla): Likewise.
91bfca59
OE
24948 (cortex_a7_fpfmad): New reservation.
24949 (cortex_a7_fpmacs): Use ffmas and update required units.
24950 (cortex_a7_fpmuld): Update required units and latency.
24951 (cortex_a7_fpmacd): Likewise.
24952 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
24953 (cortex_a7_neon). Likewise.
24954 (bypass) Update participating units.
697a3325 24955
29637783
GY
249562013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24957
91bfca59
OE
24958 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
24959 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
24960 from fmac to ffma.
24961 * config/arm/vfp11.md (vfp_farith): Use ffmas.
24962 (vfp_fmul): Use ffmad.
24963 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
24964 (cortex_r4_fmacd): Use ffmad.
24965 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
24966 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
24967 (cortex_a9_fmacd): Use ffmad.
24968 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
24969 (cortex_a8_vfp_macd): Use ffmad.
24970 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
24971 (cortex_a5_fpmacd): Use ffmad.
24972 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
24973 (cortex_a15_vfp_macd): Use ffmad.
24974 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
29637783 24975
2ee8a2d5
JM
249762013-01-29 Jason Merrill <jason@redhat.com>
24977
24978 PR libstdc++/54314
24979 * varasm.c (default_assemble_visibility): Don't warn about
24980 visibility on artificial decls.
24981
78d087bc
RB
249822013-01-29 Richard Biener <rguenther@suse.de>
24983
24984 PR tree-optimization/56113
24985 * tree-ssa-structalias.c (equiv_class_lookup): Also return
24986 the bitmap leader.
24987 (label_visit): Free duplicate bitmaps and record the leader instead.
24988 (perform_var_substitution): Adjust.
24989
83ba4d6f
RB
249902013-01-29 Richard Biener <rguenther@suse.de>
24991
24992 PR tree-optimization/55270
24993 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
24994 the CFG, schedule loops for fixup.
24995
66dfe13f
NC
249962013-01-29 Nick Clifton <nickc@redhat.com>
24997
24998 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
24999 SP_REG.
25000
5a579c3b
LE
250012013-01-28 Leif Ekblad <leif@rdos.net>
25002
25003 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
25004 * config/i386/i386.h (TARGET_RDOS): New macro.
25005 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
25006 * config/i386/i386.c (ix86_option_override_internal): For 64bit
25007 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
25008 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
25009 DEFAULT_LARGE_SECTION_THRESHOLD.
25010 * config/i386/i386.md (R14_REG, R15_REG): New constants.
25011 * config/i386/rdos.h: New file.
25012 * config/i386/rdos64.h: New file.
25013
51e44392
BS
250142013-01-28 Bernd Schmidt <bernds@codesourcery.com>
25015
25016 PR other/54814
25017 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
25018 TEST_HARD_REG_BIT.
25019
db1fb332
JJ
250202013-01-28 Jakub Jelinek <jakub@redhat.com>
25021
25022 PR rtl-optimization/56117
25023 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
25024 call cselib_lookup_from_insn on the MEM before calling
25025 add_insn_mem_dependence.
25026
16917761
RB
250272013-01-28 Richard Biener <rguenther@suse.de>
25028
25029 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
25030 to a stmt that didn't have one.
25031 (copy_phis_for_bb): Likewise for PHI arguments.
25032 (copy_debug_stmt): Likewise for debug stmts.
25033
b9fc0497
RB
250342013-01-28 Richard Biener <rguenther@suse.de>
25035
25036 PR tree-optimization/56034
6edc3e32 25037 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
b9fc0497
RB
25038 (partition_builtin_p): Adjust.
25039 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25040 it is the last partition.
25041 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25042 up the vertex for the definition.
25043 (classify_partition): Classify whether a partition is a
25044 PKIND_REDUCTION, thus has uses outside of the loop.
25045 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25046 Merge all PKIND_REDUCTION partitions into the last partition.
25047 (tree_loop_distribution): Seed partitions from reductions as well.
25048
aa710d25
JJ
250492013-01-28 Jakub Jelinek <jakub@redhat.com>
25050
0bfbca58
JJ
25051 PR tree-optimization/56125
25052 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25053 pow(x,c) into sqrt(x) * powi(x, n/2) or
25054 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25055 optimizing for size.
25056 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25057 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25058 integer.
25059
aa710d25
JJ
25060 PR tree-optimization/56094
25061 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25062 to UNKNOWN_LOCATION while gimplifying expr.
25063
77dc5297
UB
250642013-01-27 Uros Bizjak <ubizjak@gmail.com>
25065
25066 PR target/56114
25067 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25068 operand 0 in movabs insn template for -masm=intel asm alternative.
25069 (*movabs<mode>_2): Ditto for operand 1.
25070
0bfbca58 250712013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
fbf0cf90
DH
25072
25073 PR target/54663
25074 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25075 of microblaze-c.o
25076
0bfbca58 250772013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
6f14eef2
EI
25078
25079 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25080 tm_file.
25081
0bfbca58 250822013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
706b2314
N
25083
25084 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25085 Undef to avoid warning.
25086
478f60f9
MH
250872013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25088
25089 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25090 * configure: Regenerate.
25091
d7fa6ee2
JJ
250922013-01-25 Jakub Jelinek <jakub@redhat.com>
25093
25094 PR tree-optimization/56098
25095 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25096 for stmts with volatile ops.
25097 (cond_store_replacement): Don't optimize if assign has volatile ops.
25098 (cond_if_else_store_replacement_1): Don't optimize if either
25099 then_assign or else_assign have volatile ops.
25100 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25101 volatile ops.
25102
f8fe87bd
GJL
251032013-01-25 Georg-Johann Lay <avr@gjlay.de>
25104
25105 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25106
28398d0d
GJL
251072013-01-25 Georg-Johann Lay <avr@gjlay.de>
25108
25109 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25110 missing ':' in asm example.
25111
b7d7d917
TB
251122013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25113
25114 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25115 entries into lane and laneq entries.
77dc5297
UB
25116 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25117 Remove AdvSIMD scalar modes.
b7d7d917
TB
25118 (aarch64_sq<r>dmulh_laneq<mode>): New.
25119 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25120 modes.
25121 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25122 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25123 * config/aarch64/iterators.md (VCOND): New.
25124 (VCONQ): New.
25125
556f9906
GJL
251262013-01-25 Georg-Johann Lay <avr@gjlay.de>
25127
25128 PR target/54222
25129 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25130 Add NULL LIBNAME argument to existing definitions.
25131 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25132 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25133 * config/avr/avr.c (DEF_BUILTIN): Same.
25134 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25135 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25136 implementation is available (DECL_ASSEMBLER_NAME is set).
25137 (avr_fold_absfx): New static function.
25138 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25139 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25140 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25141 AVR_BUILTIN_ABSLLK.
25142 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25143 (abshk, absk, abslk, absllk): Provide as static inline functions.
25144
1f546bbb
MP
251452013-01-25 Marek Polacek <polacek@redhat.com>
25146
25147 PR tree-optimization/56035
25148 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25149
7e184bd7
UB
251502012-01-24 Uros Bizjak <ubizjak@gmail.com>
25151
25152 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25153 (*movtf_internal_rex64): Add (!o,C) alternative
25154 (*movxf_internal_rex64): Ditto.
25155 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25156
f5ba49ac
SM
251572013-01-24 Shenghou Ma <minux.ma@gmail.com>
25158
25159 * doc/invoke.texi: fix typo.
25160 * doc/objc.texi: fix typo.
25161
a6343728
RS
251622013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25163
25164 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25165 for the first two alternatives.
25166
cd030c07
DN
251672013-01-24 Diego Novillo <dnovillo@google.com>
25168
77dc5297 25169 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
cd030c07
DN
25170 (ggc-zone.o): Remove.
25171 * configure.ac: Remove option --with-gc.
25172 * configure: Re-generate.
25173 * doc/install.texi: Remove documentation for --with-gc.
25174 * gengtype.c (write_enum_defn): Remove. Update all users.
25175 (write_Types_process_field): Remove generation of gt_e_* argument.
25176 (output_type_enum): Remove. Update all users.
25177 (write_enum_defn): Remove. Update all users.
25178 (enum alloc_zone): Remove. Update all users.
77dc5297 25179 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
cd030c07
DN
25180 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25181 Update all callers.
25182 (struct ptr_data): Remove field TYPE. Update all users.
77dc5297 25183 (gt_pch_note_object): Remove argument TYPE. Update all users.
cd030c07
DN
25184 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25185 Update all users.
aaf1e810 25186 * ggc-none.c (ggc_alloc_typed_stat): Remove.
cd030c07
DN
25187 (struct alloc_zone): Remove.
25188 (ggc_internal_alloc_zone_stat): Remove.
25189 (ggc_internal_cleared_alloc_zone_stat): Remove.
25190 * ggc-page.c (ggc_alloc_typed_stat): Remove.
77dc5297
UB
25191 (ggc_pch_count_object): Remove last argument. Update all users.
25192 (ggc_pch_alloc_object): Remove last argument. Update all users.
cd030c07
DN
25193 (struct alloc_zone): Remove.
25194 * ggc-zone.c: Remove.
77dc5297 25195 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
cd030c07
DN
25196 (struct alloc_zone): Remove.
25197 (ggc_alloc_typed_stat): Remove.
25198 (ggc_alloc_typed): Remove.
25199 (ggc_splay_alloc): Remove first argument.
25200 (rtl_zone): Remove. Update all users.
25201 (tree_zone): Remove. Update all users.
25202 (tree_id_zone): Remove. Update all users.
25203 (ggc_internal_zone_alloc_stat): Remove. Update all users.
77dc5297 25204 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
cd030c07 25205 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
77dc5297 25206 * tree-ssanames.c: Remove references to zone allocator in comments.
cd030c07 25207
a861ffa4
GJL
252082013-01-24 Georg-Johann Lay <avr@gjlay.de>
25209
25210 * config/avr/avr.c (avr_out_fract): Make register numbers that
25211 might be outside of source operand signed.
25212
593c0ddd
UB
252132013-01-24 Uros Bizjak <ubizjak@gmail.com>
25214
25215 * config/i386/constraints.md (Yf): New constraint.
25216 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25217 of f constraint to conditionaly disable x87 register preferences.
25218 (*movdf_internal): Ditto.
25219 (*movsf_internal): Ditto.
25220
e86c0101
SB
252212013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25222
25223 PR inline-asm/55934
25224 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25225 that have operands with impossible constraints.
25226 Add a FIXME for a speed-up opportunity.
25227 * lra-constraints.c (process_alt_operands): Verify that a class
25228 selected from constraints on asms is valid for the operand mode.
25229 (curr_insn_transform): Remove incorrect comment.
25230
f6fee35f
DE
252312013-01-23 David Edelsohn <dje.gcc@gmail.com>
25232
25233 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25234 TOC operand is a valid symbol ref in the constant pool.
25235
252362013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
c7729184 25237
aaf1e810 25238 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
c7729184 25239
dc62d7d1
GJL
252402013-01-23 Georg-Johann Lay <avr@gjlay.de>
25241
25242 PR target/54222
25243 * config/avr/stdfix.h: New file.
25244 * t-avr (stdfix-gcc.h): New rule to build it.
25245 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25246
2660d12d
KS
252472013-01-23 Kostya Serebryany <kcc@google.com>
25248
77dc5297
UB
25249 * config/darwin.h: remove dependency on
25250 CoreFoundation (asan on Mac OS).
2660d12d 25251
a70418fc
JJ
252522013-01-23 Jakub Jelinek <jakub@redhat.com>
25253
25254 PR target/49069
25255 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25256 instead of cmpdi_operand for first comparison operand.
25257 Don't assert that comparison operands aren't both constants.
25258
47876a2a
JW
252592013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25260
25261 * doc/install.texi (Downloading the Source): Update references to
25262 downloading separate components.
25263
252642013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
4ebd56a5
JW
25265
25266 * doc/extend.texi (__int128): Improve grammar.
25267
47876a2a 252682013-01-22 Uros Bizjak <ubizjak@gmail.com>
32fc5b8a
UB
25269
25270 PR target/56028
25271 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25272 alternative to (o,r).
25273 (*movdi_internal_rex64): Remove (!o,n) alternative.
25274 (DImode immediate->memory splitter): Remove.
25275 (DImode immediate->memory peephole2): Remove.
25276 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25277 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25278 alternative to (!o,*r).
25279 (*movtf_internal_sse): New pattern.
25280 (*movxf_internal_rex64): New pattern.
25281 (*movxf_internal): Disable for TARGET_64BIT.
25282 (*movdf_internal_rex64): Remove (!o,F) alternative.
25283
3a984f10
JJ
252842013-01-22 Jakub Jelinek <jakub@redhat.com>
25285
502498d5
JJ
25286 PR middle-end/56074
25287 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25288 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25289 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25290 stmt locations where LOCATION_LOCUS of the stmt location is
25291 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25292
3a984f10
JJ
25293 PR target/55686
25294 * config/i386/i386.md (UNSPEC_STOS): New.
25295 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25296 *strsetqi_1): Add UNSPEC_STOS.
25297
fa817f7f
PC
252982013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25299
25300 PR c++/56067
25301 * doc/invoke.texi: Remove left over -Wsynth example.
25302
8f498c1b
JJ
253032013-01-21 Jakub Jelinek <jakub@redhat.com>
25304
25305 PR tree-optimization/56051
25306 * fold-const.c (fold_binary_loc): Don't fold
25307 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25308 a narrowing conversion, or widening conversion from signed
25309 to unsigned.
25310
47876a2a 253112013-01-21 Uros Bizjak <ubizjak@gmail.com>
35d59da7
UB
25312
25313 PR rtl-optimization/56023
25314 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25315 dependent on debug instruction.
25316
5b9db1bc
MJ
253172013-01-21 Martin Jambor <mjambor@suse.cz>
25318
25319 PR middle-end/56022
25320 * function.c (allocate_struct_function): Call
25321 invoke_set_current_function_hook earlier.
25322
e8bb7d68
JJ
253232013-01-21 Jakub Jelinek <jakub@redhat.com>
25324
25325 * reload1.c (init_reload): Only initialize reload_obstack
25326 during the first call.
25327
616a4e32
MP
253282013-01-21 Marek Polacek <polacek@redhat.com>
25329
25330 * cfgloop.c (verify_loop_structure): Fix up grammar.
25331
4401981b
YHH
253322013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25333
25334 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25335 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25336
8e87740b
RR
253372013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25338
25339 PR target/56058
25340 * config/arm/marvell-pj4.md: Update copyright year.
25341 Fix up use of alu to alu_reg and simple_alu_imm.
25342
47876a2a 253432013-01-21 Uros Bizjak <ubizjak@gmail.com>
1f6bc337
UB
25344
25345 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25346
89d56d79
VM
253472013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25348
25349 PR target/55433
25350 * lra-constraints.c (curr_insn_transform): Don't reuse original
1f6bc337 25351 insn for secondary memory move when memory mode should be different.
89d56d79 25352
fe603553
JDA
253532013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25354
25355 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25356 atomic_storedi_1): New patterns.
25357
01284895
VK
253582013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25359
25360 btver2 pipeline descriptions.
25361 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25362 descriptions.
25363 * config/i386/i386.md (btver2_decode): New type attributes.
1f6bc337 25364 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
01284895
VK
25365 type attributes.
25366 * config/i386/btver2.md: New file describing btver2 pipelines.
25367
5630e3e1
JL
253682013-01-19 Andrew Pinski <apinski@cavium.com>
25369
25370 PR tree-optimization/52631
25371 * tree-ssa-sccvn (visit_use): Before looking up the original
25372 statement, try looking up the simplified expression.
25373
650ae806
AG
253742013-01-19 Anthony Green <green@moxielogic.com>
25375
25376 * config/moxie/moxie.c (moxie_expand_prologue): Set
25377 current_function_static_stack_size.
25378
e300ec2d
JJ
253792013-01-18 Jakub Jelinek <jakub@redhat.com>
25380
25381 PR tree-optimization/56029
25382 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25383 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25384
a3d7af04
SS
253852013-01-18 Sharad Singhai <singhai@google.com>
25386
25387 PR tree-optimization/55995
25388 * dumpfile.c (dump_loc): Print location only if available.
25389 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25390
66aa7879
VM
253912013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25392
25393 PR target/55433
25394 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25395 secondary memory move.
25396 (inherit_reload_reg): Use rclass instead of cl for
25397 check_secondary_memory_needed_p.
25398
3f0fee7b
JJ
253992013-01-18 Jakub Jelinek <jakub@redhat.com>
25400
25401 PR middle-end/56015
25402 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
1f6bc337 25403 the case where writing real complex part of target modifies op1.
3f0fee7b 25404
70c67693
JG
254052013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25406
25407 * config/aarch64/aarch64-simd.md
25408 (aarch64_vcond_internal<mode>): Handle unordered cases.
25409 * config/aarch64/iterators.md (v_cmp_result): New.
25410
df8de9b3
YHH
254112013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25412 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25413
25414 * config/arm/marvell-pj4.md: New file.
25415 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25416 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25417 (generic_vfp): Likewise.
25418 * config/arm/arm-cores.def: Add marvell-pj4.
25419 * config/arm/arm-tune.md: Regenerate.
25420 * config/arm/arm-tables.opt: Regenerate.
25421 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25422 * doc/invoke.texi: Document marvell-pj4.
25423
be30c356
TB
254242013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25425
25426 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25427
0bfbca58 254282013-01-18 Alexandre Oliva <aoliva@redhat.com>
af6236c1
AO
25429
25430 PR debug/54114
25431 PR debug/54402
25432 PR debug/49888
25433 * var-tracking.c (negative_power_of_two_p): New.
25434 (global_get_addr_cache, local_get_addr_cache): New.
25435 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25436 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25437 heading comment.
25438 (vt_stack_offset_p): Remove.
25439 (vt_canon_true_dep): Always canonicalize loc's address.
25440 (clobber_overlapping_mems): Make sure we have a MEM.
25441 (local_get_addr_clear_given_value): New.
25442 (val_reset): Clear local cached entries.
25443 (compute_bb_dataflow): Create and release the local cache.
25444 Disable duplicate MEMs clobbering.
25445 (emit_notes_in_bb): Clobber MEMs likewise.
25446 (vt_emit_notes): Create and release the local cache.
25447 (vt_initialize, vt_finalize): Create and release the global
25448 cache, respectively.
1f6bc337 25449 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
af6236c1 25450
0bfbca58 254512013-01-18 Alexandre Oliva <aoliva@redhat.com>
7ff37ffa
AO
25452
25453 PR libmudflap/53359
25454 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25455 not found in the symtab.
25456
0bfbca58 254572013-01-18 Alexandre Oliva <aoliva@redhat.com>
3aa03517 25458
c350ba53 25459 PR debug/56006
3aa03517
AO
25460 PR rtl-optimization/55547
25461 PR rtl-optimization/53827
25462 PR debug/53671
25463 PR debug/49888
25464 * alias.c (offset_overlap_p): New, factored out of...
25465 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25466 the conservative special case for symbolic constants. Don't
25467 adjust zero sizes on alignment.
25468
c664546f
JL
254692013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25470
25471 PR rtl-optimization/52573
25472 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25473 REG_UNUSED for the same register.
25474
1bd3f750
MP
254752013-01-17 Richard Biener <rguenther@suse.de>
25476 Marek Polacek <polacek@redhat.com>
25477
25478 PR rtl-optimization/55833
25479 * loop-unswitch.c (unswitch_loops): Move loop verification...
25480 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25481 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25482 Set it to true when we're removing a loop from hierarchy tree in
25483 an irreducible region.
25484 (fix_bb_placements): Adjust caller.
25485 (fix_loop_placements): Likewise.
25486
e52a8b71
GJL
254872013-01-17 Georg-Johann Lay <avr@gjlay.de>
25488
25489 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25490 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25491 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25492 Remove ID. Adjust comments.
25493 * config/avr/avr-c.c (avr_builtin_name): Remove.
25494 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25495 * config/avr/avr.c (avr_tolower): New static function.
25496 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25497 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25498 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25499 default expansion.
25500
8386a7ea
JH
255012013-01-17 Jan Hubicka <jh@suse.cz>
25502
610fb637 25503 PR tree-optimization/55273
8386a7ea
JH
25504 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25505
47876a2a 255062013-01-17 Uros Bizjak <ubizjak@gmail.com>
8abaebcd
UB
25507
25508 PR target/55981
25509 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25510 store through atomic_store<mode>_1.
25511 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25512
8222c37e
MJ
255132013-01-17 Martin Jambor <mjambor@suse.cz>
25514
25515 PR tree-optimizations/55264
25516 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25517 for virtual methods.
25518 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25519 virtual methods before inlining is over.
25520 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25521 virtual functions.
25522 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25523 non-virtual.
25524
79f01c76
VM
255252013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25526
25527 PR rtl-optimization/56005
25528 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25529 pending reads for prefetch.
25530
d0b6bb1b
IB
255312013-01-16 Ian Bolton <ian.bolton@arm.com>
25532
aaf1e810 25533 * config/aarch64/aarch64.md
d0b6bb1b
IB
25534 (*cstoresi_neg_uxtw): New pattern.
25535 (*cmovsi_insn_uxtw): New pattern.
25536 (*<optab>si3_uxtw): New pattern.
25537 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25538 (*<optab>si3_insn_uxtw): New pattern.
25539 (*bswapsi2_uxtw): New pattern.
25540
cb9cf03b
RB
255412013-01-16 Richard Biener <rguenther@suse.de>
25542
25543 * tree-inline.c (tree_function_versioning): Remove set but
25544 never used variable.
25545
2cfc56b9
RB
255462013-01-16 Richard Biener <rguenther@suse.de>
25547
25548 PR tree-optimization/55964
25549 * tree-flow.h (rename_variables_in_loop): Remove.
25550 (rename_variables_in_bb): Likewise.
25551 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25552 (copy_loop_before): Adjust and delete update-ssa status.
25553 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
8abaebcd 25554 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
2cfc56b9
RB
25555 (rename_variables_in_loop): Remove.
25556 (slpeel_update_phis_for_duplicate_loop): Likewise.
25557 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25558 use available cfg machinery instead of duplicating it.
25559 Update PHI nodes and perform poor-mans SSA update here.
25560 (slpeel_tree_peel_loop_to_edge): Adjust.
25561
c25a0c60
RB
255622013-01-16 Richard Biener <rguenther@suse.de>
25563
25564 PR tree-optimization/54767
25565 PR tree-optimization/53465
25566 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25567 (vrp_visit_phi_node): For PHI arguments coming via backedges
25568 drop all symbolical range information.
25569 (execute_vrp): Compute backedges.
25570
04b535af
RB
255712013-01-16 Richard Biener <rguenther@suse.de>
25572
25573 * doc/install.texi: Update CLooG and ISL requirements to
25574 0.18.0 and 0.11.1.
25575
8b0a1e0b
CB
255762013-01-16 Christian Bruel <christian.bruel@st.com>
25577
25578 PR target/55301
25579 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25580 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25581 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25582
255832013-01-16 DJ Delorie <dj@redhat.com>
25584
25585 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25586 (UNSPECV_SP_SWITCH_E): New.
25587 (sp_switch_1): Change to an unspec.
25588 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25589 replace $r15.
25590
255912013-01-16 Uros Bizjak <ubizjak@gmail.com>
88e784e6
UB
25592
25593 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25594 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25595 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25596 (expand_mem_thread_fence): Ditto.
25597 (expand_mem_signal_fence): Ditto.
25598 (expand_atomic_load): Ditto.
25599 (expand_atomic_store): Ditto.
25600
0bfbca58 256012013-01-16 Alexandre Oliva <aoliva@redhat.com>
5147bf6a
AO
25602
25603 PR rtl-optimization/55547
25604 PR rtl-optimization/53827
25605 PR debug/53671
25606 PR debug/49888
25607 * alias.c (memrefs_conflict_p): Set sizes to negative after
25608 AND adjustments.
25609
305e3ac1
JJ
256102013-01-15 Jakub Jelinek <jakub@redhat.com>
25611
25612 PR target/55940
25613 * function.c (thread_prologue_and_epilogue_insns): Always
25614 add crtl->drap_reg to set_up_by_prologue.set, even if
25615 stack_realign_drap is false.
25616
f78ac4f2
JBG
256172013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25618
25619 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25620 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25621 *call): Fix indention.
25622
a78a8cc4
TV
256232013-01-15 Tom de Vries <tom@codesourcery.com>
25624
25625 PR target/55876
25626 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25627 Update comment.
25628
0e80383f
VM
256292013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25630
305e3ac1 25631 PR rtl-optimization/55153
0e80383f
VM
25632 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25633
207b5956
MJ
256342013-01-15 Martin Jambor <mjambor@suse.cz>
25635
25636 PR tree-optimization/55920
25637 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25638 accesses as grp_to_be_debug_replaced.
25639
a7818b54
JJ
256402013-01-15 Jakub Jelinek <jakub@redhat.com>
25641
25642 PR tree-optimization/55920
25643 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25644 there is non-useless type conversion needed from debug rhs to lhs,
25645 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25646
b0fe107e
JM
256472013-01-15 Joseph Myers <joseph@codesourcery.com>
25648 Mikael Pettersson <mikpe@it.uu.se>
25649
25650 PR target/43961
25651 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25652 Thumb.
25653 (ASM_OUTPUT_CASE_LABEL): Remove.
25654 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25655 * final.c (shorten_branches): Update alignment of labels before
25656 jump tables if CASE_VECTOR_SHORTEN_MODE.
25657
34ab62ee
RB
256582013-01-15 Richard Biener <rguenther@suse.de>
25659
25660 PR bootstrap/55961
25661 * system.h: Do not include gmp.h for building host tools.
25662
783a3a05
RB
256632013-01-15 Richard Biener <rguenther@suse.de>
25664
25665 PR middle-end/55882
25666 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25667 account for bitpos when computing alignment.
25668
3a579e09
VY
256692013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25670
25671 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25672 (ix86_target_macros_internal): Likewise.
25673
25674 * config/i386/i386.c (m_CORE2I7): Removed.
25675 (m_CORE_HASWELL): New macro.
25676 (m_CORE_ALL): Likewise.
25677 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25678 (initial_ix86_arch_features): Likewise.
25679 (processor_target_table): Initializations for Core avx2.
25680 (cpu_names): New names "core-avx2".
25681 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25682 PROCESSOR_CORE_HASWELL.
25683 (ix86_issue_rate): New case.
25684 (ia32_multipass_dfa_lookahead): Likewise.
25685 (ix86_sched_init_global): Likewise.
25686
25687 * config/i386/i386.h (TARGET_HASWELL): New macro.
25688 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25689 (processor_type): New PROCESSOR_HASWELL.
25690
955f5a07
JJ
256912013-01-15 Jakub Jelinek <jakub@redhat.com>
25692
ff784829
JJ
25693 PR tree-optimization/55955
25694 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25695 *SHIFT_EXPR and *ROTATE_EXPR codes.
25696
955f5a07
JJ
25697 PR tree-optimization/48766
25698 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25699 -ftrapv disable -fwrapv.
25700
aeb8b4e9
GJL
257012013-01-14 Georg-Johann Lay <avr@gjlay.de>
25702
25703 PR target/55974
25704 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25705 etc. to 1 and not to __flash.
25706 Use LL suffix for __INT24_MAX__ with -mint8.
25707 Use ULL suffix for __UINT24_MAX__ with -mint8.
25708
1c494c6a
GJL
257092013-01-14 Georg-Johann Lay <avr@gjlay.de>
25710
25711 * config/avr/avr-arch.h
25712 (struct base_arch_s): Use typedef avr_arch_t instead.
25713 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25714 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25715 * config/avr/avr.c: Same.
25716 * config/avr/avr-devices.c: Same.
25717 * config/avr/driver-avr.c: Same.
25718 * config/avr/gen-avr-mmcu-texi.c: Same.
25719 * config/avr/avr-mcus.def: Adjust comment.
25720
a50344cb
TB
257212013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25722
88e784e6
UB
25723 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25724 * config/aarch64/iterators.md (VALLDI): New.
a50344cb 25725
47876a2a 257262013-01-14 Uros Bizjak <ubizjak@gmail.com>
e6f0e052
UB
25727 Andi Kleen <ak@linux.intel.com>
25728
25729 PR target/55948
25730 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25731 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25732 memmodel flag.
25733
00892272
GJL
257342013-01-14 Georg-Johann Lay <avr@gjlay.de>
25735
25736 * config/avr/avr-stdint.h: Remove trailing blanks.
25737 * config/avr/avr-log.h: Same.
25738 * config/avr/avr-arch.h: Same.
25739 * config/avr/avr-devices.c: Same.
25740 * config/avr/avr-dimode.md: Same.
25741 * config/avr/predicates.md: Same.
25742 * config/avr/avr-c.c: Same. And fix typo.
e6f0e052 25743
00892272
GJL
25744 * config/avr/avr-protos.h: Same. And:
25745 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25746 (init_cumulative_args): Rename to avr_init_cumulative_args.
25747 (expand_prologue): Rename to avr_expand_prologue.
25748 (expand_epilogue): Rename to avr_expand_epilogue.
25749 (adjust_insn_length): Rename to avr_adjust_insn_length.
25750 (notice_update_cc): Rename to avr_notice_update_cc.
25751 (final_prescan_insn): Rename to avr_final_prescan_insn.
25752 * config/avr/avr.c: Same.
25753 * config/avr/avr.h: Same.
25754 * config/avr/avr.md: Remove trailing blanks.
25755 (prologue): Use avr_expand_prologue.
25756 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
25757
7bb01996
RB
257582013-01-14 Richard Biener <rguenther@suse.de>
25759
25760 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
25761 verify_location, collect_subblocks): New functions.
25762 (verify_gimple_in_cfg): Verify that locations only reference
25763 BLOCKs in the functions BLOCK tree.
25764
2724573f
RB
257652013-01-14 Richard Biener <rguenther@suse.de>
25766
25767 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
25768 PHI argument.
25769 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
25770 unshare reference.
25771 (insert_out_of_ssa_copy_on_edge): Likewise.
25772 (rewrite_close_phi_out_of_ssa): Likewise.
25773 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
25774 debug expressions.
25775 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
25776 propagated constants.
25777 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
25778 can not be shared.
25779
9a0bbab6
GJL
257802013-01-14 Georg-Johann Lay <avr@gjlay.de>
25781
25782 * config/avr/avr-modes.def: Add GPL copyright notice.
25783
45805f17
UB
257842013-01-13 Uros Bizjak <ubizjak@gmail.com>
25785
25786 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
25787 MEMMODEL_MASK to determine memory model.
25788 (atomic_store<mode>): Ditto from operands[2].
25789 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
25790
9d60be38
JJ
257912013-01-13 Jakub Jelinek <jakub@redhat.com>
25792
25793 PR fortran/55935
45805f17 25794 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
9d60be38
JJ
25795 (fold_gimple_assign): Don't call unshare_expr here.
25796 (fold_ctor_reference): Call unshare_expr.
25797
e7f49d92
TG
257982013-01-13 Terry Guo <terry.guo@arm.com>
25799
aaf1e810
EB
25800 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
25801 * doc/fragments.texi: Document MULTILIB_REUSE.
25802 * gcc.c (multilib_reuse): New internal spec.
e7f49d92 25803 (set_multilib_dir): Also search multilib from multilib_reuse.
aaf1e810 25804 * genmultilib (tmpmultilib3): Refactor code.
e7f49d92
TG
25805 (tmpmultilib4): Ditto.
25806 (multilib_reuse): New multilib argument.
25807
fbd03a27
RS
258082013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
25809
25810 * Makefile.in: Update copyright.
25811
b3681f13
TV
258122013-01-12 Tom de Vries <tom@codesourcery.com>
25813
25814 PR middle-end/55890
25815 * calls.c (expand_call): Check if arg_nr is valid.
25816
3f287e4b
MM
258172013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
25818
25819 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
25820 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
5bc35113
MM
25821 documentation. Add missing '__' in front of
25822 __builtin_ia32_packssdw256.
3f287e4b 25823
1abcd5eb
AK
258242013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25825
25826 PR target/55719
25827 * config/s390/s390.c (s390_preferred_reload_class): Do not return
25828 NO_REGS for larl operands.
25829 (s390_reload_larl_operand): Use s390_load_address instead of
25830 emit_move_insn.
25831
980d0812
RB
258322013-01-11 Richard Biener <rguenther@suse.de>
25833
25834 * tree-cfg.c (verify_node_sharing_1): Split out from ...
25835 (verify_node_sharing): ... here.
25836 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
25837
7cb9fd07
EB
258382013-01-11 Eric Botcazou <ebotcazou@adacore.com>
25839
25840 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
25841 Substitute TREECHECKING.
25842 * configure: Regenerate.
25843 * Makefile.in (TREECHECKING): New.
25844
47876a2a 258452013-01-11 Richard Guenther <rguenther@suse.de>
c7ab2530
RG
25846
25847 PR tree-optimization/44061
25848 * tree-vrp.c (extract_range_basic): Compute zero as
25849 value-range for __builtin_constant_p of function parameters.
25850
d1e082c2
RS
258512013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
25852
45805f17 25853 Update copyright years.
d1e082c2 25854
f85021b0
VM
258552013-01-10 Vladimir Makarov <vmakarov@redhat.com>
25856
9d60be38 25857 PR rtl-optimization/55672
0160442c
VM
25858 * lra-eliminations.c (mark_not_eliminable): Permit addition with
25859 const to be eliminable.
f85021b0 25860
7a8b1ec4
DE
258612013-01-10 David Edelsohn <dje.gcc@gmail.com>
25862
25863 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
25864 * configure: Regenerate.
25865
ad2e5b71
RB
258662013-01-10 Richard Biener <rguenther@suse.de>
25867
ddf9322d 25868 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
ad2e5b71 25869
ffc5b2bb
RB
258702013-01-10 Richard Biener <rguenther@suse.de>
25871
25872 PR bootstrap/55792
25873 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
25874 locations for virtual PHI arguments.
25875 (rewrite_update_phi_arguments): Likewise.
25876
e1f674e4
JS
258772013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
25878
25879 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
25880 on to assembler.
25881
a57fc743
JJ
258822013-01-10 Jakub Jelinek <jakub@redhat.com>
25883
25884 PR tree-optimization/55921
25885 * tree-complex.c (expand_complex_asm): New function.
25886 (expand_complex_operations_1): Call it for GIMPLE_ASM.
25887
0ff4390d
AK
258882013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25889
25890 PR target/55718
25891 * config/s390/s390.c (s390_symref_operand_p)
25892 (s390_loadrelative_operand_p): Merge the two functions.
25893 (s390_check_qrst_address, print_operand_address): Add parameters
25894 to s390_loadrelative_operand_p invokation.
25895 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
25896 (s390_reload_larl_operand, s390_secondary_reload): Use
25897 s390_loadrelative_operand_p instead of s390_symref_operand_p.
25898 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
25899
6f557e0e
MS
259002013-01-09 Mike Stump <mikestump@comcast.net>
25901
25902 * dse.c (record_store): Remove unnecessary assert.
25903
7770c9e9
JH
259042013-01-09 Jan Hubicka <jh@suse.cz>
25905
25906 PR tree-optimization/55569
25907 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
25908 * cfgloop.h (scale_loop_profile): Likewise.
25909
a19b1432
JH
259102013-01-09 Jan Hubicka <jh@suse.cz>
25911
25912 PR lto/45375
ddf9322d
UB
25913 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
25914 functions.
a19b1432
JH
25915 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
25916
9a002da8
RS
259172013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
25918
25919 PR middle-end/55114
25920 * expr.h (maybe_emit_group_store): Declare.
25921 * expr.c (maybe_emit_group_store): New function.
25922 * builtins.c (expand_builtin_int_roundingfn): Call it.
25923 (expand_builtin_int_roundingfn_2): Likewise.
25924
511dcace
VM
259252013-01-09 Vladimir Makarov <vmakarov@redhat.com>
25926
e1f2b729 25927 PR rtl-optimization/55829
511dcace
VM
25928 * lra-constraints.c (match_reload): Add code for absent output.
25929 (curr_insn_transform): Add code for reloads of matched inputs
25930 without output.
25931
7b0fe4f4
UB
259322013-01-09 Uros Bizjak <ubizjak@gmail.com>
25933
25934 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
25935 attribute of movddup insn to DF.
25936 (*vec_interleave_lowv2df): Ditto.
25937 (vec_dupv2df): Ditto.
25938
870ca331
JH
259392013-01-09 Jan Hubicka <jh@suse.cz>
25940
25941 PR tree-optimiation/55875
25942 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
25943 EVERY_ITERATION parameter.
7b0fe4f4 25944 (number_of_iterations_exit): Check if exit is executed every iteration.
870ca331
JH
25945 (idx_infer_loop_bounds): Similarly here.
25946 (n_of_executions_at_most): Simplify
25947 to only test for cases where statement is dominated by the
7b0fe4f4 25948 particular bound; handle correctly the "postdominance" test.
870ca331
JH
25949 (scev_probably_wraps_p): Use max loop iterations info
25950 as a global bound first.
25951
6f575fe4 259522013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
3cd232b5
NDD
25953 Nick Clifton <nickc@redhat.com>
25954
25955 * config/v850/v850.md (cbranchsf4): New pattern.
25956 (cstoresf4): New pattern.
25957 (cbranchdf4): New pattern.
25958 (cstoredf4): New pattern.
25959 (movsicc): Disallow floating point comparisons.
25960 (cmpsf_le_insn): Fix order of operators.
25961 (cmpsf_lt_insn): Likewise.
25962 (cmpsf_eq_insn): Likewise.
25963 (cmpdf_le_insn): Likewise.
25964 (cmpdf_lt_insn): Likewise.
25965 (cmpdf_eq_insn): Likewise.
25966 (cmpsf_ge_insn): Use LE comparison.
25967 (cmpdf_ge_insn): Likewise.
25968 (cmpsf_gt_insn): Use LT comparison.
25969 (cmpdf_gt_insn): Likewise.
25970 (cmpsf_ne_insn): Delete pattern.
25971 (cmpdf_ne_insn): Delete pattern.
25972 * config/v850/v850.c (v850_gen_float_compare): Use
25973 gen_cmpdf_eq_insn for NE comparison.
25974 (v850_float_z_comparison_operator)
25975 (v850_float_nz_comparison_operator): Move from here ...
25976 * config/v850/predicates.md: ... to here. Move GT and GE
25977 comparisons into v850_float_z_comparison_operator.
25978 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
25979 Delete prototype.
25980 (v850_float_nz_comparison_operator): Likewise.
25981
f0d54148
JDA
259822013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25983
25984 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
25985 with calls to gen_insvsi/gen_insvdi.
25986
8f01beca
VK
259872013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25988
25989 * config/i386/i386.c (initial_ix86_tune_features): Set up
25990 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
25991
2aa579ad
JJ
259922013-01-09 Steven Bosscher <steven@gcc.gnu.org>
25993 Jakub Jelinek <jakub@redhat.com>
25994
25995 PR tree-optimization/48189
25996 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
25997 If nitercst is 0, don't predict the exit edge.
25998
6edc3e32 259992013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
01a3a324
N
26000
26001 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
26002 in asm_fprintf with reg_names.
26003 (aarch64_print_operand_address): Likewise.
26004 (aarch64_return_addr): Likewise.
26005 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
26006
f6f94d94
JDA
260072013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26008
26009 * config/pa/pa.h (VAL_U6_BITS_P): Define.
26010 (INT_U6_BITS): Likewise.
26011 * config/pa/predicates.md (uint6_operand): New predicate.
26012 (shift5_operand, shift6_operand): Likewise.
26013 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
26014 arith32_operand.
26015 (lshrdi3): Use shift6_operand.
26016 (shrpsi4, shrpdi4): New insn patterns.
26017 (extzv): Delete expander.
26018 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
26019 predicates in unamed zero extract patterns. Tighten common constraint.
26020 (extv): Delete expander.
26021 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
26022 predicates in unamed sign extract patterns. Tighten common constraint.
26023 (insv): Delete expander.
26024 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
26025 predicates in unamed insert patterns. Tighten common constraint.
26026 Change uint32_operand predicate to uint6_operand predicate in unamed
26027 DImode pattern to insert constant values of type 1...1xxxx.
26028
36b72910
JH
260292013-01-04 Jan Hubicka <jh@suse.cz>
26030
26031 PR tree-optimization/55823
7b0fe4f4
UB
26032 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
26033 issue.
36b72910 26034
47876a2a 260352013-01-08 Jakub Jelinek <jakub@redhat.com>
c6d851b9
JJ
26036 Uros Bizjak <ubizjak@gmail.com>
26037
26038 PR rtl-optimization/55845
26039 * df-problems.c (can_move_insns_across): Stop scanning at
26040 volatile_insn_p source instruction or give up if
26041 across_from .. across_to range contains any volatile_insn_p
26042 instructions.
26043
4369c11e
TB
260442013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26045
7b0fe4f4
UB
26046 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26047 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26048 Declare.
4369c11e 26049 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
7b0fe4f4 26050 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
4369c11e 26051
aeb7e7c1
JJ
260522013-01-08 Jakub Jelinek <jakub@redhat.com>
26053
26054 PR fortran/55341
26055 * asan.c (asan_clear_shadow): New function.
26056 (asan_emit_stack_protection): Use it.
26057
a02ad1aa
TB
260582013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26059
26060 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26061 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26062 with tab instead of space.
26063
f90d87f5
NC
260642013-01-08 Nick Clifton <nickc@redhat.com>
26065
26066 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26067 register bank 0 at the start of an interrupt handler.
83ffd964
NC
26068 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26069 MDBH registers.
f90d87f5 26070
385eb93d
JG
260712013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26072
26073 * config/aarch64/aarch64-simd.md
26074 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26075 (aarch64_simd_bsl): Likewise.
26076 (aarch64_vcond_internal<mode>): Likewise.
26077 (vcond<mode><mode>): Likewise.
26078 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26079 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26080
4dcd1054
JG
260812013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26082
26083 * config/aarch64/aarch64-builtins.c
26084 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26085
4267a4a6
MJ
260862013-01-08 Martin Jambor <mjambor@suse.cz>
26087
26088 PR debug/55579
26089 * tree-sra.c (analyze_access_subtree): Return true also after
26090 potentially creating a debug-only replacement.
26091
5f4e6de3
JJ
260922013-01-08 Jakub Jelinek <jakub@redhat.com>
26093
3138f224
JJ
26094 PR middle-end/55890
26095 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26096
5f4e6de3
JJ
26097 PR tree-optimization/54120
26098 * tree-vrp.c (range_fits_type_p): Don't allow
26099 src_precision < precision from signed vr to unsigned_p
26100 if vr->min or vr->max is negative.
26101 (simplify_float_conversion_using_ranges): Test can_float_p
26102 against CODE_FOR_nothing.
26103
f3ef18ff
JJ
261042013-01-08 Jakub Jelinek <jakub@redhat.com>
26105 Richard Biener <rguenther@suse.de>
26106
26107 PR middle-end/55851
26108 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26109 types instead of just INTEGER_TYPE types.
26110
25c210f9
MK
261112013-01-07 Mark Kettenis <kettenis@openbsd.org>
26112
26113 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26114 TF_SIZE): Define.
f90d87f5 26115
81826a7b
SE
261162013-01-07 Steve Ellcey <sellcey@mips.com>
26117
26118 PR target/42661
26119 * config/mips/mips.opt: Change mad to mmad to match documentation.
26120
35678514
GJL
261212013-01-07 Georg-Johann Lay <avr@gjlay.de>
26122
26123 PR target/55897
26124 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26125 .progmemx.data now.
26126
e5669488
GJL
261272013-01-07 Georg-Johann Lay <avr@gjlay.de>
26128
26129 PR target/55897
26130 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26131 (avr_addrspace_t): Add .section_name field.
26132 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26133 array size.
26134 (avr_addrspace): Same. Initialize .section_name. Remove last
26135 NULL entry. Put __memx into .progmemx.data.
26136 (progmem_section_prefix): Remove.
26137 (avr_asm_init_sections): No need to initialize progmem_section.
26138 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26139 section name prefix.
26140 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26141 retrieve the progmem section.
26142 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26143 boundary to run over avr_addrspace[].
26144 (avr_register_target_pragmas): Ditto.
26145
656e6f37
JJ
261462013-01-06 Jakub Jelinek <jakub@redhat.com>
26147
26148 * varasm.c (output_constant_def_contents): For asan_protect_global
26149 protected strings, adjust DECL_ALIGN if needed, before testing for
26150 anchored symbols.
26151 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26152 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26153 normal decls.
26154 (output_object_block): For asan protected decls, emit asan padding
26155 after their contents.
26156 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26157 (asan_finish_file): Test it here instead.
26158
6edc3e32
UB
261592013-01-07 Nick Clifton <nickc@redhat.com>
26160 Matthias Klose <doko@debian.org>
26161 Doug Kwan <dougkwan@google.com>
26162 H.J. Lu <hongjiu.lu@intel.com>
b352afba
NC
26163
26164 PR driver/55470
26165 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26166
26167 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26168
26169 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26170
7b0fe4f4 26171 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
b352afba
NC
26172
26173 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26174
2a095093
GJL
261752013-01-07 Georg-Johann Lay <avr@gjlay.de>
26176
f3b503f4 26177 PR target/54461
2a095093
GJL
26178 * doc/install.texi (Cross-Compiler-Specific Options): Document
26179 --with-avrlibc.
26180
383f9b34
TB
261812013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26182
26183 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26184 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26185 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26186 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26187 vqmovun_high_s64): Fix source operand number and update copyright.
26188
3626621a
RB
261892013-01-07 Richard Biener <rguenther@suse.de>
26190
26191 PR middle-end/55890
26192 * gimple.h (gimple_call_builtin_p): New overload.
26193 * gimple.c (validate_call): New function.
26194 (gimple_call_builtin_p): Likewise.
26195 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26196 Use gimple_call_builtin_p.
26197 (find_func_clobbers): Likewise.
26198 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26199 (strlen_optimize_stmt): Likewise.
26200
8b2ea410
JG
262012013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26202
26203 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26204 (vld1q_dup_*): Likewise.
26205 (vld1_*): Likewise.
26206 (vld1q_*): Likewise.
26207 (vld1_lane_*): Likewise.
26208 (vld1q_lane_*): Likewise.
26209
9713d329
RB
262102013-01-07 Richard Biener <rguenther@suse.de>
26211
26212 * lto-streamer.h (LTO_minor_version): Bump to 2.
26213
3520f7cc
JG
262142013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26215
26216 * config/aarch64/aarch64-protos.h
26217 (aarch64_const_double_zero_rtx_p): Rename to...
26218 (aarch64_float_const_zero_rtx_p): ...this.
26219 (aarch64_float_const_representable_p): New.
26220 (aarch64_output_simd_mov_immediate): Likewise.
26221 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26222 move immediate case.
26223 * config/aarch64/aarch64.c
26224 (aarch64_const_double_zero_rtx_p): Rename to...
26225 (aarch64_float_const_zero_rtx_p): ...this.
26226 (aarch64_print_operand): Allow printing of new constants.
26227 (aarch64_valid_floating_const): New.
26228 (aarch64_legitimate_constant_p): Check for valid floating-point
26229 constants.
26230 (aarch64_simd_valid_immediate): Likewise.
26231 (aarch64_vect_float_const_representable_p): New.
26232 (aarch64_float_const_representable_p): Likewise.
26233 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26234 (aarch64_output_simd_mov_immediate): New.
26235 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26236 (*movdf_aarch64): Likewise.
26237 * config/aarch64/constraints.md (Ufc): New.
26238 (Y): call aarch64_float_const_zero_rtx.
26239 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26240
e167c04d
RB
262412013-01-07 Richard Biener <rguenther@suse.de>
26242
26243 PR tree-optimization/55888
26244 PR tree-optimization/55862
26245 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26246 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26247 not if it is contained therein.
26248
0139aaab
GJL
262492013-01-07 Georg-Johann Lay <avr@gjlay.de>
26250
26251 * config/avr/t-avr: Typo.
26252
4a176b23
GJL
262532013-01-07 Georg-Johann Lay <avr@gjlay.de>
26254
26255 PR55243
26256 * config/avr/t-avr: Don't automatically rebuild
26257 $(srcdir)/config/avr/t-multilib
26258 $(srcdir)/config/avr/avr-tables.opt
26259 $(srcdir)/doc/avr-mmcu.texi
26260 (avr-mcus): New phony target to build them on request.
26261 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26262 * avr/avr-mcus.def: Adjust comments.
26263
c7afdc98
UB
262642013-01-07 Uros Bizjak <ubizjak@gmail.com>
26265
26266 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26267
1ab05c31
RS
262682013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26269
26270 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26271
488e3acc
RS
262722013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26273
26274 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26275
a89599d2
DE
262762013-01-05 David Edelsohn <dje.gcc@gmail.com>
26277
26278 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26279 to generate profiling.
26280 * config/rs6000/aix64.h (LIB_SPEC): Same.
26281
70f09188
AP
262822013-01-04 Andrew Pinski <apinski@cavium.com>
26283
26284 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26285 New function.
26286 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26287
918d445f
UB
262882013-01-04 Uros Bizjak <ubizjak@gmail.com>
26289
26290 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26291 unconditionally.
26292 (ix86_expand_move): Ditto.
26293 (ix86_zero_extend_to_Pmode): Ditto.
26294 (ix86_expand_call): Ditto.
26295 (ix86_expand_special_args_builtin): Ditto.
26296 (ix86_expand_builtin): Ditto.
26297
361618ec
RB
262982013-01-04 Richard Biener <rguenther@suse.de>
26299
26300 PR tree-optimization/55862
26301 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26302 translating them through PHI nodes.
26303
87eab554
MJ
263042013-01-04 Martin Jambor <mjambor@suse.cz>
26305
26306 PR tree-optimization/55755
26307 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26308 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26309
33879b9f
MP
263102013-01-04 Marek Polacek <polacek@redhat.com>
26311
26312 PR middle-end/55859
26313 * opts.c (default_options_optimization): Clarify error message.
26314
3068819a
RB
263152013-01-04 Richard Biener <rguenther@suse.de>
26316
26317 PR middle-end/55863
26318 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26319 reassociation.
26320
0e4ae794
JDA
263212013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26322
26323 PR target/53789
26324 * config/pa/pa.md (movsi): Revert previous change.
e58d2e48
JDA
26325 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26326 references.
0e4ae794 26327
65c88cef
RH
263282013-01-03 Richard Henderson <rth@redhat.com>
26329
26330 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26331 after eliminating TLS symbols.
26332
8ac16127
MG
263332013-01-03 Marc Glisse <marc.glisse@inria.fr>
26334
26335 PR bootstrap/50167
26336 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26337 * graphite-poly.c (debug_gmp_value): Likewise.
26338
bb664f09
UB
263392013-01-03 Uros Bizjak <ubizjak@gmail.com>
26340
26341 PR target/55712
26342 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26343 selected code model, define __code_mode_small__, __code_model_medium__,
26344 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26345 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26346 xchg temporary register with %k. Declare temporary register as
26347 early clobbered.
26348 [__x86_64__]: For medium and large code models, preserve %rbx register.
26349
bcf1ef00
RB
263502013-01-03 Richard Biener <rguenther@suse.de>
26351
0506634a 26352 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
bcf1ef00
RB
26353 (dump_subscript): Adjust.
26354 (finalize_ddr_dependent): Do not dump redundant info.
26355 (analyze_siv_subscript): Adjust.
26356 (subscript_dependence_tester): Likewise.
26357 (compute_affine_dependence): Likewise.
26358
59fd17e3
RB
263592013-01-03 Richard Biener <rguenther@suse.de>
26360
26361 Revert
26362 2013-01-03 Richard Biener <rguenther@suse.de>
26363
26364 PR tree-optimization/55857
26365 * tree-vect-stmts.c (vectorizable_load): Do not setup
26366 re-alignment for invariant loads.
26367
26368 2013-01-02 Richard Biener <rguenther@suse.de>
26369
26370 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26371 invariant load do not generate a vector load from the scalar location.
59fd17e3 26372
595c2679
RB
263732013-01-03 Richard Biener <rguenther@suse.de>
26374
26375 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26376 for not vectorizing.
26377 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26378 not build INDIRECT_REFs, call get_name once only.
26379 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26380 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26381
90bb7d7a
RB
263822013-01-03 Richard Biener <rguenther@suse.de>
26383
26384 PR tree-optimization/55857
26385 * tree-vect-stmts.c (vectorizable_load): Do not setup
26386 re-alignment for invariant loads.
26387
f09b77ca
RB
263882013-01-03 Richard Biener <rguenther@suse.de>
26389
26390 PR lto/55848
26391 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26392 prefer a built-in decl.
26393
3461a16e
JJ
263942013-01-03 Jakub Jelinek <jakub@redhat.com>
26395
df8e2b4f
JJ
26396 * gcc.c (process_command): Update copyright notice dates.
26397 * gcov.c (print_version): Likewise.
26398 * gcov-dump.c (print_version): Likewise.
26399
3461a16e
JJ
26400 PR rtl-optimization/55838
26401 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26402 iv0.step, iv1.step and step.
26403
8b5546d6
JJ
264042013-01-03 Jakub Jelinek <jakub@redhat.com>
26405 Marc Glisse <marc.glisse@inria.fr>
26406
26407 PR tree-optimization/55832
26408 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26409 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26410 integer_{one,zero}_node.
26411
8ab1d2e9
JJ
264122013-01-03 Jakub Jelinek <jakub@redhat.com>
26413
26414 PR debug/54402
26415 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26416 * var-tracking.c (reverse_op): Don't add reverse ops to
26417 VALUEs that have already
0506634a 26418 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
8ab1d2e9 26419
5b9ad1d4
GP
264202013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26421
26422 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26423
e25a6711
TJ
264242013-01-02 Teresa Johnson <tejohnson@google.com>
26425
26426 * dumpfile.c (dump_loc): Print filename with location.
26427 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26428 new location_t parameter to emit complete unroll message with
26429 new dump framework.
26430 (canonicalize_loop_induction_variables): Compute loops location
26431 and pass to try_unroll_loop_completely.
26432 * loop-unroll.c (report_unroll_peel): New function.
26433 (peel_loops_completely): Use new dump format with location
26434 for main dumpfile message, and invoke report_unroll_peel on success.
26435 (decide_unrolling_and_peeling): Ditto.
26436 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26437 by report_unroll_peel.
26438 (decide_peel_completely): Ditto.
26439 (decide_unroll_constant_iterations): Ditto.
26440 (decide_unroll_runtime_iterations): Ditto.
26441 (decide_peel_simple): Ditto.
26442 (decide_unroll_stupid): Ditto.
26443 * cfgloop.c (get_loop_location): New function.
26444 * cfgloop.h (get_loop_location): Declare.
26445
77878621
ST
264462013-01-02 Sriraman Tallam <tmsriram@google.com>
26447
26448 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26449 NULL.
26450
9e65d03e
JDA
264512013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26452
26453 PR middle-end/55198
26454 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26455 BLKmode objects when EXPAND_MEMORY is specified.
26456
6a7da30f
ST
264572013-01-02 Sriraman Tallam <tmsriram@google.com>
26458
26459 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26460 in loop predicate.
26461 (fold_builtin_cpu): Do not share cpu model decls across statements.
26462
e78167f2
JM
264632013-01-02 Jason Merrill <jason@redhat.com>
26464
26465 PR c++/55804
26466 * tree.c (build_array_type_1): Revert earlier change.
26467
8c075fb4
YZ
264682013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26469
26470 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26471 "cortex-a57".
26472 * config/aarch64/aarch64-tune.md: Re-generate.
26473
0682ed3e
RB
264742013-01-02 Richard Biener <rguenther@suse.de>
26475
26476 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
0506634a 26477 invariant load do not generate a vector load from the scalar location.
0682ed3e 26478
b8f6a302
RB
264792013-01-02 Richard Biener <rguenther@suse.de>
26480
26481 PR bootstrap/55784
26482 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26483 * configure: Regenerate.
26484
04b80dbb
RS
264852013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26486
26487 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26488 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26489 (expand_builtin_int_roundingfn_2): Keep the original target around
26490 for the fallback case.
26491
635b0b0c
RS
264922013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26493
26494 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26495 to be clear for sign changes.
26496
42d57399
JH
264972013-01-01 Jan Hubicka <jh@suse.cz>
26498
26499 * ipa-inline-analysis.c: Fix formatting.
26500
5bb6669d
JJ
265012013-01-01 Jakub Jelinek <jakub@redhat.com>
26502
26503 PR tree-optimization/55831
26504 * tree-vect-loop.c (get_initial_def_for_induction): Use
26505 gsi_after_labels instead of gsi_start_bb.
ad41bd84 26506\f
86a2db33 26507Copyright (C) 2013 Free Software Foundation, Inc.
ad41bd84
JM
26508
26509Copying and distribution of this file, with or without modification,
26510are permitted in any medium without royalty provided the copyright
26511notice and this notice are preserved.