]> git.ipfire.org Git - thirdparty/gcc.git/blob - gcc/ChangeLog
Move Martin's most recent entry into the right ChangeLog file
[thirdparty/gcc.git] / gcc / ChangeLog
1 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
2
3 * config/arm/neon-schedgen.ml: Remove.
4 * config/arm/cortex-a9-neon.md: Remove comment regarding
5 neon-schedgen.ml.
6
7 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
8
9 * config/arm/types: Remove old neon types.
10
11 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
12
13 * config/arm/cortex-a7.md
14 (cortex_a7_neon_type): New.
15 (cortex_a7_neon_mul): Update for new types.
16 (cortex_a7_neon_mla): Likewise.
17 (cortex_a7_neon): Likewise.
18
19 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
20
21 * config/arm/cortex-a15-neon.md
22 (cortex_a15_neon_type): New,
23
24 (cortex_a15_neon_int_1): Remove.
25 (cortex_a15_neon_int_2): Likewise.
26 (cortex_a15_neon_int_3): Likewise.
27 (cortex_a15_neon_int_4): Likewise.
28 (cortex_a15_neon_int_5): Likewise.
29 (cortex_a15_neon_vqneg_vqabs): Likewise.
30 (cortex_a15_neon_vmov): Likewise.
31 (cortex_a15_neon_vaba): Likewise.
32 (cortex_a15_neon_vaba_qqq): Likewise.
33 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
34 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
35 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
36 Likewise.
37 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
38 (cortex_a15_neon_mla_qqq_8_16): Likewise.
39 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar): Likewise.
40 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
41 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
42 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
43 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
44 (cortex_a15_neon_shift_1): Likewise.
45 (cortex_a15_neon_shift_2): Likewise.
46 (cortex_a15_neon_shift_3): Likewise.
47 (cortex_a15_neon_vshl_ddd): Likewise.
48 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
49 (cortex_a15_neon_vsra_vrsra): Likewise.
50 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
51 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
52 (cortex_a15_neon_bp_3cycle): Likewise.
53 (cortex_a15_neon_ldm_2): Likewise.
54 (cortex_a15_neon_stm_2): Likewise.
55 (cortex_a15_neon_mcr): Likewise.
56 (cortex_a15_neon_mrc): Likewise.
57 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
58 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
59 (cortex_a15_neon_fp_vmul_ddd): Likewise.
60 (cortex_a15_neon_fp_vmul_qqd): Likewise.
61 (cortex_a15_neon_fp_vmla_ddd): Likewise.
62 (cortex_a15_neon_fp_vmla_qqq): Likewise.
63 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
64 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
65 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
66 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
67 (cortex_a15_neon_bp_simple): Likewise.
68 (cortex_a15_neon_bp_2cycle): Likewise.
69 (cortex_a15_neon_bp_3cycle): Likewise.
70 (cortex_a15_neon_vld1_1_2_regs): Likewise.
71 (cortex_a15_neon_vld1_3_4_regs): Likewise.
72 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
73 (cortex_a15_neon_vld2_4_regs): Likewise.
74 (cortex_a15_neon_vld3_vld4): Likewise.
75 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
76 (cortex_a15_neon_vst1_3_4_regs): Likewise.
77 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Rename to...
78 (cortex_a15_neon_vst2_4_regs_vst3): ...This, update for new attributes.
79 (cortex_a15_neon_vst3_vst4): Rename to...
80 (cortex_a15_neon_vst4): This, update for new attributes.
81 (cortex_a15_neon_vld1_vld2_lane): Update for new attributes.
82 (cortex_a15_neon_vld3_vld4_lane): Likewise.
83 (cortex_a15_neon_vst1_vst2_lane): Likewise.
84 (cortex_a15_neon_vst3_vst4_lane): Likewise.
85 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
86 (cortex_a15_neon_ldm_2): Likewise.
87 (cortex_a15_neon_stm_2): Likewise.
88 (cortex_a15_neon_mcr): Likewise.
89 (cortex_a15_neon_mcr_2_mcrr): Likewise.
90 (cortex_a15_neon_mrc): Likewise.
91 (cortex_a15_neon_mrrc): Likewise.
92
93 (cortex_a15_neon_abd): New.
94 (cortex_a15_neon_abd_q): Likewise.
95 (cortex_a15_neon_aba): Likewise.
96 (cortex_a15_neon_aba_q): Likewise.
97 (cortex_a15_neon_acc): Likewise.
98 (cortex_a15_neon_acc_q): Likewise.
99 (cortex_a15_neon_arith_basic): Likewise.
100 (cortex_a15_neon_arith_complex): Likewise.
101 (cortex_a15_neon_multiply): Likewise.
102 (cortex_a15_neon_multiply_q): Likewise.
103 (cortex_a15_neon_mla): Likewise.
104 (cortex_a15_neon_mla_q): Likewise.
105 (cortex_a15_neon_sat_mla_long): Likewise.
106 (cortex_a15_neon_shift_acc): Likewise.
107 (cortex_a15_neon_shift_imm_basic): Likewise.
108 (cortex_a15_neon_shift_imm_complex): Likewise.
109 (cortex_a15_neon_shift_reg_basic): Likewise.
110 (cortex_a15_neon_shift_reg_basic_q): Likewise.
111 (cortex_a15_neon_shift_reg_complex): Likewise.
112 (cortex_a15_neon_shift_reg_complex_q): Likewise.
113 (cortex_a15_neon_fp_negabs): Likewise
114 (cortex_a15_neon_fp_arith): Likewise
115 (cortex_a15_neon_fp_arith_q): Likewise
116 (cortex_a15_neon_fp_cvt_int): Likewise
117 (cortex_a15_neon_fp_cvt_int_q): Likewise
118 (cortex_a15_neon_fp_cvt_16): Likewise
119 (cortex_a15_neon_fp_mul): Likewise
120 (cortex_a15_neon_fp_mul_q): Likewise
121 (cortex_a15_neon_fp_mla): Likewise
122 (cortex_a15_neon_fp_mla_q): Likewise
123 (cortex_a15_neon_fp_recps_rsqrte): Likewise.
124 (cortex_a15_neon_fp_recps_rsqrte_q): Likewise.
125 (cortex_a15_neon_bitops): Likewise.
126 (cortex_a15_neon_bitops_q): Likewise.
127 (cortex_a15_neon_from_gp): Likewise.
128 (cortex_a15_neon_from_gp_q): Likewise.
129 (cortex_a15_neon_tbl3_tbl4): Likewise.
130 (cortex_a15_neon_zip_q): Likewise.
131 (cortex_a15_neon_to_gp): Likewise.
132 (cortex_a15_neon_load_a): Likewise.
133 (cortex_a15_neon_load_b): Likewise.
134 (cortex_a15_neon_load_c): Likewise.
135 (cortex_a15_neon_load_d): Likewise.
136 (cortex_a15_neon_load_e): Likewise.
137 (cortex_a15_neon_load_f): Likewise.
138 (cortex_a15_neon_store_a): Likewise.
139 (cortex_a15_neon_store_b): Likewise.
140 (cortex_a15_neon_store_c): Likewise.
141 (cortex_a15_neon_store_d): Likewise.
142 (cortex_a15_neon_store_e): Likewise.
143 (cortex_a15_neon_store_f): Likewise.
144 (cortex_a15_neon_store_g): Likewise.
145 (cortex_a15_neon_store_h): Likewise.
146 (cortex_a15_vfp_to_from_gp): Likewise.
147
148 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
149
150 * config/arm/cortex-a9-neon.md (cortex_a9_neon_type): New.
151
152 (cortex_a9_neon_vshl_ddd): Remove.
153 (cortex_a9_neon_vst3_vst4): Likewise.
154 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
155
156 (cortex_a9_neon_bit_ops_q): New.
157
158 (cortex_a9_neon_int_1): Use cortex_a8_neon_type.
159 (cortex_a9_neon_int_2): Likewise.
160 (cortex_a9_neon_int_3): Likewise.
161 (cortex_a9_neon_int_4): Likewise.
162 (cortex_a9_neon_int_5): Likewise.
163 (cortex_a9_neon_vqneg_vqabs): Likewise.
164 (cortex_a9_neon_vmov): Likewise.
165 (cortex_a9_neon_vaba): Likewise.
166 (cortex_a9_neon_vaba_qqq): Likewise.
167 (cortex_a9_neon_shift_1): Likewise.
168 (cortex_a9_neon_shift_2): Likewise.
169 (cortex_a9_neon_shift_3): Likewise.
170 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
171 (cortex_a9_neon_vsra_vrsra): Likewise.
172 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
173 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
174 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
175 Likewise.
176 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
177 (cortex_a9_neon_mla_qqq_8_16): Likewise.
178 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
179 Likewise.
180 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
181 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
182 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
183 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
184 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
185 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
186 (cortex_a9_neon_fp_vsum): Likewise.
187 (cortex_a9_neon_fp_vmul_ddd): Likewise.
188 (cortex_a9_neon_fp_vmul_qqd): Likewise.
189 (cortex_a9_neon_fp_vmla_ddd): Likewise.
190 (cortex_a9_neon_fp_vmla_qqq): Likewise.
191 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
192 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
193 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
194 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
195 (cortex_a9_neon_bp_simple): Likewise.
196 (cortex_a9_neon_bp_2cycle): Likewise.
197 (cortex_a9_neon_bp_3cycle): Likewise.
198 (cortex_a9_neon_ldr): Likewise.
199 (cortex_a9_neon_str): Likewise.
200 (cortex_a9_neon_vld1_1_2_regs): Likewise.
201 (cortex_a9_neon_vld1_3_4_regs): Likewise.
202 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
203 (cortex_a9_neon_vld2_4_regs): Likewise.
204 (cortex_a9_neon_vld3_vld4): Likewise.
205 (cortex_a9_neon_vld1_vld2_lane): Likewise.
206 (cortex_a9_neon_vld3_vld4_lane): Likewise.
207 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
208 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
209 (cortex_a9_neon_vst1_3_4_regs): Likewise.
210 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
211 (cortex_a9_neon_vst1_vst2_lane): Likewise.
212 (cortex_a9_neon_vst3_vst4_lane): Likewise.
213 (cortex_a9_neon_mcr): Likewise.
214 (cortex_a9_neon_mcr_2_mcrr): Likewise.
215 (cortex_a9_neon_mrc): Likewise.
216 (cortex_a9_neon_mrrc): Likewise.
217
218 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
219
220 * config/arm/cortex-a8-neon.md (cortex_a8_neon_type): New.
221
222 (cortex_a8_neon_vshl_ddd): Remove.
223 (cortex_a8_neon_vst3_vst4): Likewise.
224 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
225
226 (cortex_a8_neon_bit_ops_q): New.
227
228 (cortex_a8_neon_int_1): Use cortex_a8_neon_type.
229 (cortex_a8_neon_int_2): Likewise..
230 (cortex_a8_neon_int_3): Likewise.
231 (cortex_a8_neon_int_5): Likewise.
232 (cortex_a8_neon_vqneg_vqabs): Likewise.
233 (cortex_a8_neon_int_4): Likewise.
234 (cortex_a8_neon_vaba): Likewise.
235 (cortex_a8_neon_vaba_qqq): Likewise.
236 (cortex_a8_neon_shift_1): Likewise.
237 (cortex_a8_neon_shift_2): Likewise.
238 (cortex_a8_neon_shift_3): Likewise.
239 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
240 (cortex_a8_neon_vsra_vrsra): Likewise.
241 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
242 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
243 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_long_scalar):
244 Likewise.
245 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
246 (cortex_a8_neon_mla_qqq_8_16): Likewise.
247 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_long_scalar_qdd_64_32_long):
248 Likewise.
249 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
250 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
251 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
252 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
253 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
254 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
255 (cortex_a8_neon_fp_vsum): Likewise.
256 (cortex_a8_neon_fp_vmul_ddd): Likewise.
257 (cortex_a8_neon_fp_vmul_qqd): Likewise.
258 (cortex_a8_neon_fp_vmla_ddd): Likewise.
259 (cortex_a8_neon_fp_vmla_qqq): Likewise.
260 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
261 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
262 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
263 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
264 (cortex_a8_neon_bp_simple): Likewise.
265 (cortex_a8_neon_bp_2cycle): Likewise.
266 (cortex_a8_neon_bp_3cycle): Likewise.
267 (cortex_a8_neon_ldr): Likewise.
268 (cortex_a8_neon_str): Likewise.
269 (cortex_a8_neon_vld1_1_2_regs): Likewise.
270 (cortex_a8_neon_vld1_3_4_regs): Likewise.
271 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
272 (cortex_a8_neon_vld2_4_regs): Likewise.
273 (cortex_a8_neon_vld3_vld4): Likewise.
274 (cortex_a8_neon_vld1_vld2_lane): Likewise.
275 (cortex_a8_neon_vld3_vld4_lane): Likewise.
276 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
277 (cortex_a8_neon_vst1_3_4_regs): Likewise.
278 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
279 (cortex_a8_neon_vst1_vst2_lane): Likewise.
280 (cortex_a8_neon_vst3_vst4_lane): Likewise.
281 (cortex_a8_neon_mcr): Likewise.
282 (cortex_a8_neon_mcr_2_mcrr): Likewise.
283 (cortex_a8_neon_mrc): Likewise.
284 (cortex_a8_neon_mrrc): Likewise.
285
286 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
287
288 * config/aarch64/iterators.md (Vetype): Add SF and DF modes.
289 (fp): New.
290 * config/aarch64/aarch64-simd.md (neon_type): Remove.
291 (aarch64_simd_dup<mode>): Add "type" attribute.
292 (aarch64_dup_lane<mode>): Likewise.
293 (aarch64_dup_lane_<vswap_width_name><mode>): Likewise.
294 (*aarch64_simd_mov<mode>): Likewise.
295 (aarch64_simd_mov_from_<mode>low): Likewise.
296 (aarch64_simd_mov_from_<mode>high): Likewise.
297 (orn<mode>3): Likewise.
298 (bic<mode>3): Likewise.
299 (add<mode>3): Likewise.
300 (sub<mode>3): Likewise.
301 (mul<mode>3): Likewise.
302 (*aarch64_mul3_elt<mode>): Likewise.
303 (*aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
304 (*aarch64_mul3_elt_to_128df): Likewise.
305 (*aarch64_mul3_elt_to_64v2df): Likewise.
306 (neg<mode>2): Likewise.
307 (abs<mode>2): Likewise.
308 (abd<mode>_3): Likewise.
309 (aba<mode>_3): Likewise.
310 (fabd<mode>_3): Likewise.
311 (*fabd_scalar<mode>3): Likewise.
312 (and<mode>3): Likewise.
313 (ior<mode>3): Likewise.
314 (xor<mode>3): Likewise.
315 (one_cmpl<mode>2): Likewise.
316 (aarch64_simd_vec_set<mode>): Likewise.
317 (aarch64_simd_lshr<mode>): Likewise.
318 (aarch64_simd_ashr<mode>): Likewise.
319 (aarch64_simd_imm_shl<mode>): Likewise.
320 (aarch64_simd_reg_sshl<mode): Likewise.
321 (aarch64_simd_reg_shl<mode>_unsigned): Likewise.
322 (aarch64_simd_reg_shl<mode>_signed): Likewise.
323 (aarch64_simd_vec_setv2di): Likewise.
324 (aarch64_simd_vec_set<mode>): Likewise.
325 (aarch64_mla<mode>): Likewise.
326 (*aarch64_mla_elt<mode>): Likewise.
327 (*aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
328 (aarch64_mls<mode>): Likewise.
329 (*aarch64_mls_elt<mode>): Likewise.
330 (*aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
331 (<su><maxmin><mode>3): Likewise.
332 (move_lo_quad_<mode>): Likewise.
333 (aarch64_simd_move_hi_quad_<mode>): Likewise.
334 (aarch64_simd_vec_pack_trunc_<mode>): Likewise.
335 (vec_pack_trunc_<mode>): Likewise.
336 (aarch64_simd_vec_unpack<su>_lo_<mode>): Likewise.
337 (aarch64_simd_vec_unpack<su>_hi_<mode>): Likewise.
338 (*aarch64_<su>mlal_lo<mode>): Likewise.
339 (*aarch64_<su>mlal_hi<mode>): Likewise.
340 (*aarch64_<su>mlsl_lo<mode>): Likewise.
341 (*aarch64_<su>mlsl_hi<mode>): Likewise.
342 (*aarch64_<su>mlal<mode>): Likewise.
343 (*aarch64_<su>mlsl<mode>): Likewise.
344 (aarch64_simd_vec_<su>mult_lo_<mode>): Likewise.
345 (aarch64_simd_vec_<su>mult_hi_<mode>): Likewise.
346 (add<mode>3): Likewise.
347 (sub<mode>3): Likewise.
348 (mul<mode>3): Likewise.
349 (div<mode>3): Likewise.
350 (neg<mode>2): Likewise.
351 (abs<mode>2): Likewise.
352 (fma<mode>4): Likewise.
353 (*aarch64_fma4_elt<mode>): Likewise.
354 (*aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
355 (*aarch64_fma4_elt_to_128df): Likewise.
356 (*aarch64_fma4_elt_to_64v2df): Likewise.
357 (fnma<mode>4): Likewise.
358 (*aarch64_fnma4_elt<mode>): Likewise.
359 (*aarch64_fnma4_elt_<vswap_width_name><mode>
360 (*aarch64_fnma4_elt_to_128df): Likewise.
361 (*aarch64_fnma4_elt_to_64v2df): Likewise.
362 (<frint_pattern><mode>2): Likewise.
363 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): Likewise.
364 (<optab><fcvt_target><VDQF:VDQF:mode>2): Likewise.
365 (vec_unpacks_lo_v4sf): Likewise.
366 (aarch64_float_extend_lo_v2df): Likewise.
367 (vec_unpacks_hi_v4sf): Likewise.
368 (aarch64_float_truncate_lo_v2sf): Likewise.
369 (aarch64_float_truncate_hi_v4sf): Likewise.
370 (aarch64_vmls<mode>): Likewise.
371 (<su><maxmin><mode>3): Likewise.
372 (<maxmin_uns><mode>3): Likewise.
373 (reduc_<sur>plus_<mode>): Likewise.
374 (reduc_<sur>plus_v2di): Likewise.
375 (reduc_<sur>plus_v2si): Likewise.
376 (reduc_<sur>plus_<mode>): Likewise.
377 (aarch64_addpv4sf): Likewise.
378 (clz<mode>2): Likewise.
379 (reduc_<maxmin_uns>_<mode>): Likewise.
380 (reduc_<maxmin_uns>_v2di): Likewise.
381 (reduc_<maxmin_uns>_v2si): Likewise.
382 (reduc_<maxmin_uns>_<mode>): Likewise.
383 (reduc_<maxmin_uns>_v4sf): Likewise.
384 (aarch64_simd_bsl<mode>_internal): Likewise.
385 (*aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): Likewise.
386 (*aarch64_get_lane_zero_extendsi<mode>): Likewise.
387 (aarch64_get_lane<mode>): Likewise.
388 (*aarch64_combinez<mode>): Likewise.
389 (aarch64_combine<mode>): Likewise.
390 (aarch64_simd_combine<mode>): Likewise.
391 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): Likewise.
392 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): Likewise.
393 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>): Likewise.
394 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w<mode>): Likewise.
395 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>w2<mode>_internal): Likewise.
396 (aarch64_<sur>h<addsub><mode>): Likewise.
397 (aarch64_<sur><addsub>hn<mode>): Likewise.
398 (aarch64_<sur><addsub>hn2<mode>): Likewise.
399 (aarch64_pmul<mode>): Likewise.
400 (aarch64_<su_optab><optab><mode>): Likewise.
401 (aarch64_<sur>qadd<mode>): Likewise.
402 (aarch64_sqmovun<mode>): Likewise.
403 (aarch64_<sur>qmovn<mode>): Likewise.
404 (aarch64_s<optab><mode>): Likewise.
405 (aarch64_sq<r>dmulh<mode>): Likewise.
406 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
407 (aarch64_sq<r>dmulh_laneq<mode>): Likewise.
408 (aarch64_sq<r>dmulh_lane<mode>): Likewise.
409 (aarch64_sqdml<SBINQOPS:as>l<mode>): Likewise.
410 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
411 (aarch64_sqdml<SBINQOPS:as>l_lane<mode>_internal): Likewise.
412 (aarch64_sqdml<SBINQOPS:as>l_n<mode>): Likewise.
413 (aarch64_sqdml<SBINQOPS:as>l2<mode>_internal): Likewise.
414 (aarch64_sqdml<SBINQOPS:as>l2_lane<mode>_internal): Likewise.
415 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
416 (aarch64_sqdmull<mode>): Likewise.
417 (aarch64_sqdmull_lane<mode>_internal): Likewise.
418 (aarch64_sqdmull_n<mode>): Likewise.
419 (aarch64_sqdmull2<mode>_internal): Likewise.
420 (aarch64_sqdmull2_lane<mode>_internal): Likewise.
421 (aarch64_sqdmull2_n<mode>_internal): Likewise.
422 (aarch64_<sur>shl<mode>): Likewise.
423 (aarch64_<sur>q<r>shl<mode>
424 (aarch64_<sur>shll_n<mode>): Likewise.
425 (aarch64_<sur>shll2_n<mode>): Likewise.
426 (aarch64_<sur>shr_n<mode>): Likewise.
427 (aarch64_<sur>sra_n<mode>): Likewise.
428 (aarch64_<sur>s<lr>i_n<mode>): Likewise.
429 (aarch64_<sur>qshl<u>_n<mode>): Likewise.
430 (aarch64_<sur>q<r>shr<u>n_n<mode>): Likewise.
431 (aarch64_cm<optab><mode>): Likewise.
432 (aarch64_cm<optab>di): Likewise.
433 (aarch64_cm<optab><mode>): Likewise.
434 (aarch64_cm<optab>di): Likewise.
435 (aarch64_cmtst<mode>): Likewise.
436 (aarch64_cmtstdi): Likewise.
437 (aarch64_cm<optab><mode>): Likewise.
438 (*aarch64_fac<optab><mode>): Likewise.
439 (aarch64_addp<mode>): Likewise.
440 (aarch64_addpdi): Likewise.
441 (sqrt<mode>2): Likewise.
442 (vec_load_lanesoi<mode>): Likewise.
443 (vec_store_lanesoi<mode>): Likewise.
444 (vec_load_lanesci<mode>): Likewise.
445 (vec_store_lanesci<mode>): Likewise.
446 (vec_load_lanesxi<mode>): Likewise.
447 (vec_store_lanesxi<mode>): Likewise.
448 (*aarch64_mov<mode>): Likewise.
449 (aarch64_ld2<mode>_dreg): Likewise.
450 (aarch64_ld2<mode>_dreg): Likewise.
451 (aarch64_ld3<mode>_dreg): Likewise.
452 (aarch64_ld3<mode>_dreg): Likewise.
453 (aarch64_ld4<mode>_dreg): Likewise.
454 (aarch64_ld4<mode>_dreg): Likewise.
455 (aarch64_tbl1<mode>): Likewise.
456 (aarch64_tbl2v16qi): Likewise.
457 (aarch64_combinev16qi): Likewise.
458 (aarch64_<PERMUTE:perm_insn><PERMUTE:perm_hilo><mode>): Likewise.
459 (aarch64_st2<mode>_dreg): Likewise.
460 (aarch64_st2<mode>_dreg): Likewise.
461 (aarch64_st3<mode>_dreg): Likewise.
462 (aarch64_st3<mode>_dreg): Likewise.
463 (aarch64_st4<mode>_dreg): Likewise.
464 (aarch64_st4<mode>_dreg): Likewise.
465 (*aarch64_simd_ld1r<mode>): Likewise.
466 (aarch64_frecpe<mode>): Likewise.
467 (aarch64_frecp<FRECP:frecp_suffix><mode>): Likewise.
468 (aarch64_frecps<mode>): Likewise.
469
470 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
471
472 * config/arm/iterators.md (V_elem_ch): New.
473 (q): Likewise.
474 (VQH_type): Likewise.
475 * config/arm/arm.md (is_neon_type): New.
476 (conds): Use is_neon_type.
477 (anddi3_insn): Update type attribute.
478 (xordi3_insn): Likewise.
479 (one_cmpldi2): Likewise.
480 * gcc/config/arm/vfp.md (movhf_vfp_neon): Update type attribute.
481 * gcc/config/arm/neon.md (neon_mov): Update type attribute.
482 (*movmisalign<mode>_neon_store): Likewise.
483 (*movmisalign<mode>_neon_load): Likewise.
484 (vec_set<mode>_internal): Likewise.
485 (vec_set<mode>_internal): Likewise.
486 (vec_setv2di_internal): Likewise.
487 (vec_extract<mode>): Likewise.
488 (vec_extract<mode>): Likewise.
489 (vec_extractv2di): Likewise.
490 (*add<mode>3_neon): Likewise.
491 (adddi3_neon): Likewise.
492 (*sub<mode>3_neon): Likewise.
493 (subdi3_neon): Likewise.
494 (fma<VCVTF:mode>4): Likewise.
495 (fma<VCVTF:mode>4_intrinsic): Likewise.
496 (*fmsub<VCVTF:mode>4): Likewise.
497 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
498 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
499 (ior<mode>3): Likewise.
500 (and<mode>3): Likewise.
501 (orn<mode>3_neon): Likewise.
502 (orndi3_neon): Likewise.
503 (bic<mode>3_neon): Likewise.
504 (bicdi3_neon): Likewise.
505 (xor<mode>3): Likewise.
506 (one_cmpl<mode>2): Likewise.
507 (abs<mode>2): Likewise.
508 (neg<mode>2): Likewise.
509 (negdi2_neon): Likewise.
510 (*umin<mode>3_neon): Likewise.
511 (*umax<mode>3_neon): Likewise.
512 (*smin<mode>3_neon): Likewise.
513 (*smax<mode>3_neon): Likewise.
514 (vashl<mode>3): Likewise.
515 (vashr<mode>3_imm): Likewise.
516 (vlshr<mode>3_imm): Likewise.
517 (ashl<mode>3_signed): Likewise.
518 (ashl<mode>3_unsigned): Likewise.
519 (neon_load_count): Likewise.
520 (ashldi3_neon_noclobber): Likewise.
521 (ashldi3_neon): Likewise.
522 (signed_shift_di3_neon): Likewise.
523 (unsigned_shift_di3_neon): Likewise.
524 (ashrdi3_neon_imm_noclobber): Likewise.
525 (lshrdi3_neon_imm_noclobber): Likewise.
526 (<shift>di3_neon): Likewise.
527 (widen_ssum<mode>3): Likewise.
528 (widen_usum<mode>3): Likewise.
529 (quad_halves_<code>v4si): Likewise.
530 (quad_halves_<code>v4sf): Likewise.
531 (quad_halves_<code>v8hi): Likewise.
532 (quad_halves_<code>v16qi): Likewise.
533 (reduc_splus_v2di): Likewise.
534 (neon_vpadd_internal<mode>): Likewise.
535 (neon_vpsmin<mode>): Likewise.
536 (neon_vpsmax<mode>): Likewise.
537 (neon_vpumin<mode>): Likewise.
538 (neon_vpumax<mode>): Likewise.
539 (*ss_add<mode>_neon): Likewise.
540 (*us_add<mode>_neon): Likewise.
541 (*ss_sub<mode>_neon): Likewise.
542 (*us_sub<mode>_neon): Likewise.
543 (neon_vadd<mode>_unspec): Likewise.
544 (neon_vaddl<mode>): Likewise.
545 (neon_vaddw<mode>): Likewise.
546 (neon_vhadd<mode>): Likewise.
547 (neon_vqadd<mode>): Likewise.
548 (neon_vaddhn<mode>): Likewise.
549 (neon_vmul<mode>): Likewise.
550 (neon_vfms<VCVTF:mode>): Likewise.
551 (neon_vmlal<mode>): Likewise.
552 (neon_vmls<mode>): Likewise.
553 (neon_vmlsl<mode>): Likewise.
554 (neon_vqdmulh<mode>): Likewise.
555 (neon_vqdmlal<mode>): Likewise.
556 (neon_vqdmlsl<mode>): Likewise.
557 (neon_vmull<mode>): Likewise.
558 (neon_vqdmull<mode>): Likewise.
559 (neon_vsub<mode>_unspec): Likewise.
560 (neon_vsubl<mode>): Likewise.
561 (neon_vsubw<mode>): Likewise.
562 (neon_vqsub<mode>): Likewise.
563 (neon_vhsub<mode>): Likewise.
564 (neon_vsubhn<mode>): Likewise.
565 (neon_vceq<mode>): Likewise.
566 (neon_vcge<mode>): Likewise.
567 (neon_vcgeu<mode>): Likewise.
568 (neon_vcgt<mode>): Likewise.
569 (neon_vcgtu<mode>): Likewise.
570 (neon_vcle<mode>): Likewise.
571 (neon_vclt<mode>): Likewise.
572 (neon_vcage<mode>): Likewise.
573 (neon_vcagt<mode>): Likewise.
574 (neon_vtst<mode>): Likewise.
575 (neon_vabd<mode>): Likewise.
576 (neon_vabdl<mode>): Likewise.
577 (neon_vaba<mode>): Likewise.
578 (neon_vabal<mode>): Likewise.
579 (neon_vmax<mode>): Likewise.
580 (neon_vmin<mode>): Likewise.
581 (neon_vpaddl<mode>): Likewise.
582 (neon_vpadal<mode>): Likewise.
583 (neon_vpmax<mode>): Likewise.
584 (neon_vpmin<mode>): Likewise.
585 (neon_vrecps<mode>): Likewise.
586 (neon_vrsqrts<mode>): Likewise.
587 (neon_vqabs<mode>): Likewise.
588 (neon_vqneg<mode>): Likewise.
589 (neon_vcls<mode>): Likewise.
590 (clz<mode>2): Likewise.
591 (popcount<mode>2): Likewise.
592 (neon_vrecpe<mode>): Likewise.
593 (neon_vrsqrte<mode>): Likewise.
594 (neon_vget_lane<mode>_sext_internal): Likewise.
595 (neon_vget_lane<mode>_zext_internal): Likewise.
596 (neon_vdup_n<mode>): Likewise.
597 (neon_vdup_n<mode>): Likewise.
598 (neon_vdup_nv2di): Likewise.
599 (neon_vdup_lane<mode>_interal): Likewise.
600 (*neon_vswp<mode>): Likewise.
601 (neon_vcombine<mode>): Likewise.
602 (float<mode><V_cvtto>2): Likewise.
603 (floatuns<mode><V_cvtto>2): Likewise.
604 (fix_trunc<mode><V_cvtto>2): Likewise.
605 (fixuns_trunc<mode><V_cvtto>2
606 (neon_vcvt<mode>): Likewise.
607 (neon_vcvt<mode>): Likewise.
608 (neon_vcvtv4sfv4hf): Likewise.
609 (neon_vcvtv4hfv4sf): Likewise.
610 (neon_vcvt_n<mode>): Likewise.
611 (neon_vcvt_n<mode>): Likewise.
612 (neon_vmovn<mode>): Likewise.
613 (neon_vqmovn<mode>): Likewise.
614 (neon_vqmovun<mode>): Likewise.
615 (neon_vmovl<mode>): Likewise.
616 (neon_vmul_lane<mode>): Likewise.
617 (neon_vmul_lane<mode>): Likewise.
618 (neon_vmull_lane<mode>): Likewise.
619 (neon_vqdmull_lane<mode>): Likewise.
620 (neon_vqdmulh_lane<mode>): Likewise.
621 (neon_vqdmulh_lane<mode>): Likewise.
622 (neon_vmla_lane<mode>): Likewise.
623 (neon_vmla_lane<mode>): Likewise.
624 (neon_vmlal_lane<mode>): Likewise.
625 (neon_vqdmlal_lane<mode>): Likewise.
626 (neon_vmls_lane<mode>): Likewise.
627 (neon_vmls_lane<mode>): Likewise.
628 (neon_vmlsl_lane<mode>): Likewise.
629 (neon_vqdmlsl_lane<mode>): Likewise.
630 (neon_vext<mode>): Likewise.
631 (neon_vrev64<mode>): Likewise.
632 (neon_vrev32<mode>): Likewise.
633 (neon_vrev16<mode>): Likewise.
634 (neon_vbsl<mode>_internal): Likewise.
635 (neon_vshl<mode>): Likewise.
636 (neon_vqshl<mode>): Likewise.
637 (neon_vshr_n<mode>): Likewise.
638 (neon_vshrn_n<mode>): Likewise.
639 (neon_vqshrn_n<mode>): Likewise.
640 (neon_vqshrun_n<mode>): Likewise.
641 (neon_vshl_n<mode>): Likewise.
642 (neon_vqshl_n<mode>): Likewise.
643 (neon_vqshlu_n<mode>): Likewise.
644 (neon_vshll_n<mode>): Likewise.
645 (neon_vsra_n<mode>): Likewise.
646 (neon_vsri_n<mode>): Likewise.
647 (neon_vsli_n<mode>): Likewise.
648 (neon_vtbl1v8qi): Likewise.
649 (neon_vtbl2v8qi): Likewise.
650 (neon_vtbl3v8qi): Likewise.
651 (neon_vtbl4v8qi): Likewise.
652 (neon_vtbl1v16qi): Likewise.
653 (neon_vtbl2v16qi): Likewise.
654 (neon_vcombinev16qi): Likewise.
655 (neon_vtbx1v8qi): Likewise.
656 (neon_vtbx2v8qi): Likewise.
657 (neon_vtbx3v8qi): Likewise.
658 (neon_vtbx4v8qi): Likewise.
659 (*neon_vtrn<mode>_insn): Likewise.
660 (*neon_vzip<mode>_insn): Likewise.
661 (*neon_vuzp<mode>_insn): Likewise.
662 (neon_vld1<mode>): Likewise.
663 (neon_vld1_lane<mode>): Likewise.
664 (neon_vld1_lane<mode>): Likewise.
665 (neon_vld1_dup<mode>): Likewise.
666 (neon_vld1_dup<mode>): Likewise.
667 (neon_vld1_dupv2di): Likewise.
668 (neon_vst1<mode>): Likewise.
669 (neon_vst1_lane<mode>): Likewise.
670 (neon_vst1_lane<mode>): Likewise.
671 (neon_vld2<mode>): Likewise.
672 (neon_vld2<mode>): Likewise.
673 (neon_vld2_lane<mode>): Likewise.
674 (neon_vld2_lane<mode>): Likewise.
675 (neon_vld2_dup<mode>): Likewise.
676 (neon_vst2<mode>): Likewise.
677 (neon_vst2<mode>): Likewise.
678 (neon_vst2_lane<mode>): Likewise.
679 (neon_vst2_lane<mode>): Likewise.
680 (neon_vld3<mode>): Likewise.
681 (neon_vld3qa<mode>): Likewise.
682 (neon_vld3qb<mode>): Likewise.
683 (neon_vld3_lane<mode>): Likewise.
684 (neon_vld3_lane<mode>): Likewise.
685 (neon_vld3_dup<mode>): Likewise.
686 (neon_vst3<mode>): Likewise.
687 (neon_vst3qa<mode>): Likewise.
688 (neon_vst3qb<mode>): Likewise.
689 (neon_vst3_lane<mode>): Likewise.
690 (neon_vst3_lane<mode>): Likewise.
691 (neon_vld4<mode>): Likewise.
692 (neon_vld4qa<mode>): Likewise.
693 (neon_vld4qb<mode>): Likewise.
694 (neon_vld4_lane<mode>): Likewise.
695 (neon_vld4_lane<mode>): Likewise.
696 (neon_vld4_dup<mode>): Likewise.
697 (neon_vst4<mode>): Likewise.
698 (neon_vst4qa<mode>): Likewise.
699 (neon_vst4qb<mode>): Likewise.
700 (neon_vst4_lane<mode>): Likewise.
701 (neon_vst4_lane<mode>): Likewise.
702 (neon_vec_unpack<US>_lo_<mode>): Likewise.
703 (neon_vec_unpack<US>_hi_<mode>): Likewise.
704 (neon_vec_<US>mult_lo_<mode>): Likewise.
705 (neon_vec_<US>mult_hi_<mode>): Likewise.
706 (neon_vec_<US>shiftl_<mode>): Likewise.
707 (neon_unpack<US>_<mode>): Likewise.
708 (neon_vec_<US>mult_<mode>): Likewise.
709 (vec_pack_trunc_<mode>): Likewise.
710 (neon_vec_pack_trunc_<mode>): Likewise.
711 (neon_vabd<mode>_2): Likewise.
712 (neon_vabd<mode>_3): Likewise.
713
714 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
715
716 * config/aarch64/aarch64.md (movtf_aarch64): Update type attribute.
717 (load_pair): Update type attribute.
718 (store_pair): Update type attribute.
719 * config/aarch64/iterators.md (q): New.
720
721 2013-10-15 James Greenhalgh <james.greenhalgh@arm.com>
722
723 * config/arm/types.md: Add new types for Neon insns.
724
725 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
726 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
727 Sergey Lega <sergey.s.lega@intel.com>
728 Anna Tikhonova <anna.tikhonova@intel.com>
729 Ilya Tocar <ilya.tocar@intel.com>
730 Andrey Turetskiy <andrey.turetskiy@intel.com>
731 Ilya Verbin <ilya.verbin@intel.com>
732 Kirill Yukhin <kirill.yukhin@intel.com>
733 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
734
735 * config/i386/sse.md (unspec): Add UNSPEC_RCP14, UNSPEC_RSQRT14,
736 UNSPEC_FIXUPIMM, UNSPEC_SCALEF, UNSPEC_GETEXP, UNSPEC_GETMANT,
737 UNSPEC_EXP2, UNSPEC_RCP28, UNSPEC_RSQRT28.
738 (rcp14<mode>): New.
739 (srcp14<mode>): Ditto.
740 (rsqrt14<mode>): Ditto.
741 (rsqrt14<mode>): Ditto.
742 (avx512f_vmscalef<mode>): Ditto.
743 (avx512f_scalef<mode>): Ditto.
744 (avx512f_getexp<mode>): Ditto.
745 (avx512f_sgetexp<mode>): Ditto.
746 (avx512f_fixupimm<mode>): Ditto.
747 (avx512f_sfixupimm<mode>): Ditto.
748 (avx512f_rndscale<mode>): Ditto.
749 (*avx512er_exp2<mode>): Ditto.
750 (*avx512er_rcp28<mode>): Ditto.
751 (avx512er_rsqrt28<mode>): Ditto.
752 (avx512f_getmant<mode>): Ditto.
753 (avx512f_getmant<mode>): Ditto.
754 (avx512f_rndscale<mode>): Fix formatting.
755
756 2013-10-15 Martin Jambor <mjambor@suse.cz>
757
758 * ipa-utils.h (ipa_edge_within_scc): Declare.
759 * ipa-cp.c (edge_within_scc): Moved...
760 * ipa-utils.c (ipa_edge_within_scc): ...here. Updated all callers.
761
762 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
763 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
764 Sergey Lega <sergey.s.lega@intel.com>
765 Anna Tikhonova <anna.tikhonova@intel.com>
766 Ilya Tocar <ilya.tocar@intel.com>
767 Andrey Turetskiy <andrey.turetskiy@intel.com>
768 Ilya Verbin <ilya.verbin@intel.com>
769 Kirill Yukhin <kirill.yukhin@intel.com>
770 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
771
772 * config/i386/predicates.md (const_8_to_15_operand): New.
773 (const_16_to_31_operand): Ditto.
774 * config/i386/sse.md (V8FI): New.
775 (V16FI): Ditto.
776 (reduc_splus_v8df): Ditto.
777 (reduc_splus_v16sf): Ditto.
778 (avx512f_vextract<shuffletype>32x4_1): Ditto.
779 (vec_extract_hi_<mode>): Ditto.
780 (avx512f_vinsert<shuffletype>32x4_1): Ditto.
781 (vec_set_lo_<mode>): Ditto.
782 (vec_set_hi_<mode>): Ditto.
783 (avx512f_shuf_<shuffletype>64x2_1): Ditto.
784 (avx512f_shuf_<shuffletype>32x4_1): Ditto.
785 (avx512f_pshufd_1): Ditto.
786 (avx512f_broadcast<mode>): Ditto.
787 (avx512f_broadcast<mode>): Ditto.
788 (define_split): Split vec_extract_lo into move.
789 (ssequartermode): Ditto.
790 (ssedoublemode): Extened with wider modes.
791 (vec_extract_lo_<mode>): Ditto.
792
793 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
794 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
795 Sergey Lega <sergey.s.lega@intel.com>
796 Anna Tikhonova <anna.tikhonova@intel.com>
797 Ilya Tocar <ilya.tocar@intel.com>
798 Andrey Turetskiy <andrey.turetskiy@intel.com>
799 Ilya Verbin <ilya.verbin@intel.com>
800 Kirill Yukhin <kirill.yukhin@intel.com>
801 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
802
803 * config/i386/predicates.md (register_or_constm1_operand): New.
804 * config/i386/sse.md (unspec): Add UNSPEC_UNSIGNED_PCMP, UNSPEC_TESTM,
805 UNSPEC_TESTNM, UNSPEC_VTERNLOG, UNSPEC_ALIGN, UNSPEC_CONFLICT,
806 UNSPEC_MASKED_EQ, UNSPEC_MASKED_GT, UNSPEC_GATHER_PREFETCH,
807 UNSPEC_SCATTER_PREFETCH
808 (VI48_512): New.
809 (avx512f_ucmp<mode>3): Ditto.
810 (avx512f_vternlog<mode>): Ditto.
811 (avx512f_align<mode>): Ditto.
812 (<shift_insn><mode>3): Ditto.
813 (avx512f_<rotate>v<mode>): Ditto.
814 (avx512f_<rotate><mode>): Ditto.
815 (avx512f_eq<mode>3): Ditto.
816 (avx512f_eq<mode>3_1): Ditto.
817 (avx512f_gt<mode>3): Ditto.
818 (avx512f_testm<mode>3): Ditto.
819 (avx512f_testnm<mode>3): Ditto.
820 (avx512pf_gatherpf<mode>): Ditto.
821 (*avx512pf_gatherpf<mode>_mask): Ditto.
822 (*avx512pf_gatherpf<mode>): Ditto.
823 (avx512pf_scatterpf<mode>): Ditto.
824 (*avx512pf_scatterpf<mode>_mask): Ditto.
825 (*avx512pf_scatterpf<mode>): Ditto.
826 (avx512f_vec_dup_gpr<mode>): Ditto.
827 (clz<mode>2): Ditto.
828 (conflict<mode>): Ditto.
829 (REDUC_SMINMAX_MODE): Extened with wider modes.
830 (reduc_<code>_<mode>): Ditto.
831 (vlshr<mode>3): Ditto.
832 (vashl<mode>3): Ditto.
833
834 2013-10-15 Alexander Ivchenko <alexander.ivchenko@intel.com>
835 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
836 Sergey Lega <sergey.s.lega@intel.com>
837 Anna Tikhonova <anna.tikhonova@intel.com>
838 Ilya Tocar <ilya.tocar@intel.com>
839 Andrey Turetskiy <andrey.turetskiy@intel.com>
840 Ilya Verbin <ilya.verbin@intel.com>
841 Kirill Yukhin <kirill.yukhin@intel.com>
842 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
843
844 * config/i386/sse.md (unspec): Added UNSPEC_VPERMI2, UNSPEC_VPERMT2,
845 UNSPEC_SCATTER.
846 (VI48F_512): New.
847 (avx512fmaskmode): Ditto.
848 (bcstscalarsuff): Ditto.
849 (avx512f_blendm<mode>): Ditto.
850 (cmp_imm_predicate): Ditto.
851 (avx512f_cmp<mode>3): Ditto.
852 (avx512f_vec_dup<mode>): Ditto.
853 (avx512f_vec_dup_mem<mode>): Ditto.
854 (avx512f_vpermi2var<mode>3): Ditto.
855 (avx512f_vpermt2var<mode>3): Ditto.
856 (vec_init<mode>): Ditto.
857 (avx512f_gathersi<mode>): Ditto.
858 (*avx512f_gathersi<mode>): Ditto.
859 (*avx512f_gathersi<mode>_2): Ditto.
860 (avx512f_gatherdi<mode>): Ditto.
861 (*avx512f_gatherdi<mode>): Ditto.
862 (*avx512f_gatherdi<mode>_2): Ditto.
863 (avx512f_scattersi<mode>): Ditto.
864 (*avx512f_scattersi<mode>): Ditto.
865 (avx512f_scatterdi<mode>): Ditto.
866 (*avx512f_scatterdi<mode>): Ditto.
867 (sseintprefix): Extened with wider modes.
868 (VEC_GATHER_IDXSI): Ditto.
869 (VEC_GATHER_IDXDI): Ditto.
870 (VEC_GATHER_SRCDI): Ditto.
871
872 2013-10-15 Matthew Gretton-Dann <matthew.gretton-dann@arm.com>
873 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
874
875 * config/arm/t-aprofile: New file.
876 * config.gcc: Handle --with-multilib-list option.
877
878 2013-10-15 Bernd Schmidt <bernds@codesourcery.com>
879
880 * reload1.c (reloads_unique_chain_p): Ensure that r1 is the input for
881 r2.
882
883 2013-10-15 Richard Biener <rguenther@suse.de>
884
885 * tree-loop-distribution.c (build_empty_rdg): Inline into
886 single user.
887 (rdg_flag_vertex): Inline into single user.
888 (rdg_flag_vertex_and_dependent): Likewise.
889 (build_rdg_partition_for_vertex): Remove processed bitmap.
890 (rdg_build_partitions): Simplify.
891
892 2013-10-15 Richard Biener <rguenther@suse.de>
893
894 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1):
895 Restructure forwarding through conversions and copies to
896 avoid performing copy-propagation the wrong way. Adjust
897 recursion invocations.
898 (forward_propagate_addr_expr): Add argument stating if we
899 are recursing from a single-use.
900 (ssa_forward_propagate_and_combine): Adjust.
901
902 2013-10-14 David Malcolm <dmalcolm@redhat.com>
903
904 * dumpfile.h (gcc::dump_manager): New class, to hold state
905 relating to dumpfile management.
906 (get_dump_file_name): Remove in favor of method of dump_manager.
907 (dump_initialized_p): Likewise.
908 (dump_start): Likewise.
909 (dump_finish): Likewise.
910 (dump_switch_p): Likewise.
911 (dump_register): Likewise.
912 (get_dump_file_info): Likewise.
913 * context.c (gcc::context::context): Construct the dump_manager
914 instance.
915 * context.h (gcc::context::get_dumps): New.
916 (gcc::context::m_dumps): New.
917 * coverage.c (coverage_init): Port to dump_manager API.
918 * dumpfile.c (extra_dump_files): Convert to field of
919 gcc::dump_manager.
920 (extra_dump_files_in_use): Likewise.
921 (extra_dump_files_alloced): Likewise.
922 (gcc::dump_manager::dump_manager): New.
923 (dump_register): Convert to...
924 (gcc::dump_manager::dump_register): ...method, replacing
925 function-static next_dump with m_next_dump field.
926 (get_dump_file_info): Convert to...
927 (gcc::dump_manager::get_dump_file_info): ...method.
928 (get_dump_file_name): Convert to...
929 (gcc::dump_manager::get_dump_file_name): ...method.
930 (dump_start): Convert to...
931 (gcc::dump_manager::dump_start): ...method.
932 (dump_finish): Convert to...
933 (gcc::dump_manager::dump_finish): ...method.
934 (dump_begin): Replace body with...
935 (gcc::dump_manager::dump_begin): ...new method.
936 (dump_phase_enabled_p): Convert to...
937 (gcc::dump_manager::dump_phase_enabled_p): ...method.
938 (dump_phase_enabled_p): Convert to...
939 (gcc::dump_manager::dump_phase_enabled_p): ...method.
940 (dump_initialized_p): Convert to...
941 (gcc::dump_manager::dump_initialized_p): ...method.
942 (dump_flag_name): Replace body with...
943 (gcc::dump_manager::dump_flag_name): ...new method.
944 (dump_enable_all): Convert to...
945 (gcc::dump_manager::dump_enable_all): ...new method.
946 (opt_info_enable_passes): Convert to...
947 (gcc::dump_manager::opt_info_enable_passes): ...new method.
948 (dump_switch_p_1): Convert to...
949 (gcc::dump_manager::dump_switch_p_1): ...new method.
950 (dump_switch_p): Convert to...
951 (gcc::dump_manager::dump_switch_p): ...new method.
952 (opt_info_switch_p): Port to dump_manager API.
953 (enable_rtl_dump_file): Likewise.
954 * opts-global.c (handle_common_deferred_options): Port to new
955 dump_manager API.
956 * passes.c (pass_manager::finish_optimization_passes): Likewise.
957 (pass_manager::register_one_dump_file): Likewise.
958 (pass_manager::register_pass): Likewise.
959 (pass_init_dump_file): Likewise.
960 (pass_fini_dump_file): Likewise.
961 * statistics.c (statistics_early_init): Likewise.
962
963 2013-10-14 Richard Biener <rguenther@suse.de>
964
965 * gimple.c (gimple_canonical_types, canonical_type_hash_cache,
966 iterative_hash_canonical_type, gimple_canonical_type_hash,
967 gimple_canonical_types_compatible_p, gimple_canonical_type_eq,
968 gimple_register_canonical_type, print_gimple_types_stats,
969 free_gimple_type_tables): Move to lto/lto.c
970 (gt-gimple.h): Do not include.
971 * gimple.h (gimple_register_canonical_type,
972 print_gimple_types_stats, free_gimple_type_tables): Remove.
973 * Makefile.in (GTFILES): Remove gimple.c.
974
975 2013-10-14 Travis Snoozy <quandary@remstate.com>
976
977 PR target/58716
978 * config/msp430/msp430.c (msp430_option_override): Correct thinko
979 scanning for msp430x targets.
980
981 2013-10-14 Eric Botcazou <ebotcazou@adacore.com>
982
983 PR bootstrap/58509
984 * config/sparc/sparc-protos.h (widen_mem_for_ldd_peep): Declare.
985 (registers_ok_for_ldd_peep): Move around.
986 * config/sparc/sparc.c (widen_mem_for_ldd_peep): New.
987 * config/sparc/sparc.md (widening peepholes): Use it.
988
989 2013-10-14 Richard Biener <rguenther@suse.de>
990
991 PR middle-end/58712
992 PR middle-end/55358
993 * gimple.c (iterative_hash_canonical_type): Make sure to
994 record the hash into the correct hashtable slot.
995
996 2013-10-13 Eric Botcazou <ebotcazou@adacore.com>
997
998 PR rtl-optimization/58662
999 * combine.c (try_combine): Take into account death nodes on I2 when
1000 splitting a PARALLEL of two independent SETs. Fix dump message.
1001
1002 2013-10-12 Oleg Endo <olegendo@gcc.gnu.org>
1003
1004 PR target/51244
1005 * config/sh/sh_treg_combine.cc: New SH specific RTL pass.
1006 * config.gcc (SH extra_objs): Add sh_ifcvt.o.
1007 * config/sh/t-sh (sh_treg_combine.o): New entry.
1008 * config/sh/sh.c (sh_fixed_condition_code_regs): New function that
1009 implements the target hook TARGET_FIXED_CONDITION_CODE_REGS.
1010 (register_sh_passes): New function. Register sh_treg_combine pass.
1011 (sh_option_override): Invoke it.
1012 (sh_canonicalize_comparison): Handle op0_preserve_value.
1013 * sh.md (*cbranch_t"): Do not try to optimize missed test and branch
1014 opportunities. Canonicalize branch condition.
1015 (nott): Allow only if pseudos can be created for non-SH2A.
1016
1017 2013-10-12 H.J. Lu <hongjiu.lu@intel.com>
1018
1019 PR target/58690
1020 * config/i386/i386.c (ix86_copy_addr_to_reg): New function.
1021 (ix86_expand_movmem): Replace copy_addr_to_reg with
1022 ix86_copy_addr_to_reg.
1023 (ix86_expand_setmem): Likewise.
1024
1025 2013-10-12 Alexander Monakov <amonakov@ispras.ru>
1026
1027 * config/i386/i386.c (ix86_expand_sse_compare_and_jump): Use mode
1028 provided by ix86_fp_compare_mode instead of CCFPUmode.
1029
1030 2013-10-12 James Greenhalgh <james.greenhalgh@arm.com>
1031
1032 * config/aarch64/arm_neon.h
1033 (vtbx<1,3>_<psu>8): Fix register constriants.
1034
1035 2013-10-11 Jeff Law <law@redhat.com>
1036
1037 PR tree-optimization/58640
1038 * tree-ssa-threadupdate.c (mark_threaded_blocks): Truncate jump
1039 threading paths that cross over two loop entry points.
1040
1041 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1042
1043 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): Generalize to
1044 handle vector float as well.
1045 (*vsx_le_perm_load_v4si): Likewise.
1046 (*vsx_le_perm_store_v2di): Likewise.
1047 (*vsx_le_perm_store_v4si): Likewise.
1048
1049 2013-10-11 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
1050
1051 * config/rs6000/vector.md (vec_realign_load<mode>): Generate vperm
1052 directly to circumvent subtract from splat{31} workaround.
1053 * config/rs6000/rs6000-protos.h (altivec_expand_vec_perm_le): New
1054 prototype.
1055 * config/rs6000/rs6000.c (altivec_expand_vec_perm_le): New.
1056 * config/rs6000/altivec.md (define_c_enum "unspec"): Add
1057 UNSPEC_VPERM_X and UNSPEC_VPERM_UNS_X.
1058 (altivec_vperm_<mode>): Convert to define_insn_and_split to
1059 separate big and little endian logic.
1060 (*altivec_vperm_<mode>_internal): New define_insn.
1061 (altivec_vperm_<mode>_uns): Convert to define_insn_and_split to
1062 separate big and little endian logic.
1063 (*altivec_vperm_<mode>_uns_internal): New define_insn.
1064 (vec_permv16qi): Add little endian logic.
1065
1066 2013-10-11 Marc Glisse <marc.glisse@inria.fr>
1067
1068 * doc/extend.texi (returns_nonnull): Remove arguments.
1069
1070 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1071 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1072 Sergey Lega <sergey.s.lega@intel.com>
1073 Anna Tikhonova <anna.tikhonova@intel.com>
1074 Ilya Tocar <ilya.tocar@intel.com>
1075 Andrey Turetskiy <andrey.turetskiy@intel.com>
1076 Ilya Verbin <ilya.verbin@intel.com>
1077 Kirill Yukhin <kirill.yukhin@intel.com>
1078 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1079
1080 * config/i386/sse.md (VI48F_256_512): New.
1081 (avx2_permvar<mode>): Change to ...
1082 (<avx2_avx512f>_permvar<mode>): This.
1083
1084 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1085 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1086 Sergey Lega <sergey.s.lega@intel.com>
1087 Anna Tikhonova <anna.tikhonova@intel.com>
1088 Ilya Tocar <ilya.tocar@intel.com>
1089 Andrey Turetskiy <andrey.turetskiy@intel.com>
1090 Ilya Verbin <ilya.verbin@intel.com>
1091 Kirill Yukhin <kirill.yukhin@intel.com>
1092 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1093
1094 * config/i386/i386.c (bdesc_args): Change corresponding pattern for
1095 __builtin_ia32_cvtps2dq, __builtin_ia32_cvtps2dq256.
1096 * config/i386/sse.md (VI4_AVX): New.
1097 (sf2simodelower): Ditto.
1098 (sse2_cvtps2dq): Change to ...
1099 (<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>): This.
1100
1101 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1102 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1103 Sergey Lega <sergey.s.lega@intel.com>
1104 Anna Tikhonova <anna.tikhonova@intel.com>
1105 Ilya Tocar <ilya.tocar@intel.com>
1106 Andrey Turetskiy <andrey.turetskiy@intel.com>
1107 Ilya Verbin <ilya.verbin@intel.com>
1108 Kirill Yukhin <kirill.yukhin@intel.com>
1109 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1110
1111 * config/i386/sse.md (V_512): New.
1112 (VI_512): Ditto.
1113 (vcond<V_512:mode><VF_512:mode>): Ditto.
1114 (vcond<V_512:mode><VI_512:mode>): Ditto.
1115 (vcondu<V_512:mode><VI_512:mode>): Ditto.
1116
1117 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1118 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1119 Sergey Lega <sergey.s.lega@intel.com>
1120 Anna Tikhonova <anna.tikhonova@intel.com>
1121 Ilya Tocar <ilya.tocar@intel.com>
1122 Andrey Turetskiy <andrey.turetskiy@intel.com>
1123 Ilya Verbin <ilya.verbin@intel.com>
1124 Kirill Yukhin <kirill.yukhin@intel.com>
1125 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1126
1127 * config/i386/i386.c (ix86_rtx_costs): Enable fma for TARGET_AVX512F.
1128 * config/i386/sse.md (FMAMODEM): Changed modes and conditions.
1129 (FMAMODE): Ditto.
1130 (fma<mode>4): Removed condition.
1131 (fms<mode>4): Ditto.
1132 (fnma<mode>4): Ditto.
1133 (fnms<mode>4): Ditto.
1134 (fma4i_fmadd_<mode>): Ditto.
1135 (*fma_fmadd_<mode>): Ditto.
1136 (*fma_fmsub_<mode>): Ditto.
1137 (*fma_fnmadd_<mode>): Ditto.
1138 (*fma_fnmsub_<mode>): Ditto.
1139 (fmaddsub_<mode>): Allow for TARGET_AVX512F.
1140 (*fma_fmaddsub_<mode>): Ditto.
1141 (*fma_fmsubadd_<mode>): Ditto.
1142 (*fmai_fmadd_<mode>): Ditto.
1143 (*fmai_fmsub_<mode>): Ditto.
1144 (*fmai_fnmadd_<mode>): Ditto.
1145 (*fmai_fnmsub_<mode>): Ditto.
1146
1147 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1148 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1149 Sergey Lega <sergey.s.lega@intel.com>
1150 Anna Tikhonova <anna.tikhonova@intel.com>
1151 Ilya Tocar <ilya.tocar@intel.com>
1152 Andrey Turetskiy <andrey.turetskiy@intel.com>
1153 Ilya Verbin <ilya.verbin@intel.com>
1154 Kirill Yukhin <kirill.yukhin@intel.com>
1155 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1156
1157 * config/i386/sse.md (VI248_AVX2_8_AVX512F): New.
1158 (VI124_256): Changed to ...
1159 (VI124_256_48_512): This.
1160 (ssepackmode): Extended with wider modes.
1161 (<code><mode>3): Changed iterator.
1162 (*avx2_<code><mode>3): Ditto.
1163 (vec_pack_trunc_<mode>): Ditto.
1164
1165 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1166 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1167 Sergey Lega <sergey.s.lega@intel.com>
1168 Anna Tikhonova <anna.tikhonova@intel.com>
1169 Ilya Tocar <ilya.tocar@intel.com>
1170 Andrey Turetskiy <andrey.turetskiy@intel.com>
1171 Ilya Verbin <ilya.verbin@intel.com>
1172 Kirill Yukhin <kirill.yukhin@intel.com>
1173 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1174
1175 * config/i386/sse.md (VI124_AVX2_48_AVX512F): New.
1176 (VI8F_256_512): Ditto.
1177 (abs<mode>2): Changed iterator.
1178 (avx2_perm<mode>): Changed to ...
1179 (<avx2_avx512f>_perm<mode>): This.
1180 (avx2_perm<mode>_1): Changed to ...
1181 (<avx2_avx512f>_perm<mode>_1): This.
1182
1183 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1184 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1185 Sergey Lega <sergey.s.lega@intel.com>
1186 Anna Tikhonova <anna.tikhonova@intel.com>
1187 Ilya Tocar <ilya.tocar@intel.com>
1188 Andrey Turetskiy <andrey.turetskiy@intel.com>
1189 Ilya Verbin <ilya.verbin@intel.com>
1190 Kirill Yukhin <kirill.yukhin@intel.com>
1191 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1192
1193 * config/i386/sse.md (VI48_AVX512F): New.
1194 (VI48_AVX2): Changed to ...
1195 (VI48_AVX2_48_AVX512F): This.
1196 (avx2_ashrv<mode>): Changed to ...
1197 (<avx2_avx512f>_ashrv<mode>): This.
1198 (avx2_<shift_insn>v<mode>): Changed to ...
1199 (<avx2_avx512f>_<shift_insn>v<mode>): This.
1200
1201 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1202 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1203 Sergey Lega <sergey.s.lega@intel.com>
1204 Anna Tikhonova <anna.tikhonova@intel.com>
1205 Ilya Tocar <ilya.tocar@intel.com>
1206 Andrey Turetskiy <andrey.turetskiy@intel.com>
1207 Ilya Verbin <ilya.verbin@intel.com>
1208 Kirill Yukhin <kirill.yukhin@intel.com>
1209 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1210
1211 * config/i386/sse.md (VI4_AVX512F): New.
1212 (VI8_AVX2_AVX512F): Ditto.
1213 (mul<mode>3): Extended with wider modes.
1214 (*<sse4_1_avx2>_mul<mode>3): Ditto.
1215 (mul<mode>3): Ditto.
1216 (vec_widen_<s>mult_odd_<mode>): Ditto.
1217
1218 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1219 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1220 Sergey Lega <sergey.s.lega@intel.com>
1221 Anna Tikhonova <anna.tikhonova@intel.com>
1222 Ilya Tocar <ilya.tocar@intel.com>
1223 Andrey Turetskiy <andrey.turetskiy@intel.com>
1224 Ilya Verbin <ilya.verbin@intel.com>
1225 Kirill Yukhin <kirill.yukhin@intel.com>
1226 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1227
1228 * config/i386/sse.md (VI2_AVX512F): New.
1229 (VI124_AVX512F): Ditto.
1230 (sseunpackmode): Extended with wider modes.
1231 (sseunpackfltmode): Ditto.
1232 (vec_unpacks_float_hi_<mode>): Ditto.
1233 (vec_unpacks_float_lo_<mode>): Ditto.
1234 (vec_unpacku_float_hi_<mode>): Ditto.
1235 (vec_unpacku_float_lo_<mode>): Ditto.
1236 (vec_unpacks_lo_<mode>): Ditto.
1237 (vec_unpacks_hi_<mode>): Ditto.
1238 (vec_unpacku_lo_<mode>): Ditto.
1239 (vec_unpacku_hi_<mode>): Ditto.
1240
1241 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1242 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1243 Sergey Lega <sergey.s.lega@intel.com>
1244 Anna Tikhonova <anna.tikhonova@intel.com>
1245 Ilya Tocar <ilya.tocar@intel.com>
1246 Andrey Turetskiy <andrey.turetskiy@intel.com>
1247 Ilya Verbin <ilya.verbin@intel.com>
1248 Kirill Yukhin <kirill.yukhin@intel.com>
1249 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1250
1251 * config/i386/i386.md (multdiv): New.
1252 (multdiv_mnemonic): Ditto.
1253 * config/i386/sse.md (<sse>_vmmul<mode>3): Changed to...
1254 (<sse>_vm<multdiv_mnemonic><mode>3): This.
1255 (<sse>_vmdiv<mode>3): Removed.
1256
1257 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1258 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1259 Sergey Lega <sergey.s.lega@intel.com>
1260 Anna Tikhonova <anna.tikhonova@intel.com>
1261 Ilya Tocar <ilya.tocar@intel.com>
1262 Andrey Turetskiy <andrey.turetskiy@intel.com>
1263 Ilya Verbin <ilya.verbin@intel.com>
1264 Kirill Yukhin <kirill.yukhin@intel.com>
1265 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1266
1267 * config/i386/sse.md (V): Extended with wider modes.
1268 (VF2): Ditto.
1269 (ssehalfvecmode): Ditto.
1270 (i128): Ditto.
1271 (ssepackfltmode): Ditto.
1272 (avx_vec_concat<mode>): Ditto.
1273 (V_256_512): New iterator.
1274 (VF2_512_256): Ditto.
1275 (si2dfmode): New attribute.
1276 (si2dfmodelower): Ditto.
1277 (sf2dfmode): Ditto.
1278 (concat_tg_mode): Ditto.
1279 (floatv4siv4df2): Changed to ...
1280 (float<si2dfmodelower><mode>2): This.
1281 (avx_cvtps2pd256): Changed to ...
1282 (<sse2_avx_avx512f>_cvtps2pd<avxsizesuffix>): This.
1283 (vec_pack_trunc_v4df): Changed to ...
1284 (vec_pack_trunc_<mode>): This.
1285 (avx_vpermil<mode>): Changed to ...
1286 (<sse2_avx_avx512f>_vpermil<mode>): This.
1287 (<fixsuffix>fix_truncv8dfv8si2): New.
1288 (vec_pack_sfix_trunc_v8df): Ditto.
1289 (avx512f_rndscale<mode>): Ditto.
1290 (avx512f_roundpd512): Ditto.
1291 (vec_pack_ufix_trunc_<mode>): Updated iterator.
1292
1293 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1294 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1295 Sergey Lega <sergey.s.lega@intel.com>
1296 Anna Tikhonova <anna.tikhonova@intel.com>
1297 Ilya Tocar <ilya.tocar@intel.com>
1298 Andrey Turetskiy <andrey.turetskiy@intel.com>
1299 Ilya Verbin <ilya.verbin@intel.com>
1300 Kirill Yukhin <kirill.yukhin@intel.com>
1301 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1302
1303 * config/i386/i386.md (any_fix): New iterator.
1304 (fixsuffix): New attribute.
1305 * config/i386/sse.md (VF1): Extened with wider modes.
1306 (VI): Ditto.
1307 (VI_AVX2): Ditto.
1308 (VI8): Ditto.
1309 (sseintvecmodelower): Ditto.
1310 (ssescalarmode): Ditto.
1311 (ssescalarnum): Ditto.
1312 (VF1_128_256): New.
1313 (ssexmmmode): Ditto.
1314 (<fixsuffix>fix_truncv16sfv16si2): Ditto.
1315 (<sse>_rcp<mode>2): Change iterator.
1316 (rsqrt<mode>2): Ditto.
1317 (<sse>_rsqrt<mode>2): Ditto.
1318 (avx2_vec_dup<mode>): Ditto.
1319 (<sse4_1>_round<ssemodesuffix>_sfix<avxsizesuffix>): Ditto.
1320 (round<mode>2_sfix): Ditto.
1321 (avx2_pbroadcast<mode>): Ditto.
1322 (*andnot<mode>3): Handle XI mode.
1323 (*<code><mode>3): Ditto.
1324 (AVXTOSSEMODE): Removed.
1325 (avx_vpermil<mode>): Changed to ...
1326 (<sse2_avx_avx512f>_vpermil<mode>): This.
1327
1328 2013-10-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
1329 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
1330 Sergey Lega <sergey.s.lega@intel.com>
1331 Anna Tikhonova <anna.tikhonova@intel.com>
1332 Ilya Tocar <ilya.tocar@intel.com>
1333 Andrey Turetskiy <andrey.turetskiy@intel.com>
1334 Ilya Verbin <ilya.verbin@intel.com>
1335 Kirill Yukhin <kirill.yukhin@intel.com>
1336 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
1337
1338 * config/i386/sse.md (<sse2>_movnt<mode>): Update constraint to "v".
1339 (<sse>_comi): Ditto.
1340 (<sse>_ucomi): Ditto.
1341 (sse_cvtss2siq_2): Ditto.
1342 (sse2_cvtsd2si): Ditto.
1343 (sse2_cvtsd2siq): Ditto.
1344 (sse2_cvttsd2si): Ditto.
1345 (sse2_cvttsd2siq): Ditto.
1346 (<shift_insn><mode>3): Ditto.
1347 (sse2_cvtsi2sdq): Update constraint and prefix.
1348 (sse_cvtsi2ss): Update prefix.
1349 (sse_cvtsi2ssq): Ditto.
1350
1351 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1352
1353 * tree-vrp.c (infer_nonnull_range): Use is_gimple_call,
1354 ignore internal calls.
1355
1356 2013-10-11 Richard Biener <rguenther@suse.de>
1357
1358 * tree-pretty-print.c (dump_generic_node): Allow to dump
1359 both (D) and (ab) for SSA_NAMEs. Mark INTEGER_CSTs with
1360 (OVF) if TREE_OVERFLOW is set.
1361
1362 2013-10-11 Thomas Schwinge <thomas@codesourcery.com>
1363
1364 * tree.h (OMP_CLAUSE_CODE): Remove duplicate definition.
1365
1366 * gimple.c: GIMPLE statements have subcodes, not sub-codes.
1367 * gimple.h: Likewise.
1368
1369 * doc/generic.texi (OpenMP): OMP_CLAUSE_* are subcodes, not sub-codes.
1370
1371 * doc/generic.texi (Adding new DECL node types): Explain *_CHECK
1372 macros.
1373
1374 * doc/gimple.texi (is_gimple_omp): Move into the correct section.
1375
1376 * acinclude.m4 (gcc_GAS_FLAGS): Add more gcc_cv_as_flags overrides.
1377 * configure: Regenerate.
1378
1379 2013-10-11 Jakub Jelinek <jakub@redhat.com>
1380
1381 * tree-pretty-print.c (dump_omp_clause): Handle OMP_CLAUSE__LOOPTEMP_
1382 and new OpenMP 4.0 clauses, handle UDR OMP_CLAUSE_REDUCTION,
1383 formatting fixes, use pp_colon instead of pp_character (..., ':'),
1384 similarly pp_right_paren.
1385 (dump_generic_node): Handle OMP_DISTRIBUTE, OMP_TEAMS,
1386 OMP_TARGET_DATA, OMP_TARGET, OMP_TARGET_UPDATE, OMP_TASKGROUP,
1387 allow OMP_FOR_INIT to be NULL, handle OMP_ATOMIC_SEQ_CST.
1388 * tree.c (omp_clause_num_ops, omp_clause_code_name): Add OpenMP 4.0
1389 clauses.
1390 (omp_declare_simd_clauses_equal,
1391 omp_remove_redundant_declare_simd_attrs): New functions.
1392 (attribute_value_equal): Use omp_declare_simd_clauses_equal.
1393 (walk_tree_1): Handle new OpenMP 4.0 clauses.
1394 * tree.h (OMP_LOOP_CHECK): Define.
1395 (OMP_FOR_BODY, OMP_FOR_CLAUSES, OMP_FOR_INIT, OMP_FOR_COND,
1396 OMP_FOR_INCR, OMP_FOR_PRE_BODY): Use it.
1397 (OMP_TASKGROUP_BODY, OMP_TEAMS_BODY, OMP_TEAMS_CLAUSES,
1398 OMP_TARGET_DATA_BODY, OMP_TARGET_DATA_CLAUSES, OMP_TARGET_BODY,
1399 OMP_TARGET_CLAUSES, OMP_TARGET_UPDATE_CLAUSES, OMP_CLAUSE_SIZE,
1400 OMP_ATOMIC_SEQ_CST, OMP_CLAUSE_DEPEND_KIND, OMP_CLAUSE_MAP_KIND,
1401 OMP_CLAUSE_MAP_ZERO_BIAS_ARRAY_SECTION, OMP_CLAUSE_PROC_BIND_KIND,
1402 OMP_CLAUSE_REDUCTION_OMP_ORIG_REF, OMP_CLAUSE_ALIGNED_ALIGNMENT,
1403 OMP_CLAUSE_NUM_TEAMS_EXPR, OMP_CLAUSE_THREAD_LIMIT_EXPR,
1404 OMP_CLAUSE_DEVICE_ID, OMP_CLAUSE_DIST_SCHEDULE_CHUNK_EXPR,
1405 OMP_CLAUSE_SIMDLEN_EXPR): Define.
1406 (OMP_CLAUSE_DECL): Change range up to OMP_CLAUSE__LOOPTEMP_.
1407 (omp_remove_redundant_declare_simd_attrs): New prototype.
1408 * gimple.def (GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET,
1409 GIMPLE_OMP_TEAMS): New codes.
1410 (GIMPLE_OMP_RETURN): Use GSS_OMP_ATOMIC_STORE instead of GSS_BASE.
1411 * omp-low.c (struct omp_context): Add cancel_label and cancellable
1412 fields.
1413 (target_nesting_level): New variable.
1414 (extract_omp_for_data): Handle GF_OMP_FOR_KIND_DISTRIBUTE and
1415 OMP_CLAUSE_DIST_SCHEDULE. Don't fallback to library implementation
1416 for collapse > 1 static schedule unless ordered.
1417 (get_ws_args_for): Add par_stmt argument. Handle combined loops.
1418 (determine_parallel_type): Adjust get_ws_args_for caller.
1419 (install_var_field): Handle mask & 4 for double indirection.
1420 (scan_sharing_clauses): Ignore shared clause on teams construct.
1421 Handle OMP_CLAUSE__LOOPTEMP_ and new OpenMP 4.0 clauses.
1422 (create_omp_child_function): If inside target or declare target
1423 constructs, set "omp declare target" attribute on the child function.
1424 (find_combined_for): New function.
1425 (scan_omp_parallel): Handle combined loops.
1426 (scan_omp_target, scan_omp_teams): New functions.
1427 (check_omp_nesting_restrictions): Check new OpenMP 4.0 nesting
1428 restrictions and set ctx->cancellable for cancellable constructs.
1429 (scan_omp_1_stmt): Call check_omp_nesting_restrictions also on
1430 selected builtin calls. Handle GIMPLE_OMP_TASKGROUP,
1431 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS.
1432 (build_omp_barrier): Add lhs argument, return gimple rather than tree.
1433 (omp_clause_aligned_alignment): New function.
1434 (lower_rec_simd_input_clauses): Only call SET_DECL_VALUE_EXPR
1435 on decls.
1436 (lower_rec_input_clauses): Add FD argument. Ignore shared clauses
1437 on teams constructs. Handle user defined reductions and new
1438 OpenMP 4.0 clauses.
1439 (lower_reduction_clauses): Don't set placeholder to address of ref
1440 if it has already the right type.
1441 (lower_send_clauses): Handle OMP_CLAUSE__LOOPTEMP_.
1442 (expand_parallel_call): Use the new non-_start suffixed builtins,
1443 handle OMP_CLAUSE_PROC_BIND, don't call the outlined function
1444 and GOMP_parallel_end after the call.
1445 (expand_task_call): Handle OMP_CLAUSE_DEPEND.
1446 (expand_omp_for_init_counts): Handle combined loops.
1447 (expand_omp_for_init_vars): Add inner_stmt argument, handle combined
1448 loops.
1449 (expand_omp_for_generic): Likewise. Use GOMP_loop_end_cancel at the
1450 end of cancellable loops.
1451 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
1452 Likewise. Handle collapse > 1 loops.
1453 (expand_omp_simd): Handle combined loops.
1454 (expand_omp_for): Add inner_stmt argument, adjust callers of
1455 expand_omp_for* functions, use expand_omp_for_static*chunk even
1456 for collapse > 1 unless ordered.
1457 (expand_omp_sections): Use GOMP_sections_end_cancel at the end
1458 of cancellable sections.
1459 (expand_omp_single): Remove need_barrier variable, just rely on
1460 gimple_omp_return_nowait_p. Adjust build_omp_barrier caller.
1461 (expand_omp_synch): Allow GIMPLE_OMP_TASKGROUP and GIMPLE_OMP_TEAMS.
1462 (expand_omp_atomic_load, expand_omp_atomic_store,
1463 expand_omp_atomic_fetch_op): Handle gimple_omp_atomic_seq_cst_p.
1464 (expand_omp_target): New function.
1465 (expand_omp): Handle combined loops. Handle GIMPLE_OMP_TASKGROUP,
1466 GIMPLE_OMP_TEAMS, GIMPLE_OMP_TARGET.
1467 (build_omp_regions_1): Immediately close region for
1468 GF_OMP_TARGET_KIND_UPDATE.
1469 (maybe_add_implicit_barrier_cancel): New function.
1470 (lower_omp_sections): Adjust lower_rec_input_clauses caller. Handle
1471 cancellation.
1472 (lower_omp_single): Likewise. Add clobber after the barrier.
1473 (lower_omp_taskgroup): New function.
1474 (lower_omp_for): Handle combined loops. Adjust
1475 lower_rec_input_clauses caller. Handle cancellation.
1476 (lower_depend_clauses): New function.
1477 (lower_omp_taskreg): Lower depend clauses. Adjust
1478 lower_rec_input_clauses caller. Add clobber after the call. Handle
1479 cancellation.
1480 (lower_omp_target, lower_omp_teams): New functions.
1481 (lower_omp_1): Handle cancellation. Handle GIMPLE_OMP_TASKGROUP,
1482 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GOMP_barrier, GOMP_cancel
1483 and GOMP_cancellation_point calls.
1484 (lower_omp): Fold stmts inside of target region.
1485 (diagnose_sb_1, diagnose_sb_2): Handle GIMPLE_OMP_TASKGROUP,
1486 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1487 * builtin-types.def (DEF_FUNCTION_TYPE_8): Document.
1488 (BT_FN_VOID_OMPFN_PTR_UINT,
1489 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG,
1490 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG,
1491 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT): Remove.
1492 (BT_FN_VOID_OMPFN_PTR_UINT_UINT_UINT,
1493 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_UINT,
1494 BT_FN_VOID_OMPFN_PTR_UINT_LONG_LONG_LONG_LONG_UINT,
1495 BT_FN_BOOL_INT, BT_FN_BOOL_INT_BOOL, BT_FN_VOID_UINT_UINT,
1496 BT_FN_VOID_INT_PTR_SIZE_PTR_PTR_PTR,
1497 BT_FN_VOID_INT_OMPFN_PTR_SIZE_PTR_PTR_PTR,
1498 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR): New.
1499 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1,
1500 call_may_clobber_ref_p_1): Handle BUILT_IN_GOMP_BARRIER_CANCEL,
1501 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_LOOP_END_CANCEL,
1502 BUILT_IN_GOMP_SECTIONS_END_CANCEL. Don't handle
1503 BUILT_IN_GOMP_PARALLEL_END.
1504 * gimple-low.c (lower_stmt): Handle GIMPLE_OMP_TASKGROUP,
1505 GIMPLE_OMP_TARGET and GIMPLE_OMP_TEAMS.
1506 * gimple-pretty-print.c (dump_gimple_omp_for): Handle
1507 GF_OMP_FOR_KIND_DISTRIBUTE.
1508 (dump_gimple_omp_target, dump_gimple_omp_teams): New functions.
1509 (dump_gimple_omp_block): Handle GIMPLE_OMP_TASKGROUP.
1510 (dump_gimple_omp_return): Print lhs if it has any.
1511 (dump_gimple_omp_atomic_load, dump_gimple_omp_atomic_store): Handle
1512 gimple_omp_atomic_seq_cst_p.
1513 (pp_gimple_stmt_1): Handle GIMPLE_OMP_TASKGROUP, GIMPLE_OMP_TARGET
1514 and GIMPLE_OMP_TEAMS.
1515 * langhooks.c (lhd_omp_mappable_type): New function.
1516 * tree-vectorizer.c (struct simd_array_to_simduid): Fix up comment.
1517 * langhooks.h (struct lang_hooks_for_types): Add omp_mappable_type
1518 hook.
1519 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_MAP,
1520 GOVD_ALIGNED and GOVD_MAP_TO_ONLY.
1521 (enum omp_region_type): Add ORT_TEAMS, ORT_TARGET_DATA and
1522 ORT_TARGET.
1523 (struct gimplify_omp_ctx): Add combined_loop field.
1524 (gimplify_call_expr, gimplify_modify_expr): Don't call fold_stmt
1525 on stmts inside of target region.
1526 (is_gimple_stmt): Return true for OMP_DISTRIBUTE and OMP_TASKGROUP.
1527 (omp_firstprivatize_variable): Handle GOVD_MAP, GOVD_ALIGNED,
1528 ORT_TARGET and ORT_TARGET_DATA.
1529 (omp_add_variable): Avoid checks on readding var for GOVD_ALIGNED.
1530 Handle GOVD_MAP.
1531 (omp_notice_threadprivate_variable): Complain about threadprivate
1532 variables in target region.
1533 (omp_notice_variable): Complain about vars with non-mappable type
1534 in target region. Handle ORT_TEAMS, ORT_TARGET and ORT_TARGET_DATA.
1535 (omp_check_private): Ignore ORT_TARGET* regions.
1536 (gimplify_scan_omp_clauses, gimplify_adjust_omp_clauses_1,
1537 gimplify_adjust_omp_clauses): Handle new OpenMP 4.0 clauses.
1538 (find_combined_omp_for): New function.
1539 (gimplify_omp_for): Handle gimplification of combined loops.
1540 (gimplify_omp_workshare): Gimplify also OMP_TARGET, OMP_TARGET_DATA,
1541 OMP_TEAMS.
1542 (gimplify_omp_target_update): New function.
1543 (gimplify_omp_atomic): Handle OMP_ATOMIC_SEQ_CST.
1544 (gimplify_expr): Handle OMP_DISTRIBUTE, OMP_TARGET, OMP_TARGET_DATA,
1545 OMP_TARGET_UPDATE, OMP_TEAMS, OMP_TASKGROUP.
1546 (gimplify_body): If fndecl has "omp declare target" attribute, add
1547 implicit ORT_TARGET context around it.
1548 * tree.def (OMP_DISTRIBUTE, OMP_TEAMS, OMP_TARGET_DATA, OMP_TARGET,
1549 OMP_TASKGROUP, OMP_TARGET_UPDATE): New tree codes.
1550 * tree-nested.c (convert_nonlocal_reference_stmt,
1551 convert_local_reference_stmt, convert_gimple_call): Handle
1552 GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1553 * omp-builtins.def (BUILT_IN_GOMP_TASK): Use
1554 BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT_PTR
1555 instead of BT_FN_VOID_OMPFN_PTR_OMPCPYFN_LONG_LONG_BOOL_UINT.
1556 (BUILT_IN_GOMP_TARGET, BUILT_IN_GOMP_TARGET_DATA,
1557 BUILT_IN_GOMP_TARGET_END_DATA, BUILT_IN_GOMP_TARGET_UPDATE,
1558 BUILT_IN_GOMP_TEAMS, BUILT_IN_BARRIER_CANCEL,
1559 BUILT_IN_GOMP_LOOP_END_CANCEL,
1560 BUILT_IN_GOMP_SECTIONS_END_CANCEL, BUILT_IN_OMP_GET_TEAM_NUM,
1561 BUILT_IN_OMP_GET_NUM_TEAMS, BUILT_IN_GOMP_TASKGROUP_START,
1562 BUILT_IN_GOMP_TASKGROUP_END, BUILT_IN_GOMP_PARALLEL_LOOP_STATIC,
1563 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC,
1564 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED,
1565 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME, BUILT_IN_GOMP_PARALLEL,
1566 BUILT_IN_GOMP_PARALLEL_SECTIONS, BUILT_IN_GOMP_CANCEL,
1567 BUILT_IN_GOMP_CANCELLATION_POINT): New built-ins.
1568 (BUILT_IN_GOMP_PARALLEL_LOOP_STATIC_START,
1569 BUILT_IN_GOMP_PARALLEL_LOOP_DYNAMIC_START,
1570 BUILT_IN_GOMP_PARALLEL_LOOP_GUIDED_START,
1571 BUILT_IN_GOMP_PARALLEL_LOOP_RUNTIME_START,
1572 BUILT_IN_GOMP_PARALLEL_START, BUILT_IN_GOMP_PARALLEL_END,
1573 BUILT_IN_GOMP_PARALLEL_SECTIONS_START): Remove.
1574 * tree-inline.c (remap_gimple_stmt, estimate_num_insns):
1575 Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1576 * gimple.c (gimple_build_omp_taskgroup, gimple_build_omp_target,
1577 gimple_build_omp_teams): New functions.
1578 (walk_gimple_op): Handle GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS and
1579 GIMPLE_OMP_TASKGROUP. Walk optional lhs on GIMPLE_OMP_RETURN.
1580 (walk_gimple_stmt, gimple_copy): Handle GIMPLE_OMP_TARGET,
1581 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1582 * gimple.h (enum gf_mask): GF_OMP_FOR_KIND_DISTRIBUTE,
1583 GF_OMP_FOR_COMBINED, GF_OMP_FOR_COMBINED_INTO,
1584 GF_OMP_TARGET_KIND_MASK, GF_OMP_TARGET_KIND_REGION,
1585 GF_OMP_TARGET_KIND_DATA, GF_OMP_TARGET_KIND_UPDATE,
1586 GF_OMP_ATOMIC_SEQ_CST): New.
1587 (gimple_build_omp_taskgroup, gimple_build_omp_target,
1588 gimple_build_omp_teams): New prototypes.
1589 (gimple_has_substatements): Handle GIMPLE_OMP_TARGET,
1590 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1591 (gimple_omp_subcode): Use GIMPLE_OMP_TEAMS instead of
1592 GIMPLE_OMP_SINGLE as end of range.
1593 (gimple_omp_return_set_lhs, gimple_omp_return_lhs,
1594 gimple_omp_return_lhs_ptr, gimple_omp_atomic_seq_cst_p,
1595 gimple_omp_atomic_set_seq_cst, gimple_omp_for_combined_p,
1596 gimple_omp_for_set_combined_p, gimple_omp_for_combined_into_p,
1597 gimple_omp_for_set_combined_into_p, gimple_omp_target_clauses,
1598 gimple_omp_target_clauses_ptr, gimple_omp_target_set_clauses,
1599 gimple_omp_target_kind, gimple_omp_target_set_kind,
1600 gimple_omp_target_child_fn, gimple_omp_target_child_fn_ptr,
1601 gimple_omp_target_set_child_fn, gimple_omp_target_data_arg,
1602 gimple_omp_target_data_arg_ptr, gimple_omp_target_set_data_arg,
1603 gimple_omp_teams_clauses, gimple_omp_teams_clauses_ptr,
1604 gimple_omp_teams_set_clauses): New inlines.
1605 (CASE_GIMPLE_OMP): Add GIMPLE_OMP_TARGET, GIMPLE_OMP_TEAMS
1606 and GIMPLE_OMP_TASKGROUP.
1607 * tree-core.h (enum omp_clause_code): Add new OpenMP 4.0 clause codes.
1608 (enum omp_clause_depend_kind, enum omp_clause_map_kind,
1609 enum omp_clause_proc_bind_kind): New.
1610 (union omp_clause_subcode): Add depend_kind, map_kind and
1611 proc_bind_kind fields.
1612 * tree-cfg.c (make_edges): Handle GIMPLE_OMP_TARGET,
1613 GIMPLE_OMP_TEAMS and GIMPLE_OMP_TASKGROUP.
1614 * langhooks-def.h (lhd_omp_mappable_type): New prototype.
1615 (LANG_HOOKS_OMP_MAPPABLE_TYPE): Define.
1616 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
1617
1618 2013-10-10 Teresa Johnson <tejohnson@google.com>
1619
1620 * predict.c (tree_estimate_probability): Add new parameter
1621 for estimate_bb_frequencies.
1622 (estimate_bb_frequencies): Add new parameter to force estimation.
1623 (rebuild_frequencies): When max frequency in function is small,
1624 recompute counts from frequencies.
1625 * predict.h (estimate_bb_frequencies): New parameter.
1626
1627 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1628
1629 * ipa-inline.c (ipa_inline): Fix leak of "order" when
1630 optimizations are disabled.
1631
1632 2013-10-10 David Malcolm <dmalcolm@redhat.com>
1633
1634 * coverage.c (coverage_finish): Fix leak of da_file_name.
1635
1636 2013-10-10 Jan Hubicka <jh@suse.cz>
1637
1638 * config/i386/x86-tune.def: Enable X86_TUNE_SSE_TYPELESS_STORES
1639 for generic, enable X86_TUNE_SSE_LOAD0_BY_PXOR for Bulldozer,
1640 Bobcat and generic.
1641
1642 2013-10-10 Jakub Jelinek <jakub@redhat.com>
1643
1644 PR middle-end/58670
1645 * stmt.c (expand_asm_operands): Add FALLTHRU_BB argument,
1646 if any labels are in FALLTHRU_BB, use a special label emitted
1647 immediately after the asm goto insn rather than label_rtx
1648 of the LABEL_DECL.
1649 (expand_asm_stmt): Adjust caller.
1650 * cfgrtl.c (commit_one_edge_insertion): Force splitting of
1651 edge if the last insn in predecessor is a jump with single successor,
1652 but it isn't simplejump_p.
1653
1654 2013-10-10 Richard Biener <rguenther@suse.de>
1655
1656 PR tree-optimization/58656
1657 * tree-ssa-pre.c (phi_translate): Do not cache failed
1658 translations.
1659
1660 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1661
1662 * gimplify.c: Include expr.h and tm_p.h for targets with special
1663 va-arg padding requirements.
1664
1665 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1666
1667 * tree-flow.h: Move some prototypes to gimple.h.
1668 (gimple_fold_indirect_ref): Move prototype to gimple-fold.h.
1669 * gimple.h: Relocate some prototypes from tree-flow.h
1670 * builtins.c (std_gimplify_va_arg_expr, build_va_arg_indirect_ref):
1671 Move to gimplify.c.
1672 * gimplify.c (gimple_fold_indirect_ref): Move to gimple-fold.c.
1673 (build_va_arg_indirect_ref): Relocate and make static.
1674 (std_gimplify_va_arg_expr): Relocate here.
1675 * gimple-fold.c (gimple_fold_indirect_ref): Relocate here.
1676 * gimple-fold.h (gimple_fold_indirect_ref): Add prototype.
1677
1678 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1679
1680 * doc/md.texi: Document the mnemonic attribute.
1681
1682 2013-10-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1683
1684 PR target/57377
1685 * gensupport.c (gen_mnemonic_attr): Handle (set (attr x) y) and
1686 (set_attr_alternative x ...) when searching for user defined
1687 mnemonic attribute.
1688
1689 2013-10-10 Andrew MacLeod <amacleod@redhat.com>
1690
1691 * config/aplha/alpha.c: Add gimple-ssa.h to include list.
1692
1693 2013-10-09 Easwaran Raman <eraman@google.com>
1694
1695 * params.def (PARAM_MIN_SIZE_FOR_STACK_SHARING): New param...
1696 * cfgexpand.c (defer_stack_allocation): ...use here
1697 * doc/invoke.texi: Add documentation for min-size-for-stack-sharing.
1698
1699 2013-10-09 Zhenqiang Chen <zhenqiang.chen@arm.com>
1700
1701 * tree-ssa-phiopts.c (rhs_is_fed_for_value_replacement): New function.
1702 (operand_equal_for_value_replacement): New function, extracted from
1703 value_replacement and enhanced to catch more cases.
1704 (value_replacement): Use operand_equal_for_value_replacement.
1705
1706 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1707
1708 * loop-doloop.c (doloop_modify, doloop_optimize): Use
1709 get_max_loop_iterations.
1710
1711 2013-10-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
1712
1713 * config/arm/aarch-common.c (arm_early_load_addr_dep):
1714 Place comment above function.
1715
1716 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1717
1718 * tree-flow.h: Remove all remaining prototypes, enums and structs that
1719 are not related to tree-cfg.c.
1720 * tree-ssa-address.h: New file. Relocate prototypes.
1721 * tree-ssa-address.c: (struct mem_address): Relocate from tree-flow.h.
1722 (addr_for_mem_ref): New. Combine call to get_address_description and
1723 return addr_for_mem_ref.
1724 * expr.c (expand_expr_real_1): Use new addr_for_mem_ref routine.
1725 * tree-ssa-live.h: Adjust prototypes.
1726 * passes.c: Include tree-ssa-live.h.
1727 * gimple-pretty-print.h (gimple_dump_bb): Add prototype.
1728 * graphite.c (graphite_transform_loops): Make static.
1729 (graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1730 make_pass_graphite, pass_data_graphite_transforms,
1731 make_pass_graphite_transforms): Relocate here from tree-ssa-loop.c.
1732 * ipa-pure-const.c (warn_function_noreturn): Make static.
1733 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1734 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1735 Relocate from tree-cfg.c
1736 * tree-cfg.c (tree_node_can_be_shared, gimple_empty_block_p): Make
1737 static.
1738 (execute_warn_function_noreturn, gate_warn_function_noreturn,
1739 class pass_warn_function_noreturn, make_pass_warn_function_noreturn):
1740 Move to ipa-pure-const.c.
1741 (execute_fixup_cfg, class pass_fixup_cfg, make_pass_fixup_cfg):
1742 Relocate from tree-optimize.c.
1743 * tree-optimize.c (execute_fixup_cfg, class pass_fixup_cfg,
1744 make_pass_fixup_cfg): Move to tree-cfg.c.
1745 * tree-chrec.h: (enum ev_direction): Relocate here from tree-flow.h.
1746 Relocate some prototypes.
1747 * tree-data-ref.h (tree_check_data_deps) Add prototype.
1748 * tree-dump.c (dump_function_to_file): Remove prototype.
1749 Add tree-flow.h to the include file.
1750 * tree-dump.h: Remove prototype.
1751 * tree-parloops.h: New File. Add prototypes.
1752 * tree-parloops.c (gate_tree_parallelize_loops, tree_parallelize_loops,
1753 pass_data_parallelize_loops, make_pass_parallelize_loops): Relocate
1754 from tree-ssa-loop.c.
1755 * tree-predcom.c (run_tree_predictive_commoning,
1756 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom):
1757 Relocate here from tree-ssa-loop.c.
1758 * tree-ssa-dom.c (tree_ssa_dominator_optimize) Don't call
1759 ssa_name_values.release ().
1760 * tree-ssa-threadedge.h: New File. Relocate prototypes here.
1761 (ssa_name_values): Relocate from tree-flow.h.
1762 * tree-ssa.h: Include tree-ssa-threadedge.h and tree-ssa-address.h.
1763 * tree-ssa-loop.c (run_tree_predictive_commoning,
1764 gate_tree_predictive_commoning, pass_data_predcom, make_pass_predcom,
1765 graphite_transforms, gate_graphite_transforms, pass_data_graphite,
1766 make_pass_graphite, pass_data_graphite_transforms,
1767 make_pass_graphite_transforms, gate_tree_parallelize_loops,
1768 tree_parallelize_loops, pass_data_parallelize_loops,
1769 make_pass_parallelize_loops): Move to other files.
1770 * tree-vectorizer.h (lpeel_tree_duplicate_loop_to_edge_cfg): Prototype
1771 moved here.
1772 * tree.h: Remove prototypes from tree-address.c.
1773
1774 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1775
1776 * tree-flow.h (tm_restart_node, gimple_df): Move to gimple-ssa.h.
1777 (struct int_tree_map): Move to tree-hasher.h
1778 (SCALE, LABEL, PERCENT): Move to gimple.h
1779 * tree-flow-inline.h: Delete. Move functions to other files.
1780 (unmodifiable_var_p, ref_contains_array_ref): Unused, so delete.
1781 * gimple-ssa.h (tm_restart_node, gimple_df): Relocate from tree-flow.h.
1782 (gimple_in_ssa_p, gimple_vop): Relocate from tree-flow-inline.h
1783 * gimple.h (imple_stmt_max_uid, set_gimple_stmt_max_uid,
1784 inc_gimple_stmt_max_uid, get_lineno): Relocate from tree-flow-inline.h.
1785 (SCALE, LABEL, PERCENT): Relocate from tree-flow.h
1786 * tree-hasher.h: Don't include tree-flow.h.
1787 (struct int_tree_map): Relocate from tree-flow.h.
1788 * tree-sra.c (contains_view_convert_expr_p): Relocate from
1789 tree-flow-inline.h and make static.
1790 * tree-ssa-alias.h (ranges_overlap_p): Relocate from
1791 tree-flow-inline.h.
1792 * tree-ssa-operands.c (gimple_ssa_operands): Relocate from
1793 tree-flow-inline.h and make static.
1794 * tree.h (is_global_var, may_be_aliased): Relocate from
1795 tree-flow-inline.h.
1796 * Makefile.in (GTFILES): Remove tree-flow.h and add gimple-ssa.h.
1797 * value-prof.c: No longer include tree-flow-inline.h.
1798 * tree-switch-conversion.c: No longer include tree-flow-inline.h.
1799
1800 2013-10-09 Andrew MacLeod <amacleod@redhat.com>
1801
1802 * tree-flow.h: Move some protoypes. Include new tree-ssa-loop.h.
1803 (struct affine_iv, struct tree_niter_desc): Move to tree-ssa-loop.h.
1804 (enum move_pos): Move to tree-ssa-loop-im.h
1805 * cfgloop.h: Move some prototypes.
1806 (gcov_type_to_double_int): relocate from tree-ssa-loop.niter.c.
1807 * tree-flow-inline.h (loop_containing_stmt): Move to tree-ssa-loop.h.
1808 * tree-ssa-loop.h: New File. Include other tree-ssa-loop-*.h files.
1809 (struct affine_iv, struct tree_niter_desc): Relocate from tree-flow.h.
1810 (loop_containing_stmt): Relocate from tree-flow-inline.h.
1811 * tree-ssa-loop-ch.c: (do_while_loop_p): Make static.
1812 * tree-ssa-loop-im.c (for_each_index): Move to tree-ssa-loop.c.
1813 (enum move_pos): Relocate here.
1814 (lsm_tmp_name_add, gen_lsm_tmp_name, get_lsm_tmp_name): Move to
1815 tree-ssa-loop.c.
1816 (execute_sm_if_changed_flag_set): Change get_lsm_tmp_name call.
1817 (tree_ssa_loop_im, gate_tree_ssa_loop_im, pass_data_lim,
1818 make_pass_lim): Relocate here from tree-ssa-loop.c.
1819 * tree-ssa-loop-ivcanon.c (tree_num_loop_insns): Move to
1820 tree-ssa-loop.c.
1821 (loop_edge_to_cancel, unloop_loops): Make static.
1822 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
1823 make_pass_iv_canon): Relocate from tree-ssa-loop.c.
1824 (tree_complete_unroll, gate_tree_complete_unroll,
1825 pass_data_complete_unroll, make_pass_complete_unroll): Relocate here.
1826 (tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
1827 pass_data_complete_unrolli, make_pass_complete_unrolli): Relocate here.
1828 * tree-ssa-loop-ivopts.c: Remove local prototypes.
1829 (stmt_invariant_in_loop_p): Remove unused function.
1830 * tree-ssa-loop-ivopts.h: New file. Add prototypes.
1831 * tree-ssa-loop-manip.h: New file. Add prototypes.
1832 * tree-ssa-loop-niter.c (record_niter_bound): Move to cfgloop.c.
1833 (gcov_type_to_double_int): Move to cfgloop.h.
1834 (double_int_cmp, bound_index,
1835 estimate_numbers_of_iterations_loop): Make static.
1836 (estimated_loop_iterations): Factor out get_estimated_loop_iterations.
1837 (max_loop_iterations): Factor out get_max_loop_iterations.
1838 (estimated_loop_iterations_int, max_stmt_executions_int): Move to
1839 cfgloop.c.
1840 * tree-ssa-loop-niter.h: New file. Add prototypes.
1841 * tree-ssa-loop-prefetch.c (tree_ssa_loop_prefetch,
1842 gate_tree_ssa_loop_prefetch, pass_data_loop_prefetch,
1843 make_pass_loop_prefetch): Relocate from tree-ssa-loop.c.
1844 * tree-ssa-loop-unswitch.c (tree_ssa_loop_unswitch,
1845 gate_tree_ssa_loop_unswitch, pass_data_tree_unswitch,
1846 make_pass_tree_unswitch): Relocate from tree-ssa-loop.c.
1847 * tree-ssa-loop.c (tree_ssa_loop_im, gate_tree_ssa_loop_im,
1848 pass_data_lim, make_pass_lim): Move to tree-ssa-loop-im.c.
1849 (tree_ssa_loop_unswitch, gate_tree_ssa_loop_unswitch,
1850 pass_data_tree_unswitch, make_pass_tree_unswitch): Move.
1851 (tree_ssa_loop_ivcanon, gate_tree_ssa_loop_ivcanon, pass_data_iv_canon,
1852 make_pass_iv_canon, tree_complete_unroll, gate_tree_complete_unroll,
1853 pass_data_complete_unroll, make_pass_complete_unroll,
1854 tree_complete_unroll_inner, gate_tree_complete_unroll_inner,
1855 pass_data_complete_unrolli, make_pass_complete_unrolli): Move to
1856 tree-ssa-loop-ivcanon.c.
1857 (tree_ssa_loop_prefetch, gate_tree_ssa_loop_prefetch,
1858 pass_data_loop_prefetch, make_pass_loop_prefetch): Move to
1859 tree-ssa-loop-prefetch.c.
1860 (for_each_index, lsm_tmp_name_add, gen_lsm_tmp_name): Relocate from
1861 tree-ssa-loop-im.c.
1862 (get_lsm_tmp_name): Relocate and add suffix parameter.
1863 (tree_num_loop_insns): Relocate from tree-ssa-ivcanon.c.
1864 * tree-scalar-evolution.h (simple_iv): Don't use affive_iv typedef.
1865 * cfgloop.c (record_niter_bound, estimated_loop_iterations_int,
1866 max_stmt_executions_int): Move from tree-ssa-loop-niter.c.
1867 (get_estimated_loop_iterations): Factor out accessor from
1868 estimated_loop_iterations in tree-ssa-loop-niter.c.
1869 (get_max_loop_iterations): Factor out accessor from
1870 _max_loop_iterations in tree-ssa-niter.c.
1871 * loop-unroll.c (decide_unroll_constant_iterations,
1872 decide_unroll_runtime_iterations, decide_peel_simple,
1873 decide_unroll_stupid): Use new get_* accessors.
1874
1875 2013-10-09 Marc Glisse <marc.glisse@inria.fr>
1876
1877 PR tree-optimization/20318
1878 * doc/extend.texi (returns_nonnull): New function attribute.
1879 * fold-const.c (tree_expr_nonzero_warnv_p): Look for returns_nonnull
1880 attribute.
1881 * tree-vrp.c (gimple_stmt_nonzero_warnv_p): Likewise.
1882 (stmt_interesting_for_vrp): Accept all GIMPLE_CALL.
1883
1884 2013-10-09 Eric Botcazou <ebotcazou@adacore.com>
1885
1886 PR middle-end/58570
1887 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): Return
1888 false if both components are bitfields.
1889
1890 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1891
1892 * config/aarch64/arm_neon.h (vclz_s8, vclz_s16, vclz_s32)
1893 (vclzq_s8, vclzq_s16, vclzq_s32, vclz_u8, vclz_u16, vclz_u32)
1894 (vclzq_u8, vclzq_u16, vclzq_u32): Replace ASM with C.
1895 * config/aarch64/aarch64.h
1896 (CLZ_DEFINED_VALUE_AT_ZERO): Macro fixed for clz.
1897 * config/aarch64/aarch64-simd-builtins.def
1898 (VAR1 (UNOP, clz, 0, v4si)): Replaced with iterator.
1899
1900 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1901
1902 * config/aarch64/arm_neon.h (vadd_f64, vsub_f64): Implementation added.
1903
1904 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1905
1906 * config/aarch64/arm_neon.h (vdiv_f64): Added.
1907
1908 2013-10-09 Alex Velenko <Alex.Velenko@arm.com>
1909
1910 * config/aarch64/arm_neon.h (vneg_f32): Asm replaced with C.
1911 (vneg_f64): New intrinsic.
1912 (vneg_s8): Asm replaced with C.
1913 (vneg_s16): Likewise.
1914 (vneg_s32): Likewise.
1915 (vneg_s64): New intrinsic.
1916 (vnegq_f32): Asm replaced with C.
1917 (vnegq_f64): Likewise.
1918 (vnegq_s8): Likewise.
1919 (vnegq_s16): Likewise.
1920 (vnegq_s32): Likewise.
1921 (vnegq_s64): Likewise.
1922
1923 2013-10-09 Renlin Li <Renlin.Li@arm.com>
1924
1925 * config/arm/arm.c (arm_output_mi_thunk): Use plus_constant.
1926
1927 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1928
1929 * config/s390/s390.c (s390_register_info_stdarg_fpr): Remove
1930 packed stack special handling.
1931 (s390_frame_info, s390_emit_prologue, s390_emit_epilogue): Switch
1932 back to fixed stack slots for FPRs saved due to stdarg.
1933
1934 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1935
1936 * config/s390/s390.c (s390_frame_info): Restructure function.
1937
1938 2013-10-09 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
1939
1940 * config/s390/s390.c (struct s390_frame_layout): New field
1941 gpr_save_slots.
1942 (cfun_save_arg_fprs_p, cfun_gpr_save_slot): New macros.
1943 (s390_reg_clobbered_rtx, s390_regs_ever_clobbered): Change type of
1944 regs_ever_clobbered to char*.
1945 (s390_regs_ever_clobbered): Check crtl->saves_all_registers instead
1946 of cfun->has_nonlocal_label. Ignore frame related restore INSNs.
1947 (s390_register_info): Enable FPR save slots. Move/Copy some
1948 functionality into ...
1949 (s390_register_info_gprtofpr, s390_register_info_stdarg_fpr)
1950 (s390_register_info_stdarg_gpr, s390_optimize_register_info): New
1951 function.
1952 (s390_frame_info): Do gpr slot allocation here now. stdarg does
1953 not imply a stack frame.
1954 (s390_init_frame_layout): Remove variable clobbered_regs.
1955 (s390_update_register_info): Remove function.
1956 (s390_hard_regno_rename_ok): Call-saved regs without a save slot
1957 cannot be used for register renaming.
1958 (s390_hard_regno_scratch_ok): New function.
1959 (TARGET_HARD_REGNO_SCRATCH_OK): Define target hook.
1960 (s390_initial_elimination_offset): Change offset calculation of
1961 the return address pointer.
1962 (save_gprs): Deal with only r6 being saved from the call-saved regs.
1963 (restore_gprs): Set frame related flag.
1964 (s390_save_gprs_to_fprs, s390_restore_gprs_from_fprs): New functions.
1965 (s390_emit_prologue): Call s390_register_info instead of
1966 s390_update_frame_layout. Call s390_save_gprs_to_fprs.
1967 (s390_emit_epilogue): Call s390_restore_gprs_from_fprs.
1968 (s390_optimize_prologue): Call s390_optimize_register_info.
1969 Try to remove also FPR slot save/restore INSNs. Remove frame
1970 related flags from restore INSNs.
1971
1972 2013-10-08 DJ Delorie <dj@redhat.com>
1973
1974 * config/rl78/rl78-expand.md (movqi): use operands[] not operandN.
1975 (movhi): Likewise.
1976
1977 * config/rl78/rl78.c (rl78_print_operand_1): Change %c to %C to
1978 avoid conflict with the MI use of %c.
1979 * config/rl78/rl78-real.md: change %c to %C throughout.
1980 * config/rl78/rl78-virt.md: Likewise.
1981
1982 2013-10-08 Jan Hubicka <jh@suse.cz>
1983
1984 * config/i386/i386.c (ix86_option_override_internal): Switch
1985 to SSE math for -ffast-math when target ISA supports SSE2.
1986
1987 2013-10-08 Andrew MacLeod <amacleod@redhat.com>
1988
1989 * tree-flow.h: Remove some prototypes.
1990 * tree.h: Remove some protypes, add a couple.
1991 * tree.c (using_eh_for_cleanups_flag, using_eh_for_cleanups,
1992 using_eh_for_cleanups_p): Add interface routines for front ends.
1993 * tree-eh.h: New file. Add protoptyes.
1994 * tree-eh.c (using_eh_for_cleanups_p, using_eh_for_cleanups): Delete.
1995 (add_stmt_to_eh_lp_fn): Make static.
1996 (lower_try_finally): Use new using_eh_for_cleanups_p.
1997 * emit-rtl.c: Include tree-eh.h.
1998 * gimple.h: Include tree-eh.h.
1999
2000 2013-10-08 Marc Glisse <marc.glisse@inria.fr>
2001
2002 PR tree-optimization/58480
2003 * tree-vrp.c (infer_nonnull_range): New function.
2004 (infer_value_range): Call infer_nonnull_range.
2005
2006 2013-10-08 Dehao Chen <dehao@google.com>
2007
2008 PR tree-optimization/58619
2009 * tree-inline.c (copy_phis_for_bb): Combine location data
2010 only if non-null.
2011
2012 2013-10-08 Zhenqiang Chen <zhenqiang.chen@linaro.org>
2013
2014 PR target/58423
2015 * config/arm/arm.c (arm_emit_ldrd_pop): Attach
2016 RTX_FRAME_RELATED_P on INSN.
2017
2018 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2019
2020 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const_le): New.
2021 (altivec_expand_vec_perm_const): Call it.
2022
2023 2013-10-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2024
2025 * config/rs6000/vector.md (mov<mode>): Emit permuted move
2026 sequences for LE VSX loads and stores at expand time.
2027 * config/rs6000/rs6000-protos.h (rs6000_emit_le_vsx_move): New
2028 prototype.
2029 * config/rs6000/rs6000.c (rs6000_const_vec): New.
2030 (rs6000_gen_le_vsx_permute): New.
2031 (rs6000_gen_le_vsx_load): New.
2032 (rs6000_gen_le_vsx_store): New.
2033 (rs6000_gen_le_vsx_move): New.
2034 * config/rs6000/vsx.md (*vsx_le_perm_load_v2di): New.
2035 (*vsx_le_perm_load_v4si): New.
2036 (*vsx_le_perm_load_v8hi): New.
2037 (*vsx_le_perm_load_v16qi): New.
2038 (*vsx_le_perm_store_v2di): New.
2039 (*vsx_le_perm_store_v4si): New.
2040 (*vsx_le_perm_store_v8hi): New.
2041 (*vsx_le_perm_store_v16qi): New.
2042 (*vsx_xxpermdi2_le_<mode>): New.
2043 (*vsx_xxpermdi4_le_<mode>): New.
2044 (*vsx_xxpermdi8_le_V8HI): New.
2045 (*vsx_xxpermdi16_le_V16QI): New.
2046 (*vsx_lxvd2x2_le_<mode>): New.
2047 (*vsx_lxvd2x4_le_<mode>): New.
2048 (*vsx_lxvd2x8_le_V8HI): New.
2049 (*vsx_lxvd2x16_le_V16QI): New.
2050 (*vsx_stxvd2x2_le_<mode>): New.
2051 (*vsx_stxvd2x4_le_<mode>): New.
2052 (*vsx_stxvd2x8_le_V8HI): New.
2053 (*vsx_stxvd2x16_le_V16QI): New.
2054
2055 2013-10-07 Renlin Li <Renlin.Li@arm.com>
2056
2057 * config/arm/arm-cores.def (cortex-a53): Use cortex tuning.
2058
2059 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2060
2061 * config/s390/s390.c (s390_register_info): Make the call-saved FPR
2062 loop to work also for 31bit ABI.
2063 Save the stack pointer for frame_size > 0.
2064
2065 2013-10-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2066
2067 * config/s390/s390.md ("tbegin", "tbegin_nofloat", "tbegin_retry")
2068 ("tbegin_retry_nofloat", "tend", "tabort", "tx_assist"): Remove
2069 constraint letters from expanders.
2070 ("tbegin_retry", "tbegin_retry_nofloat"): Change predicate of the
2071 retry count to general_operand.
2072 ("tabort"): Give operand 0 a mode.
2073 ("tabort_1"): Add mode and constraint letter for operand 0.
2074 * doc/extend.texi: Fix protoype of __builtin_non_tx_store.
2075
2076 2013-10-04 Jeff Law <law@redhat.com>
2077
2078 * tree-ssa-threadedge.c: Fix some trailing whitespace problems.
2079
2080 * tree-ssa-threadedge.c (thread_through_normal_block): Broken
2081 out of ...
2082 (thread_across_edge): Here. Call it.
2083
2084 2013-10-04 Cary Coutant <ccoutant@google.com>
2085
2086 * dwarf2out.c (dw_sra_loc_expr): Release addr_table entries when
2087 discarding a location list expression (or a piece of one).
2088
2089 2013-10-03 Jan Hubicka <jh@suse.cz>
2090
2091 * config/i386/i386.c (ix86_issue_rate): Pentium4, Nocona has issue
2092 rate of 2. Core2, Corei7 and Haswell has issue rate of 4.
2093 (ix86_adjust_cost): Remove Atom case; fix core2/corei7/Haswell case.
2094
2095 2013-10-03 Jan Hubicka <jh@suse.cz>
2096
2097 * config/i386/i386.c (ix86_option_override_internal): Do not enable
2098 accumulate-outgoing-args when producing unwind info.
2099
2100 2013-10-03 Wei Mi <wmi@google.com>
2101
2102 * lra-constraints.c (insert_move_for_subreg): New function
2103 extracted from simplify_operand_subreg.
2104 (simplify_operand_subreg): Add reload for paradoxical subreg.
2105
2106 2013-10-03 Rong Xu <xur@google.com>
2107
2108 * ipa-inline-analysis.c (find_foldable_builtin_expect): Find
2109 the candidate of builtin_expect such that we should fix the
2110 size/time estimation.
2111 (estimate_function_body_sizes): Do the acutally size/time fix-up
2112 for builtin_expect.
2113
2114 2013-10-03 Rong Xu <xur@google.com>
2115
2116 * predict.c (tree_predict_by_opcode): Get the probability
2117 for builtin_expect from param builtin_expect_probability.
2118 * params.def (BUILTIN_EXPECT_PROBABILITY): New parameter.
2119 * predict.def (PRED_BUILTIN_EXPECT_RELAXED): Fix comments.
2120 * doc/invoke.texi: Add documentation for builtin-expect-probability.
2121
2122 2013-10-03 Marc Glisse <marc.glisse@inria.fr>
2123
2124 PR c++/19476
2125 * common.opt (fcheck-new): Moved from c.opt. Make it 'Common'.
2126 * calls.c (alloca_call_p): Use get_callee_fndecl.
2127 * fold-const.c (tree_expr_nonzero_warnv_p): Handle operator new.
2128 * tree-vrp.c (gimple_stmt_nonzero_warnv_p, stmt_interesting_for_vrp):
2129 Likewise.
2130 (vrp_visit_stmt): Remove duplicated code.
2131
2132 2013-10-03 Michael Meissner <meissner@linux.vnet.ibm.com>
2133
2134 * config/rs6000/rs6000-builtin.def (XSRDPIM): Use floatdf2,
2135 ceildf2, btruncdf2, instead of vsx_* name.
2136
2137 * config/rs6000/vsx.md (vsx_add<mode>3): Change arithmetic
2138 iterators to only do V2DF and V4SF here. Move the DF code to
2139 rs6000.md where it is combined with SF mode. Replace <VSv> with
2140 just 'v' since only vector operations are handled with these insns
2141 after moving the DF support to rs6000.md.
2142 (vsx_sub<mode>3): Likewise.
2143 (vsx_mul<mode>3): Likewise.
2144 (vsx_div<mode>3): Likewise.
2145 (vsx_fre<mode>2): Likewise.
2146 (vsx_neg<mode>2): Likewise.
2147 (vsx_abs<mode>2): Likewise.
2148 (vsx_nabs<mode>2): Likewise.
2149 (vsx_smax<mode>3): Likewise.
2150 (vsx_smin<mode>3): Likewise.
2151 (vsx_sqrt<mode>2): Likewise.
2152 (vsx_rsqrte<mode>2): Likewise.
2153 (vsx_fms<mode>4): Likewise.
2154 (vsx_nfma<mode>4): Likewise.
2155 (vsx_copysign<mode>3): Likewise.
2156 (vsx_btrunc<mode>2): Likewise.
2157 (vsx_floor<mode>2): Likewise.
2158 (vsx_ceil<mode>2): Likewise.
2159 (vsx_smaxsf3): Delete scalar ops that were moved to rs6000.md.
2160 (vsx_sminsf3): Likewise.
2161 (vsx_fmadf4): Likewise.
2162 (vsx_fmsdf4): Likewise.
2163 (vsx_nfmadf4): Likewise.
2164 (vsx_nfmsdf4): Likewise.
2165 (vsx_cmpdf_internal1): Likewise.
2166
2167 * config/rs6000/rs6000.h (TARGET_SF_SPE): Define macros to make it
2168 simpler to select whether a target has SPE or traditional floating
2169 point support in iterators.
2170 (TARGET_DF_SPE): Likewise.
2171 (TARGET_SF_FPR): Likewise.
2172 (TARGET_DF_FPR): Likewise.
2173 (TARGET_SF_INSN): Macros to say whether floating point support
2174 exists for a given operation for expanders.
2175 (TARGET_DF_INSN): Likewise.
2176
2177 * config/rs6000/rs6000.c (Ftrad): New mode attributes to allow
2178 combining of SF/DF mode operations, using both traditional and VSX
2179 registers.
2180 (Fvsx): Likewise.
2181 (Ff): Likewise.
2182 (Fv): Likewise.
2183 (Fs): Likewise.
2184 (Ffre): Likewise.
2185 (FFRE): Likewise.
2186 (abs<mode>2): Combine SF/DF modes using traditional floating point
2187 instructions. Add support for using the upper DF registers with
2188 VSX support, and SF registers with power8-vector support. Update
2189 expanders for operations supported by both the SPE and traditional
2190 floating point units.
2191 (abs<mode>2_fpr): Likewise.
2192 (nabs<mode>2): Likewise.
2193 (nabs<mode>2_fpr): Likewise.
2194 (neg<mode>2): Likewise.
2195 (neg<mode>2_fpr): Likewise.
2196 (add<mode>3): Likewise.
2197 (add<mode>3_fpr): Likewise.
2198 (sub<mode>3): Likewise.
2199 (sub<mode>3_fpr): Likewise.
2200 (mul<mode>3): Likewise.
2201 (mul<mode>3_fpr): Likewise.
2202 (div<mode>3): Likewise.
2203 (div<mode>3_fpr): Likewise.
2204 (sqrt<mode>3): Likewise.
2205 (sqrt<mode>3_fpr): Likewise.
2206 (fre<Fs>): Likewise.
2207 (rsqrt<mode>2): Likewise.
2208 (cmp<mode>_fpr): Likewise.
2209 (smax<mode>3): Likewise.
2210 (smin<mode>3): Likewise.
2211 (smax<mode>3_vsx): Likewise.
2212 (smin<mode>3_vsx): Likewise.
2213 (negsf2): Delete SF operations that are merged with DF.
2214 (abssf2): Likewise.
2215 (addsf3): Likewise.
2216 (subsf3): Likewise.
2217 (mulsf3): Likewise.
2218 (divsf3): Likewise.
2219 (fres): Likewise.
2220 (fmasf4_fpr): Likewise.
2221 (fmssf4_fpr): Likewise.
2222 (nfmasf4_fpr): Likewise.
2223 (nfmssf4_fpr): Likewise.
2224 (sqrtsf2): Likewise.
2225 (rsqrtsf_internal1): Likewise.
2226 (smaxsf3): Likewise.
2227 (sminsf3): Likewise.
2228 (cmpsf_internal1): Likewise.
2229 (copysign<mode>3_fcpsgn): Add VSX/power8-vector support.
2230 (negdf2): Delete DF operations that are merged with SF.
2231 (absdf2): Likewise.
2232 (nabsdf2): Likewise.
2233 (adddf3): Likewise.
2234 (subdf3): Likewise.
2235 (muldf3): Likewise.
2236 (divdf3): Likewise.
2237 (fred): Likewise.
2238 (rsqrtdf_internal1): Likewise.
2239 (fmadf4_fpr): Likewise.
2240 (fmsdf4_fpr): Likewise.
2241 (nfmadf4_fpr): Likewise.
2242 (nfmsdf4_fpr): Likewise.
2243 (sqrtdf2): Likewise.
2244 (smaxdf3): Likewise.
2245 (smindf3): Likewise.
2246 (cmpdf_internal1): Likewise.
2247 (lrint<mode>di2): Use TARGET_<MODE>_FPR macro.
2248 (btrunc<mode>2): Delete separate expander, and combine with the
2249 insn and add VSX instruction support. Use TARGET_<MODE>_FPR.
2250 (btrunc<mode>2_fpr): Likewise.
2251 (ceil<mode>2): Likewise.
2252 (ceil<mode>2_fpr): Likewise.
2253 (floor<mode>2): Likewise.
2254 (floor<mode>2_fpr): Likewise.
2255 (fma<mode>4_fpr): Combine SF and DF fused multiply/add support.
2256 Add support for using the upper registers with VSX and
2257 power8-vector. Move insns to be closer to the define_expands. On
2258 VSX systems, prefer the traditional form of FMA over the VSX
2259 version, since the traditional form allows the target not to
2260 overlap with the inputs.
2261 (fms<mode>4_fpr): Likewise.
2262 (nfma<mode>4_fpr): Likewise.
2263 (nfms<mode>4_fpr): Likewise.
2264
2265 2013-10-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2266 Richard Earnshaw <richard.earnshaw@arm.com>
2267
2268 * config/arm/aarch-common-protos.h (struct alu_cost_table): New.
2269 (struct mult_cost_table): Likewise.
2270 (struct mem_cost_table): Likewise.
2271 (struct fp_cost_table): Likewise.
2272 (struct vector_cost_table): Likewise.
2273 (cpu_cost_table): Likewise.
2274 * config/arm/arm.opt (mold-rts-costs): New option.
2275 (mnew-generic-costs): Likewise.
2276 * config/arm/arm.c (generic_extra_costs): New table.
2277 (cortexa15_extra_costs): Likewise.
2278 (arm_slowmul_tune): Use NULL as new costs.
2279 (arm_fastmul_tune): Likewise.
2280 (arm_strongarm_tune): Likewise.
2281 (arm_xscale_tune): Likewise.
2282 (arm_9e_tune): Likewise.
2283 (arm_v6t2_tune): Likewise.
2284 (arm_cortex_a5_tune): Likewise.
2285 (arm_cortex_a9_tune): Likewise.
2286 (arm_v6m_tune): Likewise.
2287 (arm_fa726te_tune): Likewise.
2288 (arm_cortex_a15_tune): Use cortex15_extra_costs.
2289 (arm_cortex_tune): Use generict_extra_costs.
2290 (shifter_op_p): New function.
2291 (arm_unspec_cost): Likewise.
2292 (LIBCALL_COST): Define.
2293 (arm_new_rtx_costs): New function.
2294 (arm_rtx_costs): Use arm_new_rtx_costs when core-specific
2295 table is available. Use old costs otherwise unless mnew-generic-costs
2296 is specified.
2297 * config/arm/arm-protos.h (tune_params): Add insn_extra_cost field.
2298 (cpu_cost_table): Declare.
2299
2300 2013-10-03 Marcus Shawcroft <marcus.shawcroft@arm.com>
2301
2302 PR target/58460
2303 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>)
2304 (*subs_mul_imm_<mode>)
2305 (*add_<shift>_<mode>, *add_<shift>_si_uxtw,*add_mul_imm_<mode>)
2306 (*sub_<shift>_<mode>)
2307 (*sub_<shift>_si_uxtw,*sub_mul_imm_<mode>, *sub_mul_imm_si_uxtw):
2308 Remove k constraint.
2309
2310 2013-10-03 Ian Bolton <ian.bolton@arm.com>
2311
2312 * config/aarch64/aarch64.c (aarch64_secondary_reload): Remove legacy
2313 code.
2314 * config/aarch64/aarch64.md (reload_sp_immediate): Likewise.
2315
2316 2013-10-02 Teresa Johnson <tejohnson@google.com>
2317
2318 * predict.c (probably_never_executed): New function.
2319 (probably_never_executed_bb_p): Invoke probably_never_executed.
2320 (probably_never_executed_edge_p): Ditto.
2321 * bb-reorder.c (find_rarely_executed_basic_blocks_and_crossing_edges):
2322 Treat profile insanities conservatively.
2323
2324 2013-10-02 John David Anglin <danglin@gcc.gnu.org>
2325
2326 * config.gcc (hppa*64*-*-linux*): Don't add pa/t-linux to tmake_file.
2327
2328 2013-10-02 Vladimir Makarov <vmakarov@redhat.com>
2329
2330 * lra-constraints.c (process_alt_operand): Calculate scratch_p and
2331 use it. Use smaller increase for scratch. Don't increase reject
2332 for early clobber scratch.
2333 * lra-eliminations.c (eliminate_regs_in_insn): Remove all insns
2334 setting eliminated regs except setting fp from hfp.
2335 (lra_eliminate): Check lra_insn_recog_data on NULL.
2336
2337 2013-10-02 Michael Meissner <meissner@linux.vnet.ibm.com>
2338
2339 PR target/58587
2340 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Turn off
2341 setting -mvsx-timode by default until the underlying problem is fixed.
2342 (RS6000_CPU, power7 defaults): Likewise.
2343
2344 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2345
2346 * config/x-linux (host-linux.o): Remove header dependencies.
2347 Use $(COMPILE) and $(POSTCOMPILE).
2348 * config/t-linux-android (linux-android.o): Ditto.
2349
2350 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2351
2352 * Makefile.in (expmed.o-warn): Remove.
2353
2354 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2355
2356 * graphite-scop-detection.c: Include tree-ssa-propagate,h.
2357 * graphite-sese-to-poly.c: Include tree-ssa-propagate.h.
2358
2359 2013-10-02 Teresa Johnson <tejohnson@google.com>
2360
2361 * dojump.c (do_jump_1): Divide probability between
2362 both conditions of a TRUTH_ANDIF_EXPR/TRUTH_ORIF_EXPR.
2363
2364 2013-10-02 Tom Tromey <tromey@redhat.com>
2365
2366 * Makefile.in (DRIVER_DEFINES): Use $(if), not $(and).
2367
2368 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2369
2370 * tree-flow.h: Remove some prototypes.
2371 * tree-ssa-dce.c (mark_virtual_operand_for_renaming,
2372 mark_virtual_phi_result_for_renaming): Move to tree-into-ssa.c.
2373 * tree-into-ssa.c (mark_virtual_operand_for_renaming,
2374 mark_virtual_phi_result_for_renaming): Relocate here.
2375 * tree-into-ssa.h: Add prototypes.
2376 * tree-ssa-phiopt.c: (tree_ssa_phiopt_worker) Use
2377 single_pred_before_succ_order.
2378 (blocks_in_phiopt_order): Rename and move to cfganal.c.
2379 (nonfreeing_call_p) Move to gimple.c.
2380 * cfganal.c (single_pred_before_succ_order): Move and renamed from
2381 tree-ssa-phiopt.c.
2382 * basic-block.h (single_pred_before_succ_order): Add prototype.
2383 * gimple.c (nonfreeing_call_p): Relocate here.
2384 * gimple.h: Add prototype.
2385 * tree-ssa-ifcombine.c: Include tree-ssa-phiopt.h.
2386 * tree-ssa-dom.h: New file. Relocate prototypes here.
2387 * tree-ssa.h: Include tree-ssa-dom.h.
2388
2389 2013-10-02 Uros Bizjak <ubizjak@gmail.com>
2390
2391 * config/i386/x-i386 (driver-i386.o): Remove header dependencies.
2392 Use $(COMPILE) and $(POSTCOMPILE).
2393
2394 * config/alpha/x-alpha (driver-alpha.o): Ditto.
2395
2396 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2397
2398 * tree-flow.h: Remove some prototypes.
2399 * gimple-fold.h: Add prototypes from gimple.h and tree-flow.h.
2400 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2401 * tree-ssa-copy.c (may_propagate*, propagate_value, replace_exp,
2402 propagate_tree_value*): Move from here to...
2403 * tree-ssa-propagate.c (may_propagate*, propagate_value, replace_exp,
2404 propagate_tree_value*): Relocate here.
2405 * tree-ssa-propagate.h: Relocate prototypes from tree-flow.h.
2406 * gimple.h: Include gimple-fold.h, move prototypes into gimple-fold.h.
2407 * gimple-fold.c: Remove gimple-fold.h from include list.
2408 * tree-vrp.c: Remove gimple-fold.h from include list.
2409 * tree-ssa-sccvn.c: Remove gimple-fold.h from include list.
2410 * tree-ssa-ccp.c: Remove gimple-fold.h from include list.
2411 * tree-scalar-evolution.c: Add tree-ssa-propagate.h to include list.
2412 * tree-ssa-pre.c: Add tree-ssa-propagate.h to include list.
2413 * sese.c: Add tree-ssa-propagate.h to include list.
2414
2415 2013-10-02 Richard Biener <rguenther@suse.de>
2416
2417 * tree-loop-distribution.c: Include tree-vectorizer.h for
2418 find_loop_location.
2419 (enum partition_kind): Remove PKIND_REDUCTION.
2420 (struct partition_s): Remove has_writes member, add reduction_p member.
2421 (partition_alloc): Adjust.
2422 (partition_builtin_p): Likewise.
2423 (partition_has_writes): Remove.
2424 (partition_reduction_p): New function.
2425 (partition_merge_into): Likewise.
2426 (generate_code_for_partition): Commonize builtin partition
2427 handling tail.
2428 (rdg_cannot_recompute_vertex_p): Remove.
2429 (already_processed_vertex_p): Likewise.
2430 (rdg_flag_vertex): Do not set has_writes.
2431 (classify_partition): Adjust.
2432 (rdg_build_partitions): Do not set has_writes, treat all
2433 partitions as useful.
2434 (distribute_loop): Record number of library calls generated. Adjust.
2435 (tree_loop_distribution): Report number of loops and library
2436 calls generated as opt-info.
2437
2438 2013-10-02 Andrew MacLeod <amacleod@redhat.com>
2439
2440 * tree-flow.h: Include new .h files. Move prototypes.
2441 * tree-cfgcleanup.h: New file. Add prototypes from tree-flow.h.
2442 * tree-dfa.h: New File. Add prototypes from tree-flow.h.
2443 (get_addr_base_and_unit_offset_1) Move from tree-flow-inline.h.
2444 * tree-pretty-print.h: Add prototypes from tree-flow.h.
2445 * tree-into-ssa.h: New File. Add prototypes from tree-flow.h.
2446 ({debug|dump}*): Move debugging prototypes out of tree-into-ssa.c.
2447 * tree-into-ssa.c ({debug|dump}*): Move prototypes to header file.
2448 * tree.h (get_ref_base_and_extent): Move prototype out.
2449 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Move to
2450 tree-dfa.h.
2451 * gimple-low.h: New File. Add prototypes from tree-flow.h.
2452 * gimple-low.c (try_catch_may_fallthru, block_may_fallthru): Move to...
2453 * tree.c (try_catch_may_fallthru, block_may_fallthru): Here.
2454 * tree-scalar-evolution.c: Include tree.h.
2455 * sese.c: Include tree.h.
2456 * dumpfile.c: Move gimple-pretty-print.h include after tree.h.
2457 * dwarf2out.c: Include tree-dfa.h.
2458 * tree-chrec.c: Include tree.h.
2459 * tree-data-ref.c: Include tree.h.
2460
2461 2013-10-02 Yufeng Zhang <yufeng.zhang@arm.com>
2462
2463 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref):
2464 Fix whitespace.
2465
2466 2013-10-02 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2467
2468 * config/t-sol2 (sol2-c.o): Remove header dependencies.
2469 Use $(COMPILE) and $(POSTCOMPILE).
2470 (sol2-cxx.o): Likewise.
2471 (sol2-stubs.o): Likewise.
2472 (sol2.o): Likewise.
2473 * config/x-solaris (host-solaris.o): Likewise.
2474
2475 * config/sparc/t-sparc (sparc.o): Remove.
2476 (sparc-c.o): Remove header dependencies.
2477 Use $(COMPILE) and $(POSTCOMPILE).
2478 * config/sparc/x-sparc: Likewise.
2479
2480 2013-10-02 Joern Rennecke <joern.rennecke@embecosm.com>
2481
2482 * config/arc/arc-opts.h: Add 2013 to Copyright years.
2483 * config/arc/arc700.md: Likewise.
2484 * config/arc/arc-modes.def: Likewise.
2485 * config/arc/arc-simd.h: Likewise.
2486 * config/arc/t-arc-uClibc: Likewise.
2487 * config/arc/t-arc-newlib: Likewise.
2488
2489 2013-10-02 Renlin Li <renlin.li@arm.com>
2490
2491 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
2492 plus_constant.
2493 (aarch64_expand_epilogue): Likewise.
2494
2495 2013-10-02 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
2496 Yufeng Zhang <yufeng.zhang@arm.com>
2497
2498 * gimple-ssa-strength-reduction.c (legal_cast_p_1): Forward
2499 declaration.
2500 (backtrace_base_for_ref): Call get_unwidened with 'base_in' if
2501 'base_in' represent a conversion and legal_cast_p_1 holds; set
2502 'base_in' with the returned value from get_unwidened.
2503
2504 2013-10-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2505
2506 * config/arm/arm.c (arm_legitimize_reload_address): Explain why
2507 plus_constant is not used.
2508
2509 2013-10-01 Wei Mi <wmi@google.com>
2510
2511 * config/i386/x86-tune.def (DEF_TUNE): Remove m_CORE_ALL.
2512 * config/i386/i386.md: Add define_peephole2 to
2513 break partial reg stall for cvtss2sd/cvtsd2ss.
2514
2515 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2516
2517 * config/arc/arc.c (pass_arc_ifcvt::clone):
2518 Update for ctxt_ -> m_ctxt change.
2519
2520 2013-10-01 Jeff Law <law@redhat.com>
2521
2522 * tree-ssa-threadupdate.c (struct redirection_data): Delete
2523 outgoing_edge and intermediate_edge fields. Instead store the path.
2524 (redirection_data::hash): Hash on the last edge's destination index.
2525 (redirection_data::equal): Check the entire thread path.
2526 (lookup_redirectio_data): Corresponding changes.
2527 (create_edge_and_update_destination_phis): Likewise.
2528 (thread_single_edge): Likewise.
2529
2530 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2531 Diego Novillo <dnovillo@google.com>
2532
2533 * config/arc/simdext.md (UNSPEC_ARC_SIMD_VLD32WH): Delete.
2534 (UNSPEC_ARC_SIMD_VLD32WL): Likewise.
2535 (vld32wh_insn, vld32wl_insn): Delete commented-out old
2536 versions of these patterns.
2537
2538 * doc/extend.texi (long_call/medium_call/short_call): Typo fix.
2539 (__builtin_arc_aligned): Likewise.
2540
2541 * config/arc/arc.md: Expand adc_0 comment stating the intended
2542 purpose and why it isn't ready.
2543 Replace commented out call_value_via_label_mixed with a
2544 plain comment about bl_s.
2545
2546 * config/arc/arc.c (stdio.h): Don't include directly.
2547 (arc_expand_epilogue): Remove [0]: Remove fp_restored_p.
2548 Remove if (1) condition.
2549 (arc_encode_section_info): Fix comment.
2550
2551 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2552
2553 * config/arc/arc.c (arc_conditional_register_usage):
2554 Use ARC_FIRST_SIMD_VR_REG / ARC_LAST_SIMD_VR_REG.
2555 Also set reg_alloc_order for DMA config regs.
2556
2557 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2558 Jeremy Bennett <jeremy.bennett@embecosm.com>
2559
2560 * doc/install.texi (--with-cpu): Mention ARC.
2561 (arc-*-elf32): New paragraph.
2562 (arc-linux-uclibc): Likewise.
2563 * doc/md.texi (Machine Constraints): Add ARC part.
2564 * doc/invoke.texi: (menu): Add ARC Options.
2565 (Machine Dependent Options) <ARC Options>: Add synopsis.
2566 (node ARC Options): Add.
2567 * doc/extend.texi (long_call / short_call attribute): Add ARC.
2568 (ARC Built-in Functions): New section defining
2569 generic ARC built-in functions.
2570 (ARC SIMD Built-in Functions): New section defining SIMD specific
2571 built-in functions.
2572 (Declaring Attributes of Functions): Extended
2573 description of short_call and long_call attributes for ARC and
2574 added index entries.
2575
2576 2013-10-01 Saurabh Verma <saurabh.verma@codito.com>
2577 Ramana Radhakrishnan <ramana.radhakrishnan@codito.com>
2578 Joern Rennecke <joern.rennecke@embecosm.com>
2579 Muhammad Khurram Riaz <khurram.riaz@arc.com>
2580 Brendan Kehoe <brendan@zen.org>
2581 Michael Eager <eager@eagercon.com>
2582 Simon Cook <simon.cook@embecosm.com>
2583 Jeremy Bennett <jeremy.bennett@embecosm.com>
2584
2585 * config/arc, common/config/arc: New directories.
2586
2587 2013-10-01 Joern Rennecke <joern.rennecke@embecosm.com>
2588 Brendan Kehoe <brendan@zen.org>
2589 Simon Cook <simon.cook@embecosm.com>
2590
2591 * config.gcc (arc*-*-elf*, arc*-*-linux-uclibc*): New configurations.
2592
2593 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2594
2595 * tree-ssa-live.h (coalesce_ssa_name): Move Prototype to...
2596 * tree-ssa-coalesce.h: New. Move prototype to here.
2597 * tree-outof-ssa.h: Include tree-ssa-coalesce.h.
2598 * tree-ssa-coalesce.c: Include tree-outof-ssa.h.
2599 (gimple_can_coalesce_p): Move to...
2600 * gimple.c (gimple_can_coalesce_p): Here.
2601
2602 2013-10-01 Andrew MacLeod <amacleod@redhat.com>
2603
2604 * tree-into-ssa.c (enum need_phi_state): Relocate from tree-flow.h.
2605 (dump_decl_set): Move to gimple.c.
2606 * gimple.h: Don't include tree-ssa-operands.h.
2607 (dump_decl_set): Add prototype.
2608 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2609 Move to gimple-ssa.h.
2610 (phi_ssa_name_p, phi_nodes, phi_nodes_ptr, gimple_phi_arg_def,
2611 gimple_phi_arg_def_ptr, gimple_phi_arg_edge, gimple_phi_arg_location,
2612 gimple_phi_arg_location_from_edge, gimple_phi_arg_set_location,
2613 gimple_phi_arg_has_location): Relocate from tree-flow-inline.h
2614 * gimple.c (walk_stmt_load_store_ops): Use gimple_phi_arg_def rather
2615 than PHI_ARG_DEF.
2616 (dump_decl_set): Relocate here.
2617 * gimple-ssa.h: New file.
2618 (gimple_vuse_op, gimple_vdef_op, update_stmt, update_stmt_if_modified):
2619 Relocate from gimple.h.
2620 * tree-cfg.c (has_zero_uses_1, single_imm_use_1): Move to...
2621 * tree-ssa-operands.c (swap_ssa_operands): Rename from
2622 swap_tree_operands and remove non-ssa path.
2623 (has_zero_uses_1, single_imm_use_1): Relocate from tree-cfg.c.
2624 * tree-ssa-reassoc.c (linearize_expr_tree, repropagate_negates): Use
2625 swap_ssa_operands.
2626 * tree-vect-loop.c (destroy_loop_vec_info, vect_is_slp_reduction,
2627 vect_is_simple_reduction_1): Use swap_ssa_operands.
2628 * tree-flow.h: Move various prototypes to tree-phinodes.h.
2629 (enum need_phi_state): Move to tree-into-ssa.c.
2630 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2631 BREAK_FROM_IMM_USE_STMT): Move to ssa-iterators.h.
2632 (swap_tree_operands): Rename and move prototype to tree-ssa-operands.h.
2633 * tree-flow-inline.h (delink_imm_use, link_imm_use_to_list,
2634 link_imm_use, set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2635 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2636 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2637 num_imm_uses): Move to ssa-iterators.h.
2638 (get_use_from_ptr, get_def_from_ptr): Move to tree-ssa-operands.h
2639 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Move to
2640 tree-phinodes.h.
2641 (op_iter_done, op_iter_next_def, op_iter_next_tree,
2642 clear_and_done_ssa_iter, op_iter_init, op_iter_init_use,
2643 op_iter_init_def, op_iter_init_tree, single_ssa_tree_operand,
2644 single_ssa_use_operand, single_ssa_def_operand, zero_ssa_operands,
2645 num_ssa_operands, delink_stmt_imm_use, single_phi_def,
2646 op_iter_init_phiuse, op_iter_init_phidef, end_imm_use_stmt_p,
2647 end_imm_use_stmt_traverse, move_use_after_head, link_use_stmts_after,
2648 first_imm_use_stmt, next_imm_use_stmt, first_imm_use_on_stmt,
2649 end_imm_use_on_stmt_p, next_imm_use_on_stmt): Move to ssa-iterators.h.
2650 (gimple_phi_arg_def, gimple_phi_arg_def_ptr, gimple_phi_arg_edge,
2651 gimple_phi_arg_location, gimple_phi_arg_location_from_edge,
2652 gimple_phi_arg_set_location, gimple_phi_arg_has_location, phi_nodes,
2653 phi_nodes_ptr, phi_ssa_name_p): Move to gimple.h.
2654 (set_phi_nodes): Move to tree-phinodes.h.
2655 * tree-ssa-operands.h (enum ssa_op_iter_type,
2656 struct ssa_operand_iterator_d, SSA_OP*, FOR_EACH_SSA*, SINGLE_SSA*,
2657 ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS): Move to ssa-iterators.h.
2658 (dump_decl_set): Remove prototype.
2659 (get_use_from_ptr, get_def_from_ptr): Relocate from tree-flow.h.
2660 * tree-phinodes.h: New file. Move some prototypes from tree-flow.h.
2661 (set_phi_nodes): Relocate from tree-flow-inline.h.
2662 (gimple_phi_arg_imm_use_ptr, phi_arg_index_from_use): Relocate from
2663 tree-flow-inline.h
2664 * tree-ssa.h: Add tree-phinodes.h, gimple-ssa.h, ssa-iterators.h to
2665 include list. Temporarily add gimple.h to include list.
2666 * ssa-iterators.h: New file.
2667 (struct immediate_use_iterator_d, FOR_EACH_IMM_*,
2668 BREAK_FROM_IMM_USE_STMT): Relocate from tree-flow.h.
2669 (enum ssa_op_iter_type, struct ssa_operand_iterator_d, SSA_OP*,
2670 FOR_EACH_SSA*, SINGLE_SSA*, ZERO_SSA_OPERANDS, NUM_SSA_OPERANDS):
2671 Relocate from tree-ssa-operands.h.
2672 (delink_imm_use, link_imm_use_to_list, link_imm_use,
2673 set_ssa_use_from_ptr, link_imm_use_stmt, relink_imm_use,
2674 relink_imm_use_stmt, end_readonly_imm_use_p, first_readonly_imm_use,
2675 next_readonly_imm_use, has_zero_uses, has_single_use, single_imm_use,
2676 num_imm_uses, get_use_from_ptr, get_def_from_ptr,
2677 phi_arg_index_from_use, op_iter_done, op_iter_next_def,
2678 op_iter_next_tree, clear_and_done_ssa_iter, op_iter_init,
2679 op_iter_init_use, op_iter_init_def, op_iter_init_tree,
2680 single_ssa_tree_operand, single_ssa_use_operand, single_ssa_def_operand,
2681 zero_ssa_operands, num_ssa_operands, delink_stmt_imm_use,
2682 single_phi_def, op_iter_init_phiuse, op_iter_init_phidef,
2683 end_imm_use_stmt_p, end_imm_use_stmt_traverse, move_use_after_head,
2684 link_use_stmts_after, first_imm_use_stmt, next_imm_use_stmt,
2685 first_imm_use_on_stmt, end_imm_use_on_stmt_p, next_imm_use_on_stmt):
2686 Relocate from tree-flow-inline.h.
2687 * tree-outof-ssa.h: Change _SSAEXPAND_H macro to GCC_TREE_OUTOF_SSA_H.
2688
2689 2013-10-01 Vidya Praveen <vidyapraveen@arm.com>
2690
2691 * aarch64-simd.md
2692 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l2<mode>_internal): Rename to ...
2693 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal): ... this;
2694 Insert '\t' to output template.
2695 (aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_lo_internal): New.
2696 (aarch64_saddl2<mode>, aarch64_uaddl2<mode>): Modify to call
2697 gen_aarch64_<ANY_EXTEND:su><ADDSUB:optab>l<mode>_hi_internal() instead.
2698 (aarch64_ssubl2<mode>, aarch64_usubl2<mode>): Ditto.
2699
2700 2013-10-01 Uros Bizjak <ubizjak@gmail.com>
2701
2702 * doc/install.texi (Host/target specific installation notes for GCC):
2703 Put @anchor before @heading.
2704 * doc/gcc.texi (titlepage): Use @uref and http:// prefix for website.
2705 Use @email for email addresses.
2706
2707 2013-10-01 Jeff Law <law@redhat.com>
2708
2709 * tree-ssa-threadedge.c (thread_across_edge): Make path a pointer to
2710 a vec. Only delete the path if we create one without successfully
2711 registering a jump thread.
2712 * tree-ssa-threadupdate.h (register_jump_thread): Pass in path vector
2713 as a pointer.
2714 * tree-ssa-threadupdate.c (threaded_edges): Remove. No longer used
2715 (paths): New vector of jump threading paths.
2716 (THREAD_TARGET, THREAD_TARGET2): Remove accessor macros.
2717 (THREAD_PATH): New accessor macro for the entire thread path.
2718 (lookup_redirection_data): Get intermediate and final outgoing edge
2719 from the thread path.
2720 (create_edge_and_update_destination_phis): Copy the threading path.
2721 (ssa_fix_duplicate_block_edges): Get edges and block types from the
2722 jump threading path.
2723 (ssa_redirect_edges): Get edges and block types from the jump threading
2724 path. Free the path vector.
2725 (thread_block): Get edges from the jump threading path. Look at the
2726 entire path to see if we thread to a loop exit. If we cancel a jump
2727 thread request, then free the path vector.
2728 (thread_single_edge): Get edges and block types from the jump threading
2729 path. Free the path vector.
2730 (thread_through_loop_header): Get edges and block types from the jump
2731 threading path. Free the path vector.
2732 (mark_threaded_blocks): Iterate over the vector of paths and store
2733 the path on the appropriate edge. Get edges and block types from the
2734 jump threading path.
2735 (mark_threaded_blocks): Get edges and block types from the jump
2736 threading path. Free the path vector.
2737 (thread_through_all_blocks): Use the vector of paths rather than
2738 a vector of 3-edge sets.
2739 (register_jump_thread): Accept pointer to a path vector rather
2740 than the path vector itself. Store the path vector for later use.
2741 Simplify.
2742
2743 2013-10-01 Jakub Jelinek <jakub@redhat.com>
2744 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
2745
2746 PR target/58574
2747 * config/s390/s390.c (s390_split_branches): Modify check for table
2748 jump insns.
2749 (s390_chunkify_start): Rearrange table jump insn check in order to
2750 deal with compare and branch insns correctly.
2751
2752 2013-10-01 Kugan Vivekanandarajah <kuganv@linaro.org>
2753
2754 PR target/58578
2755 Revert
2756 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
2757 * config/arm/arm.md (arm_ashldi3_1bit): define_insn into
2758 define_insn_and_split.
2759 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
2760 (shiftsi3_compare): New pattern.
2761 (rrx): New pattern.
2762 * config/arm/unspecs.md (UNSPEC_RRX): New.
2763
2764 2013-10-01 Alan Modra <amodra@gmail.com>
2765
2766 * stmt.c (expand_asm_operands): Revert part of 2013-09-24 special
2767 casing inout operands.
2768
2769 2013-10-01 Richard Biener <rguenther@suse.de>
2770
2771 PR tree-optimization/58553
2772 * tree-loop-distribution.c (struct partition_s): Add niter member.
2773 (classify_partition): Populate niter member for the partition
2774 and properly identify whether the relevant store happens before
2775 or after the loop exit.
2776 (generate_memset_builtin): Use niter member from the partition.
2777 (generate_memcpy_builtin): Likewise.
2778
2779 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
2780
2781 * vec.h (vec_prefix, vec): Prefix member names with "m_".
2782 * vec.c (vec_prefix::calculate_allocation): Update accordingly.
2783
2784 2013-09-30 Richard Sandiford <rdsandiford@googlemail.com>
2785
2786 * basic-block.h (edge_list): Prefix member names with "m_".
2787 * context.h (context): Likewise.
2788 * domwalk.h (dom_walker): Likewise.
2789 * gengtype-state.c (s_expr_writer, state_writer): Likewise.
2790 * graphite-sese-to-poly.c (sese_dom_walker): Likewise.
2791 * hash-table.h (hash_table): Likewise.
2792 * machmode.h (bit_field_mode_iterator): Likewise.
2793 * pass_manager.h (pass_list): Likewise.
2794 * tree-into-ssa.c (mark_def_dom_walker): Likewise.
2795 * tree-pass.h (pass_data): Likewise.
2796 * tree-ssa-dom.c (dom_opt_dom_walker): Likewise.
2797 * tree-ssa-phiopt.c (nontrapping_dom_walker): Likewise,
2798 * tree-ssa-uncprop.c (uncprop_dom_walker): Likewise.
2799 * asan.c (pass_data_asan): Update accordingly.
2800 * cfganal.c (control_dependences::find_control_dependence): Likewise.
2801 (control_dependences::control_dependences): Likewise.
2802 (control_dependences::~control_dependences): Likewise.
2803 (control_dependences::~control_dependences): Likewise.
2804 (control_dependences::get_edges_dependent_on): Likewise.
2805 * cgraphbuild.c (pass_data_rebuild_cgraph_edges::clone): Likewise.
2806 (pass_data_remove_cgraph_callee_edges::clone): Likewise.
2807 * context.c (gcc::context::context): Likewise.
2808 * cprop.c (pass_rtl_cprop::clone): Likewise.
2809 * domwalk.c (dom_walker::walk): Likewise.
2810 * ipa-inline-analysis.c (pass_inline_parameters::clone): Likewise.
2811 * ipa-pure-const.c (pass_local_pure_const::clone): Likewise.
2812 * mode-switching.c (pass_mode_switching::clone): Likewise.
2813 * passes.c (opt_pass::opt_pass): Likewise.
2814 (pass_manager::pass_manager): Likewise.
2815 * predict.c (pass_strip_predict_hints::clone): Likewise.
2816 * recog.c (pass_data pass_data_peephole2::clone): Likewise.
2817 (pass_split_all_insns::clone): Likewise.
2818 * stor-layout.c (bit_field_mode_iterator::bit_field_mode_iterator):
2819 Likewise.
2820 (bit_field_mode_iterator::next_mode): Likewise.
2821 (bit_field_mode_iterator::prefer_smaller_modes): Likewise.
2822 * tree-cfg.c (pass_split_crit_edges::clone): Likewise.
2823 * tree-cfgcleanup.c (pass_merge_phi::clone): Likewise.
2824 * tree-complex.c (pass_lower_complex::clone): Likewise.
2825 * tree-eh.c (pass_cleanup_eh::clone): Likewise.
2826 * tree-object-size.c (pass_object_sizes::clone): Likewise.
2827 * tree-optimize.c (pass_fixup_cfg::clone): Likewise.
2828 * tree-ssa-ccp.c (pass_data_ccp::clone): Likewise.
2829 (pass_fold_builtins::clone): Likewise.
2830 * tree-ssa-copy.c (pass_data_copy_prop::clone): Likewise.
2831 * tree-ssa-copyrename.c (pass_rename_ssa_copies::clone): Likewise.
2832 * tree-ssa-dce.c (pass_dce::clone, pass_dce_loop::clone): Likewise.
2833 (pass_cd_dce::clone): Likewise.
2834 * tree-ssa-dom.c (pass_dominator::clone): Likewise.
2835 (pass_phi_only_cprop::clone): Likewise.
2836 * tree-ssa-dse.c (pass_dse::clone): Likewise.
2837 * tree-ssa-forwprop.c (pass_forwprop::clone): Likewise.
2838 * tree-ssa-loop.c (pass_lim::clone): Likewise.
2839 * tree-ssa-phiopt.c (pass_phiopt::clone): Likewise.
2840 * tree-ssa-pre.c (pass_fre::clone): Likewise.
2841 * tree-ssa-reassoc.c (pass_reassoc::clone): Likewise.
2842 * tree-ssa-uninit.c (pass_late_warn_uninitialized::clone): Likewise.
2843 * tree-tailcall.c (pass_tail_recursion::clone): Likewise.
2844 * tree-vect-generic.c (pass_lower_vector_ssa::clone): Likewise.
2845 * tree-vrp.c (pass_vrp::clone): Likewise.
2846 * tsan.c (pass_tsan::clone): Likewise.
2847
2848 2013-09-30 Jakub Jelinek <jakub@redhat.com>
2849
2850 PR middle-end/58564
2851 * fold-const.c (tree_unary_nonnegative_warnv_p): Use
2852 INTEGRAL_TYPE_P (t) instead of TREE_CODE (t) == INTEGER_TYPE.
2853
2854 PR middle-end/58564
2855 * fold-const.c (fold_ternary_loc): For A < 0 : <sign bit of A> : 0
2856 optimization, punt if sign_bit_p looked through any zero extension.
2857
2858 2013-09-30 Teresa Johnson <tejohnson@google.com>
2859
2860 * tree-ssa-threadupdate.c (ssa_fix_duplicate_block_edges):
2861 Update redirected out edge count in joiner case.
2862 (ssa_redirect_edges): Common the joiner and non-joiner cases
2863 so that joiner case gets profile updates.
2864
2865 2013-09-30 Richard Biener <rguenther@suse.de>
2866
2867 PR tree-optimization/58554
2868 * tree-loop-distribution.c (classify_partition): Require
2869 unconditionally executed stores for memcpy and memset recognition.
2870 (tree_loop_distribution): Calculate dominance info.
2871
2872 2013-09-30 Venkataramanan Kumar <venkataramanan.kumar@linaro.org>
2873
2874 * config/aarch64/aarch64.h (MCOUNT_NAME): Define.
2875 (NO_PROFILE_COUNTERS): Likewise.
2876 (PROFILE_HOOK): Likewise.
2877 (FUNCTION_PROFILER): Likewise.
2878 * config/aarch64/aarch64.c (aarch64_function_profiler): Remove.
2879
2880 2013-09-30 Iain Sandoe <iain@codesourcery.com>
2881
2882 * config/rs6000/darwin.md (load_macho_picbase_si): Wrap machopic
2883 calls and defines in TARGET_MACHO conditional.
2884 (load_macho_picbase_di): Likewise.
2885 (reload_macho_picbase): Likewise.
2886 (reload_macho_picbase_si): Likewise.
2887 (reload_macho_picbase_di): Likewise.
2888 (nonlocal_goto_receiver): Likewise.
2889
2890 2013-09-30 Nick Clifton <nickc@redhat.com>
2891
2892 * config/msp430/msp430.c (msp430x_names): New array. Lists MCUs
2893 that use the MSP430X ISA.
2894 (msp430_option_override): Scan -mmcu command line option for any
2895 MCU name that supports the MSP430X ISA.
2896 * config/msp430/t-msp430 (MULTILIB_MATCHES): Add matches for known
2897 -mmcu options which enable the MSP430X ISA.
2898
2899 2013-09-30 Richard Biener <rguenther@suse.de>
2900
2901 PR middle-end/58532
2902 * tree-cfg.c (make_abnormal_goto_edges): Skip debug statements
2903 before looking for setjmp-like calls.
2904
2905 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2906
2907 PR target/10901
2908 * config/darwin-protos.h (machopic_get_function_picbase): New.
2909 * config/darwin.c (machopic_get_function_picbase): New.
2910 * config/rs6000/darwin.md (load_macho_picbase_si): Update picbase
2911 label for a new func. (load_macho_picbase_di): Likewise.
2912 (reload_macho_picbase): New expand.
2913 (reload_macho_picbase_si): New insn.
2914 (reload_macho_picbase_di): New insn.
2915 (nonlocal_goto_receiver): New define and split.
2916 * config/rs6000/rs6000.md (unspec enum): Add UNSPEC_RELD_MPIC.
2917 (unspecv enum): Add UNSPECV_NLGR.
2918
2919 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2920
2921 * config/rs6000/rs6000.c (rs6000_init_dwarf_reg_sizes_extra): Ensure
2922 that altivec registers are correctly sized on Darwin.
2923
2924 2013-09-29 Iain Sandoe <iain@codesourcery.com>
2925
2926 * config/t-darwin (darwin.o, darwin-c.o, darwin-f.o,
2927 darwin-driver.o): Use COMPILE and POSTCOMPILE.
2928 * config/x-darwin (host-darwin.o): Likewise.
2929 * config/i386/x-darwin (host-i386-darwin.o): Likewise.
2930 * config/rs6000/x-darwin (host-ppc-darwin.o): Likewise.
2931 * config/rs6000/x-darwin64 (host-ppc64-darwin.o): Likewise.
2932
2933 2013-09-29 Uros Bizjak <ubizjak@gmail.com>
2934
2935 * doc/invoke.texi: Fix usage of @tie{} command.
2936
2937 2013-09-29 Eric Botcazou <ebotcazou@adacore.com>
2938
2939 * config/sparc/sync.md: Add peephole for consecutive memory barriers.
2940
2941 2013-09-28 Jan Hubicka <jh@suse.cz>
2942
2943 * config/i386/x86-tune.def: Add documentation for each of the options;
2944 add whitespace.
2945
2946 2013-09-28 Jan Hubicka <jh@suse.cz>
2947
2948 * x86-tune.def (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Enable for
2949 generic.
2950 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Likewise.
2951 (X86_TUNE_FOUR_JUMP_LIMIT): Drop for generic and buldozer.
2952 (X86_TUNE_PAD_RETURNS): Drop for buldozer chips.
2953 (X86_TUNE_AVOID_VECTOR_DECODE): Drop for generic.
2954 (X86_TUNE_REASSOC_FP_TO_PARALLEL): Enable for generic.
2955
2956 2013-09-28 Richard Sandiford <rdsandiford@googlemail.com>
2957
2958 * alloc-pool.c, asan.c, auto-inc-dec.c, basic-block.h, bb-reorder.c,
2959 bitmap.c, bitmap.h, bt-load.c, builtins.c, calls.c, cfgcleanup.c,
2960 cfgexpand.c, cfghooks.c, cfgloop.c, cfgloopmanip.c, cfgrtl.c, cgraph.c,
2961 cgraph.h, cgraphbuild.c, cgraphclones.c, cgraphunit.c, collect2.c,
2962 combine-stack-adj.c, combine.c, compare-elim.c, context.c, context.h,
2963 cprop.c, cse.c, cselib.c, dbxout.c, dce.c, defaults.h, df-core.c,
2964 df-problems.c, df-scan.c, df.h, diagnostic.c, double-int.c, dse.c,
2965 dumpfile.c, dwarf2asm.c, dwarf2cfi.c, dwarf2out.c, emit-rtl.c,
2966 errors.c, except.c, expmed.c, expr.c, file-find.c, final.c,
2967 fixed-value.c, fold-const.c, function.c, fwprop.c, gcc-ar.c, gcc.c,
2968 gcov-io.c, gcov-io.h, gcov.c, gcse.c, genattr-common.c, genattr.c,
2969 genattrtab.c, genautomata.c, genconfig.c, genemit.c, genextract.c,
2970 genflags.c, gengenrtl.c, gengtype-state.c, gengtype.c, genmodes.c,
2971 genopinit.c, genoutput.c, genpeep.c, genpreds.c, genrecog.c,
2972 gensupport.c, ggc-common.c, ggc-page.c, gimple-fold.c, gimple-low.c,
2973 gimple-pretty-print.c, gimple-ssa-strength-reduction.c, gimple.c,
2974 gimple.h, godump.c, graphite-clast-to-gimple.c,
2975 graphite-optimize-isl.c, graphite-poly.h, graphite-sese-to-poly.c,
2976 graphite.c, haifa-sched.c, hash-table.c, hash-table.h, hwint.c,
2977 hwint.h, ifcvt.c, incpath.c, init-regs.c, input.h, intl.c, intl.h,
2978 ipa-cp.c, ipa-devirt.c, ipa-inline-analysis.c, ipa-inline.c,
2979 ipa-profile.c, ipa-pure-const.c, ipa-reference.c, ipa-split.c,
2980 ipa-utils.c, ipa.c, ira-build.c, ira.c, jump.c, loop-doloop.c,
2981 loop-init.c, loop-invariant.c, loop-iv.c, lower-subreg.c, lto-cgraph.c,
2982 lto-streamer-in.c, lto-streamer-out.c, lto-wrapper.c, mcf.c,
2983 mode-switching.c, modulo-sched.c, omp-low.c, optabs.c, opts.c,
2984 pass_manager.h, passes.c, plugin.c, postreload-gcse.c, postreload.c,
2985 predict.c, prefix.c, pretty-print.c, print-rtl.c, print-tree.c,
2986 profile.c, read-md.c, real.c, real.h, recog.c, ree.c, reg-stack.c,
2987 regcprop.c, reginfo.c, regmove.c, regrename.c, regs.h, regstat.c,
2988 reload1.c, reorg.c, rtl.c, rtl.h, rtlanal.c, sbitmap.c, sched-rgn.c,
2989 sdbout.c, sel-sched-ir.c, sel-sched.c, sparseset.c, stack-ptr-mod.c,
2990 statistics.c, stmt.c, stor-layout.c, store-motion.c, streamer-hooks.h,
2991 system.h, target-hooks-macros.h, targhooks.c, targhooks.h, toplev.c,
2992 tracer.c, trans-mem.c, tree-browser.c, tree-call-cdce.c, tree-cfg.c,
2993 tree-cfgcleanup.c, tree-complex.c, tree-data-ref.c, tree-data-ref.h,
2994 tree-eh.c, tree-emutls.c, tree-flow.h, tree-if-conv.c, tree-into-ssa.c,
2995 tree-iterator.c, tree-loop-distribution.c, tree-mudflap.c,
2996 tree-nested.c, tree-nomudflap.c, tree-nrv.c, tree-object-size.c,
2997 tree-optimize.c, tree-pass.h, tree-pretty-print.c, tree-profile.c,
2998 tree-scalar-evolution.c, tree-sra.c, tree-ssa-ccp.c,
2999 tree-ssa-coalesce.c, tree-ssa-copy.c, tree-ssa-copyrename.c,
3000 tree-ssa-dce.c, tree-ssa-dom.c, tree-ssa-dse.c, tree-ssa-forwprop.c,
3001 tree-ssa-ifcombine.c, tree-ssa-live.c, tree-ssa-loop-ch.c,
3002 tree-ssa-loop-im.c, tree-ssa-loop-ivopts.c, tree-ssa-loop-prefetch.c,
3003 tree-ssa-loop.c, tree-ssa-math-opts.c, tree-ssa-operands.c,
3004 tree-ssa-phiopt.c, tree-ssa-phiprop.c, tree-ssa-pre.c,
3005 tree-ssa-reassoc.c, tree-ssa-sink.c, tree-ssa-strlen.c,
3006 tree-ssa-structalias.c, tree-ssa-threadedge.c, tree-ssa-threadupdate.c,
3007 tree-ssa-uncprop.c, tree-ssa-uninit.c, tree-ssa.c, tree-ssanames.c,
3008 tree-stdarg.c, tree-switch-conversion.c, tree-tailcall.c,
3009 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
3010 tree-vect-stmts.c, tree-vectorizer.c, tree-vectorizer.h, tree-vrp.c,
3011 tree.c, tree.h, tsan.c, tsystem.h, value-prof.c, var-tracking.c,
3012 varasm.c, vec.h, vmsdbgout.c, vtable-verify.c, web.c: Add missing
3013 whitespace before "(".
3014
3015 2013-09-28 Sandra Loosemore <sandra@codesourcery.com>
3016
3017 * expr.h (extract_bit_field): Remove packedp parameter.
3018 * expmed.c (extract_fixed_bit_field): Remove packedp parameter
3019 from forward declaration.
3020 (store_split_bit_field): Remove packedp arg from calls to
3021 extract_fixed_bit_field.
3022 (extract_bit_field_1): Remove packedp parameter and packedp
3023 argument from recursive calls and calls to extract_fixed_bit_field.
3024 (extract_bit_field): Remove packedp parameter and corresponding
3025 arg to extract_bit_field_1.
3026 (extract_fixed_bit_field): Remove packedp parameter. Remove code
3027 to issue warnings.
3028 (extract_split_bit_field): Remove packedp arg from call to
3029 extract_fixed_bit_field.
3030 * expr.c (emit_group_load_1): Adjust calls to extract_bit_field.
3031 (copy_blkmode_from_reg): Likewise.
3032 (copy_blkmode_to_reg): Likewise.
3033 (read_complex_part): Likewise.
3034 (store_field): Likewise.
3035 (expand_expr_real_1): Likewise.
3036 * calls.c (store_unaligned_arguments_into_pseudos): Adjust call
3037 to extract_bit_field.
3038 * config/tilegx/tilegx.c (tilegx_expand_unaligned_load): Adjust
3039 call to extract_bit_field.
3040 * config/tilepro/tilepro.c (tilepro_expand_unaligned_load): Adjust
3041 call to extract_bit_field.
3042 * doc/invoke.texi (Code Gen Options): Remove mention of warnings
3043 and special packedp behavior from -fstrict-volatile-bitfields
3044 documentation.
3045
3046 2013-09-27 Jan-Benedict Glaw <jbglaw@lug-owl.de>
3047
3048 * lra-eliminations.c (init_elim_table): Guard value_p.
3049
3050 2013-09-27 Michael Meissner <meissner@linux.vnet.ibm.com>
3051
3052 * config/rs6000/rs6000.c (rs6000_hard_regno_mode_ok): Allow
3053 DFmode, DImode, and SFmode in the upper VSX registers based on the
3054 -mupper-regs-{df,sf} flags. Fix wu constraint to be ALTIVEC_REGS
3055 if -mpower8-vector. Combine -mvsx-timode handling with the rest
3056 of the VSX register handling.
3057
3058 * config/rs6000/rs6000.md (f32_lv): Use %x0 for VSX regsters.
3059 (f32_sv): Likewise.
3060 (zero_extendsidi2_lfiwzx): Add support for loading into the
3061 Altivec registers with -mpower8-vector. Use wu/wv constraints to
3062 only do VSX memory options on Altivec registers.
3063 (extendsidi2_lfiwax): Likewise.
3064 (extendsfdf2_fpr): Likewise.
3065 (mov<mode>_hardfloat, SF/SD modes): Likewise.
3066 (mov<mode>_hardfloat32, DF/DD modes): Likewise.
3067 (mov<mode>_hardfloat64, DF/DD modes): Likewise.
3068 (movdi_internal64): Likewise.
3069
3070 2013-09-27 Xinliang David Li <davidxl@google.com>
3071
3072 * opts.c (finish_options): Adjust parameters
3073 according to vect cost model.
3074 (common_handle_option): Set dynamic vect cost
3075 model for FDO.
3076 targhooks.c (default_add_stmt_cost): Compute stmt cost
3077 unconditionally.
3078 * tree-vect-loop.c (vect_estimate_min_profitable_iters):
3079 Use helper function.
3080 * tree-vectorizer.h (unlimited_cost_model): New function.
3081 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use helper function.
3082 * tree-vect-data-refs.c (vect_peeling_hash_insert): Use helper
3083 function.
3084 (vect_enhance_data_refs_alignment): Ditto.
3085 * flag-types.h: New enum.
3086 * common/config/i386/i386-common.c (ix86_option_init_struct):
3087 No need to initialize vect_cost_model flag.
3088 * config/i386/i386.c (ix86_add_stmt_cost): Compute stmt cost
3089 unconditionally.
3090
3091 2013-09-27 Diego Novillo <dnovillo@google.com>
3092
3093 * gimple.h (enum ssa_mode): Remove.
3094
3095 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3096
3097 PR middle-end/58463
3098 * gcc.dg/pr58463.c: New test.
3099
3100 2013-09-27 Paulo Matos <pmatos@broadcom.com>
3101
3102 * cfgloop.h (number_of_loops): Fix typo in check for null.
3103
3104 2013-09-27 Jakub Jelinek <jakub@redhat.com>
3105
3106 PR middle-end/58551
3107 * tree-cfg.c (move_sese_region_to_fn): Also move loops that
3108 are children of outermost saved_cfun's loop, and set it up to
3109 be moved to dest_cfun's outermost loop. Fix up num_nodes adjustments
3110 if loop != loop0 and SESE region contains bbs that belong to loop0.
3111
3112 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3113
3114 * rtlanal.c (must_be_base_p, must_be_index_p): Delete.
3115 (binary_scale_code_p, get_base_term, get_index_term): New functions.
3116 (set_address_segment, set_address_base, set_address_index)
3117 (set_address_disp): Accept the argument unconditionally.
3118 (baseness): Remove must_be_base_p and must_be_index_p checks.
3119 (decompose_normal_address): Classify as much as possible in the
3120 main loop.
3121
3122 2013-09-27 Richard Sandiford <rdsandiford@googlemail.com>
3123
3124 * cse.c (count_reg_usage): Handle INT_LIST.
3125 * lra-eliminations.c (lra_eliminate_regs_1): Likewise.
3126 * reginfo.c (reg_scan_mark_refs): Likewise.
3127 * reload1.c (eliminate_regs_1): Likewise.
3128
3129 2013-09-27 Iain Sandoe <iain@codesourcery.com>
3130
3131 PR middle-end/58547
3132 * rtlanal.c (lsb_bitfield_op_p): Make both parts of the comparison
3133 signed.
3134
3135 2013-09-27 Richard Biener <rguenther@suse.de>
3136
3137 PR tree-optimization/58459
3138 * tree-ssa-forwprop.c (forward_propagate_addr_expr): Remove
3139 restriction not propagating into loops.
3140
3141 2013-09-26 Florian Weimer <fw@deneb.enyo.de>
3142
3143 * tree-ssa.h (walk_use_def_chains_fn, walk_use_def_chains): Delete.
3144 * tree-ssa.c (walk_use_def_chains_1, walk_use_def_chains): Delete.
3145 * doc/tree-ssa.texi (Walking use-def chains): Delete.
3146
3147 2013-09-26 Richard Biener <rguenther@suse.de>
3148
3149 * tree-into-ssa.c (rewrite_into_ssa): Make more SSA names to anonymous.
3150
3151 2013-09-26 Richard Biener <rguenther@suse.de>
3152
3153 * alias.h (component_uses_parent_alias_set): Rename to ...
3154 (component_uses_parent_alias_set_from): ... this.
3155 * alias.c (component_uses_parent_alias_set): Rename to ...
3156 (component_uses_parent_alias_set_from): ... this and return
3157 the desired parent.
3158 (reference_alias_ptr_type_1): Use the result from
3159 component_uses_parent_alias_set_from instead of stripping
3160 components one at a time.
3161 * emit-rtl.c (set_mem_attributes_minus_bitpos): Adjust.
3162
3163 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3164
3165 * tree-ssa-live.h (find_replaceable_exprs, dump_replaceable_exprs):
3166 Move prototypes to...
3167 * tree-ssa-ter.h: New File. Move prototypes here.
3168 * tree-flow.h (stmt_is_replaceable_p): Remove prototype.
3169 * tree-outof-ssa.h: New. Rename ssaexpand.h, include tree-ssa-ter.h.
3170 * tree-outof-ssa.c (ssa_is_replaceable_p): New. Refactor common bits
3171 from is_replaceable_p.
3172 * tree-ssa-ter.c (is_replaceable_p, stmt_is_replaceable_p): Delete.
3173 (ter_is_replaceable_p): New. Use new refactored ssa_is_replaceable_p.
3174 (process_replaceable): Use ter_is_replaceable_p.
3175 (find_replaceable_in_bb): Use ter_is_replaceable_p.
3176 * expr.c (stmt_is_replaceable_p): Relocate from tree-ssa-ter.c. Use
3177 newly refactored ssa_is_replaceable_p.
3178 * cfgexpand.c: Include tree-outof-ssa.h.
3179 * ssaexpand.h: Delete.
3180
3181 2013-09-26 Andrew MacLeod <amacleod@redhat.com>
3182
3183 * gimple.c (gimple_replace_lhs): Move to tree-ssa.c and rename.
3184 (struct count_ptr_d, count_ptr_derefs, count_uses_and_derefs): Move to
3185 tree-ssa.c
3186 (create_gimple_tmp): Delete.
3187 (get_expr_type, build_assign, build_type_cast): Move to...
3188 * gimple-builder.c: New File.
3189 (get_expr_type): Relocate from gimple.c.
3190 (build_assign, build_type_cast): Change to only create ssanames.
3191 * gimple.h: Move prototypes to...
3192 * gimple-builder.h: New File. Here.
3193 * tree-ssa.h: And here.
3194 * tree-ssa.c (struct count_ptr_d, count_ptr_derefs,
3195 count_uses_and_derefs): Relocate from gimple.c.
3196 (gimple_replace_ssa_lhs): Renamed gimple_replace_ssa from gimple.c
3197 * tree-ssa-reassoc.c (repropagate_negates): Use gimple_replace_ssa_lhs.
3198 * tree-ssa-math-opts (execute_cse_reciprocals): Use
3199 gimple_replace_ssa_lhs.
3200 * asan.c: Include gimple-builder.h.
3201 * Makefile.in: Add gimple-builder.o.
3202
3203 2013-09-26 Richard Biener <rguenther@suse.de>
3204
3205 * tree-ssa-live.c (var_map_base_init): Handle SSA names with
3206 DECL_IGNORED_P base VAR_DECLs like anonymous SSA names.
3207 (loe_visit_block): Use gcc_checking_assert.
3208 * tree-ssa-coalesce.c (create_outofssa_var_map): Use
3209 gimple_assign_ssa_name_copy_p.
3210 (gimple_can_coalesce_p): Adjust according to the var_map_base_init
3211 change.
3212
3213 2013-09-26 David Edelsohn <dje.gcc@gmail.com>
3214
3215 * config/rs6000/t-rs6000 (rs6000.o): Remove.
3216 (rs6000-c.o): Use COMPILE and POSTCOMPILE.
3217
3218 2013-09-26 Richard Biener <rguenther@suse.de>
3219
3220 PR tree-optimization/58539
3221 * tree-vect-loop.c (vect_create_epilog_for_reduction): Honor
3222 the fact that debug statements are not taking part in loop-closed
3223 SSA construction.
3224
3225 2013-09-26 Nick Clifton <nickc@redhat.com>
3226
3227 * config/msp430/msp430.c (msp430_expand_epilogue): Fix compile
3228 time warning message.
3229 (msp430_print_operand_raw): Delete unused letter parameter.
3230 (TARGET_PRINT_OPERAND_ADDRESS): Define.
3231 (msp430_print_operand_address): New function.
3232 (msp430_print_operand): Move address printing code from here to
3233 new function.
3234 * config/msp430/msp430.md (movsipsi2): Add comment in generated
3235 assembler.
3236 (zero_extendpsisi2): Likewise.
3237 (extendpsisi2): New pattern.
3238 (andneghi3): New pattern.
3239
3240 2013-09-26 Yvan Roux <yvan.roux@linaro.org>
3241
3242 * config/aarch64/aarch64.opt (mlra): New option.
3243 * config/aarch64/aarch64.c (aarch64_lra_p): New function.
3244 (TARGET_LRA_P): Define.
3245
3246 2013-09-26 Eric Botcazou <ebotcazou@adacore.com>
3247
3248 * expr.c (expand_assignment): Remove obsolete comment.
3249
3250 2013-09-25 Jeff Law <law@redhat.com>
3251
3252 * tree-flow.h (thread_through_all_blocks): Prototype moved into
3253 tree-ssa-threadupdate.h.
3254 (register_jump_thread): Similarly.
3255 * tree-ssa-threadupdate.h: New header file.
3256 * tree-ssa-dom.c: Include tree-ssa-threadupdate.h.
3257 * tree-vrp.c: Likewise.
3258 * tree-ssa-threadedge.c: Include tree-ssa-threadupdate.h.
3259 (thread_around_empty_blocks): Change type of path vector argument to
3260 an edge,type pair from just an edge. Initialize both elements when
3261 appending to a jump threading path. Tweak references to elements
3262 appropriately.
3263 (thread_across_edge): Similarly. Release memory for the elements
3264 as needed.
3265 * tree-ssa-threadupdate.c: Include tree-ssa-threadupdate.h.
3266 (dump_jump_thread_path): New function broken out from
3267 register_jump_thread.
3268 (register_jump_thread): Use dump_jump_thread_path. Change type of
3269 path vector entries. Search the path for NULL edges and dump
3270 the path if one is found. Tweak the conversion of path to 3-edge
3271 form to use the block copy type information embedded in the path.
3272
3273 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3274
3275 * lra.c (update_inc_notes): Remove all REG_DEAD and REG_UNUSED notes.
3276
3277 2013-09-25 Yvan Roux <yvan.roux@linaro.org>
3278 Vladimir Makarov <vmakarov@redhat.com>
3279
3280 * rtlanal.c (lsb_bitfield_op_p): New predicate for bitfield operations
3281 from the least significant bit.
3282 (strip_address_mutations): Add bitfield operations handling.
3283 (must_be_index_p): Add shifting and rotate operations handling.
3284 (set_address_base): Use must_be_base_p predicate.
3285 (set_address_index): Use must_be_index_p predicate.
3286
3287 2013-09-25 Alexander Ivchenko <alexander.ivchenko@intel.com>
3288 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
3289 Sergey Lega <sergey.s.lega@intel.com>
3290 Anna Tikhonova <anna.tikhonova@intel.com>
3291 Ilya Tocar <ilya.tocar@intel.com>
3292 Andrey Turetskiy <andrey.turetskiy@intel.com>
3293 Ilya Verbin <ilya.verbin@intel.com>
3294 Kirill Yukhin <kirill.yukhin@intel.com>
3295 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
3296
3297 * config/i386/i386.c (ix86_avx256_split_vector_move_misalign):
3298 Use new names.
3299 (ix86_expand_vector_move_misalign): Support new unaligned load and
3300 stores and use new names.
3301 (CODE_FOR_sse2_storedqu): Rename to ...
3302 (CODE_FOR_sse2_storedquv16qi): ... this.
3303 (CODE_FOR_sse2_loaddqu): Rename to ...
3304 (CODE_FOR_sse2_loaddquv16qi): ... this.
3305 (CODE_FOR_avx_loaddqu256): Rename to ...
3306 (CODE_FOR_avx_loaddquv32qi): ... this.
3307 (CODE_FOR_avx_storedqu256): Rename to ...
3308 (CODE_FOR_avx_storedquv32qi): ... this.
3309 * config/i386/i386.md (fpint_logic): New.
3310 * config/i386/sse.md (VMOVE): Extend for AVX512.
3311 (VF): Ditto.
3312 (VF_128_256): New.
3313 (VF_512): Ditto.
3314 (VI_UNALIGNED_LOADSTORE): Ditto.
3315 (sse2_avx_avx512f): Ditto.
3316 (sse2_avx2): Extend for AVX512.
3317 (sse4_1_avx2): Ditto.
3318 (avx2_avx512f): New.
3319 (sse): Extend for AVX512.
3320 (sse2): Ditto.
3321 (sse4_1): Ditto.
3322 (avxsizesuffix): Ditto.
3323 (sseintvecmode): Ditto.
3324 (ssePSmode): Ditto.
3325 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Ditto.
3326 (<sse>_storeu<ssemodesuffix><avxsizesuffix>): Ditto.
3327 (<sse2>_loaddqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3328 (<sse2_avx_avx512f>_loaddqu<mode>): ... this.
3329 (<sse2>_storedqu<avxsizesuffix>): Extend for AVX512 and rename to ...
3330 (<sse2_avx_avx512f>_storedqu<mode): ... this.
3331 (<sse>_movnt<mode>): Replace constraint "x" with "v".
3332 (STORENT_MODE): Extend for AVX512.
3333 (*absneg<mode>2): Replace constraint "x" with "v".
3334 (*mul<mode>3): Ditto.
3335 (*ieee_smin<mode>3): Ditto.
3336 (*ieee_smax<mode>3): Ditto.
3337 (avx_cmp<mode>3): Replace VF with VF_128_256.
3338 (*<sse>_maskcmp<mode>3_comm): Ditto.
3339 (<sse>_maskcmp<mode>3): Ditto.
3340 (<sse>_andnot<mode>3): Extend for AVX512.
3341 (<code><mode>3, anylogic): Replace VF with VF_128_256.
3342 (<code><mode>3, fpint_logic): New.
3343 (*<code><mode>3): Extend for AVX512.
3344 (avx512flogicsuff): New.
3345 (avx512f_<logic><mode>): Ditto.
3346 (<sse>_movmsk<ssemodesuffix><avxsizesuffix>): Replace VF with
3347 VF_128_256.
3348 (<sse4_1>_blend<ssemodesuffix><avxsizesuffix>): Ditto.
3349 (<sse4_1>_blendv<ssemodesuffix><avxsizesuffix>): Ditto.
3350 (<sse4_1>_dp<ssemodesuffix><avxsizesuffix>): Ditto.
3351 (avx_vtest<ssemodesuffix><avxsizesuffix>): Ditto.
3352 (<sse4_1>_round<ssemodesuffix><avxsizesuffix>): Ditto.
3353 (xop_vpermil2<mode>3): Ditto.
3354 (*avx_vpermilp<mode>): Extend for AVX512 and rename to ...
3355 (*<sse2_avx_avx512f>_vpermilp<mode>): ... this.
3356 (avx_vpermilvar<mode>3): Extend for AVX512 and rename to ...
3357 (<sse2_avx_avx512f>_vpermilvar<mode>3): ... this.
3358
3359 2013-09-25 Tom Tromey <tromey@redhat.com>
3360
3361 * Makefile.in (PARTITION_H, LTO_SYMTAB_H, COMMON_TARGET_DEF_H)
3362 (RTL_ERROR_H, TRANS_MEM_H, COVERAGE_H, DEMANGLE_H, ALIAS_H)
3363 (SCHED_INT_H, SEL_SCHED_IR_H, SEL_SCHED_DUMP_H, VALTRACK_H, DDG_H)
3364 (GGC_INTERNAL_H, DECNUM_H, BACKTRACE_H, MKDEPS_H, TREE_HASHER_H)
3365 (TREE_SSA_LIVE_H, SSAEXPAND_H, DWARF2OUT_H, SCEV_H, OMEGA_H)
3366 (TREE_DATA_REF_H, IRA_INT_H, LRA_INT_H, DBGCNT_H, DATA_STREAMER_H)
3367 (GIMPLE_STREAMER_H, TREE_STREAMER_H, STREAMER_HOOKS_H)
3368 (TREE_VECTORIZER_H, IPA_INLINE_H, GSTAB_H, LIBFUNCS_H)
3369 (GRAPHITE_HTAB_H): Remove.
3370
3371 2013-09-25 Tom Tromey <tromey@redhat.com>
3372
3373 * config/mcore/t-mcore (CROSS_FLOAT_H): Remove.
3374
3375 2013-09-25 Tom Tromey <tromey@redhat.com>
3376
3377 * config/t-glibc (glibc-c.o): Use COMPILE and POSTCOMPILE.
3378
3379 2013-09-25 Tom Tromey <tromey@redhat.com>
3380
3381 * config/i386/t-i386 (i386.o): Remove.
3382 (i386-c.o): Use COMPILE and POSTCOMPILE.
3383
3384 2013-09-25 Tom Tromey <tromey@redhat.com>
3385
3386 * Makefile.in ($(out_object_file)): Use COMPILE and POSTCOMPILE.
3387
3388 2013-09-25 Tom Tromey <tromey@redhat.com>
3389
3390 * Makefile.in (graph.o, sbitmap.o, sparseset.o, gcc-ar.o)
3391 (gcc-ranlib.o, gcc-nm.o, collect2.o, collect2-aix.o, tlink.o)
3392 (lto-wrapper.o, default-c.o, attribs.o, incpath.o, prefix.o)
3393 (gcc.o, options.o, options-save.o, version.o, gtype-desc.o)
3394 (trans-mem.o, ggc-common.o, ggc-page.o, ggc-none.o, stringpool.o)
3395 (convert.o, double-int.o, lto-compress.o, data-streamer-in.o)
3396 (data-streamer-out.o, data-streamer.o, gimple-streamer-in.o)
3397 (gimple-streamer-out.o, tree-streamer.o, tree-streamer-in.o)
3398 (tree-streamer-out.o, streamer-hooks.o, lto-cgraph.o)
3399 (lto-streamer-in.o, lto-streamer-out.o, lto-section-in.o)
3400 (lto-section-out.o, lto-opts.o, lto-streamer.o, langhooks.o)
3401 (test-dump.o, tree.o, tree-dump.o, tree-inline.o, print-tree.o)
3402 (stor-layout.o, asan.o, tsan.o, ubsan.o, tree-ssa-tail-merge.o)
3403 (tree-ssa-structalias.o, tree-ssa-uninit.o, tree-ssa.o)
3404 (tree-into-ssa.o, tree-ssa-ter.o, tree-ssa-coalesce.o)
3405 (tree-outof-ssa.o, tree-ssa-dse.o, tree-ssa-forwprop.o)
3406 (tree-ssa-phiprop.o, tree-ssa-ifcombine.o, tree-ssa-phiopt.o)
3407 (tree-nrv.o, tree-ssa-copy.o, tree-ssa-propagate.o)
3408 (tree-ssa-dom.o, tree-ssa-uncprop.o, tree-ssa-threadedge.o)
3409 (tree-ssa-threadupdate.o, tree-ssanames.o, tree-phinodes.o)
3410 (domwalk.o, tree-ssa-live.o, tree-ssa-copyrename.o)
3411 (tree-ssa-pre.o, tree-ssa-sccvn.o)
3412 (gimple-ssa-strength-reduction.o, tree-vrp.o, tree-cfg.o)
3413 (tree-cfgcleanup.o, tree-tailcall.o, tree-ssa-sink.o)
3414 (tree-nested.o, tree-if-conv.o, tree-iterator.o, tree-dfa.o)
3415 (tree-ssa-operands.o, tree-eh.o, tree-ssa-loop.o)
3416 (tree-ssa-loop-unswitch.o, tree-ssa-address.o)
3417 (tree-ssa-loop-niter.o, tree-ssa-loop-ivcanon.o)
3418 (tree-ssa-loop-ch.o, tree-ssa-loop-prefetch.o, tree-predcom.o)
3419 (tree-ssa-loop-ivopts.o, tree-affine.o, tree-ssa-loop-manip.o)
3420 (tree-ssa-loop-im.o, tree-ssa-math-opts.o, tree-ssa-alias.o)
3421 (tree-ssa-reassoc.o, tree-optimize.o, gimplify.o)
3422 (gimple-iterator.o, gimple-fold.o, gimple-low.o, omp-low.o)
3423 (tree-browser.o, omega.o, tree-chrec.o, tree-scalar-evolution.o)
3424 (tree-data-ref.o, sese.o, graphite.o, graphite-blocking.o)
3425 (graphite-clast-to-gimple.o, graphite-dependences.o)
3426 (graphite-interchange.o, graphite-poly.o)
3427 (graphite-scop-detection.o, graphite-sese-to-poly.o)
3428 (graphite-optimize-isl.o, tree-vect-loop.o)
3429 (tree-vect-loop-manip.o, tree-vect-patterns.o, tree-vect-slp.o)
3430 (tree-vect-stmts.o, tree-vect-data-refs.o, tree-vectorizer.o)
3431 (vtable-verify.o, tree-loop-distribution.o, tree-parloops.o)
3432 (tree-stdarg.o, tree-object-size.o, internal-fn.o, gimple.o)
3433 (gimple-pretty-print.o, tree-mudflap.o, tree-nomudflap.o)
3434 (tree-pretty-print.o, tree-diagnostic.o, fold-const.o)
3435 (diagnostic.o, diagnostic-color.o, opts.o, opts-global.o)
3436 (opts-common.o, targhooks.o, common/common-targhooks.o, input.o)
3437 (toplev.o, hwint.o, passes.o, plugin.o, main.o, host-default.o)
3438 (rtl-error.o, rtl.o, print-rtl.o, rtlanal.o, varasm.o, function.o)
3439 (statistics.o, stmt.o, except.o, expr.o, dojump.o, builtins.o)
3440 (calls.o, expmed.o, explow.o, optabs.o, dbxout.o, debug.o)
3441 (sdbout.o, dwarf2out.o, dwarf2cfi.o, dwarf2asm.o, vmsdbgout.o)
3442 (xcoffout.o, godump.o, emit-rtl.o, real.o, realmpfr.o, dfp.o)
3443 (fixed-value.o, jump.o, simplify-rtx.o, symtab.o, cgraph.o)
3444 (cgraphunit.o, cgraphclones.o, cgraphbuild.o, varpool.o, ipa.o)
3445 (ipa-profile.o, ipa-devirt.o, ipa-prop.o, ipa-ref.o, ipa-cp.o)
3446 (ipa-split.o, ipa-inline.o, ipa-inline-analysis.o)
3447 (ipa-inline-transform.o, ipa-utils.o, ipa-reference.o)
3448 (ipa-pure-const.o, coverage.o, cselib.o, cse.o, dce.o, dumpfile.o)
3449 (dse.o, fwprop.o, web.o, ree.o, cprop.o, gcse.o, store-motion.o)
3450 (resource.o, lcm.o, mode-switching.o, tree-ssa-dce.o)
3451 (tree-call-cdce.o, tree-ssa-ccp.o, tree-ssa-strlen.o, tree-sra.o)
3452 (tree-switch-conversion.o, tree-complex.o, tree-emutls.o)
3453 (tree-vect-generic.o, df-core.o, df-problems.o, df-scan.o)
3454 (regstat.o, valtrack.o, var-tracking.o, profile.o, mcf.o)
3455 (tree-profile.o, value-prof.o, loop-doloop.o, alloc-pool.o)
3456 (auto-inc-dec.o, cfg.o, cfghooks.o, cfgexpand.o, cfgrtl.o)
3457 (cfganal.o, cfgbuild.o, cfgcleanup.o, cfgloop.o, cfgloopanal.o)
3458 (graphds.o, loop-iv.o, loop-invariant.o, cfgloopmanip.o)
3459 (loop-init.o, loop-unswitch.o, loop-unroll.o, dominance.o)
3460 (et-forest.o, combine.o, reginfo.o, bitmap.o, vec.o, hash-table.o)
3461 (reload.o, reload1.o, rtlhooks.o, postreload.o, postreload-gcse.o)
3462 (caller-save.o, bt-load.o, reorg.o, alias.o, stack-ptr-mod.o)
3463 (init-regs.o, ira-build.o, ira-costs.o, ira-conflicts.o)
3464 (ira-color.o, ira-emit.o, ira-lives.o, ira.o, lra.o)
3465 (lra-assigns.o, lra-coalesce.o, lra-constraints.o)
3466 (lra-eliminations.o, lra-lives.o, lra-spills.o, regmove.o)
3467 (combine-stack-adj.o, compare-elim.o, ddg.o, modulo-sched.o)
3468 (haifa-sched.o, sched-deps.o, sched-rgn.o, sched-ebb.o)
3469 (sched-vis.o, sel-sched.o, sel-sched-dump.o, sel-sched-ir.o)
3470 (final.o, recog.o, reg-stack.o, sreal.o, predict.o, lists.o)
3471 (bb-reorder.o, tracer.o, timevar.o, regcprop.o, regrename.o)
3472 (ifcvt.o, params.o, pointer-set.o, hooks.o, pretty-print.o)
3473 (errors.o, dbgcnt.o, lower-subreg.o, target-globals.o)
3474 (hw-doloop.o, file-find.o, context.o, $(common_out_object_file))
3475 (insn-attrtab.o, insn-automata.o, insn-dfatab.o, insn-emit.o)
3476 (insn-enums.o, insn-extract.o, insn-latencytab.o, insn-modes.o)
3477 (insn-opinit.o, insn-output.o, insn-peep.o, insn-preds.o)
3478 (insn-recog.o, intl.o, cppbuiltin.o, cppdefault.o, gcov.o)
3479 (gcov-dump.o): Remove.
3480 (default-c.o): Use COMPILE and POSTCOMPILE.
3481 (CFLAGS-gcc.o): New variable.
3482 ($(common_out_object_file)): Use COMPILE and POSTCOMPILE.
3483
3484 2013-09-25 Tom Tromey <tromey@redhat.com>
3485
3486 * Makefile.in (c-family/cppspec.o, c-family/c-common.o)
3487 (c-family/c-cppbuiltin.o, c-family/c-dump.o, c-family/c-format.o)
3488 (c-family/c-gimplify.o, c-family/c-lex.o, c-family/c-omp.o)
3489 (c-family/c-opts.o, c-family/c-pch.o, c-family/c-ppoutput.o)
3490 (c-family/c-pragma.o, c-family/c-pretty-print.o)
3491 (c-family/c-semantics.o, c-family/c-ada-spec.o)
3492 (c-family/array-notation-common.o, c-family/stub-objc.o)
3493 (c-family/c-ubsan.o): Remove.
3494
3495 2013-09-25 Tom Tromey <tromey@redhat.com>
3496
3497 * Makefile.in (C_TREE_H): Reference c/c-tree.h.
3498
3499 2013-09-25 Tom Tromey <tromey@redhat.com>
3500
3501 * Makefile.in (DRIVER_DEFINES): Use $(and), not shell code,
3502 to add -DENABLE_SHARED_LIBGCC.
3503 (gcc.o): Don't use subshell.
3504
3505 2013-09-25 Tom Tromey <tromey@redhat.com>
3506
3507 * Makefile.in (OUTPUT_OPTION): Define as "-o $@".
3508 * configure.ac: Don't invoke AM_PROG_CC_C_O.
3509 (NO_MINUS_C_MINUS_O, OUTPUT_OPTION): Don't subst.
3510 * configure, config.in: Rebuild.
3511
3512 2013-09-25 Tom Tromey <tromey@redhat.com>
3513
3514 * Makefile.in (CCDEPMODE, DEPDIR, depcomp, COMPILE.base)
3515 (COMPILE, POSTCOMPILE): New variables.
3516 (.cc.o .c.o): Use COMPILE, POSTCOMPILE.
3517 (DEPFILES): New variable.
3518 Include ".Po" files.
3519 * configure.ac: Add checks for dependency checking.
3520 * configure, aclocal.m4: Regenerate.
3521
3522 2013-09-25 Tom Tromey <tromey@redhat.com>
3523
3524 * Makefile.in (ALL_HOST_BACKEND_OBJS): Add lto-wrapper.o.
3525 ($(ALL_HOST_OBJS)): Move order-only dependency to end of file.
3526
3527 2013-09-25 Tom Tromey <tromey@redhat.com>
3528
3529 * Makefile.in (generated_files): Add options.h,
3530 target-hooks-def.h, insn-opinit.h,
3531 common/common-target-hooks-def.h, pass-instances.def,
3532 c-family/c-target-hooks-def.h.
3533
3534 2013-09-25 Jeff Law <law@redhat.com>
3535
3536 * tree-ssa-threadedge.c (thread_across_edge): Use foo.last () rather
3537 than foo[foo.length () - 1] to access last member in a vec.
3538 * tree-ssa-threadupdate.c (register_jump_thread): Similarly.
3539
3540 2013-09-25 Richard Biener <rguenther@suse.de>
3541
3542 PR middle-end/58521
3543 * tree.c (iterative_hash_expr): Remove MEM_REF special handling.
3544
3545 2013-09-25 Jan Hubicka <jh@suse.cz>
3546
3547 * cgraph.c (cgraph_resolve_speculation): Use semantical equivalency
3548 test.
3549
3550 2013-09-25 Marek Polacek <polacek@redhat.com>
3551
3552 PR sanitizer/58420
3553 * ubsan.c (ubsan_type_descriptor): Handle IDENTIFIER_NODEs
3554 when determining the type name.
3555
3556 2013-09-24 Oleg Endo <olegendo@gcc.gnu.org>
3557
3558 * config/sh/sh.md: Fix formatting.
3559
3560 2013-09-24 Xinliang David Li <davidxl@google.com>
3561
3562 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Check
3563 max peel iterations parameter.
3564 * param.def: New parameter.
3565 * doc/invoke.texi: Document New parameter.
3566
3567 2013-09-24 Christophe Lyon <christophe.lyon@linaro.org>
3568
3569 * gimple-pretty-print.c: Various whitespace tweaks.
3570 * tree-core.h: Likewise.
3571 * tree-pretty-print.c: Likewise.
3572 * tree-ssa-alias.c: Likewise.
3573 * tree-ssa-copy.c: Likewise.
3574 * tree-ssanames.c: Likewise.
3575 * tree-ssanames.h: Likewise.
3576 * tree-vrp.c: Likewise.
3577
3578 2013-09-24 Alan Modra <amodra@gmail.com>
3579
3580 PR middle-end/57134
3581 PR middle-end/57586
3582 * stmt.c (expand_asm_operands): Call expand_expr with EXPAND_MEMORY
3583 for output operands that disallow regs. Don't use EXPAND_WRITE on
3584 inout operands.
3585
3586 2013-09-24 Richard Biener <rguenther@suse.de>
3587
3588 PR middle-end/58513
3589 * tree.c (reference_alias_ptr_type): Move ...
3590 * alias.c (reference_alias_ptr_type): ... here and implement
3591 in terms of the new reference_alias_ptr_type_1.
3592 (ref_all_alias_ptr_type_p): New helper.
3593 (get_deref_alias_set_1): Drop flag_strict_aliasing here,
3594 use ref_all_alias_ptr_type_p.
3595 (get_deref_alias_set): Add flag_strict_aliasing check here.
3596 (reference_alias_ptr_type_1): New function, split out from ...
3597 (get_alias_set): ... here.
3598 (alias_ptr_types_compatible_p): New function.
3599 * alias.h (reference_alias_ptr_type): Declare.
3600 (alias_ptr_types_compatible_p): Likewise.
3601 * tree.h (reference_alias_ptr_type): Remove.
3602 * fold-const.c (operand_equal_p): Use alias_ptr_types_compatible_p
3603 to compare MEM_REF alias types.
3604
3605 2013-09-24 Richard Biener <rguenther@suse.de>
3606
3607 * tree-vrp.c (vrp_finalize): Check for SSA name presence.
3608
3609 2013-09-23 Michael Meissner <meissner@linux.vnet.ibm.com>
3610
3611 * config/rs6000/rs6000.c (rs6000_vector_reload): Delete, combine
3612 reload helper function arrays into a single array reg_addr.
3613 (reload_fpr_gpr): Likewise.
3614 (reload_gpr_vsx): Likewise.
3615 (reload_vsx_gpr): Likewise.
3616 (struct rs6000_reg_addr): Likewise.
3617 (reg_addr): Likewise.
3618 (rs6000_debug_reg_global): Change rs6000_vector_reload,
3619 reload_fpr_gpr, reload_gpr_vsx, reload_vsx_gpr uses to reg_addr.
3620 (rs6000_init_hard_regno_mode_ok): Likewise.
3621 (rs6000_secondary_reload_direct_move): Likewise.
3622 (rs6000_secondary_reload): Likewise.
3623
3624 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add new
3625 constraints: wu, ww, and wy. Repurpose wv constraint added during
3626 power8 changes. Put wg constraint in alphabetical order.
3627
3628 * config/rs6000/rs6000.opt (-mvsx-scalar-float): New debug switch
3629 for future work to add ISA 2.07 VSX single precision support.
3630 (-mvsx-scalar-double): Change default from -1 to 1, update
3631 documentation comment.
3632 (-mvsx-scalar-memory): Rename debug switch to -mupper-regs-df.
3633 (-mupper-regs-df): New debug switch to control whether DF values
3634 can go in the traditional Altivec registers.
3635 (-mupper-regs-sf): New debug switch to control whether SF values
3636 can go in the traditional Altivec registers.
3637
3638 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print wu, ww,
3639 and wy constraints.
3640 (rs6000_init_hard_regno_mode_ok): Use ssize_t instead of int for
3641 loop variables. Rename -mvsx-scalar-memory to -mupper-regs-df.
3642 Add new constraints, wu/ww/wy. Repurpose wv constraint.
3643 (rs6000_debug_legitimate_address_p): Print if we are running
3644 before, during, or after reload.
3645 (rs6000_secondary_reload): Add a comment.
3646 (rs6000_opt_masks): Add -mupper-regs-df, -mupper-regs-sf.
3647
3648 * config/rs6000/constraints.md (wa constraint): Sort w<x>
3649 constraints. Update documentation string.
3650 (wd constraint): Likewise.
3651 (wf constraint): Likewise.
3652 (wg constraint): Likewise.
3653 (wn constraint): Likewise.
3654 (ws constraint): Likewise.
3655 (wt constraint): Likewise.
3656 (wx constraint): Likewise.
3657 (wz constraint): Likewise.
3658 (wu constraint): New constraint for ISA 2.07 SFmode scalar
3659 instructions.
3660 (ww constraint): Likewise.
3661 (wy constraint): Likewise.
3662 (wv constraint): Repurpose ISA 2.07 constraint that we not used in
3663 the previous submissions.
3664 * doc/md.texi (PowerPC and IBM RS6000): Likewise.
3665
3666 2013-09-23 Richard Sandiford <rdsandiford@googlemail.com>
3667
3668 * doc/rtl.texi (REG_NOTES): Say that int_list can also be used.
3669 (REG_BR_PROB): Say that the probability is stored in an int_list.
3670 * reg-notes.def: Update commentary to mention INT_LIST.
3671 * rtl.def (EXPR_LIST, INSN_LIST): Capitalize comments.
3672 (INT_LIST): New rtx.
3673 * rtl.h (add_int_reg_note, add_shallow_copy_of_reg_note): Declare.
3674 * rtlanal.c (int_reg_note_p): New function.
3675 (alloc_reg_note): Assert that the note does not have an int argument.
3676 (add_int_reg_note, add_shallow_copy_of_reg_note): New functions.
3677 * combine.c (distribute_notes): Use add_shallow_copy_of_rtx.
3678 * cse.c (cse_process_notes_1): Expect REG_EQUAL to be an EXPR_LIST
3679 rather than an INSN_LIST. Handle INT_LIST.
3680 * ifcvt.c (cond_exec_process_insns): Take the probability as an int
3681 rather than an rtx. Use gen_rtx_INT_LIST to create a REG_BR_PROB note.
3682 (cond_exec_process_if_block): Use XINT to extract REG_BR_PROB values.
3683 Manipulate them as ints rather than rtxes.
3684 * reg-stack.c (subst_asm_stack_regs): Only handle EXPR_LIST notes.
3685 * regmove.c (copy_src_to_dest): Likewise.
3686 * sched-vis.c (print_insn_with_notes): Handle INT_LIST.
3687
3688 * config/i386/winnt.c (i386_pe_seh_unwind_emit): Sink pat assignment
3689 into the cases that need it.
3690 * config/arm/arm.c (arm_unwind_emit): Likewise.
3691
3692 * asan.c (asan_clear_shadow): Use add_int_reg_note for REG_BR_PROB.
3693 * emit-rtl.c (try_split, emit_copy_of_insn_after): Likewise.
3694 * loop-doloop.c (add_test, doloop_modify): Likewise.
3695 * loop-unswitch.c (compare_and_jump_seq): Likewise.
3696 * optabs.c (emit_cmp_and_jump_insn_1): Likewise.
3697 * predict.c (combine_predictions_for_insn): Likewise.
3698 * print-rtl.c (print_rtx): Handle INT_LIST.
3699 * config/aarch64/aarch64.c (aarch64_emit_unlikely_jump): Likewise.
3700 * config/alpha/alpha.c (emit_unlikely_jump): Likewise.
3701 * config/arm/arm.c (emit_unlikely_jump): Likewise.
3702 * config/i386/i386.c (ix86_expand_split_stack_prologue): Likewise.
3703 (ix86_split_fp_branch, predict_jump): Likewise.
3704 * config/rs6000/rs6000.c (emit_unlikely_jump): Likewise.
3705 * config/sh/sh.c (expand_cbranchsi4): Likewise.
3706 * config/spu/spu.c (ea_load_store_inline): Likewise.
3707
3708 * cfgbuild.c (compute_outgoing_frequencies): Use XINT to access the
3709 value of a REG_BR_PROB note.
3710 * cfgrtl.c (force_nonfallthru_and_redirect): Likewise.
3711 (update_br_prob_note, rtl_verify_edges, purge_dead_edges): Likewise.
3712 * emit-rtl.c (try_split): Likewise.
3713 * predict.c (br_prob_note_reliable_p): Likewise.
3714 (invert_br_probabilities, combine_predictions_for_insn): Likewise.
3715 * reorg.c (mostly_true_jump): Likewise.
3716 * config/bfin/bfin.c (cbranch_predicted_taken_p): Likewise.
3717 * config/frv/frv.c (frv_print_operand_jump_hint): Likewise.
3718 * config/i386/i386.c (ix86_print_operand): Likewise.
3719 * config/ia64/ia64.c (ia64_print_operand): Likewise.
3720 * config/mmix/mmix.c (mmix_print_operand): Likewise.
3721 * config/rs6000/rs6000.c (output_cbranch): Likewise.
3722 * config/s390/s390.c (s390_expand_tbegin): Likewise.
3723 * config/sh/sh.c (sh_print_operand, sh_adjust_cost): Likewise.
3724 * config/sparc/sparc.c (output_cbranch): Likewise.
3725 * config/spu/spu.c (get_branch_target): Likewise.
3726 * config/tilegx/tilegx.c (cbranch_predicted_p): Likewise.
3727 * config/tilepro/tilepro.c (cbranch_predicted_p): Likewise.
3728
3729 2013-09-23 Jan Hubicka <jh@suse.cz>
3730
3731 * ipa-cp.c (ipa_get_indirect_edge_target_1): Add sanity check
3732 for ipa-devirt.
3733 * ipa-utils.h (possible_polymorphic_call_target_p): New function.
3734 * ipa-devirt.c (possible_polymorphic_call_target_p): Be tolerant
3735 of external calls
3736 * gimple-fold.c: Include ipa-utils.h and gimple-pretty-print.h
3737 (gimple_fold_call): Dump inconsistent devirtualizations; add
3738 sanity check for type based devirtualizations.
3739 * ipa-prop.c: Include ipa-utils.h
3740 (ipa_intraprocedural_devirtualization): Add sanity check.
3741 (try_make_edge_direct_virtual_call): Likewise.
3742
3743 2013-09-23 Eric Botcazou <ebotcazou@adacore.com>
3744
3745 * tree-ssa-ccp.c (insert_clobber_before_stack_restore): Recurse on copy
3746 assignment statements.
3747
3748 2013-09-23 Kugan Vivekanandarajah <kuganv@linaro.org>
3749
3750 * gimple-pretty-print.c (dump_ssaname_info): New function.
3751 (dump_gimple_phi): Call it.
3752 (pp_gimple_stmt_1): Likewise.
3753 * tree-core.h (tree_ssa_name): New union ssa_name_info_type field.
3754 (range_info_def): Declare.
3755 * tree-pretty-print.c (pp_double_int): New function.
3756 (dump_generic_node): Call it.
3757 * tree-pretty-print.h (pp_double_int): Declare.
3758 * tree-ssa-alias.c (dump_alias_info): Check pointer type.
3759 * tree-ssanames.h (range_info_def): New structure.
3760 (value_range_type): Move definition here.
3761 (set_range_info, value_range_type, duplicate_ssa_name_range_info):
3762 Declare.
3763 * tree-ssanames.c (make_ssa_name_fn): Check pointer type at
3764 initialization.
3765 (set_range_info): New function.
3766 (get_range_info): Likewise.
3767 (duplicate_ssa_name_range_info): Likewise.
3768 (duplicate_ssa_name_fn): Check pointer type and call
3769 duplicate_ssa_name_range_info.
3770 * tree-ssa-copy.c (fini_copy_prop): Likewise.
3771 * tree-vrp.c (value_range_type): Remove definition, now in
3772 tree-ssanames.h.
3773 (vrp_finalize): Call set_range_info to update value range of SSA_NAMEs.
3774 * tree.h (SSA_NAME_PTR_INFO): Macro changed to access via union.
3775 (SSA_NAME_RANGE_INFO): New macro.
3776
3777 2013-09-23 Richard Biener <rguenther@suse.de>
3778
3779 PR tree-optimization/58464
3780 * tree-ssa-pre.c (phi_trans_lookup): Remove.
3781 (phi_trans_add): Change to add conditionally on being not present.
3782 (phi_translate_1): Remove recursion detection here.
3783 (phi_translate): Pre-seed the cache with NULL to catch
3784 recursion here in a more generic way.
3785 (bitmap_find_leader): Adjust comment.
3786 (get_representative_for): Dump value-numbers.
3787 (create_expression_by_pieces): Likewise.
3788 (insert_into_preds_of_block): Likewise.
3789
3790 2013-09-23 Christian Bruel <christian.bruel@st.com>
3791
3792 PR target/58475
3793 * config/sh/sh.md (movsf_ie): Allow fpul_operand.
3794 * config/sh/predicate.md (arith_reg_operand): Disallow FPUL_REG.
3795
3796 2013-09-23 James Greenhalgh <james.greenhalgh@arm.com>
3797
3798 Revert r202780:
3799 2013-09-20 Renlin Li <renlin.li@arm.com>
3800
3801 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use
3802 plus_constant.
3803 (aarch64_expand_epilogue): Likewise.
3804 (aarch64_legitimize_reload_address): Likewise.
3805
3806 2013-09-22 Eric Botcazou <ebotcazou@adacore.com>
3807
3808 * gimplify.c (gimplify_asm_expr): Reset the TREE_CHAIN of clobbers to
3809 NULL_TREE before pushing them onto the vector. Likewise for labels.
3810
3811 2013-09-21 Eric Botcazou <ebotcazou@adacore.com>
3812
3813 * config/ia64/predicates.md (ia64_cbranch_operator): Accept unordered
3814 comparison operators when -fno-trapping-math is in effect.
3815 * config/ia64/ia64.c (ia64_expand_compare): Add support for unordered
3816 comparison operators in TFmode and assert that unsupported operators
3817 cannot reach here.
3818 (ia64_print_operand): Likewise.
3819
3820 2013-09-21 Jan Hubicka <jh@suse.cz>
3821
3822 * x86-tune.def (partial_reg_stall): Disable for CoreI7 and newer.
3823 (sse_typeless_stores): Enable for core
3824 (sse_load0_by_pxor): Likewise.
3825 (four_jump_limit): Disable for core.
3826 (pad_returns): Likewise.
3827 (avoid_vector_decode): Likewise.
3828 (fuse_cmp_and_branch): Enable for cores.
3829 * i386.c (x86_accumulate_outgoing_args): Disable for cores.
3830
3831 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
3832
3833 PR middle-end/56791
3834 * config/pa/pa.c (pa_option_override): Disable auto increment and
3835 decrement instructions until reload is completed.
3836
3837 * config/pa/pa-linux.h (TARGET_OS_CPP_BUILTINS): Define
3838 __GCC_HAVE_SYNC_COMPARE_AND_SWAP_1, __GCC_HAVE_SYNC_COMPARE_AND_SWAP_2,
3839 and __GCC_HAVE_SYNC_COMPARE_AND_SWAP_4.
3840
3841 2013-09-20 DJ Delorie <dj@redhat.com>
3842 Nick Clifton <nickc@redhat.com>
3843
3844 * config/rl78/rl78.c: Various whitespace and comment tweaks.
3845 (need_to_save): Save bank 0 on interrupts.
3846 (characterize_address): Strip far address wrappers.
3847 (rl78_as_legitimate_address): Likewise.
3848 (transcode_memory_rtx): Likewise.
3849 (rl78_peep_movhi_p): Disable this peephole after devirt.
3850 (rl78_propogate_register_origins): Forget all origins when a
3851 CLOBBER is seen.
3852 * config/rl78/rl78-virt.md: Various whitespace tweaks.
3853 * config/rl78/rl78-real.md: Various whitespace tweaks. Additional
3854 peephole2's.
3855 * config/rl78/rl78.md (sel_rb): Disable for G10 just in case.
3856 * config/rl78/rl78-expand.md (movqi): Check for subregs of consts.
3857 * config/rl78/rl78.h (LINK_SPEC): Pass -gc-sections unless
3858 relocating.
3859 * config/rl78/constraints.md: Various whitespace and paren tweaks.
3860
3861 2013-09-20 John David Anglin <danglin@gcc.gnu.org>
3862
3863 * config/pa/pa.md: In "scc" insn patterns, change output template to
3864 handle const0_rtx in reg_or_0_operand operands.
3865
3866 2013-09-20 Martin Husemann <martin@NetBSD.org>
3867
3868 PR target/56875
3869 * config/vax/vax.c (vax_output_int_move): Use D format specifier.
3870 * config/vax/vax.md (ashldi3, <unnamed>): Ditto.
3871
3872 2013-09-20 Richard Biener <rguenther@suse.de>
3873
3874 PR middle-end/58484
3875 * tree-scalar-evolution.c (struct scev_info_str): Shrink by
3876 remembering SSA name version and block index.
3877 (new_scev_info_str): Adjust.
3878 (hash_scev_info): Likewise. Also hash the block index.
3879 (eq_scev_info): Adjust.
3880 (find_var_scev_info): Likewise.
3881 (struct instantiate_cache_entry): Remove.
3882 (struct instantiate_cache_type): Use a htab to map name, block
3883 to chrec.
3884 (instantiate_cache_type::~instantiate_cache_type): Adjust.
3885 (get_instantiated_value_entry): Likewise.
3886 (hash_idx_scev_info, eq_idx_scev_info): New functions.
3887 (instantiate_scev_name): Adjust.
3888
3889 2013-09-20 Jeff Law <law@redhat.com>
3890
3891 * tree-ssa-dom.c (record_temporary_equivalences): Add comment.
3892
3893 2013-09-20 Yufeng Zhang <yufeng.zhang@arm.com>
3894
3895 * config/aarch64/aarch64-builtins.c (aarch64_simd_expand_args):
3896 Call aarch64_simd_expand_args to update op[argc].
3897
3898 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3899
3900 * plugin.c (parse_plugin_arg_opt): Accept equal sign inside
3901 plugin argument.
3902
3903 2013-09-20 Basile Starynkevitch <basile@starynkevitch.net>
3904
3905 * gengtype.c (file_rules): Added rule for *.cc files.
3906 (get_output_file_with_visibility): Give fatal message when no
3907 rules found.
3908
3909 2013-09-20 Renlin Li <renlin.li@arm.com>
3910
3911 * config/aarch64/aarch64.c (aarch64_expand_prologue): Use plus_constant.
3912 (aarch64_expand_epilogue): Likewise.
3913 (aarch64_legitimize_reload_address): Likewise.
3914
3915 2013-09-20 Bernd Edlinger <bernd.edlinger@hotmail.de>
3916
3917 PR middle-end/57748
3918 * expr.c (expand_assignment): Remove misalignp code path.
3919
3920 2013-09-20 Marek Polacek <polacek@redhat.com>
3921
3922 PR sanitizer/58413
3923 * ubsan.c (get_ubsan_type_info_for_type): Use TYPE_SIZE instead of
3924 TYPE_PRECISION. Add asserts.
3925
3926 2013-09-20 Richard Biener <rguenther@suse.de>
3927
3928 PR tree-optimization/58453
3929 * tree-loop-distribution.c (distribute_loop): Apply the cost
3930 model for -ftree-loop-distribute-patterns, too.
3931
3932 2013-09-20 Richard Biener <rguenther@suse.de>
3933
3934 PR middle-end/58473
3935 * tree-chrec.h (build_polynomial_chrec): Use gcc_checking_assert,
3936 make type comparison less strict.
3937
3938 2013-09-20 Alan Modra <amodra@gmail.com>
3939
3940 * configure: Regenerate.
3941 * aclocal.m4: Regenerate.
3942
3943 2013-09-20 Marek Polacek <polacek@redhat.com>
3944
3945 PR other/58467
3946 * doc/extend.texi: Document that attribute used is meant to be used
3947 on variables with static storage duration.
3948
3949 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3950
3951 PR tree-optimization/58472
3952 * tree-vect-stmts.c (vectorizable_store, vectorizable_load): For
3953 simd_lane_access set inv_p = false.
3954 * omp-low.c (lower_rec_input_clauses): Set TREE_NO_WARNING on
3955 the simduid magic VAR_DECL.
3956
3957 2013-09-19 Jan Hubicka <jh@suse.cz>
3958
3959 * i386.c (generic_memcpy, generic_memset): Fix 32bit template.
3960
3961 2013-09-17 Jeff Law <law@redhat.com>
3962
3963 * tree-ssa-dom.c (record_temporary_equivalences): New function
3964 split out of dom_opt_dom_walker::after_dom_children.
3965 (dom_opt_dom_walker::thread_across_edge): Move common code
3966 in here from dom_opt_dom_walker::after_dom_children.
3967 (dom_opt_dom_walker::after_dom_children): Corresponding simplifictions.
3968
3969 2013-09-19 Jan Hubicka <jh@suse.cz>
3970
3971 * i386.h (TARGET_GENERIC32, TARGET_GENERIC64): Remove.
3972 (TARGET_GENERIC): Use PROCESOR_GENERIC
3973 (enum processor_type): Unify generic32 and 64.
3974 * i386.md (cpu): Likewise.
3975 * x86-tune.def (use_leave): Enable for generic32.
3976 (avoid_vector_decode, slow_imul_imm32_mem, slow_imul_imm8): Likewise.
3977 * athlon.md: Change generic64 to generic in all occurences.
3978 * i386-c.c (ix86_target_macros_internal): Unify generic64 and 32.
3979 (ix86_target_macros_internal): Likewise.
3980 * driver-i386.c (host_detect_local_cpu): Likewise.
3981 * i386.c (generic64_memcpy, generic64_memset, generic64_cost): Rename
3982 to ..
3983 (generic_memcpy, generic_memset, generic_cost): This one.
3984 (generic32_memcpy, generic32_memset, generic32_cost): Remove.
3985 (m_GENERIC32, m_GENERIC64): Remove.
3986 (m_GENERIC): Turn into one flag.
3987 (processor_target): Unify generic tunnings.
3988 (ix86_option_override_internal): Replace generic32/64 by generic.
3989 (ix86_issue_rate): Likewise.
3990 (ix86_adjust_cost): Likewise.
3991
3992 2013-09-19 Jan Hubicka <jh@suse.cz>
3993
3994 * cgraph.c (cgraph_create_edge_1): Avoid uninitialized read
3995 of speculative flag.
3996
3997 2013-09-19 Jakub Jelinek <jakub@redhat.com>
3998
3999 * omp-low.c (expand_omp_sections): Always pass len - 1 to
4000 GOMP_sections_start, even if !exit_reachable.
4001
4002 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4003
4004 * lra-constraints.c (need_for_all_save_p): Use macro
4005 HARD_REGNO_CALL_PART_CLOBBERED.
4006 * lra-lives.c (check_pseudos_live_through_calls): Use the macro to
4007 set up pseudo conflict hard regs.
4008
4009 2013-09-18 Michael Meissner <meissner@linux.vnet.ibm.com>
4010
4011 PR target/58452
4012 * config/rs6000/paired.md (movmisalignv2sf): Fix to allow memory
4013 operands.
4014
4015 2013-09-18 Vladimir Makarov <vmakarov@redhat.com>
4016
4017 PR rtl-optimization/58438
4018 * lra.c (lra): Clear lra_optional_reload_pseudos in upper loop.
4019 * lra-constraints.c (undo_optional_reloads): Keep optional reloads
4020 from previous subpasses.
4021
4022 2013-09-18 Richard Earnshaw <rearnsha@arm.com>
4023
4024 * arm.c (arm_get_frame_offsets): Validate architecture supports
4025 LDRD/STRD before accepting the tuning preference.
4026 (arm_expand_prologue): Likewise.
4027 (arm_expand_epilogue): Likewise.
4028
4029 2013-09-18 Richard Biener <rguenther@suse.de>
4030
4031 PR tree-optimization/58417
4032 * tree-chrec.c (chrec_fold_plus_1): Assert that we do not
4033 have chrecs with symbols defined in the loop as operands.
4034 (chrec_fold_multiply): Likewise.
4035 * tree-scalar-evolution.c (interpret_rhs_expr): Instantiate
4036 parameters before folding binary operations.
4037 (struct instantiate_cache_entry_hasher): Remove.
4038 (struct instantiate_cache_type): Use a pointer-map.
4039 (instantiate_cache_type::instantiate_cache_type): New function.
4040 (instantiate_cache_type::get): Likewise.
4041 (instantiate_cache_type::set): Likewise.
4042 (instantiate_cache_type::~instantiate_cache_type): Adjust.
4043 (get_instantiated_value_entry): Likewise.
4044 (global_cache): New global.
4045 (instantiate_scev_r, instantiate_scev_poly, instantiate_scev_binary,
4046 instantiate_array_ref, instantiate_scev_convert, instantiate_scev_3,
4047 instantiate_scev_2, instantiate_scev_1): Do not pass along cache.
4048 (instantiate_scev_name): Adjust.
4049 (instantiate_scev): Construct global instead of local cache.
4050 (resolve_mixers): Likewise.
4051
4052 2013-09-18 Daniel Morris <danielm@ecoscentric.com>
4053 Paolo Carlini <paolo.carlini@oracle.com>
4054
4055 PR c++/58458
4056 * doc/implement-cxx.texi: Fix references to the C++ standards.
4057
4058 2013-09-18 Jakub Jelinek <jakub@redhat.com>
4059
4060 * omp-low.c (copy_var_decl): Copy DECL_ATTRIBUTES.
4061 * tree-vect-data-refs.c (vect_analyze_data_refs): For
4062 simd_lane_access drs, update also DR_ALIGNED_TO.
4063
4064 2013-09-18 Marek Polacek <polacek@redhat.com>
4065
4066 PR sanitizer/58411
4067 * doc/extend.texi: Document no_sanitize_undefined attribute.
4068 * builtins.c (fold_builtin_0): Don't sanitize function if it has the
4069 no_sanitize_undefined attribute.
4070
4071 2013-09-18 Nick Clifton <nickc@redhat.com>
4072
4073 * config/msp430/msp430.h (ASM_SPEC): Pass -md on to the assembler.
4074 (ASM_DECLARE_FUNCTION_NAME): Define.
4075
4076 2013-09-17 Trevor Saunders <tsaunders@mozilla.com>
4077
4078 * compare-elim.c (find_comparison_dom_walker): New class
4079 (find_comparisons_in_bb): Rename to
4080 find_comparison_dom_walker::before_dom_children
4081 (find_comparisons): Adjust
4082 * domwalk.c (walk_dominator_tree): Rename to dom_walker::walk, and
4083 adjust.
4084 (init_walk_dominator_tree, fini_walk_dominator_tree): Remove
4085 * domwalk.h (dom_walk_data): Convert it To a class dom_walker.
4086 (init_walk_dominator_tree): Remove declaration.
4087 (fini_walk_dominator_tree): Remove declaration.
4088 * fwprop.c (single_def_use_dom_walker): New class
4089 (single_def_use_enter_block): Convert to
4090 single_def_use_dom_walker::before_dom_children.
4091 (single_def_use_leave_block): Convert to
4092 single_def_use_dom_walker::after_dom_children.
4093 (build_single_def_use_links): Adjust.
4094 * gimple-ssa-strength-reduction.c (find_candidates_dom_walker): New
4095 class.
4096 (find_candidates_in_block): Convert to
4097 find_candidates_dom_walker::before_dom_children.
4098 (execute_strength_reduction): Adjust.
4099 * graphite-sese-to-poly.c (struct bsc, build_sese_conditions): Remove.
4100 (sese_dom_walker): New class.
4101 (sese_dom_walker::sese_dom_walker): New constructor.
4102 (sese_dom_walker::~sese_dom_walker): New destructor.
4103 (build_sese_conditions_before): Convert to
4104 sese_dom_walker::before_dom_children.
4105 (build_sese_conditions_after): Convert to
4106 sese_dom_walker::after_dom_children.
4107 (build_poly_scop): Adjust
4108 * tree-into-ssa.c (rewrite_dom_walker): New class
4109 (rewrite_enter_block): Convert to
4110 rewrite_dom_walker::before_dom_children.
4111 (rewrite_leave_block): Convert to
4112 rewrite_dom_walker::after_dom_children.
4113 (rewrite_update_dom_walker): New class.
4114 (rewrite_update_enter_block): Convert to
4115 rewrite_update_dom_walker::before_dom_children.
4116 (rewrite_update_leave_block): Convert to
4117 rewrite_update_dom_walker::after_dom_children.
4118 (rewrite_blocks, rewrite_into_ssa): Adjust.
4119 (mark_def_dom_walker): New class.
4120 (mark_def_dom_walker::mark_def_dom_walker): New constructor.
4121 (mark_def_dom_walker::~mark_def_dom_walker): New destructor.
4122 (mark_def_sites_blocks): Convert to
4123 mark_def_dom_walker::before_dom_children.
4124 (mark_def_site_blocks): Remove.
4125 * tree-ssa-dom.c (dom_opt_dom_walker): New class.
4126 (tree_ssa_dominator_optimize): Adjust.
4127 (dom_thread_across_edge): Convert to method
4128 dom_opt_dom_walker::thread_across_edge.
4129 (dom_opt_enter_block): Convert to member function
4130 dom_opt_dom_walker::before_dom_children.
4131 (dom_opt_leave_block): Convert to member function
4132 dom_opt_dom_walker::after_dom_children.
4133 * tree-ssa-dse.c (dse_dom_walker): New class.
4134 (dse_enter_block): Convert to member function
4135 dse_dom_walker::before_dom_children.
4136 (tree_ssa_dse): Adjust.
4137 * tree-ssa-loop-im.c (invariantness_dom_walker): New class.
4138 (determine_invariantness_stmt): Convert to method
4139 invariantness_dom_walker::before_dom_children.
4140 (determine_invariantness): Remove
4141 (move_computations_dom_walker): New class.
4142 (move_computations_stmt): Convert to method
4143 move_computations_dom_walker::before_dom_children.
4144 (move_computations, tree_ssa_lim): Adjust.
4145 * tree-ssa-phiopt.c (nontrapping_dom_walker): new class
4146 (nt_init_block): Make method
4147 notrappping_dom_walker::before_dom_children.
4148 (nt_fini_block): Make
4149 method nontrapping_dom_walker::after_dom_children.
4150 (get_non_trapping): Adjust.
4151 * tree-ssa-pre.c (eliminate_dom_walker): New class.
4152 (eliminate_bb): Make method eliminate_dom_walker::before_dom_children.
4153 (eliminate_leave_block): Make method.
4154 eliminate_dom_walker::after_dom_children.
4155 (eliminate): Adjust
4156 * tree-ssa-strlen.c (strlen_dom_walker): New class.
4157 (strlen_enter_block): Make method
4158 strlen_dom_walker::before_dom_children.
4159 (strlen_leave_block): Make
4160 method strlen_dom_walker::after_dom_children.
4161 (tree_ssa_strlen): Adjust.
4162 * tree-ssa-uncprop.c (uncprop_dom_walker): New class.
4163 (tree_ssa_uncprop): Adjust.
4164 (uncprop_leave_block): Make method
4165 uncprop_dom_walker::after_dom_children.
4166 (uncprop_leave_block): Make method
4167 uncprop_dom_walker::before_dom_children.
4168
4169 2013-09-18 Bin Cheng <bin.cheng@arm.com>
4170
4171 * config/arm/arm.c (thumb1_reorg): Search for flag setting insn before
4172 branch in same basic block. Check both src and dest of the move insn.
4173
4174 2013-09-17 Nick Clifton <nickc@redhat.com>
4175
4176 * config/rl78/rl78-real.md (bf): New pattern.
4177 (bt): New pattern.
4178 * config/rl78/rl78.c (rl78_print_operand_1): Handle %B.
4179 (rl78_print_operand): Do not put a # before a %B.
4180 * config/rl78/rl78.opt: Tweak doc strings.
4181
4182 2013-09-17 DJ Delorie <dj@redhat.com>
4183
4184 * config/rl78/constraints.md (Wcv): Allow up to $r31.
4185 * config/rl78/rl78.c (rl78_asm_file_start: Likewise.
4186 (rl78_option_override): Likewise, if -mallregs.
4187 (is_virtual_register): Likewise.
4188 * config/rl78/rl78.h (reg_class): Extend VREGS to $r31.
4189 (REGNO_OK_FOR_BASE_P): Likewise.
4190 * config/rl78/rl78.opt (-mallregs): New.
4191
4192 2013-09-17 Nick Clifton <nickc@redhat.com>
4193
4194 * config/rl78/rl78.c (need_to_save): Change return type to bool.
4195 For interrupt functions: save all call clobbered registers if the
4196 interrupt handler is not a leaf function.
4197 (rl78_expand_prologue): Always recompute the frame information.
4198 For interrupt functions: only select bank 0 if one of the bank 0
4199 registers is going to be psuhed.
4200
4201 2013-09-17 DJ Delorie <dj@redhat.com>
4202
4203 * config/rl78/constraints.md: For each W* constraint, rename to C*
4204 and create a W* constraint that checks for an optional ES: prefix
4205 pattern also.
4206 * config/rl78/rl78.md (UNS_ES_ADDR): New.
4207 (es_addr): New. Used to wrap far addresses.
4208 * config/rl78/rl78-protos.h (rl78_es_addr): New.
4209 (rl78_es_base): New.
4210 * config/rl78/rl78.c (rl78_as_legitimate_address): Accept "unspec"
4211 wrapped far addresses.
4212 (rl78_print_operand_1): Unwrap far addresses before processing.
4213 (rl78_lo16): Wrap far addresses in unspecs.
4214 (rl78_es_addr): New.
4215 (rl78_es_base): New.
4216 (insn_ok_now): Check for not-yet-wrapped far addresses.
4217 (transcode_memory_rtx): Properly re-wrap far addresses.
4218
4219 2013-09-17 Sebastian Huber <sebastian.huber@embedded-brains.de>
4220
4221 * config/sparc/t-rtems: Add leon3 multilibs.
4222
4223 2013-09-17 Cong Hou <congh@google.com>
4224
4225 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Fix a bug
4226 when checking the dot production pattern. The type of rhs operand
4227 of multiply is now checked correctly.
4228
4229 2013-09-17 Jeff Law <law@redhat.com>
4230
4231 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4232 edge implied equivalences into successor phis.
4233 * tree-ssa-threadupdate.c (phi_args_equal_on_edges): Moved into
4234 here from tree-ssa-threadedge.c.
4235 (mark_threaded_blocks): When threading through a joiner, if both
4236 successors of the joiner's clone reach the same block, verify the
4237 PHI arguments are equal. If not, cancel the jump threading request.
4238 * tree-ssa-threadedge.c (phi_args_equal_on_edges): Moved into
4239 tree-ssa-threadupdate.c
4240 (thread_across_edge): Don't check PHI argument equality when
4241 threading through joiner block here.
4242
4243 2013-09-17 Andrew MacLeod <amacleod@redhat.com>
4244
4245 * tree-flow.h (ssa_undefined_value_p): Remove prototype.
4246 * tree-ssa.c (ssa_undefined_value_p): Move pass independent parts here.
4247 (warn_uninit, warn_uninitialized_vars,
4248 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4249 Move to tree-ssa-uninit.c.
4250 * tree-ssa-uninit.c (ssa_undefined_value_p): Move to tree-ssa.c.
4251 (has_undefined_value_p): New. Pass dependant parts of
4252 ssa_undefined_value_p.
4253 (uninit_undefined_value_p): Use has_undefined_value_p.
4254 (warn_uninit, warn_uninitialized_vars,
4255 execute_early_warn_uninitialized, make_pass_early_warn_uninitialized):
4256 Move from tree-ssa.c.
4257 * tree-ssa.h: Adjust prototypes.
4258
4259 2013-09-17 Jan Hubicka <jh@suse.cz>
4260
4261 PR middle-end/58332
4262 * cif-code.def (FUNCTION_NOT_OPTIMIZED): New CIF code.
4263 * ipa-inline.c (can_inline_edge_p): Do not downgrade
4264 FUNCTION_NOT_OPTIMIZED.
4265 * ipa-inline-analysis.c (compute_inline_parameters): Function
4266 not optimized is not inlinable unless it is alwaysinline.
4267 (inline_analyze_function): Force calls in not optimized
4268 function not inlinable.
4269
4270 2013-09-17 Jan Hubicka <jh@suse.cz>
4271
4272 PR middle-end/58329
4273 * ipa-devirt.c (ipa_devirt): Be ready for symtab_nonoverwritable_alias
4274 to return NULL.
4275 * ipa.c (function_and_variable_visibility): Likewise.
4276 * ipa-profile.c (ipa_profile): Likewise.
4277
4278 2013-09-17 Bernd Edlinger <bernd.edlinger@hotmail.de>
4279
4280 PR ipa/58398
4281 * cgraph.c (cgraph_function_body_availability): Check for ifunc
4282 attribute, and don't inline the resolver in this case.
4283
4284 2013-09-17 Teresa Johnson <tejohnson@google.com>
4285
4286 * coverage.c (get_coverage_counts): Add missing newline.
4287
4288 2013-09-17 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4289
4290 PR tree-optimization/58088
4291 * fold-const.c (mask_with_trailing_zeros): New function.
4292 (fold_binary_loc): Make sure we don't recurse infinitely
4293 when the X in (X & C1) | C2 is a tree of the form (Y * K1) & K2.
4294 Use mask_with_trailing_zeros where appropriate.
4295
4296 2013-09-17 Yuri Rumyantsev <ysrumyan@gmail.com>
4297
4298 * config/i386/i386.c (distance_agu_use_in_bb) : Proper initialization
4299 of 'prev' var to get better distance estimation.
4300
4301 2013-09-17 Eric Botcazou <ebotcazou@adacore.com>
4302
4303 * tree-inline.h (struct copy_body_data): Add transform_parameter.
4304 * tree-inline.c (is_parameter_of): New predicate.
4305 (remap_gimple_op_r): Do not propagate TREE_THIS_NOTRAP on MEM_REF if
4306 a parameter has been remapped.
4307 (copy_tree_body_r): Likewise on INDIRECT_REF and MEM_REF.
4308 (optimize_inline_calls): Initialize transform_parameter.
4309 (copy_gimple_seq_and_replace_locals): Likewise.
4310 (tree_function_versioning): Likewise.
4311 (maybe_inline_call_in_expr): Likewise.
4312
4313 2013-09-17 Nick Clifton <nickc@redhat.com>
4314
4315 * config/msp430/msp430-protos.h: Add prototypes for new functions.
4316 * config/msp430/msp430.c (msp430_preserve_reg_p): Add support for
4317 interrupt handlers.
4318 (is_attr_func): New function.
4319 (msp430_is_interrupt_func): New function.
4320 (is_naked_func): New function.
4321 (is_reentrant_func): New function.
4322 (is_critical_func): New function.
4323 (msp430_start_function): Add annotations for function attributes.
4324 (msp430_attr): New function.
4325 (msp430_attribute_table): New.
4326 (msp430_function_section): New function.
4327 (TARGET_ASM_FUNCTION_SECTION): Define.
4328 (msp430_builtin): New enum.
4329 (msp430_init_builtins): New function.
4330 (msp430_builtin_devl): New function.
4331 (msp430_expand_builtin): New function.
4332 (TARGET_INIT_BUILTINS): Define.
4333 (TARGET_EXPAND_BUILTINS): Define.
4334 (TARGET_BUILTIN_DECL): Define.
4335 (msp430_expand_prologue): Add support for naked, interrupt,
4336 critical and reentrant functions.
4337 (msp430_expand_epilogue): Likewise.
4338 (msp430_print_operand): Handle 'O' character.
4339 * config/msp430/msp430.h (TARGET_CPU_CPP_BUILTINS): Define
4340 NO_TRAMPOLINES.
4341 * config/msp430/msp430.md (unspec): Add UNS_DINT, UNS_EINT,
4342 UNS_PUSH_INTR, UNS_POP_INTR, UNS_BIC_SR, UNS_BIS_SR.
4343 (pushm): Use a 'n' rather than an 'i' constraint.
4344 (msp_return): Add generation of the interrupt return instruction.
4345 (disable_interrupts): New pattern.
4346 (enable_interrupts): New pattern.
4347 (push_intr_state): New pattern.
4348 (pop_intr_state): New pattern.
4349 (bic_SR): New pattern.
4350 (bis_SR): New pattern.
4351 * doc/extend.texi: Document MSP430 function attributes and builtin
4352 functions.
4353
4354 2013-09-17 Richard Biener <rguenther@suse.de>
4355
4356 PR tree-optimization/58432
4357 * tree-loop-distribution.c (tree_loop_distribution): Also
4358 scan PHIs for outside loop uses and seed a partition from them.
4359
4360 2013-09-17 Bin Cheng <bin.cheng@arm.com>
4361
4362 * gimple-ssa-strength-reduction.c (backtrace_base_for_ref): New.
4363 (restructure_reference): Call backtrace_base_for_ref.
4364
4365 2013-09-17 Alan Modra <amodra@gmail.com>
4366
4367 PR target/57589
4368 * config/rs6000/driver-rs6000.c (elf_platform): Revert 2013-06-11
4369 patch.
4370
4371 2013-09-16 DJ Delorie <dj@redhat.com>
4372
4373 * config/rl78/rl78.c (rl78_asm_file_start): Specify alternate
4374 vregs location for RL78/G10.
4375 (rl78_expand_prologue): Avoid SEL on G10.
4376 (rl78_expand_epilogue): Likewise.
4377 (rl78_peep_movhi_p): Can't move a constant to memory in HImode.
4378 * config/rl78/rl78.h (TARGET_CPU_CPP_BUILTINS): Define
4379 __RL78_G10__ when appropriate.
4380 (ASM_SPEC): Pass -mg10 along to the assembler.
4381 * config/rl78/rl78.md (sel_rb): Disable for G10.
4382 * config/rl78/rl78.opt: Add -mg10 option.
4383 * config/rl78/t-rl78: Add -mg10 multilib.
4384
4385 2013-09-16 Xinliang David Li <davidxl@google.com>
4386
4387 * tree-if-conv.c (main_tree_if_conversion): Check new flag.
4388 * omp-low.c (omp_max_vf): Ditto.
4389 (expand_omp_simd): Ditto.
4390 * tree-vectorizer.c (vectorize_loops): Ditto.
4391 (gate_vect_slp): Ditto.
4392 (gate_increase_alignment): Ditto.
4393 * tree-ssa-pre.c (inhibit_phi_insertion): Ditto.
4394 * tree-ssa-loop.c (gate_tree_vectorize): Ditto.
4395 (gate_tree_vectorize): Name change.
4396 (tree_vectorize): Ditto.
4397 (pass_vectorize::gate): Call new function.
4398 (pass_vectorize::execute): Ditto.
4399 * opts.c: O3 default setting change.
4400 (finish_options): Check new flag.
4401 * doc/invoke.texi: Document new flags.
4402 * common.opt: New flags.
4403
4404 2013-09-16 Andreas Schwab <schwab@linux-m68k.org>
4405
4406 * doc/tm.texi.in (Cond Exec Macros): Remove node.
4407 (Condition Code): Don't reference it.
4408 * doc/tm.texi: Regenerate.
4409
4410 2013-09-16 Vladimir Makarov <vmakarov@redhat.com>
4411
4412 PR middle-end/58418
4413 * lra-constraints.c (undo_optional_reloads): Consider all optional
4414 reload even if it did not get a hard reg.
4415
4416 2013-09-16 Teresa Johnson <tejohnson@google.com>
4417
4418 * dumpfile.c (dump_loc): Remove newline emission.
4419 * tree-vect-data-refs.c (vect_lanes_optab_supported_p): Add newline
4420 emission to dump_printf_loc calls where missing.
4421 (vect_mark_for_runtime_alias_test): Ditto.
4422 (vect_analyze_data_ref_dependence): Ditto.
4423 (vect_analyze_data_ref_dependences): Ditto.
4424 (vect_slp_analyze_data_ref_dependence): Ditto.
4425 (vect_slp_analyze_data_ref_dependences): Ditto.
4426 (vect_compute_data_ref_alignment): Ditto.
4427 (vect_update_misalignment_for_peel): Ditto.
4428 (vect_verify_datarefs_alignment): Ditto.
4429 (vector_alignment_reachable_p): Ditto.
4430 (vect_get_data_access_cost): Ditto.
4431 (vect_enhance_data_refs_alignment): Ditto.
4432 (vect_find_same_alignment_drs): Ditto.
4433 (vect_analyze_data_refs_alignment): Ditto.
4434 (vect_analyze_group_access): Ditto.
4435 (vect_analyze_data_ref_access): Ditto.
4436 (vect_analyze_data_ref_accesses): Ditto.
4437 (vect_prune_runtime_alias_test_list): Ditto.
4438 (vect_analyze_data_refs): Ditto.
4439 (vect_create_addr_base_for_vector_ref): Ditto.
4440 (vect_create_data_ref_ptr): Ditto.
4441 (vect_grouped_store_supported): Ditto.
4442 (vect_grouped_load_supported): Ditto.
4443 * value-prof.c (check_counter): Ditto.
4444 (check_ic_target): Ditto.
4445 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Ditto.
4446 (vect_recog_widen_mult_pattern): Ditto.
4447 (vect_recog_widen_sum_pattern): Ditto.
4448 (vect_recog_over_widening_pattern): Ditto.
4449 (vect_recog_widen_shift_pattern): Ditto.
4450 (vect_recog_rotate_pattern): Ditto.
4451 (vect_recog_vector_vector_shift_pattern): Ditto.
4452 (vect_recog_divmod_pattern): Ditto.
4453 (vect_recog_mixed_size_cond_pattern): Ditto.
4454 (vect_recog_bool_pattern): Ditto.
4455 (vect_pattern_recog_1): Ditto.
4456 (vect_pattern_recog): Ditto.
4457 * tree-vect-loop.c (vect_determine_vectorization_factor): Ditto.
4458 (vect_is_simple_iv_evolution): Ditto.
4459 (vect_analyze_scalar_cycles_1): Ditto.
4460 (vect_get_loop_niters): Ditto.
4461 (vect_analyze_loop_1): Ditto.
4462 (vect_analyze_loop_form): Ditto.
4463 (vect_analyze_loop_operations): Ditto.
4464 (vect_analyze_loop_2): Ditto.
4465 (vect_analyze_loop): Ditto.
4466 (report_vect_op): Ditto.
4467 (vect_is_slp_reduction): Ditto.
4468 (vect_is_simple_reduction_1): Ditto.
4469 (vect_get_known_peeling_cost): Ditto.
4470 (vect_estimate_min_profitable_iters): Ditto.
4471 (vect_model_reduction_cost): Ditto.
4472 (vect_model_induction_cost): Ditto.
4473 (get_initial_def_for_induction): Ditto.
4474 (vect_create_epilog_for_reduction): Ditto.
4475 (vectorizable_reduction): Ditto.
4476 (vectorizable_induction): Ditto.
4477 (vectorizable_live_operation): Ditto.
4478 (vect_loop_kill_debug_uses): Ditto.
4479 (vect_transform_loop): Ditto.
4480 * tree-vect-stmts.c (vect_mark_relevant): Ditto.
4481 (vect_stmt_relevant_p): Ditto.
4482 (process_use): Ditto.
4483 (vect_mark_stmts_to_be_vectorized): Ditto.
4484 (vect_model_simple_cost): Ditto.
4485 (vect_model_promotion_demotion_cost): Ditto.
4486 (vect_model_store_cost): Ditto.
4487 (vect_get_store_cost): Ditto.
4488 (vect_model_load_cost): Ditto.
4489 (vect_get_load_cost): Ditto.
4490 (vect_init_vector_1): Ditto.
4491 (vect_get_vec_def_for_operand): Ditto.
4492 (vect_finish_stmt_generation): Ditto.
4493 (vectorizable_call): Ditto.
4494 (vectorizable_conversion): Ditto.
4495 (vectorizable_assignment): Ditto.
4496 (vectorizable_shift): Ditto.
4497 (vectorizable_operation): Ditto.
4498 (vectorizable_store): Ditto.
4499 (vectorizable_load): Ditto.
4500 (vectorizable_condition): Ditto.
4501 (vect_analyze_stmt): Ditto.
4502 (vect_transform_stmt): Ditto.
4503 (vect_is_simple_use): Ditto.
4504 * tree-vect-loop-manip.c (slpeel_make_loop_iterate_ntimes): Ditto.
4505 (vect_can_advance_ivs_p): Ditto.
4506 (vect_update_ivs_after_vectorizer): Ditto.
4507 (vect_do_peeling_for_loop_bound): Ditto.
4508 (vect_gen_niters_for_prolog_loop): Ditto.
4509 (vect_update_inits_of_drs): Ditto.
4510 (vect_create_cond_for_alias_checks): Ditto.
4511 * tree-vect-slp.c (vect_get_and_check_slp_defs): Ditto.
4512 (vect_build_slp_tree_1): Ditto.
4513 (vect_supported_load_permutation_p): Ditto.
4514 (vect_analyze_slp_instance): Ditto.
4515 (vect_analyze_slp): Ditto.
4516 (vect_make_slp_decision): Ditto.
4517 (vect_detect_hybrid_slp): Ditto.
4518 (vect_bb_vectorization_profitable_p): Ditto.
4519 (vect_slp_analyze_bb_1): Ditto.
4520 (vect_update_slp_costs_according_to_vf): Ditto.
4521 (vect_get_mask_element): Ditto.
4522 (vect_transform_slp_perm_load): Ditto.
4523 (vect_schedule_slp_instance): Ditto.
4524 (vect_schedule_slp): Ditto.
4525 (vect_slp_transform_bb): Ditto.
4526 * profile.c (read_profile_edge_counts): Ditto.
4527 (compute_branch_probabilities): Ditto.
4528 * coverage.c (get_coverage_counts): Ditto.
4529
4530 2013-09-16 Diego Novillo <dnovillo@google.com>
4531
4532 * tree-core.h: Add missing comment lines from refactoring of tree.h.
4533
4534 2013-09-16 Jan Hubicka <jh@suse.cz>
4535
4536 * gimple-fold.c (can_refer_decl_in_current_unit_p): Do not accept
4537 abstract functions; for static functions check the presence of body.
4538
4539 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4540
4541 * config/aarch64/aarch64-simd-builtins.def (fma): New.
4542 * config/aarch64/aarch64-simd.md
4543 (aarch64_mla_elt<mode>): New.
4544 (aarch64_mla_elt_<vswap_width_name><mode>): Likewise.
4545 (aarch64_mls_elt<mode>): Likewise.
4546 (aarch64_mls_elt_<vswap_width_name><mode>): Likewise.
4547 (aarch64_fma4_elt<mode>): Likewise.
4548 (aarch64_fma4_elt_<vswap_width_name><mode>): Likewise.
4549 (aarch64_fma4_elt_to_128v2df): Likewise.
4550 (aarch64_fma4_elt_to_64df): Likewise.
4551 (fnma<mode>4): Likewise.
4552 (aarch64_fnma4_elt<mode>): Likewise.
4553 (aarch64_fnma4_elt_<vswap_width_name><mode>): Likewise.
4554 (aarch64_fnma4_elt_to_128v2df): Likewise.
4555 (aarch64_fnma4_elt_to_64df): Likewise.
4556 * config/aarch64/iterators.md (VDQSF): New.
4557 * config/aarch64/arm_neon.h
4558 (vfm<as><sdq>_lane<q>_f<32, 64>): Convert to C implementation.
4559 (vml<sa><q>_lane<q>_<fsu><16, 32, 64>): Likewise.
4560
4561 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4562
4563 * config/aarch64/aarch64-simd.md (aarch64_mul3_elt<mode>): New.
4564 (aarch64_mul3_elt_<vswap_width_name><mode>): Likewise.
4565 (aarch64_mul3_elt_to_128df): Likewise.
4566 (aarch64_mul3_elt_to_64v2df): Likewise.
4567 * config/aarch64/iterators.md (VEL): Also handle DFmode.
4568 (VMUL): New.
4569 (VMUL_CHANGE_NLANES) Likewise.
4570 (h_con): Likewise.
4571 (f): Likewise.
4572 * config/aarch64/arm_neon.h
4573 (vmul<q>_lane<q>_<suf><16,32,64>): Convert to C implementation.
4574
4575 2013-09-16 James Greenhalgh <james.greenhalgh@arm.com>
4576
4577 * config/aarch64/arm_neon.h
4578 (vcvtx_high_f32_f64): Fix parameters.
4579
4580 2013-09-16 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4581 Uros Bizjak <ubizjak@gmail.com>
4582
4583 * config/alpha.c: Include tree-ssanames.h.
4584
4585 2013-09-16 Richard Biener <rguenther@suse.de>
4586
4587 * tree-loop-distribution.c (enum rdg_dep_type): Add control_dd.
4588 (dot_rdg_1): Handle control_dd.
4589 (create_edge_for_control_dependence): New function.
4590 (create_rdg_edges): Add control dependences if asked for.
4591 (build_rdg): Likewise.
4592 (generate_loops_for_partition): If there are not necessary
4593 control stmts remove all their dependencies.
4594 (collect_condition_stmts, rdg_flag_loop_exits): Remove.
4595 (distribute_loop): Pass on control dependences.
4596 (tree_loop_distribution): Compute control dependences and remove
4597 restriction on number of loop nodes.
4598
4599 2013-09-16 Jakub Jelinek <jakub@redhat.com>
4600
4601 * ipa-prop.c (ipa_compute_jump_functions_for_edge): Return early
4602 for internal calls.
4603
4604 2013-09-16 Richard Sandiford <rdsandiford@googlemail.com>
4605
4606 * cse.c (try_const_anchors): Punt on CC modes.
4607
4608 2013-09-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4609
4610 * config/vax/constraints.md (T): Add missing CONSTANT_P check.
4611
4612 2013-09-14 John David Anglin <danglin@gcc.gnu.org>
4613
4614 PR target/58382
4615 * config/pa/pa.c (pa_expand_prologue): Change mode in gen_rtx_POST_INC
4616 calls to word_mode.
4617
4618 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4619
4620 PR target/48094
4621 * config/darwin.c (darwin_objc2_section): Note if ObjC Metadata is
4622 seen.
4623 (darwin_objc1_section): Likewise.
4624 (darwin_file_end): Emit Image Info section when required.
4625
4626 2013-09-14 Jan Hubicka <jh@suse.cz>
4627
4628 * tree-into-ssa.c (gate_into_ssa): New.
4629 (pass_data_build_ssa): Use it.
4630 * cgraph.h (expand_thunk): Update prototype.
4631 * cgraphunit.c (analyze_function): Expand thunks early.
4632 (expand_thunk): Fix DECL_CONTEXT of reust_decl;
4633 build proper cgraph; set in_ssa_p; clear bogus TREE_ASM_WRITTEN;
4634 set lowered flag; do not add new function.
4635 (assemble_thunks_and_aliases): Update.
4636 * tree-ssa.c (gate_init_datastructures): New gate.
4637 (pass_data_init_datastructures): Use it.
4638
4639 2013-09-14 Iain Sandoe <iain@codesourcery.com>
4640
4641 PR target/58269
4642 * config/i386/i386.c (ix86_function_arg_regno_p): Make Darwin use the
4643 xmm register set described in the psABI.
4644
4645 2013-09-13 Evgeny Gavrin <e.gavrin@samsung.com>
4646
4647 * dwarf2out.c (should_emit_struct_debug): Add check
4648 for type_decl variable is not NULL.
4649
4650 2013-09-13 Jacek Caban <jacek@codeweavers.com>
4651
4652 * config.gcc: Use new winnt-c.c target hooks
4653 * config/t-winnt: New file
4654 * config/winnt-c.c: New file
4655 * doc/tm.texi.in: Document new hook
4656 * doc/tm.texi: Regenerated
4657
4658 2013-09-13 Jan Hubicka <jh@suse.cz>
4659
4660 PR middle-end/58094
4661 * ipa-inline.c (check_callers): New function.
4662 (check_caller_edge): Remove.
4663 (want_inline_function_to_all_callers_p): Also permit alises that are
4664 called dirrectly.
4665 (inline_to_all_callers): Terminate the walk when devirtualization
4666 introduce new calls.
4667
4668 2013-09-13 Jan Hubicka <jh@suse.cz>
4669
4670 * ipa-inline-analysis.c (struct growth_data): Add node.
4671 (do_estimate_growth_1): Fix detection of recursion.
4672
4673 2013-09-13 Jakub Jelinek <jakub@redhat.com>
4674
4675 PR tree-optimization/58392
4676 * tree-cfg.c (move_sese_region_to_fn): Rename loop variable
4677 to avoid shadowing of outer loop variable. If
4678 saved_cfun->has_simduid_loops or saved_cfun->has_force_vect_loops,
4679 replace_by_duplicate_decl simduid of loops that have it set and
4680 set dest_cfun->has_simduid_loops and/or
4681 dest_cfun->has_force_vect_loops.
4682 * omp-low.c (build_outer_var_ref): Call maybe_lookup_decl_in_outer_ctx
4683 instead of maybe_lookup_decl.
4684 * tree-inline.c (copy_loops): Change blocks_to_copy argument to id.
4685 Use id->blocks_to_copy instead of blocks_to_copy. Adjust recursive
4686 call. Copy over force_vect and copy and remap simduid. Set
4687 cfun->has_simduid_loops and/or cfun->has_force_vect_loops.
4688 (copy_cfg_body): Remove blocks_to_copy argument. Use
4689 id->blocks_to_copy instead of blocks_to_copy. Adjust copy_loops
4690 caller. Don't set cfun->has_simduid_loops and/or
4691 cfun->has_force_vect_loops here.
4692 (copy_body): Remove blocks_to_copy argument. Adjust copy_cfg_body
4693 caller.
4694 (expand_call_inline, tree_function_versioning): Adjust copy_body
4695 callers.
4696
4697 2013-09-13 Martin Jambor <mjambor@suse.cz>
4698
4699 PR bootstrap/58388
4700 * ipa-prop.c (try_make_edge_direct_simple_call): Be less strict in
4701 the assert if the edge was a speculative one.
4702
4703 2013-09-13 Richard Biener <rguenther@suse.de>
4704
4705 * tree-data-ref.h (known_dependences_p): Move here ...
4706 * tree-loop-distribution.c (known_dependences_p): ... from here.
4707 (dump_rdg_component, debug_rdg_component): Remove.
4708 (dump_rdg): Adjust.
4709 (generate_loops_for_partition): Use gimple_uid instead of
4710 relying on matching stmt visit order.
4711 (rdg_build_partitions): Take starting stmt vector.
4712 (ldist_gen): Merge into ...
4713 (distribute_loop): ... this function. Do not compute starting
4714 vertices vector.
4715 * tree-cfg.c (gimple_duplicate_bb): Copy UID for PHIs.
4716
4717 2013-09-13 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4718
4719 * config/arm/arm.md (arm_cmpsi_insn): Split rI alternative.
4720 Set type attribute correctly. Set predicable_short_it attribute.
4721 (cmpsi_shiftsi): Remove %? from output template.
4722
4723 2013-09-13 Richard Biener <rguenther@suse.de>
4724
4725 * tree-loop-distribution.c (struct rdg_component,
4726 rdg_defs_used_in_other_loops_p, free_rdg_components,
4727 rdg_build_components): Remove.
4728 (stmts_from_loop): Do not record virtual PHIs.
4729 (generate_loops_for_partition): Skip virtual PHIs.
4730 (build_rdg_partition_for_component): Rename to ...
4731 (build_rdg_partition_for_vertex): ... this and adjust.
4732 (rdg_build_partitions): Take a vector of starting vertices
4733 instead of components. Remove unnecessary leftover handling.
4734 (ldist_gen): Do not build components or record other stores.
4735 (distribute_loop): Do not distribute loops containing stmts
4736 with side-effects.
4737
4738 2013-09-13 Christian Bruel <christian.bruel@st.com>
4739
4740 PR target/58314
4741 * config/sh/sh.md (mov<mode>_reg_reg): Allow memory reloads.
4742
4743 2013-09-13 Kai Tietz <ktietz@redhat.com>
4744
4745 * config.gcc: Separate cases for mingw and cygwin targets,
4746 and add 64-bit cygwin target case.
4747
4748 * config/i386/winnt-cxx.c (i386_pe_type_dllexport_p): Don't
4749 dll-export inline-functions.
4750 * config/i386/winnt.c (i386_pe_determine_dllexport_p): Likewise.
4751
4752
4753 2013-09-13 Jeff Law <law@redhat.com>
4754
4755 PR middle-end/58387
4756 Revert:
4757 2013-09-06 Jeff Law <law@redhat.com>
4758
4759 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
4760 edge implied equivalences into successor phis.
4761
4762 2013-09-12 DJ Delorie <dj@redhat.com>
4763
4764 * config/rl78/rl78-virt.md: Change from | to \; for asm line
4765 separators.
4766
4767 2013-09-12 Brooks Moses <bmoses@google.com>
4768
4769 PR driver/42955
4770 * Makefile.in: Do not install driver binaries in $(target)/bin.
4771
4772 2013-09-12 DJ Delorie <dj@redhat.com>
4773
4774 * config/rl78/rl78.opt (mrelax): New.
4775 * config/rl78/rl78.h (ASM_SPEC): New, pass on -mrelax to gas.
4776 * config/rl78/rl78.h (LINK_SPEC): New, pass on -mrelax to ld.
4777
4778 * config/rl78/rl78.c (rl78_expand_prologue): Use AX to copy
4779 between SP and FP.
4780 (rl78_expand_epilogue): Likewise.
4781
4782 2013-09-12 Vladimir Makarov <vmakarov@redhat.com>
4783
4784 PR middle-end/58335
4785 * lra-eliminations.c (remove_reg_equal_offset_note): New.
4786 (eliminate_regs_in_insn): Rewrite frame pointer to hard frame
4787 pointer elimination with using remove_reg_equal_offset_note.
4788
4789 2013-09-12 DJ Delorie <dj@redhat.com>
4790
4791 * config/msp430/: New port.
4792 * config.gcc (msp430): Added.
4793 * doc/invoke.texi: Document MSP430 options.
4794 * doc/install.texi: Document msp430-elf
4795 * doc/md.texi: Document msp430-elf
4796 * doc/contrib.texi: Document msp430-elf
4797
4798 * cfgexpand.c (expand_debug_expr): Avoid sign-extending SImode to
4799 PSImode.
4800
4801 2013-09-12 Martin Jambor <mjambor@suse.cz>
4802
4803 PR ipa/58389
4804 * ipa-prop.c (remove_described_reference): Give up if the edge in the
4805 reference descriptor is NULL.
4806 (ipa_edge_removal_hook): If owning a reference descriptor, set its
4807 edge to NULL.
4808
4809 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
4810
4811 * tree-flow.h (FREE_SSANAMES): Move to tree-ssanames.c
4812 (SSANAMES, MODIFIED_NORETURN_CALLS, DEFAULT_DEFS, ptr_info_def,
4813 num_ssa_names, ssa_name): Move to tree-ssanames.h + prototypes.
4814 * tree-flow-inline.h (make_ssa_name, copy_ssa_name, duplicate_ssa_name,
4815 make_temp_ssa_name): move to tree-ssanames.h
4816 * tree-ssa-alias.h: Move prototype.
4817 * tree-ssa.h: Include tree-ssanames.h.
4818 * tree-ssanames.c (FREE_SSANAMES): Move to here.
4819 * tree-ssanames.h: New. Move items from tree-flow*.h
4820 * Makefile.in (tree-ssanames.h): Add to tree-ssanames.o and GTFILES.
4821
4822 2013-09-12 Richard Biener <rguenther@suse.de>
4823
4824 PR tree-optimization/58404
4825 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Also
4826 propagate non-invariant addresses into dereferences wrapped
4827 in component references.
4828
4829 2013-09-12 Richard Biener <rguenther@suse.de>
4830
4831 PR tree-optimization/58402
4832 * passes.def: Move pass_late_warn_uninitialized later.
4833
4834 2013-09-12 Andrew MacLeod <amacleod@redhat.com>
4835
4836 * tree-ssa.h: New. Move content from tree-flow.h and
4837 tree-flow-inline.h.
4838 * tree-flow.h (_edge_var_map, edge_var_map_vector): Move to tree-ssa.h.
4839 Move prototypes belonging to tree-ssa.c.
4840 * tree-flow-inline.h (redirect_edge_var_map_def,
4841 redirect_edge_var_map_result, redirect_edge_var_map_location): Move to
4842 tree-ssa.h.
4843 * gimple.h: Adjust prototypes.
4844 * tree-ssa.c (useless_type_conversion_p, types_compatible_p): Move
4845 to...
4846 * gimple.c (useless_type_conversion_p, types_compatible_p): Here.
4847 * tree.h: Move prototype to tree-ssa.h.
4848 * gengtype.c (open_base_files): Replace tree-flow.h with tree-ssa.h.
4849 * Makefile.in: (TREE_SSA_H, TREE_FLOW_H): Adjust dependencies.
4850 * alias.c, asan.c, builtins.c, calls.c, cfgexpand.c, cfghooks.c,
4851 cfgloop.c, cfgloopmanip.c, cgraph.c, cgraphbuild.c, cgraphclones.c,
4852 cgraphunit.c, dse.c, except.c, expr.c, final.c, fold-const.c,
4853 ggc-page.c, gimple-fold.c, gimple-iterator.c, gimple-low.c,
4854 gimple-pretty-print.c, gimple-ssa-strength-reduction.c,
4855 gimple-streamer-in.c, gimple-streamer-out.c, gimple.c, gimplify.c,
4856 graphite-blocking.c, graphite-clast-to-gimple.c,
4857 graphite-dependences.c, graphite-interchange.c,
4858 graphite-optimize-isl.c, graphite-poly.c, graphite-scop-detection.c,
4859 graphite-sese-to-poly.c, graphite.c, ipa-cp.c, ipa-inline-analysis.c,
4860 ipa-inline-transform.c, ipa-inline.c, ipa-prop.c, ipa-pure-const.c,
4861 ipa-reference.c, ipa-split.c, ipa-utils.c,
4862 loop-init.c, lto-cgraph.c, lto-section-in.c, lto-section-out.c,
4863 lto-streamer-in.c, lto-streamer-out.c, lto-streamer.c, omp-low.c,
4864 passes.c, predict.c, print-tree.c, profile.c, sese.c, targhooks.c,
4865 tracer.c, trans-mem.c, tree-call-cdce.c, tree-cfg.c, tree-cfgcleanup.c,
4866 tree-chrec.c, tree-complex.c, tree-data-ref.c, tree-dfa.c, tree-eh.c,
4867 tree-emutls.c, tree-if-conv.c, tree-inline.c, tree-into-ssa.c,
4868 tree-loop-distribution.c, tree-mudflap.c, tree-nested.c, tree-nrv.c,
4869 tree-object-size.c, tree-optimize.c, tree-outof-ssa.c, tree-parloops.c,
4870 tree-phinodes.c, tree-predcom.c, tree-pretty-print.c, tree-profile.c,
4871 tree-scalar-evolution.c, tree-sra.c, tree-ssa*.c, tree-stdarg.c,
4872 tree-streamer-in.c, tree-switch-conversion.c, tree-tailcall.c,
4873 tree-vect-data-refs.c, tree-vect-generic.c, tree-vect-loop-manip.c,
4874 tree-vect-loop.c, tree-vect-patterns.c, tree-vect-slp.c,
4875 tree-vect-stmts.c, tree-vectorizer.c, tree-vrp.c, tsan.c,
4876 value-prof.c, var-tracking.c,
4877 varpool.c, vtable-verify.c: Replace tree-flow.h with tree-ssa.h
4878
4879 2013-09-12 Richard Biener <rguenther@suse.de>
4880
4881 PR tree-optimization/58396
4882 * tree-loop-distribution.c (create_rdg_edges): Free unused DDRs.
4883 (build_rdg): Take a loop-nest parameter, fix memleaks.
4884 (distribute_loop): Compute loop-nest here and pass it to build_rdg.
4885
4886 2013-09-12 Yuri Rumyantsev <ysrumyan@gmail.com>
4887
4888 * config/i386/x86-tune.def: Turn on X86_TUNE_AVOID_MEM_OPND_FOR_CMOVE
4889 for SLM.
4890
4891 2013-09-12 Cameron McInally <cameron.mcinally@nyu.edu>
4892
4893 * doc/extend.texi: Fix errors in x86 FMA builtin naming.
4894 The FMA instruction names should have a 'v' prefix.
4895
4896 2013-09-12 Richard Biener <rguenther@suse.de>
4897
4898 * tree-loop-distribution.c (dot_rdg_1): Make graph prettier.
4899 (dot_rdg): Use popen instead of system in optional code.
4900 (remaining_stmts, upstream_mem_writes): Remove global bitmaps.
4901 (already_processed_vertex_p): Adjust.
4902 (has_anti_or_output_dependence, predecessor_has_mem_write,
4903 mark_nodes_having_upstream_mem_writes, has_upstream_mem_writes,
4904 rdg_flag_uses): Remove.
4905 (rdg_flag_vertex): Simplify.
4906 (rdg_flag_vertex_and_dependent): Rely on a correct RDG and
4907 remove recursion.
4908 (build_rdg_partition_for_component): Process the first vertex
4909 of a component only.
4910 (ldist_gen): Do not compute remaining_stmts or upstream_mem_writes.
4911
4912 2013-09-12 Alan Modra <amodra@gmail.com>
4913
4914 * config/rs6000/rs6000.c (toc_relative_expr_p): Use add_cint_operand.
4915
4916 2013-09-11 DJ Delorie <dj@redhat.com>
4917 Nick Clifton <nickc@redhat.com>
4918
4919 * config/rl78/predicates.md (rl78_cmp_operator_signed): New.
4920 (rl78_stack_based_mem): New.
4921 * config/rl78/constraints.md (Iv08): New.
4922 (Iv16): New.
4923 (Iv24): New.
4924 (Is09): New.
4925 (Is17): New.
4926 (Is25): New.
4927 (ISsi): New.
4928 (IShi): New.
4929 (ISqi): New.
4930 * config/rl78/rl78-expand.md (movqi): Reject more SUBREG operands.
4931 (movhi): Likewise.
4932 (movsi): Change from expand to insn-and-split.
4933 (ashrsi3): Clobber AX.
4934 (lshrsi3): New.
4935 (ashlsi3): New.
4936 (cbranchsi4): New.
4937 * config/rl78/rl78.md (CC_REG): Fix.
4938 (addsi3): Allow memory and immediate operands.
4939 (addsi3_internal): Split into...
4940 (addsi3_internal_virt): ...new, and ...
4941 (addsi3_internal_real): ...new.
4942 (subsi): New.
4943 (subsi3_internal_virt): New.
4944 (subsi3_internal_real): New.
4945 (mulsi3): Add memory operand.
4946 (mulsi3_rl78): Likewise.
4947 (mulsi3_g13): Likewise.
4948 * config/rl78/rl78-real.md (cbranchqi4_real_signed): New.
4949 (cbranchqi4_real): Add more constraint options.
4950 (cbranchhi4_real): Expand pattern.
4951 (cbranchhi4_real_signed): New.
4952 (cbranchhi4_real_inverted): New.
4953 (cbranchsi4_real_lt): New.
4954 (cbranchsi4_real_ge): New.
4955 (cbranchsi4_real_signed): New.
4956 (cbranchsi4_real): New.
4957 (peephole2): New.
4958 * config/rl78/rl78-virt.md (ashrsi3_virt): Add custom cases for
4959 constant shifts.
4960 (lshrsi3_virt): Likewise.
4961 (ashlsi3_virt): Likewise.
4962 (cbranchqi4_virt_signed): New.
4963 (cbranchhi4_virt_signed): New.
4964 (cbranchsi4_virt): New.
4965 * config/rl78/rl78.c: Whitespace fixes throughout.
4966 (move_elim_pass): New.
4967 (pass_data_rl78_move_elim): New.
4968 (pass_rl78_move_elim): New.
4969 (make_pass_rl78_move_elim): New.
4970 (rl78_devirt_info): Run devirt earlier.
4971 (rl78_move_elim_info): New.
4972 (rl78_asm_file_start): Register it.
4973 (rl78_split_movsi): New.
4974 (rl78_as_legitimate_address): Allow virtual base registers when
4975 appropriate.
4976 (rl78_addr_space_convert): Remove spurious debug stuff.
4977 (rl78_print_operand_1): Add z,s,S,r,E modifiers.
4978 (rl78_print_operand): More cases for not printing '#'.
4979 (rl78_expand_compare): Remove most of the logic.
4980 (content_memory): New.
4981 (clear_content_memory): New.
4982 (get_content_index): New.
4983 (get_content_name): New.
4984 (display_content_memory): New.
4985 (update_content): New.
4986 (record_content): New.
4987 (already_contains): New.
4988 (insn_ok_now): Re-recog insns with virtual registers.
4989 (add_postponed_content_update): New.
4990 (process_postponed_content_update): New.
4991 (gen_and_emit_move): New.
4992 (transcode_memory_rtx): Record new location content.
4993 Use gen_and_emit_move.
4994 (force_into_acc): New.
4995 (move_to_acc): Use gen_and_emit_move.
4996 (move_from_acc): Likewise.
4997 (move_acc_to_reg): Likewise.
4998 (move_to_x): Likewise.
4999 (move_to_hl): Likewise.
5000 (move_to_de): Likewise.
5001 (rl78_alloc_physical_registers_op1): Record location content.
5002 (has_constraint): New.
5003 (rl78_alloc_physical_registers_op2): Record location content.
5004 Optimize use of HL.
5005 (rl78_alloc_physical_registers_ro1): Likewise.
5006 (rl78_alloc_physical_registers_cmp): Likewise.
5007 (rl78_alloc_physical_registers_umul): Likewise.
5008 (rl78_alloc_address_registers_macax): New.
5009 (rl78_alloc_physical_registers): Initialize and set location
5010 content memory as needed.
5011 (rl78_reorg): Make sure split2 is called.
5012 (rl78_rtx_costs): New.
5013
5014 2013-09-11 Richard Sandiford <rdsandiford@googlemail.com>
5015
5016 * simplify-rtx.c (simplify_unary_operation_1): Use simplify_gen_binary
5017 for (not (neg ...)) and (neg (not ...)) cases.
5018
5019 2013-09-11 Richard Biener <rguenther@suse.de>
5020
5021 PR middle-end/58377
5022 * passes.def: Split critical edges before late uninit warning passes.
5023 * tree-cfg.c (pass_split_crit_edges): Implement clone method.
5024
5025 2013-09-11 Jakub Jelinek <jakub@redhat.com>
5026
5027 PR tree-optimization/58385
5028 * fold-const.c (build_range_check): If both low and high are NULL,
5029 use omit_one_operand_loc to preserve exp side-effects.
5030
5031 2013-09-11 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5032
5033 * config/arm/arm.md (arm_shiftsi3): New alternative l/l/M.
5034
5035 2013-09-11 Richard Biener <rguenther@suse.de>
5036
5037 * tree-data-ref.c (dump_rdg_vertex, debug_rdg_vertex,
5038 dump_rdg_component, debug_rdg_component, dump_rdg, debug_rdg,
5039 dot_rdg_1, dot_rdg, rdg_vertex_for_stmt, create_rdg_edge_for_ddr,
5040 create_rdg_edges_for_scalar, create_rdg_edges, create_rdg_vertices,
5041 stmts_from_loop, known_dependences_p, build_empty_rdg,
5042 build_rdg, free_rdg, rdg_defs_used_in_other_loops_p): Move ...
5043 * tree-loop-distribution.c: ... here.
5044 * tree-data-ref.h (struct rdg_vertex, RDGV_STMT, RDGV_DATAREFS,
5045 RDGV_HAS_MEM_WRITE, RDGV_HAS_MEM_READS, RDG_STMT, RDG_DATAREFS,
5046 RDG_MEM_WRITE_STMT, RDG_MEM_READS_STMT, enum rdg_dep_type,
5047 struct rdg_edge, RDGE_TYPE, RDGE_LEVEL, RDGE_RELATION): Move ...
5048 * tree-loop-distribution.c: ... here.
5049 * tree-loop-distribution.c: Include gimple-pretty-print.h.
5050 (struct partition_s): Add loops member.
5051 (partition_alloc, partition_free, rdg_flag_uses, rdg_flag_vertex,
5052 rdg_flag_vertex_and_dependent, rdg_flag_loop_exits,
5053 build_rdg_partition_for_component, rdg_build_partitions): Adjust.
5054
5055 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5056 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
5057 Sergey Lega <sergey.s.lega@intel.com>
5058 Anna Tikhonova <anna.tikhonova@intel.com>
5059 Ilya Tocar <ilya.tocar@intel.com>
5060 Andrey Turetskiy <andrey.turetskiy@intel.com>
5061 Ilya Verbin <ilya.verbin@intel.com>
5062 Kirill Yukhin <kirill.yukhin@intel.com>
5063 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
5064
5065 * config/i386/constraints.md (k): New.
5066 (Yk): Ditto.
5067 * config/i386/i386.c (const regclass_map): Add new mask registers.
5068 (dbx_register_map): Ditto.
5069 (dbx64_register_map): Ditto.
5070 (svr4_dbx_register_map): Ditto.
5071 (ix86_conditional_register_usage): Squash mask registers if AVX512F is
5072 disabled.
5073 (ix86_preferred_reload_class): Disable constants for mask registers.
5074 (ix86_secondary_reload): Do spill of mask register using 32-bit insn.
5075 (ix86_hard_regno_mode_ok): Support new mask registers.
5076 (x86_order_regs_for_local_alloc): Ditto.
5077 * config/i386/i386.h (FIRST_PSEUDO_REGISTER): Update.
5078 (FIXED_REGISTERS): Add new mask registers.
5079 (CALL_USED_REGISTERS): Ditto.
5080 (REG_ALLOC_ORDER): Ditto.
5081 (VALID_MASK_REG_MODE): New.
5082 (FIRST_MASK_REG): Ditto.
5083 (LAST_MASK_REG): Ditto.
5084 (reg_class): Add MASK_EVEX_REGS, MASK_REGS.
5085 (MAYBE_MASK_CLASS_P): New.
5086 (REG_CLASS_NAMES): Add MASK_EVEX_REGS, MASK_REGS.
5087 (REG_CLASS_CONTENTS): Ditto.
5088 (MASK_REGNO_P): New.
5089 (ANY_MASK_REG_P): Ditto.
5090 (HI_REGISTER_NAMES): Add new mask registers.
5091 * config/i386/i386.md (MASK0_REG, MASK1_REG, MASK2_REG, MASK3_REG,
5092 MASK4_REG, MASK5_REG, MASK6_REG, MASK7_REG): Constants for new
5093 mask registers.
5094 (attribute "type"): Add mskmov, msklog.
5095 (attribute "length_immediate"): Support them.
5096 (attribute "memory"): Ditto.
5097 (attribute "prefix_0f"): Ditto.
5098 (*movhi_internal): Support new mask registers.
5099 (*movqi_internal): Ditto.
5100 (define_split): Split out clobber pattern is a logic
5101 insn on mask registers.
5102 (*k<logic><mode>): New.
5103 (*andhi_1): Extend to support mask regs.
5104 (*andqi_1): Extend to support mask regs.
5105 (kandn<mode>): New.
5106 (define_split): Split and-not to and and not if operands
5107 are not mask regs.
5108 (*<code><mode>_1): Separate HI mode to new pattern...
5109 (*<code>hi_1): This.
5110 (*<code>qi_1): Extend to support mask regs.
5111 (kxnor<mode>): New.
5112 (kortestzhi): Ditto.
5113 (kortestchi): Ditto.
5114 (kunpckhi): Ditto.
5115 (*one_cmpl<mode>2_1): Remove HImode and handle it...
5116 (*one_cmplhi2_1): ...Here, now with mask registers support.
5117 (*one_cmplqi2_1): Support new mask registers.
5118 (HI/QImode arithmetics splitter): Don't split if mask registers
5119 are used.
5120 (HI/QImode not splitter): Ditto.
5121 * config/i386/predicated.md (mask_reg_operand): New.
5122 (general_reg_operand): Ditto.
5123
5124 2013-09-11 Alexander Ivchenko <alexander.ivchenko@intel.com>
5125
5126 * doc/invoke.texi: Document fxsr, xsave and xsaveopt options.
5127 * doc/extend.texi: Document fxsr, xsave and xsaveopt builtins.
5128
5129 2013-09-10 Jeff Law <law@redhat.com>
5130
5131 PR tree-optimization/58380
5132 * tree-ssa-threadupdate.c (thread_block): Recognize another case
5133 of threading through a buried loop header.
5134
5135 * tree-ssa-threadedge.c (thread_around_empty_blocks): Correct
5136 return value for single successor case.
5137
5138 2013-09-10 Jan Hubicka <jh@suse.cz>
5139
5140 * ipa-devirt.c (ipa_devirt): Enable with LTO.
5141
5142 2013-09-10 Richard Earnshaw <rearnsha@arm.com>
5143
5144 PR target/58361
5145 * arm/vfp.md (combine_vcvt_f32_<FCVTI32typename>): Fix pattern to
5146 support conditional execution.
5147 (combine_vcvt_f64_<FCVTI32typename>): Likewise.
5148
5149 2013-09-10 Vladimir Makarov <vmakarov@redhat.com>
5150
5151 * lra.c (lra): Clear lra_optional_reload_pseudos before every
5152 constraint pass.
5153 * lra-constraints.c (curr_insn_transform): Switch on optional reloads.
5154 Check destination too to check move insn.
5155 (undo_optional_reloads): Add check that the original peudo did not
5156 changed its allocation and the optional reload was inherited on last
5157 inheritance pass. Break loop after deciding to keep optional reload.
5158 (lra_undo_inheritance): Add check that inherited pseudo still in
5159 memory.
5160
5161 2013-09-10 James Greenhalgh <james.greenhalgh@arm.com>
5162
5163 * config/aarch64/aarch64.md (generic_sched): New.
5164 * config/aarch64/aarch64-generic.md (load): Make conditional
5165 on generic_sched attribute.
5166 (nonload): Likewise.
5167
5168 2013-09-10 Jan Hubicka <jh@suse.cz>
5169
5170 * lto-cgraph.c: Include ipa-utils.h.
5171 (compute_ltrans_boundary): Also add possible targets into the boundary.
5172
5173 2013-09-10 Jan Hubicka <jh@suse.cz>
5174
5175 * gimple-fold.c (gimple_get_virt_method_for_binfo): Pass real
5176 VAR_DECL of vtable rather than full expression.
5177
5178 2013-09-10 Jan Hubicka <jh@suse.cz>
5179 Paolo Carlini <paolo.carlini@oracle.com>
5180
5181 * cgraphunit.c (analyze_functions): Save input_location, set it
5182 to UNKNOWN_LOCATION and restore it at the end.
5183
5184 2013-09-10 Martin Jambor <mjambor@suse.cz>
5185
5186 * ipa-cp.c (propagate_constants_topo): Do not ignore SCC
5187 represented by a thunk.
5188
5189 2013-09-10 Jeff Law <law@redhat.com>
5190
5191 PR tree-optimization/58343
5192 * tree-ssa-threadupdate.c (thread_block): Identify and disable
5193 jump threading requests through loop headers buried in the middle
5194 of a jump threading path.
5195
5196 * tree-ssa-threadedge.c (thread_around_empty_blocks): Fix thinko
5197 in return value/type.
5198
5199 2013-09-10 Jakub Jelinek <jakub@redhat.com>
5200
5201 PR rtl-optimization/58365
5202 * cfgcleanup.c (merge_memattrs): Also clear MEM_READONLY_P
5203 resp. MEM_NOTRAP_P if they differ, or set MEM_VOLATILE_P if
5204 it differs.
5205
5206 2013-09-10 Richard Biener <rguenther@suse.de>
5207
5208 * tree-data-ref.h (build_rdg): Drop all parameters but loop.
5209 * tree-data-ref.c (create_rdg_vertices): Collect all data
5210 references, signal failure to the caller, use data-ref API.
5211 (build_rdg): Compute data references only once. Maintain lifetime
5212 of data references and data dependences from within RDG.
5213 (free_rdg): Free dependence relations.
5214 * tree-loop-distribution.c (rdg_flag_uses): Drop weird code
5215 inventing extra dependences.
5216 (distribute_loop): Update for RDG API changes.
5217
5218 2013-09-10 Kai Tietz <ktietz@redhat.com>
5219
5220 * doc/invoke.texi (fms-extensions): Document changed
5221 behavior for ms-abi targets.
5222 * config/i386/i386.c (ix86_option_override_internal):
5223 Set default value of option -fms-extension for ms-abi targets.
5224
5225 2013-09-10 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
5226
5227 * config/i386/i386.c (ix86_expand_movmem): Fix epilogue generation.
5228
5229 2013-09-10 Alan Modra <amodra@gmail.com>
5230
5231 PR target/58330
5232 * config/rs6000/rs6000.md (bswapdi2_64bit): Disable for volatile mems.
5233
5234 2013-09-10 Alan Modra <amodra@gmail.com>
5235
5236 * config/rs6000/predicates.md (add_cint_operand): New.
5237 (reg_or_add_cint_operand, small_toc_ref): Use add_cint_operand.
5238 * config/rs6000/rs6000.md (largetoc_high_plus): Restrict offset
5239 using add_cint_operand.
5240 (largetoc_high_plus_aix): Likewise.
5241
5242 2013-09-09 Jakub Jelinek <jakub@redhat.com>
5243
5244 PR tree-optimization/58364
5245 * tree-ssa-reassoc.c (init_range_entry): For BIT_NOT_EXPR on
5246 BOOLEAN_TYPE, only invert in_p and continue with arg0 if
5247 the current range can't be an unconditional true or false.
5248
5249 2013-09-09 James Greenhalgh <james.greenhalgh@arm.com>
5250
5251 * config/aarch64/arm_neon.h (vrsqrte_f64): Fix parameter type.
5252
5253 2013-09-09 Uros Bizjak <ubizjak@gmail.com>
5254
5255 * ipa-prop.c (ipa_modify_call_arguments): Initialize deref_align.
5256
5257 2013-09-09 Paolo Carlini <paolo.carlini@oracle.com>
5258
5259 PR c++/43452
5260 * doc/invoke.texi (-Wdelete-incomplete): Document it.
5261
5262 2013-09-09 Ian Bolton <ian.bolton@arm.com>
5263
5264 * config/aarch64/aarch64.c (aarch64_preferred_reload_class): Return
5265 NO_REGS for immediate that can't be moved directly into FP_REGS.
5266
5267 2013-09-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5268
5269 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return CC_SWP for
5270 comparison with negated operand.
5271 * config/aarch64/aarch64.md (compare_neg<mode>): Match canonical
5272 RTL form.
5273
5274 2013-09-09 Richard Biener <rguenther@suse.de>
5275
5276 PR middle-end/58326
5277 * cfgloopmanip.c (fix_bb_placements): When fixing the placement
5278 of a subloop record all its block as affecting loop-closed SSA form.
5279
5280 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5281
5282 * expmed.c (lshift_value): Take an unsigned HOST_WIDE_INT instead
5283 of an rtx/bitpos pair.
5284 (store_fixed_bit_field): Update accordingly.
5285
5286 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5287
5288 * asan.c (asan_emit_stack_protection): Use gen_int_mode instead of
5289 GEN_INT.
5290 * builtins.c (expand_errno_check): Likewise.
5291 * dwarf2cfi.c (init_return_column_size): Likewise.
5292 * except.c (sjlj_mark_call_sites): Likewise.
5293 * expr.c (move_by_pieces_1, store_by_pieces_2): Likewise.
5294 * lra-constraints.c (emit_inc): Likewise.
5295 * ree.c (combine_set_extension): Likewise.
5296 * regmove.c (fixup_match_2): Likewise.
5297 * reload1.c (inc_for_reload): Likewise.
5298
5299 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5300
5301 * combine.c (simplify_set, expand_field_assignment, extract_left_shift)
5302 (force_to_mode, simplify_shift_const_1, simplify_comparison):
5303 Use gen_int_mode with the mode of the associated simplify_* call.
5304 * explow.c (probe_stack_range, anti_adjust_stack_and_probe): Likewise.
5305 * expmed.c (expand_shift_1): Likewise.
5306 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5307 * loop-iv.c (iv_number_of_iterations): Likewise.
5308 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5309 * simplify-rtx.c (simplify_binary_operation_1): Likewise.
5310
5311 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5312
5313 * asan.c (asan_clear_shadow): Use gen_int_mode with the mode
5314 of the associated expand_* call.
5315 (asan_emit_stack_protection): Likewise.
5316 * builtins.c (round_trampoline_addr): Likewise.
5317 * explow.c (allocate_dynamic_stack_space, probe_stack_range): Likewise.
5318 * expmed.c (expand_smod_pow2, expand_sdiv_pow2, expand_divmod)
5319 (emit_store_flag): Likewise.
5320 * expr.c (emit_move_resolve_push, push_block, emit_single_push_insn_1)
5321 (emit_push_insn, optimize_bitfield_assignment_op, expand_expr_real_1):
5322 Likewise.
5323 * function.c (instantiate_virtual_regs_in_insn): Likewise.
5324 * ifcvt.c (noce_try_store_flag_constants): Likewise.
5325 * loop-unroll.c (unroll_loop_runtime_iterations): Likewise.
5326 * modulo-sched.c (generate_prolog_epilog): Likewise.
5327 * optabs.c (expand_binop, widen_leading, expand_doubleword_clz)
5328 (expand_ctz, expand_ffs, expand_unop): Likewise.
5329
5330 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5331
5332 * alias.c (addr_side_effect_eval): Use gen_int_mode with the mode
5333 of the associated gen_rtx_* call.
5334 * caller-save.c (init_caller_save): Likewise.
5335 * combine.c (find_split_point, make_extraction): Likewise.
5336 (make_compound_operation): Likewise.
5337 * dwarf2out.c (mem_loc_descriptor): Likewise.
5338 * explow.c (plus_constant, probe_stack_range): Likewise.
5339 * expmed.c (expand_mult_const): Likewise.
5340 * expr.c (emit_single_push_insn_1, do_tablejump): Likewise.
5341 * reload1.c (init_reload): Likewise.
5342 * valtrack.c (cleanup_auto_inc_dec): Likewise.
5343 * var-tracking.c (adjust_mems): Likewise.
5344 * modulo-sched.c (sms_schedule): Likewise, but use gen_rtx_GT
5345 rather than gen_rtx_fmt_ee.
5346
5347 2013-09-09 Jan Hubicka <jh@suse.cz>
5348
5349 PR middle-end/58294
5350 * value-prof.c (gimple_ic): Copy also abnormal edges.
5351
5352 2013-09-09 Richard Sandiford <rdsandiford@googlemail.com>
5353
5354 * asan.c (asan_shadow_cst): Use gen_int_mode.
5355
5356 2013-09-08 Jan Hubicka <jh@suse.cz>
5357
5358 * ipa-profile.c: Add toplevel comment.
5359 (ipa_propagate_frequency_1): Be more conservative when profile is read.
5360 (contains_hot_call_p): New function.
5361 (ipa_propagate_frequency): Set frequencies based on counts when
5362 profile is read.
5363 * predict.c (compute_function_frequency): Use PROFILE_READ gueard for
5364 profile; do not tamper with profile after inlining if it is read.
5365
5366 2013-09-08 Jan Hubicka <jh@suse.cz>
5367
5368 * ipa-prop.c (try_make_edge_direct_simple_call): Do not special case
5369 speculative edges.
5370
5371 2013-09-08 Jan Hubicka <jh@suse.cz>
5372
5373 * ipa.c (walk_polymorphic_call_targets): Fix redirection before IPA
5374 summary generation.
5375
5376 2013-09-08 Jeff Law <law@redhat.com>
5377
5378 PR bootstrap/58340
5379 * tree-ssa-threadedge.c (thread_across_edge): Fix initialization
5380 of 'found'.
5381
5382 2013-09-08 Andi Kleen <ak@linux.intel.com>
5383
5384 * tree-inline.c (estimate_num_insns): Limit asm cost to 1000.
5385
5386 2013-09-08 Jan Hubicka <jh@suse.cz>
5387
5388 * ipa.c (walk_polymorphic_call_targets): Fix inliner summary update.
5389
5390 2013-09-08 Richard Sandiford <rdsandiford@googlemail.com>
5391
5392 * ira.c (update_equiv_regs): Only call set_paradoxical_subreg
5393 for non-debug insns.
5394 * lra.c (new_insn_reg): Take the containing insn as a parameter.
5395 Only modify lra_reg_info[].biggest_mode if it's non-debug insn.
5396 (collect_non_operand_hard_regs, add_regs_to_insn_regno_info): Update
5397 accordingly.
5398
5399 2013-09-08 Jan Hubicka <jh@suse.cz>
5400
5401 * cgraphunit.c (walk_polymorphic_call_targets): Permit 0 possible
5402 targets and devirtualize to BUILT_IN_UNREACHABLE.
5403 * timevar.def (TV_IPA_UNREACHABLE): New timevar.
5404 * ipa.c (walk_polymorphic_call_targets): New function.
5405 (symtab_remove_unreachable_nodes): Use it; do not keep all virtual
5406 functions; use the new timevar.
5407 * ipa-devirt.c (maybe_record_node): Do not insert static nodes that
5408 was removed from the program.
5409 (record_binfo): If BINFO corresponds to an anonymous namespace, we may
5410 not consider it in the walk when its vtable is dead.
5411 (possible_polymorphic_call_targets_1): Pass anonymous flag to
5412 record_binfo.
5413 (devirt_variable_node_removal_hook): New function.
5414 (possible_polymorphic_call_targets): Also register
5415 devirt_variable_node_removal_hook.
5416 (ipa_devirt): Do not do non-speculative devirtualization.
5417 (gate_ipa_devirt): One execute if devirtualizing speculatively.
5418
5419 2013-09-08 Jan Hubicka <jh@suse.cz>
5420
5421 * cgraph.h (varpool_node_hook, varpool_node_hook_list,
5422 varpool_add_node_removal_hook, varpool_add_variable_insertion_hook,
5423 varpool_remove_variable_insertion_hook): Declare.
5424 * varpool.c (varpool_node_hook_list): New structure.
5425 (first_varpool_node_removal_hook,
5426 first_varpool_variable_insertion_hook): New variables.
5427 (varpool_add_node_removal_hook, varpool_remove_node_removal_hook,
5428 varpool_call_node_removal_hooks, varpool_add_variable_insertion_hook,
5429 varpool_remove_variable_insertion_hook,
5430 varpool_call_variable_insertion_hooks): New functions.
5431 (varpool_remove_node): Use it.
5432
5433 2013-09-08 Paolo Carlini <paolo.carlini@oracle.com>
5434
5435 PR c++/54941
5436 * diagnostic.c (diagnostic_build_prefix): When s.file is
5437 "<built-in>" don't output line and column numbers.
5438
5439 2013-09-06 Jan Hubicka <jh@suse.cz>
5440
5441 * cgraphunit.c (expand_thunk): Get body before touching arguments.
5442 * lto-streamer-out.c: Stream thunks, too.
5443 * lto-streamer-in.c (input_function): Pop cfun here
5444 (lto_read_body): Instead of here.
5445
5446 2013-09-06 Caroline Tice <cmtice@google.com>
5447
5448 * doc/install.texi: Add documentation for the --enable-vtable-verify
5449 and the --disable-libvtv configure options.
5450
5451 2013-09-06 Jeff Law <law@redhat.com>
5452
5453 * tree-ssa-dom.c (cprop_into_successor_phis): Also propagate
5454 edge implied equivalences into successor phis.
5455
5456 2013-09-06 Joern Rennecke <joern.rennecke@embecosm.com>
5457
5458 * resource.c (mark_referenced_resources): Handle COND_EXEC.
5459
5460 2013-09-06 Claudiu Zissulescu <claziss@synopsys.com>
5461
5462 * resource.c (mark_target_live_regs): Compute resources taking
5463 into account if a call is predicated or not.
5464
5465 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5466
5467 * toplev.c (output_stack_usage): Be prepared for suffixes created by
5468 the compiler in the function names.
5469
5470 2013-09-06 Jan Hubicka <jh@suse.cz>
5471
5472 PR middle-end/58094
5473 * ipa-inline.c (has_caller_p): New function.
5474 (want_inline_function_to_all_callers_p): Use it.
5475 (sum_callers, inline_to_all_callers): Break out from ...
5476 (ipa_inline): ... here.
5477
5478 2013-09-06 Jan Hubicka <jh@suse.cz>
5479
5480 * config/i386/i386.c (ix86_hard_regno_mode_ok): AVX modes are valid
5481 only when AVX is enabled.
5482
5483 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5484
5485 * config/aarch64/aarch64.md
5486 (*movtf_aarch64): Use neon_<ls>dm_2 as type where v8type
5487 is fpsimd_<load/store>2.
5488 (load_pair<mode>): Likewise.
5489 (store_pair<mode>): Likewise.
5490
5491 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5492
5493 * config/arm/types.md (type): Add "mrs" type.
5494 * config/aarch64/aarch64.md
5495 (aarch64_load_tp_hard): Make type "mrs".
5496 * config/arm/arm.md
5497 (load_tp_hard): Make type "mrs".
5498 * config/arm/cortex-a15.md: Update with new attributes.
5499 * config/arm/cortex-a5.md: Update with new attributes.
5500 * config/arm/cortex-a53.md: Update with new attributes.
5501 * config/arm/cortex-a7.md: Update with new attributes.
5502 * config/arm/cortex-a8.md: Update with new attributes.
5503 * config/arm/cortex-a9.md: Update with new attributes.
5504 * config/arm/cortex-m4.md: Update with new attributes.
5505 * config/arm/cortex-r4.md: Update with new attributes.
5506 * config/arm/fa526.md: Update with new attributes.
5507 * config/arm/fa606te.md: Update with new attributes.
5508 * config/arm/fa626te.md: Update with new attributes.
5509 * config/arm/fa726te.md: Update with new attributes.
5510
5511 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5512
5513 * config/aarch64/aarch64.md
5514 (*movti_aarch64): Use "multiple" for type where v8type is "move2".
5515 (*movtf_aarch64): Likewise.
5516 * config/arm/arm.md
5517 (thumb1_movdi_insn): Use "multiple" for type where more than one
5518 instruction is used for a move.
5519 (*arm32_movhf): Likewise.
5520 (*thumb_movdf_insn): Likewise.
5521
5522 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5523
5524 * config/arm/types.md (type): Rename fcpys to fmov.
5525 * config/arm/vfp.md
5526 (*arm_movsi_vfp): Rename type fcpys as fmov.
5527 (*thumb2_movsi_vfp): Likewise
5528 (*movhf_vfp_neon): Likewise
5529 (*movhf_vfp): Likewise
5530 (*movsf_vfp): Likewise
5531 (*thumb2_movsf_vfp): Likewise
5532 (*movsfcc_vfp): Likewise
5533 (*thumb2_movsfcc_vfp): Likewise
5534 * config/aarch64/aarch64-simd.md
5535 (move_lo_quad_<mode>): Replace type mov_reg with fmovs.
5536 * config/aarch64/aarch64.md
5537 (*movsi_aarch64): Replace type mov_reg with fmovs.
5538 (*movdi_aarch64): Likewise
5539 (*movsf_aarch64): Likewise
5540 (*movdf_aarch64): Likewise
5541 * config/arm/arm.c
5542 (cortexa7_older_only): Rename TYPE_FCPYS to TYPE_FMOV.
5543 * config/arm/iwmmxt.md
5544 (*iwmmxt_movsi_insn): Rename type fcpys as fmov.
5545 * config/arm/arm1020e.md: Update with new attributes.
5546 * config/arm/cortex-a15-neon.md: Update with new attributes.
5547 * config/arm/cortex-a5.md: Update with new attributes.
5548 * config/arm/cortex-a53.md: Update with new attributes.
5549 * config/arm/cortex-a7.md: Update with new attributes.
5550 * config/arm/cortex-a8-neon.md: Update with new attributes.
5551 * config/arm/cortex-a9.md: Update with new attributes.
5552 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5553 * config/arm/cortex-r4f.md: Update with new attributes.
5554 * config/arm/marvell-pj4.md: Update with new attributes.
5555 * config/arm/vfp11.md: Update with new attributes.
5556
5557 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5558
5559 * config/aarch64/aarch64.md
5560 (*madd<mode>): Fix type attribute.
5561 (*maddsi_uxtw): Likewise.
5562 (*msub<mode>): Likewise.
5563 (*msubsi_uxtw): Likewise.
5564 (<su_optab>maddsidi4): Likewise.
5565 (<su_optab>msubsidi4): Likewise.
5566
5567 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5568
5569 * config/arm/types.md: Split fdiv<sd> as fsqrt<sd>, fdiv<sd>.
5570 * config/arm/arm.md (core_cycles): Remove fdiv.
5571 * config/arm/vfp.md:
5572 (*sqrtsf2_vfp): Update for attribute changes.
5573 (*sqrtdf2_vfp): Likewise.
5574 * config/aarch64/aarch64.md:
5575 (sqrt<mode>2): Update for attribute changes.
5576 * config/arm/arm1020e.md: Update with new attributes.
5577 * config/arm/cortex-a15-neon.md: Update with new attributes.
5578 * config/arm/cortex-a5.md: Update with new attributes.
5579 * config/arm/cortex-a53.md: Update with new attributes.
5580 * config/arm/cortex-a7.md: Update with new attributes.
5581 * config/arm/cortex-a8-neon.md: Update with new attributes.
5582 * config/arm/cortex-a9.md: Update with new attributes.
5583 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5584 * config/arm/cortex-r4f.md: Update with new attributes.
5585 * config/arm/marvell-pj4.md: Update with new attributes.
5586 * config/arm/vfp11.md: Update with new attributes.
5587
5588 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5589
5590 * config/arm/types.md
5591 (type): Split f_cvt as f_cvt, f_cvtf2i, f_cvti2f.
5592 * config/aarch64/aarch64.md
5593 (l<fcvt_pattern><su_optab><GPF:mode><GPI:mode>2): Update with
5594 new attributes.
5595 (fix_trunc<GPF:mode><GPI:mode>2): Likewise.
5596 (fixuns_trunc<GPF:mode><GPI:mode>2): Likewise.
5597 (float<GPI:mode><GPF:mode>2): Likewise.
5598 * config/arm/vfp.md
5599 (*truncsisf2_vfp): Update with new attributes.
5600 (*truncsidf2_vfp): Likewise.
5601 (fixuns_truncsfsi2): Likewise.
5602 (fixuns_truncdfsi2): Likewise.
5603 (*floatsisf2_vfp): Likewise.
5604 (*floatsidf2_vfp): Likewise.
5605 (floatunssisf2): Likewise.
5606 (floatunssidf2): Likewise.
5607 (*combine_vcvt_f32_<FCVTI32typename>): Likewise.
5608 (*combine_vcvt_f64_<FCVTI32typename>): Likewise.
5609 * config/arm/arm1020e.md: Update with new attributes.
5610 * config/arm/cortex-a15-neon.md: Update with new attributes.
5611 * config/arm/cortex-a5.md: Update with new attributes.
5612 * config/arm/cortex-a53.md: Update with new attributes.
5613 * config/arm/cortex-a7.md: Update with new attributes.
5614 * config/arm/cortex-a8-neon.md: Update with new attributes.
5615 * config/arm/cortex-a9.md: Update with new attributes.
5616 * config/arm/cortex-m4-fpu.md: Update with new attributes.
5617 * config/arm/cortex-r4f.md: Update with new attributes.
5618 * config/arm/marvell-pj4.md: Update with new attributes.
5619 * config/arm/vfp11.md: Update with new attributes.
5620
5621 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5622
5623 * config/aarch64/arm_neon.h
5624 (vqtbl<1,2,3,4><q>_s8): Fix control vector parameter type.
5625 (vqtbx<1,2,3,4><q>_s8): Likewise.
5626
5627 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5628
5629 * config/arm/types.md: Add "no_insn", "multiple" and "untyped" types.
5630 * config/arm/arm-fixed.md: Add type attribute to all insn patterns.
5631 (add<mode>3): Add type attribute.
5632 (add<mode>3): Likewise.
5633 (usadd<mode>3): Likewise.
5634 (ssadd<mode>3): Likewise.
5635 (sub<mode>3): Likewise.
5636 (sub<mode>3): Likewise.
5637 (ussub<mode>3): Likewise.
5638 (sssub<mode>3): Likewise.
5639 (ssmulsa3): Likewise.
5640 (usmulusa3): Likewise.
5641 (arm_usatsihi): Likewise.
5642 * config/arm/vfp.md
5643 (*movdi_vfp): Add types for all instructions.
5644 (*movdi_vfp_cortexa8): Likewise.
5645 (*movhf_vfp_neon): Likewise.
5646 (*movhf_vfp): Likewise.
5647 (*movdf_vfp): Likewise.
5648 (*thumb2_movdf_vfp): Likewise.
5649 (*thumb2_movdfcc_vfp): Likewise.
5650 * config/arm/arm.md: Add type attribute to all insn patterns.
5651 (*thumb1_adddi3): Add type attribute.
5652 (*arm_adddi3): Likewise.
5653 (*adddi_sesidi_di): Likewise.
5654 (*adddi_zesidi_di): Likewise.
5655 (*thumb1_addsi3): Likewise.
5656 (addsi3_compare0): Likewise.
5657 (*addsi3_compare0_scratch): Likewise.
5658 (*compare_negsi_si): Likewise.
5659 (cmpsi2_addneg): Likewise.
5660 (*addsi3_carryin_<optab>): Likewise.
5661 (*addsi3_carryin_alt2_<optab>): Likewise.
5662 (*addsi3_carryin_clobercc_<optab>): Likewise.
5663 (*subsi3_carryin): Likewise.
5664 (*subsi3_carryin_const): Likewise.
5665 (*subsi3_carryin_compare): Likewise.
5666 (*subsi3_carryin_compare_const): Likewise.
5667 (*arm_subdi3): Likewise.
5668 (*thumb_subdi3): Likewise.
5669 (*subdi_di_zesidi): Likewise.
5670 (*subdi_di_sesidi): Likewise.
5671 (*subdi_zesidi_di): Likewise.
5672 (*subdi_sesidi_di): Likewise.
5673 (*subdi_zesidi_ze): Likewise.
5674 (thumb1_subsi3_insn): Likewise.
5675 (*arm_subsi3_insn): Likewise.
5676 (*anddi3_insn): Likewise.
5677 (*anddi_zesidi_di): Likewise.
5678 (*anddi_sesdi_di): Likewise.
5679 (*ne_zeroextracts): Likewise.
5680 (*ne_zeroextracts): Likewise.
5681 (*ite_ne_zeroextr): Likewise.
5682 (*ite_ne_zeroextr): Likewise.
5683 (*anddi_notdi_di): Likewise.
5684 (*anddi_notzesidi): Likewise.
5685 (*anddi_notsesidi): Likewise.
5686 (andsi_notsi_si): Likewise.
5687 (thumb1_bicsi3): Likewise.
5688 (*iordi3_insn): Likewise.
5689 (*iordi_zesidi_di): Likewise.
5690 (*iordi_sesidi_di): Likewise.
5691 (*thumb1_iorsi3_insn): Likewise.
5692 (*xordi3_insn): Likewise.
5693 (*xordi_zesidi_di): Likewise.
5694 (*xordi_sesidi_di): Likewise.
5695 (*arm_xorsi3): Likewise.
5696 (*andsi_iorsi3_no): Likewise.
5697 (*smax_0): Likewise.
5698 (*smax_m1): Likewise.
5699 (*arm_smax_insn): Likewise.
5700 (*smin_0): Likewise.
5701 (*arm_smin_insn): Likewise.
5702 (*arm_umaxsi3): Likewise.
5703 (*arm_uminsi3): Likewise.
5704 (*minmax_arithsi): Likewise.
5705 (*minmax_arithsi_): Likewise.
5706 (*satsi_<SAT:code>): Likewise.
5707 (arm_ashldi3_1bit): Likewise.
5708 (arm_ashrdi3_1bit): Likewise.
5709 (arm_lshrdi3_1bit): Likewise.
5710 (*arm_negdi2): Likewise.
5711 (*thumb1_negdi2): Likewise.
5712 (*arm_negsi2): Likewise.
5713 (*thumb1_negsi2): Likewise.
5714 (*negdi_extendsid): Likewise.
5715 (*negdi_zero_extend): Likewise.
5716 (*arm_abssi2): Likewise.
5717 (*thumb1_abssi2): Likewise.
5718 (*arm_neg_abssi2): Likewise.
5719 (*thumb1_neg_abss): Likewise.
5720 (one_cmpldi2): Likewise.
5721 (extend<mode>di2): Likewise.
5722 (*compareqi_eq0): Likewise.
5723 (*arm_extendhisi2addsi): Likewise.
5724 (*arm_movdi): Likewise.
5725 (*thumb1_movdi_insn): Likewise.
5726 (*arm_movt): Likewise.
5727 (*thumb1_movsi_insn): Likewise.
5728 (pic_add_dot_plus_four): Likewise.
5729 (pic_add_dot_plus_eight): Likewise.
5730 (tls_load_dot_plus_eight): Likewise.
5731 (*thumb1_movhi_insn): Likewise.
5732 (*thumb1_movsf_insn): Likewise.
5733 (*movdf_soft_insn): Likewise.
5734 (*thumb_movdf_insn): Likewise.
5735 (cbranchsi4_insn): Likewise.
5736 (cbranchsi4_scratch): Likewise.
5737 (*negated_cbranchsi4): Likewise.
5738 (*tbit_cbranch): Likewise.
5739 (*tlobits_cbranch): Likewise.
5740 (*tstsi3_cbranch): Likewise.
5741 (*cbranchne_decr1): Likewise.
5742 (*addsi3_cbranch): Likewise.
5743 (*addsi3_cbranch_scratch): Likewise.
5744 (*arm_cmpdi_insn): Likewise.
5745 (*arm_cmpdi_unsig): Likewise.
5746 (*arm_cmpdi_zero): Likewise.
5747 (*thumb_cmpdi_zero): Likewise.
5748 (*deleted_compare): Likewise.
5749 (*mov_scc): Likewise.
5750 (*mov_negscc): Likewise.
5751 (*mov_notscc): Likewise.
5752 (*cstoresi_eq0_thumb1_insn): Likewise.
5753 (cstoresi_nltu_thumb1): Likewise.
5754 (cstoresi_ltu_thu): Likewise.
5755 (thumb1_addsi3_addgeu): Likewise.
5756 (*arm_jump): Likewise.
5757 (*thumb_jump): Likewise.
5758 (*check_arch2): Likewise.
5759 (arm_casesi_internal): Likewise.
5760 (thumb1_casesi_dispatch): Likewise.
5761 (*arm_indirect_jump): Likewise.
5762 (*thumb1_indirect_jump): Likewise.
5763 (nop): Likewise.
5764 (*and_scc): Likewise.
5765 (*ior_scc): Likewise.
5766 (*compare_scc): Likewise.
5767 (*cond_move): Likewise.
5768 (*cond_arith): Likewise.
5769 (*cond_sub): Likewise.
5770 (*cmp_ite0): Likewise.
5771 (*cmp_ite1): Likewise.
5772 (*cmp_and): Likewise.
5773 (*cmp_ior): Likewise.
5774 (*ior_scc_scc): Likewise.
5775 (*ior_scc_scc_cmp): Likewise.
5776 (*and_scc_scc): Likewise.
5777 (*and_scc_scc_cmp): Likewise.
5778 (*and_scc_scc_nod): Likewise.
5779 (*negscc): Likewise.
5780 (movcond_addsi): Likewise.
5781 (movcond): Likewise.
5782 (*ifcompare_plus_move): Likewise.
5783 (*if_plus_move): Likewise.
5784 (*ifcompare_move_plus): Likewise.
5785 (*if_move_plus): Likewise.
5786 (*ifcompare_arith_arith): Likewise.
5787 (*if_arith_arith): Likewise.
5788 (*ifcompare_arith_move): Likewise.
5789 (*if_arith_move): Likewise.
5790 (*ifcompare_move_arith): Likewise.
5791 (*if_move_arith): Likewise.
5792 (*ifcompare_move_not): Likewise.
5793 (*if_move_not): Likewise.
5794 (*ifcompare_not_move): Likewise.
5795 (*if_not_move): Likewise.
5796 (*ifcompare_shift_move): Likewise.
5797 (*if_shift_move): Likewise.
5798 (*ifcompare_move_shift): Likewise.
5799 (*if_move_shift): Likewise.
5800 (*ifcompare_shift_shift): Likewise.
5801 (*ifcompare_not_arith): Likewise.
5802 (*ifcompare_arith_not): Likewise.
5803 (*if_arith_not): Likewise.
5804 (*ifcompare_neg_move): Likewise.
5805 (*if_neg_move): Likewise.
5806 (*ifcompare_move_neg): Likewise.
5807 (*if_move_neg): Likewise.
5808 (prologue_thumb1_interwork): Likewise.
5809 (*cond_move_not): Likewise.
5810 (*sign_extract_onebit): Likewise.
5811 (*not_signextract_onebit): Likewise.
5812 (stack_tie): Likewise.
5813 (align_4): Likewise.
5814 (align_8): Likewise.
5815 (consttable_end): Likewise.
5816 (consttable_1): Likewise.
5817 (consttable_2): Likewise.
5818 (consttable_4): Likewise.
5819 (consttable_8): Likewise.
5820 (consttable_16): Likewise.
5821 (*thumb1_tablejump): Likewise.
5822 (prefetch): Likewise.
5823 (force_register_use): Likewise.
5824 (thumb_eh_return): Likewise.
5825 (load_tp_hard): Likewise.
5826 (load_tp_soft): Likewise.
5827 (tlscall): Likewise.
5828 (*arm_movtas_ze): Likewise.
5829 (*arm_rev): Likewise.
5830 (*arm_revsh): Likewise.
5831 (*arm_rev16): Likewise.
5832 * config/arm/thumb2.md
5833 (*thumb2_smaxsi3): Likewise.
5834 (*thumb2_sminsi3): Likewise.
5835 (*thumb32_umaxsi3): Likewise.
5836 (*thumb2_uminsi3): Likewise.
5837 (*thumb2_negdi2): Likewise.
5838 (*thumb2_abssi2): Likewise.
5839 (*thumb2_neg_abss): Likewise.
5840 (*thumb2_movsi_insn): Likewise.
5841 (tls_load_dot_plus_four): Likewise.
5842 (*thumb2_movhi_insn): Likewise.
5843 (*thumb2_mov_scc): Likewise.
5844 (*thumb2_mov_negs): Likewise.
5845 (*thumb2_mov_negs): Likewise.
5846 (*thumb2_mov_nots): Likewise.
5847 (*thumb2_mov_nots): Likewise.
5848 (*thumb2_movsicc_): Likewise.
5849 (*thumb2_movsfcc_soft_insn): Likewise.
5850 (*thumb2_indirect_jump): Likewise.
5851 (*thumb2_and_scc): Likewise.
5852 (*thumb2_ior_scc): Likewise.
5853 (*thumb2_ior_scc_strict_it): Likewise.
5854 (*thumb2_cond_move): Likewise.
5855 (*thumb2_cond_arith): Likewise.
5856 (*thumb2_cond_ari): Likewise.
5857 (*thumb2_cond_sub): Likewise.
5858 (*thumb2_negscc): Likewise.
5859 (*thumb2_movcond): Likewise.
5860 (thumb2_casesi_internal): Likewise.
5861 (thumb2_casesi_internal_pic): Likewise.
5862 (*thumb2_alusi3_short): Likewise.
5863 (*thumb2_mov<mode>_shortim): Likewise.
5864 (*thumb2_addsi_short): Likewise.
5865 (*thumb2_subsi_short): Likewise.
5866 (thumb2_addsi3_compare0): Likewise.
5867 (*thumb2_cbz): Likewise.
5868 (*thumb2_cbnz): Likewise.
5869 (*thumb2_one_cmplsi2_short): Likewise.
5870 (*thumb2_negsi2_short): Likewise.
5871 (*orsi_notsi_si): Likewise.
5872 * config/arm/arm1020e.md: Update with new attributes.
5873 * config/arm/arm1026ejs.md: Update with new attributes.
5874 * config/arm/arm1136jfs.md: Update with new attributes.
5875 * config/arm/arm926ejs.md: Update with new attributes.
5876 * config/arm/cortex-a15.md: Update with new attributes.
5877 * config/arm/cortex-a5.md: Update with new attributes.
5878 * config/arm/cortex-a53.md: Update with new attributes.
5879 * config/arm/cortex-a7.md: Update with new attributes.
5880 * config/arm/cortex-a8.md: Update with new attributes.
5881 * config/arm/cortex-a9.md: Update with new attributes.
5882 * config/arm/cortex-m4.md: Update with new attributes.
5883 * config/arm/cortex-r4.md: Update with new attributes.
5884 * config/arm/fa526.md: Update with new attributes.
5885 * config/arm/fa606te.md: Update with new attributes.
5886 * config/arm/fa626te.md: Update with new attributes.
5887 * config/arm/fa726te.md: Update with new attributes.
5888
5889 2013-09-06 James Greenhalgh <james.greenhalgh@arm.com>
5890
5891 * config/aarch64/aarch64-simd.md
5892 (aarch64_sqdml<SBINQOPS:as>l_n<mode>_internal): Use
5893 <vwx> iterator to ensure correct register choice.
5894 (aarch64_sqdml<SBINQOPS:as>l2_n<mode>_internal): Likewise.
5895 (aarch64_sqdmull_n<mode>): Likewise.
5896 (aarch64_sqdmull2_n<mode>_internal): Likewise.
5897 * config/aarch64/arm_neon.h
5898 (vml<as><q>_lane<q>_<su>16): Use 'x' constraint for element vector.
5899 (vml<as><q>_n_<su>16): Likewise.
5900 (vml<as>l_high_lane<q>_<su>16): Likewise.
5901 (vml<as>l_high_n_<su>16): Likewise.
5902 (vml<as>l_lane<q>_<su>16): Likewise.
5903 (vml<as>l_n_<su>16): Likewise.
5904 (vmul<q>_lane<q>_<su>16): Likewise.
5905 (vmul<q>_n_<su>16): Likewise.
5906 (vmull_lane<q>_<su>16): Likewise.
5907 (vmull_n_<su>16): Likewise.
5908 (vmull_high_lane<q>_<su>16): Likewise.
5909 (vmull_high_n_<su>16): Likewise.
5910 (vqrdmulh<q>_n_s16): Likewise.
5911
5912 2013-09-06 Tejas Belagod <tejas.belagod@arm.com>
5913
5914 * config/aarch64/arm_neon.h: Fix all vdup<bhsd_lane<q> intrinsics to
5915 have the correct lane parameter.
5916
5917 2013-09-06 Richard Biener <rguenther@suse.de>
5918
5919 * cfganal.c (control_dependences::~control_dependences):
5920 Properly free all of the vector.
5921
5922 2013-09-06 Kirill Yukhin <kirill.yukhin@intel.com>
5923
5924 PR target/58269
5925 * config/i386/i386.c (ix86_conditional_register_usage):
5926 Proper initialize extended SSE registers.
5927
5928 2013-09-06 Jan Hubicka <jh@suse.cz>
5929
5930 PR tree-optimization/58311
5931 * ipa-devirt.c (gate_ipa_devirt): Only execute when optimizing.
5932
5933 2013-09-06 Jan Hubicka <jh@suse.cz>
5934
5935 * Makefile.in (tree-sra.o): Update dependencies.
5936 * tree-sra.c: Include ipa-utils.h
5937 (scan_function): Use recursive_call_p.
5938 (has_caller_p): New function.
5939 (cgraph_for_node_and_aliases): Count also callers of aliases.
5940
5941 2013-09-06 Jan Hubicka <jh@suse.cz>
5942
5943 PR middle-end/58094
5944 * cgraph.h (symtab_semantically_equivalent_p): Declare.
5945 * tree-tailcall.c: Include ipa-utils.h.
5946 (find_tail_calls): Use it.
5947 * ipa-pure-const.c (check_call): Likewise.
5948 * ipa-utils.c (recursive_call_p): New function.
5949 * ipa-utils.h (recursive_call_p): Dclare.
5950 * symtab.c (symtab_nonoverwritable_alias): Fix formatting.
5951 (symtab_semantically_equivalent_p): New function.
5952 * Makefile.in (tree-tailcall.o): Update dependencies.
5953
5954 2013-09-06 Eric Botcazou <ebotcazou@adacore.com>
5955
5956 * ipa-split.c (split_function): Set DECL_NO_INLINE_WARNING_P on the
5957 non-inlinable part.
5958
5959 2013-09-06 Richard Biener <rguenther@suse.de>
5960
5961 * lto-streamer.h (lto_global_var_decls): Remove.
5962 * Makefile.in (OBJS): Remove lto-symtab.o.
5963 (lto-symtab.o): Remove.
5964 (GTFILES): Remove lto-symtab.c
5965 * lto-symtab.c: Move to lto/
5966
5967 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5968
5969 * config/s390/s390.md (UNSPEC_FPINT_FLOOR, UNSPEC_FPINT_BTRUNC)
5970 (UNSPEC_FPINT_ROUND, UNSPEC_FPINT_CEIL, UNSPEC_FPINT_NEARBYINT)
5971 (UNSPEC_FPINT_RINT): New constant definitions.
5972 (FPINT, fpint_name, fpint_roundingmode): New integer iterator
5973 definition with 2 attributes.
5974 ("<FPINT:fpint_name><BFP:mode>2", "rint<BFP:mode>2")
5975 ("<FPINT:fpint_name><DFP:mode>2", "rint<DFP:mode>2"): New pattern
5976 definitions.
5977
5978 2013-09-06 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
5979
5980 * config/s390/s390.md: Add "bcr_flush" value to mnemonic attribute.
5981 ("mem_thread_fence_1"): Use bcr 14,0 for z196 and later.
5982 Set the mnemonic attribute to "bcr_flush". Set the "z196prop"
5983 attribute to "z196_alone".
5984 * config/s390/2827.md: Add "bcr_flush" to "ooo_groupalone" and
5985 "zEC12_simple".
5986
5987 2013-09-06 Richard Biener <rguenther@suse.de>
5988
5989 * basic-block.h (class control_dependences): New.
5990 * tree-ssa-dce.c (control_dependence_map): Remove.
5991 (cd): New global.
5992 (EXECUTE_IF_CONTROL_DEPENDENT): Remove.
5993 (set_control_dependence_map_bit, clear_control_dependence_bitmap,
5994 find_pdom, find_control_dependence, find_all_control_dependences):
5995 Move to cfganal.c.
5996 (mark_control_dependent_edges_necessary,
5997 find_obviously_necessary_stmts, propagate_necessity, tree_dce_init,
5998 tree_dce_done, perform_tree_ssa_dce): Adjust.
5999 * cfganal.c (set_control_dependence_map_bit,
6000 clear_control_dependence_bitmap, find_pdom, find_control_dependence,
6001 find_all_control_dependences): Move from tree-ssa-dce.c and
6002 implement as methods of control_dependences class.
6003 (control_dependences::control_dependences): New.
6004 (control_dependences::~control_dependences): Likewise.
6005 (control_dependences::get_edges_dependent_on): Likewise.
6006 (control_dependences::get_edge): Likewise.
6007
6008 2013-09-04 Jan Hubicka <jh@suse.cz>
6009
6010 * tree.c (types_same_for_odr): Drop overactive check.
6011 * ipa-devirt.c (hash_type_name): Likewise.
6012
6013 2013-09-04 Jan Hubicka <jh@suse.cz>
6014
6015 * cgraphunit.c (walk_polymorphic_call_targets): Break out from ...
6016 (analyze_functions): ... here.
6017
6018 2013-09-04 Jan Hubicka <jh@suse.cz>
6019
6020 PR middle-end/58201
6021 * cgraphunit.c (analyze_functions): Clear AUX fields
6022 after processing; initialize assembler name has.
6023
6024 2013-09-05 Jeff Law <law@redhat.com>
6025
6026 * tree-ssa-threadedge.c (thread_around_empty_blocks): Renamed
6027 from thread_around_empty_block. Record threading path into PATH.
6028 Recurse if threading through the initial block is successful.
6029 (thread_across_edge): Corresponding changes to slightly simplify.
6030
6031 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6032
6033 * config/aarch64/aarch64.md
6034 (type): Remove frecpe, frecps, frecpx.
6035 (aarch64_frecp<FRECP:frecp_suffix><mode>): Move to aarch64-simd.md,
6036 fix to be a TARGET_SIMD instruction.
6037 (aarch64_frecps): Remove.
6038 * config/aarch64/aarch64-simd.md
6039 (aarch64_frecp<FRECP:frecp_suffix><mode>): New, moved from aarch64.md
6040 (aarch64_frecps<mode>): Handle all float/vector of float modes.
6041
6042 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6043 Sofiane Naci <sofiane.naci@arm.com>
6044
6045 * config/arm/types.md (define_attr "type"): Expand "arlo_imm"
6046 into "adr", "alu_imm", "alus_imm", "logic_imm", "logics_imm".
6047 Expand "arlo_reg" into "adc_reg", "adc_imm", "adcs_reg", "adcs_imm",
6048 "alu_ext", "alu_reg", "alus_ext", "alus_reg", "bfm", "csel",
6049 "logic_reg", "logics_reg", "rev". Expand "arlo_shift" into
6050 "alu_shift_imm", "alus_shift_imm", "logic_shift_imm",
6051 "logics_shift_imm". Expand "arlo_shift_reg" into "alu_shift_reg",
6052 "alus_shift_reg", "logic_shift_reg", "logics_shift_reg". Expand "clz"
6053 into "clz, "rbit". Rename "shift" to "shift_imm".
6054 * config/arm/arm.md (define_attr "core_cycles"): Update for attribute
6055 changes. Update for attribute changes all occurrences of arlo_* and
6056 shift* types.
6057 * config/arm/arm-fixed.md: Update for attribute changes
6058 all occurrences of arlo_* types.
6059 * config/arm/thumb2.md: Update for attribute changes all occurrences
6060 of arlo_* types.
6061 * config/arm/arm.c (xscale_sched_adjust_cost): (rtx insn, rtx
6062 (cortexa7_older_only): Likewise.
6063 (cortexa7_younger): Likewise.
6064 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
6065 (1020alu_shift_op): Likewise.
6066 (1020alu_shift_reg_op): Likewise.
6067 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
6068 (alu_shift_op): Likewise.
6069 (alu_shift_reg_op): Likewise.
6070 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
6071 (11_alu_shift_op): Likewise.
6072 (11_alu_shift_reg_op): Likewise.
6073 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
6074 (9_alu_shift_reg_op): Likewise.
6075 * config/arm/cortex-a15.md (cortex_a15_alu): Update for
6076 attribute changes.
6077 (cortex_a15_alu_shift): Likewise.
6078 (cortex_a15_alu_shift_reg): Likewise.
6079 * config/arm/cortex-a5.md (cortex_a5_alu): Update for
6080 attribute changes.
6081 (cortex_a5_alu_shift): Likewise.
6082 * config/arm/cortex-a53.md (cortex_a53_alu): Update for
6083 attribute changes.
6084 (cortex_a53_alu_shift): Likewise.
6085 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for
6086 attribute changes.
6087 (cortex_a7_alu_reg): Likewise.
6088 (cortex_a7_alu_shift): Likewise.
6089 * config/arm/cortex-a8.md (cortex_a8_alu): Update for
6090 attribute changes.
6091 (cortex_a8_alu_shift): Likewise.
6092 (cortex_a8_alu_shift_reg): Likewise.
6093 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
6094 (cortex_a9_dp_shift): Likewise.
6095 * config/arm/cortex-m4.md (cortex_m4_alu): Update for
6096 attribute changes.
6097 * config/arm/cortex-r4.md
6098 (cortex_r4_alu): Update for attribute changes.
6099 (cortex_r4_mov): Likewise.
6100 (cortex_r4_alu_shift_reg): Likewise.
6101 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
6102 (526_alu_shift_op): Likewise.
6103 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
6104 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
6105 (626te_alu_shift_op): Likewise.
6106 * config/arm/fa726te.md (726te_alu_op): Update for attribute changes.
6107 (726te_alu_shift_op): Likewise.
6108 (726te_alu_shift_reg_op): Likewise.
6109 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
6110 (mp626_alu_shift_op): Likewise.
6111 * config/arm/marvell-pj4.md (pj4_alu): Update for attribute changes.
6112 (pj4_alu_conds): Likewise.
6113 (pj4_shift): Likewise.
6114 (pj4_shift_conds): Likewise.
6115 (pj4_alu_shift): Likewise.
6116 (pj4_alu_shift_conds): Likewise.
6117 * config/aarch64/aarch64.md: Update for attribute change
6118 all occurrences of arlo_* and shift* types.
6119
6120 2013-09-05 Mike Stump <mikestump@comcast.net>
6121
6122 * tree.h: Move documentation for tree_function_decl to tree-core.h
6123 with the declaration.
6124
6125 2013-09-05 Peter Bergner <bergner@vnet.ibm.com>
6126
6127 PR target/58139
6128 * reginfo.c (choose_hard_reg_mode): Scan through all mode classes
6129 looking for widest mode.
6130
6131 2013-09-05 Eric Botcazou <ebotcazou@adacore.com>
6132
6133 * config.gcc (*-*-vxworks*): Do not override an existing extra_objs.
6134
6135 2013-09-05 Richard Biener <rguenther@suse.de>
6136
6137 PR tree-optimization/58137
6138 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size):
6139 Do not create vectors of pointers.
6140 * tree-vect-loop.c (get_initial_def_for_induction): Use proper
6141 types for the components of the vector initializer.
6142 * tree-cfg.c (verify_gimple_assign_binary): Remove special-casing
6143 allowing pointer vectors with PLUS_EXPR/MINUS_EXPR.
6144
6145 2013-09-05 Martin Jambor <mjambor@suse.cz>
6146
6147 * ipa-prop.c (remove_described_reference): Accept missing references,
6148 return false if that hppens, otherwise return true.
6149 (cgraph_node_for_jfunc): New function.
6150 (try_decrement_rdesc_refcount): Likewise.
6151 (try_make_edge_direct_simple_call): Use them.
6152 (ipa_edge_removal_hook): Remove references from rdescs.
6153 (ipa_edge_duplication_hook): Clone rdescs and their references
6154 when the new edge has the same caller as the old one.
6155 * cgraph.c (cgraph_resolve_speculation): Remove speculative
6156 reference before removing any edges.
6157
6158 2013-09-05 Richard Earnshaw <rearnsha@arm.com>
6159
6160 * arm.c (thumb2_emit_strd_push): Rewrite to use pre-decrement on
6161 initial store.
6162 * thumb2.md (thumb2_storewb_parisi): New pattern.
6163
6164 2013-09-05 Yufeng Zhang <yufeng.zhang@arm.com>
6165
6166 * config/aarch64/aarch64-option-extensions.def: Add
6167 AARCH64_OPT_EXTENSION of 'crc'.
6168 * config/aarch64/aarch64.h (AARCH64_FL_CRC): New define.
6169 (AARCH64_ISA_CRC): Ditto.
6170 * doc/invoke.texi (-march and -mcpu feature modifiers): Add
6171 description of the CRC extension.
6172
6173 2013-09-05 Alexander Ivchenko <alexander.ivchenko@intel.com>
6174
6175 * config/rs6000/linux64.h: Define OPTION_BIONIC and OPTION_UCLIBC.
6176 * config/rs6000/linux.h: Ditto.
6177 * alpha/linux.h: Ditto.
6178 * config/bfin/uclinux.h: Define TARGET_LIBC_HAS_FUNCTION as
6179 no_c99_libc_has_function.
6180 * config/c6x/uclinux-elf.h: Ditto.
6181 * config/lm32/uclinux-elf.h: Ditto.
6182 * config/m68k/uclinux.h: Ditto.
6183 * config/moxie/uclinux.h: Ditto.
6184 * config.gcc (bfin*-linux-uclibc*): Add t-linux-android to tmake_file.
6185 (crisv32-*-linux*, cris-*-linux*): Ditto.
6186 * config/bfin/bfin.c: Include "tm_p.h".
6187
6188 2013-09-05 Richard Biener <rguenther@suse.de>
6189
6190 * tree-vect-loop.c (vect_analyze_loop_operations): Properly
6191 check for a definition without a basic-block.
6192
6193 2013-09-05 James Greenhalgh <james.greenhalgh@arm.com>
6194 Sofiane Naci <sofiane.naci@arm.com>
6195
6196 * config/aarch64/aarch64.md
6197 (*movti_aarch64): Rename r_2_f and f_2_r.
6198 (*movsf_aarch64): Likewise.
6199 (*movdf_aarch64): Likewise.
6200 (*movtf_aarch64): Likewise.
6201 (aarch64_movdi_<mode>low): Likewise.
6202 (aarch64_movdi_<mode>high): Likewise.
6203 (aarch64_mov<mode>high_di): Likewise.
6204 (aarch64_mov<mode>low_di): Likewise.
6205 (aarch64_movtilow_tilow): Likewise.
6206 * config/arm/arm.md (attribute "neon_type"): Delete. Move attribute
6207 values to config/arm/types.md
6208 (attribute "conds"): Update for attribute change.
6209 (anddi3_insn): Likewise.
6210 (iordi3_insn): Likewise.
6211 (xordi3_insn): Likewise.
6212 (one_cmpldi2): Likewise.
6213 * config/arm/types.md (type): Add Neon types.
6214 * config/arm/neon.md (neon_mov<mode>): Remove "neon_type" attribute,
6215 use "type" attribute.
6216 (movmisalign<mode>_neon_store): Likewise.
6217 (movmisalign<mode>_neon_load): Likewise.
6218 (vec_set<mode>_internal): Likewise.
6219 (vec_setv2di_internal): Likewise.
6220 (vec_extract<mode>): Likewise.
6221 (vec_extractv2di): Likewise.
6222 (add<mode>3_neon): Likewise.
6223 (adddi3_neon): Likewise.
6224 (sub<mode>3_neon): Likewise.
6225 (subdi3_neon): Likewise.
6226 (mul<mode>3_neon): Likewise.
6227 (mul<mode>3add<mode>_neon): Likewise.
6228 (mul<mode>3neg<mode>add<mode>_neon): Likewise.
6229 (fma<VCVTF:mode>4)): Likewise.
6230 (fma<VCVTF:mode>4_intrinsic): Likewise.
6231 (fmsub<VCVTF:mode>4)): Likewise.
6232 (fmsub<VCVTF:mode>4_intrinsic): Likewise.
6233 (neon_vrint<NEON_VRINT:nvrint_variant><VCVTF:mode>): Likewise.
6234 (ior<mode>3): Likewise.
6235 (and<mode>3): Likewise.
6236 (anddi3_neon): Likewise.
6237 (orn<mode>3_neon): Likewise.
6238 (orndi3_neon): Likewise.
6239 (bic<mode>3_neon): Likewise.
6240 (bicdi3_neon): Likewise.
6241 (xor<mode>3): Likewise.
6242 (one_cmpl<mode>2): Likewise.
6243 (abs<mode>2): Likewise.
6244 (neg<mode>2): Likewise.
6245 (umin<mode>3_neon): Likewise.
6246 (umax<mode>3_neon): Likewise.
6247 (smin<mode>3_neon): Likewise.
6248 (smax<mode>3_neon): Likewise.
6249 (vashl<mode>3): Likewise.
6250 (vashr<mode>3_imm): Likewise.
6251 (vlshr<mode>3_imm): Likewise.
6252 (ashl<mode>3_signed): Likewise.
6253 (ashl<mode>3_unsigned): Likewise.
6254 (neon_load_count): Likewise.
6255 (ashldi3_neon_noclobber): Likewise.
6256 (signed_shift_di3_neon): Likewise.
6257 (unsigned_shift_di3_neon): Likewise.
6258 (ashrdi3_neon_imm_noclobber): Likewise.
6259 (lshrdi3_neon_imm_noclobber): Likewise.
6260 (widen_ssum<mode>3): Likewise.
6261 (widen_usum<mode>3): Likewise.
6262 (quad_halves_<code>v4si): Likewise.
6263 (quad_halves_<code>v4sf): Likewise.
6264 (quad_halves_<code>v8hi): Likewise.
6265 (quad_halves_<code>v16qi): Likewise.
6266 (reduc_splus_v2di): Likewise.
6267 (neon_vpadd_internal<mode>): Likewise.
6268 (neon_vpsmin<mode>): Likewise.
6269 (neon_vpsmax<mode>): Likewise.
6270 (neon_vpumin<mode>): Likewise.
6271 (neon_vpumax<mode>): Likewise.
6272 (ss_add<mode>_neon): Likewise.
6273 (us_add<mode>_neon): Likewise.
6274 (ss_sub<mode>_neon): Likewise.
6275 (us_sub<mode>_neon): Likewise.
6276 (neon_vadd<mode>_unspec): Likewise.
6277 (neon_vaddl<mode>): Likewise.
6278 (neon_vaddw<mode>): Likewise.
6279 (neon_vhadd<mode>): Likewise.
6280 (neon_vqadd<mode>): Likewise.
6281 (neon_vaddhn<mode>): Likewise.
6282 (neon_vmul<mode>): Likewise.
6283 (neon_vmla<mode>): Likewise.
6284 (neon_vmlal<mode>): Likewise.
6285 (neon_vmls<mode>): Likewise.
6286 (neon_vmlsl<mode>): Likewise.
6287 (neon_vqdmulh<mode>): Likewise.
6288 (neon_vqdmlal<mode>): Likewise.
6289 (neon_vqdmlsl<mode>): Likewise.
6290 (neon_vmull<mode>): Likewise.
6291 (neon_vqdmull<mode>): Likewise.
6292 (neon_vsub<mode>_unspec): Likewise.
6293 (neon_vsubl<mode>): Likewise.
6294 (neon_vsubw<mode>): Likewise.
6295 (neon_vqsub<mode>): Likewise.
6296 (neon_vhsub<mode>): Likewise.
6297 (neon_vsubhn<mode>): Likewise.
6298 (neon_vceq<mode>): Likewise.
6299 (neon_vcge<mode>): Likewise.
6300 (neon_vcgeu<mode>): Likewise.
6301 (neon_vcgt<mode>): Likewise.
6302 (neon_vcgtu<mode>): Likewise.
6303 (neon_vcle<mode>): Likewise.
6304 (neon_vclt<mode>): Likewise.
6305 (neon_vcage<mode>): Likewise.
6306 (neon_vcagt<mode>): Likewise.
6307 (neon_vtst<mode>): Likewise.
6308 (neon_vabd<mode>): Likewise.
6309 (neon_vabdl<mode>): Likewise.
6310 (neon_vaba<mode>): Likewise.
6311 (neon_vabal<mode>): Likewise.
6312 (neon_vmax<mode>): Likewise.
6313 (neon_vmin<mode>): Likewise.
6314 (neon_vpaddl<mode>): Likewise.
6315 (neon_vpadal<mode>): Likewise.
6316 (neon_vpmax<mode>): Likewise.
6317 (neon_vpmin<mode>): Likewise.
6318 (neon_vrecps<mode>): Likewise.
6319 (neon_vrsqrts<mode>): Likewise.
6320 (neon_vqabs<mode>): Likewise.
6321 (neon_vqneg<mode>): Likewise.
6322 (neon_vcls<mode>): Likewise.
6323 (clz<mode>2): Likewise.
6324 (popcount<mode>2): Likewise.
6325 (neon_vrecpe): Likewise.
6326 (neon_vrsqrte): Likewise.
6327 (neon_vget_lane<mode>_sext_internal): Likewise.
6328 (neon_vget_lane<mode>_zext_internal): Likewise.
6329 (neon_vdup_n<mode>): Likewise.
6330 (neon_vdup_nv2di): Likewise.
6331 (neon_vdpu_lane<mode>_internal): Likewise.
6332 (neon_vswp<mode>): Likewise.
6333 (float<mode><V_cvtto>2): Likewise.
6334 (floatuns<mode><V_cvtto>2): Likewise.
6335 (fix_trunc<mode><V_cvtto>)2): Likewise
6336 (fixuns_trunc<mode><V_cvtto)2): Likewise.
6337 (neon_vcvt<mode>): Likewise.
6338 (neon_vcvtv4sfv4hf): Likewise.
6339 (neon_vcvtv4hfv4sf): Likewise.
6340 (neon_vcvt_n<mode>): Likewise.
6341 (neon_vmovn<mode>): Likewise.
6342 (neon_vqmovn<mode>): Likewise.
6343 (neon_vqmovun<mode>): Likewise.
6344 (neon_vmovl<mode>): Likewise.
6345 (neon_vmul_lane<mode>): Likewise.
6346 (neon_vmull_lane<mode>): Likewise.
6347 (neon_vqdmull_lane<mode>): Likewise.
6348 (neon_vqdmulh_lane<mode>): Likewise.
6349 (neon_vmla_lane<mode>): Likewise.
6350 (neon_vmlal_lane<mode>): Likewise.
6351 (neon_vqdmlal_lane<mode>): Likewise.
6352 (neon_vmls_lane<mode>): Likewise.
6353 (neon_vmlsl_lane<mode>): Likewise.
6354 (neon_vqdmlsl_lane<mode>): Likewise.
6355 (neon_vext<mode>): Likewise.
6356 (neon_vrev64<mode>): Likewise.
6357 (neon_vrev32<mode>): Likewise.
6358 (neon_vrev16<mode>): Likewise.
6359 (neon_vbsl<mode>_internal): Likewise.
6360 (neon_vshl<mode>): Likewise.
6361 (neon_vqshl<mode>): Likewise.
6362 (neon_vshr_n<mode>): Likewise.
6363 (neon_vshrn_n<mode>): Likewise.
6364 (neon_vqshrn_n<mode>): Likewise.
6365 (neon_vqshrun_n<mode>): Likewise.
6366 (neon_vshl_n<mode>): Likewise.
6367 (neon_vqshl_n<mode>): Likewise.
6368 (neon_vqshlu_n<mode>): Likewise.
6369 (neon_vshll_n<mode>): Likewise.
6370 (neon_vsra_n<mode>): Likewise.
6371 (neon_vsri_n<mode>): Likewise.
6372 (neon_vsli_n<mode>): Likewise.
6373 (neon_vtbl1v8qi): Likewise.
6374 (neon_vtbl2v8qi): Likewise.
6375 (neon_vtbl3v8qi): Likewise.
6376 (neon_vtbl4v8qi): Likewise.
6377 (neon_vtbx1v8qi): Likewise.
6378 (neon_vtbx2v8qi): Likewise.
6379 (neon_vtbx3v8qi): Likewise.
6380 (neon_vtbx4v8qi): Likewise.
6381 (neon_vtrn<mode>_internal): Likewise.
6382 (neon_vzip<mode>_internal): Likewise.
6383 (neon_vuzp<mode>_internal): Likewise.
6384 (neon_vld1<mode>): Likewise.
6385 (neon_vld1_lane<mode>): Likewise.
6386 (neon_vld1_dup<mode>): Likewise.
6387 (neon_vld1_dupv2di): Likewise.
6388 (neon_vst1<mode>): Likewise.
6389 (neon_vst1_lane<mode>): Likewise.
6390 (neon_vld2<mode>): Likewise.
6391 (neon_vld2_lane<mode>): Likewise.
6392 (neon_vld2_dup<mode>): Likewise.
6393 (neon_vst2<mode>): Likewise.
6394 (neon_vst2_lane<mode>): Likewise.
6395 (neon_vld3<mode>): Likewise.
6396 (neon_vld3qa<mode>): Likewise.
6397 (neon_vld3qb<mode>): Likewise.
6398 (neon_vld3_lane<mode>): Likewise.
6399 (neon_vld3_dup<mode>): Likewise.
6400 (neon_vst3<mode>): Likewise.
6401 (neon_vst3qa<mode>): Likewise.
6402 (neon_vst3qb<mode>): Likewise.
6403 (neon_vst3_lane<mode>): Likewise.
6404 (neon_vld4<mode>): Likewise.
6405 (neon_vld4qa<mode>): Likewise.
6406 (neon_vld4qb<mode>): Likewise.
6407 (neon_vld4_lane<mode>): Likewise.
6408 (neon_vld4_dup<mode>): Likewise.
6409 (neon_vst4<mode>): Likewise.
6410 (neon_vst4qa<mode>): Likewise.
6411 (neon_vst4qb<mode>): Likewise.
6412 (neon_vst4_lane<mode>): Likewise.
6413 (neon_vec_unpack<US>_lo_<mode>): Likewise.
6414 (neon_vec_unpack<US>_hi_<mode>): Likewise.
6415 (neon_vec_<US>mult_lo_<mode>): Likewise.
6416 (neon_vec_<US>mult_hi_<mode>): Likewise.
6417 (neon_vec_<US>shiftl_<mode>): Likewise.
6418 (neon_unpack<US>_<mode>): Likewise.
6419 (neon_vec_<US>mult_<mode>): Likewise.
6420 (vec_pack_trunc_<mode>): Likewise.
6421 (neon_vec_pack_trunk_<mode>): Likewise.
6422 (neon_vabd<mode>_2): Likewise.
6423 (neon_vabd<mode>_3): Likewise.
6424 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
6425 (thumb2_movsi_vfp): Likewise.
6426 (movdi_vfp): Likewise.
6427 (movdi_vfp_cortexa8): Likewise.
6428 (movhf_vfp_neon): Likewise.
6429 (movhf_vfp): Likewiwse.
6430 (movsf_vfp): Likewiwse.
6431 (thumb2_movsf_vfp): Likewiwse.
6432 (movdf_vfp): Likewise.
6433 (thumb2_movdf_vfp): Likewise.
6434 (movsfcc_vfp): Likewise.
6435 (thumb2_movsfcc_vfp): Likewise.
6436 (movdfcc_vfp): Likewise.
6437 (thumb2_movdfcc_vfp): Likewise.
6438 * config/arm/arm.c (cortexa7_older_only): Update for attribute change.
6439 * config/arm/arm1020e.md (v10_c2v): Update for attribute change.
6440 (v10_v2c): Likewise.
6441 * config/arm/cortex-a15-neon.md (cortex_a15_neon_int_1): Update for
6442 attribute change.
6443 (cortex_a15_neon_int_2): Likewise.
6444 (cortex_a15_neon_int_3): Likewise.
6445 (cortex_a15_neon_int_4): Likewise.
6446 (cortex_a15_neon_int_5): Likewise.
6447 (cortex_a15_neon_vqneg_vqabs): Likewise.
6448 (cortex_a15_neon_vmov): Likewise.
6449 (cortex_a15_neon_vaba): Likewise.
6450 (cortex_a15_neon_vaba_qqq): Likewise.
6451 (cortex_a15_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6452 (cortex_a15_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6453 (cortex_a15_neon_mul_qdd_64_32_long_qqd_16_ddd_32_\
6454 scalar_64_32_long_scalar): Likewise.
6455 (cortex_a15_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6456 (cortex_a15_neon_mla_qqq_8_16): Likewise.
6457 (cortex_a15_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6458 lotype_qdd_64_32_long): Likewise.
6459 (cortex_a15_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6460 (cortex_a15_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6461 (cortex_a15_neon_mul_qqd_32_scalar): Likewise.
6462 (cortex_a15_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6463 (cortex_a15_neon_shift_1): Likewise.
6464 (cortex_a15_neon_shift_2): Likewise.
6465 (cortex_a15_neon_shift_3): Likewise.
6466 (cortex_a15_neon_vshl_ddd): Likewise.
6467 (cortex_a15_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6468 (cortex_a15_neon_vsra_vrsra): Likewise.
6469 (cortex_a15_neon_fp_vadd_ddd_vabs_dd): Likewise.
6470 (cortex_a15_neon_fp_vadd_qqq_vabs_qq): Likewise.
6471 (cortex_a15_neon_fp_vmul_ddd): Likewise.
6472 (cortex_a15_neon_fp_vmul_qqd): Likewise.
6473 (cortex_a15_neon_fp_vmla_ddd): Likewise.
6474 (cortex_a15_neon_fp_vmla_qqq): Likewise.
6475 (cortex_a15_neon_fp_vmla_ddd_scalar): Likewise.
6476 (cortex_a15_neon_fp_vmla_qqq_scalar): Likewise.
6477 (cortex_a15_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6478 (cortex_a15_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6479 (cortex_a15_neon_bp_simple): Likewise.
6480 (cortex_a15_neon_bp_2cycle): Likewise.
6481 (cortex_a15_neon_bp_3cycle): Likewise.
6482 (cortex_a15_neon_vld1_1_2_regs): Likewise.
6483 (cortex_a15_neon_vld1_3_4_regs): Likewise.
6484 (cortex_a15_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6485 (cortex_a15_neon_vld2_4_regs): Likewise.
6486 (cortex_a15_neon_vld3_vld4): Likewise.
6487 (cortex_a15_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6488 (cortex_a15_neon_vst1_3_4_regs): Likewise.
6489 (cortex_a15_neon_vst2_4_regs_vst3_vst4): Likewise.
6490 (cortex_a15_neon_vst3_vst4): Likewise.
6491 (cortex_a15_neon_vld1_vld2_lane): Likewise.
6492 (cortex_a15_neon_vld3_vld4_lane" 10
6493 (cortex_a15_neon_vst1_vst2_lane): Likewise.
6494 (cortex_a15_neon_vst3_vst4_lane): Likewise.
6495 (cortex_a15_neon_vld3_vld4_all_lanes): Likewise.
6496 (cortex_a15_neon_ldm_2): Likewise.0
6497 (cortex_a15_neon_stm_2): Likewise.
6498 (cortex_a15_neon_mcr): Likewise.
6499 (cortex_a15_neon_mcr_2_mcrr): Likewise.
6500 (cortex_a15_neon_mrc): Likewise.
6501 (cortex_a15_neon_mrrc): Likewise.
6502 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
6503 change.
6504 (cortex_a15_alu_shift): Likewise.
6505 (cortex_a15_alu_shift_reg): Likewise.
6506 (cortex_a15_mult32): Likewise.
6507 (cortex_a15_mult64): Likewise.
6508 (cortex_a15_block): Likewise.
6509 (cortex_a15_branch): Likewise.
6510 (cortex_a15_load1): Likewise.
6511 (cortex_a15_load3): Likewise.
6512 (cortex_a15_store1): Likewise.
6513 (cortex_a15_store3): Likewise.
6514 (cortex_a15_call): Likewise.
6515 * config/arm/cortex-a5.md (cortex_a5_r2f): Update for attribute change.
6516 (cortex_a5_f2r): Likewise.
6517 * config/arm/cortex-a53.md (cortex_a53_r2f): Update for attribute
6518 change.
6519 (cortex_a53_f2r): Likewise.
6520 * config/arm/cortex-a7.md
6521 (cortex_a7_branch): Update for attribute change.
6522 (cortex_a7_call): Likewise.
6523 (cortex_a7_alu_imm): Likewise.
6524 (cortex_a7_alu_reg): Likewise.
6525 (cortex_a7_alu_shift): Likewise.
6526 (cortex_a7_mul): Likewise.
6527 (cortex_a7_load1): Likewise.
6528 (cortex_a7_store1): Likewise.
6529 (cortex_a7_load2): Likewise.
6530 (cortex_a7_store2): Likewise.
6531 (cortex_a7_load3): Likewise.
6532 (cortex_a7_store3): Likewise.
6533 (cortex_a7_load4): Likewise.
6534 (cortex_a7_store4): Likewise.
6535 (cortex_a7_fpalu): Likewise.
6536 (cortex_a7_fconst): Likewise.
6537 (cortex_a7_fpmuls): Likewise.
6538 (cortex_a7_neon_mul): Likewise.
6539 (cortex_a7_fpmacs): Likewise.
6540 (cortex_a7_neon_mla: Likewise.
6541 (cortex_a7_fpmuld: Likewise.
6542 (cortex_a7_fpmacd: Likewise.
6543 (cortex_a7_fpfmad: Likewise.
6544 (cortex_a7_fdivs: Likewise.
6545 (cortex_a7_fdivd: Likewise.
6546 (cortex_a7_r2f: Likewise.
6547 (cortex_a7_f2r: Likewise.
6548 (cortex_a7_f_flags: Likewise.
6549 (cortex_a7_f_loads: Likewise.
6550 (cortex_a7_f_loadd: Likewise.
6551 (cortex_a7_f_stores: Likewise.
6552 (cortex_a7_f_stored: Likewise.
6553 (cortex_a7_neon): Likewise.
6554 * config/arm/cortex-a8-neon.md
6555 (cortex_a8_neon_mrc): Update for attribute change.
6556 (cortex_a8_neon_mrrc): Likewise.
6557 (cortex_a8_neon_int_1): Likewise.
6558 (cortex_a8_neon_int_2): Likewise.
6559 (cortex_a8_neon_int_3): Likewise.
6560 (cortex_a8_neon_int_4): Likewise.
6561 (cortex_a8_neon_int_5): Likewise.
6562 (cortex_a8_neon_vqneg_vqabs): Likewise.
6563 (cortex_a8_neon_vmov): Likewise.
6564 (cortex_a8_neon_vaba): Likewise.
6565 (cortex_a8_neon_vaba_qqq): Likewise.
6566 (cortex_a8_neon_vsma): Likewise.
6567 (cortex_a8_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6568 (cortex_a8_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6569 (cortex_a8_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6570 long_scalar): Likewise.
6571 (cortex_a8_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6572 (cortex_a8_neon_mla_qqq_8_16): Likewise.
6573 (cortex_a8_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6574 long_scalar_qdd_64_32_long): Likewise.
6575 (cortex_a8_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6576 (cortex_a8_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6577 (cortex_a8_neon_mul_qqd_32_scalar): Likewise.
6578 (cortex_a8_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6579 (cortex_a8_neon_shift_1): Likewise.
6580 (cortex_a8_neon_shift_2): Likewise.
6581 (cortex_a8_neon_shift_3): Likewise.
6582 (cortex_a8_neon_vshl_ddd): Likewise.
6583 (cortex_a8_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6584 (cortex_a8_neon_vsra_vrsra): Likewise.
6585 (cortex_a8_neon_fp_vadd_ddd_vabs_dd): Likewise.
6586 (cortex_a8_neon_fp_vadd_qqq_vabs_qq): Likewise.
6587 (cortex_a8_neon_fp_vsum): Likewise.
6588 (cortex_a8_neon_fp_vmul_ddd): Likewise.
6589 (cortex_a8_neon_fp_vmul_qqd): Likewise.
6590 (cortex_a8_neon_fp_vmla_ddd): Likewise.
6591 (cortex_a8_neon_fp_vmla_qqq): Likewise.
6592 (cortex_a8_neon_fp_vmla_ddd_scalar): Likewise.
6593 (cortex_a8_neon_fp_vmla_qqq_scalar): Likewise.
6594 (cortex_a8_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6595 (cortex_a8_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6596 (cortex_a8_neon_bp_simple): Likewise.
6597 (cortex_a8_neon_bp_2cycle): Likewise.
6598 (cortex_a8_neon_bp_3cycle): Likewise.
6599 (cortex_a8_neon_ldr): Likewise.
6600 (cortex_a8_neon_str): Likewise.
6601 (cortex_a8_neon_vld1_1_2_regs): Likewise.
6602 (cortex_a8_neon_vld1_3_4_regs): Likewise.
6603 (cortex_a8_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6604 (cortex_a8_neon_vld2_4_regs): Likewise.
6605 (cortex_a8_neon_vld3_vld4): Likewise.
6606 (cortex_a8_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6607 (cortex_a8_neon_vst1_3_4_regs): Likewise.
6608 (cortex_a8_neon_vst2_4_regs_vst3_vst4): Likewise.
6609 (cortex_a8_neon_vst3_vst4): Likewise.
6610 (cortex_a8_neon_vld1_vld2_lane): Likewise.
6611 (cortex_a8_neon_vld3_vld4_lane): Likewise.
6612 (cortex_a8_neon_vst1_vst2_lane): Likewise.
6613 (cortex_a8_neon_vst3_vst4_lane): Likewise.
6614 (cortex_a8_neon_vld3_vld4_all_lanes): Likewise.
6615 (cortex_a8_neon_mcr): Likewise.
6616 (cortex_a8_neon_mcr_2_mcrr): Likewise.
6617 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
6618 * config/arm/cortex-a9-neon.md (ca9_neon_mrc): Update for attribute
6619 change.
6620 (ca9_neon_mrrc): Likewise.
6621 (cortex_a9_neon_int_1): Likewise.
6622 (cortex_a9_neon_int_2): Likewise.
6623 (cortex_a9_neon_int_3): Likewise.
6624 (cortex_a9_neon_int_4): Likewise.
6625 (cortex_a9_neon_int_5): Likewise.
6626 (cortex_a9_neon_vqneg_vqabs): Likewise.
6627 (cortex_a9_neon_vmov): Likewise.
6628 (cortex_a9_neon_vaba): Likewise.
6629 (cortex_a9_neon_vaba_qqq): Likewise.
6630 (cortex_a9_neon_vsma): Likewise.
6631 (cortex_a9_neon_mul_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6632 (cortex_a9_neon_mul_qqq_8_16_32_ddd_32): Likewise.
6633 (cortex_a9_neon_mul_qdd_64_32_long_qqd_16_ddd_32_scalar_64_32_\
6634 long_scalar): Likewise.
6635 (cortex_a9_neon_mla_ddd_8_16_qdd_16_8_long_32_16_long): Likewise.
6636 (cortex_a9_neon_mla_qqq_8_16): Likewise.
6637 (cortex_a9_neon_mla_ddd_32_qqd_16_ddd_32_scalar_qdd_64_32_\
6638 long_scalar_qdd_64_32_long): Likewise.
6639 (cortex_a9_neon_mla_qqq_32_qqd_32_scalar): Likewise.
6640 (cortex_a9_neon_mul_ddd_16_scalar_32_16_long_scalar): Likewise.
6641 (cortex_a9_neon_mul_qqd_32_scalar): Likewise.
6642 (cortex_a9_neon_mla_ddd_16_scalar_qdd_32_16_long_scalar): Likewise.
6643 (cortex_a9_neon_shift_1): Likewise.
6644 (cortex_a9_neon_shift_2): Likewise.
6645 (cortex_a9_neon_shift_3): Likewise.
6646 (cortex_a9_neon_vshl_ddd): Likewise.
6647 (cortex_a9_neon_vqshl_vrshl_vqrshl_qqq): Likewise.
6648 (cortex_a9_neon_vsra_vrsra): Likewise.
6649 (cortex_a9_neon_fp_vadd_ddd_vabs_dd): Likewise.
6650 (cortex_a9_neon_fp_vadd_qqq_vabs_qq): Likewise.
6651 (cortex_a9_neon_fp_vsum): Likewise.
6652 (cortex_a9_neon_fp_vmul_ddd): Likewise.
6653 (cortex_a9_neon_fp_vmul_qqd): Likewise.
6654 (cortex_a9_neon_fp_vmla_ddd): Likewise.
6655 (cortex_a9_neon_fp_vmla_qqq): Likewise.
6656 (cortex_a9_neon_fp_vmla_ddd_scalar): Likewise.
6657 (cortex_a9_neon_fp_vmla_qqq_scalar): Likewise.
6658 (cortex_a9_neon_fp_vrecps_vrsqrts_ddd): Likewise.
6659 (cortex_a9_neon_fp_vrecps_vrsqrts_qqq): Likewise.
6660 (cortex_a9_neon_bp_simple): Likewise.
6661 (cortex_a9_neon_bp_2cycle): Likewise.
6662 (cortex_a9_neon_bp_3cycle): Likewise.
6663 (cortex_a9_neon_ldr): Likewise.
6664 (cortex_a9_neon_str): Likewise.
6665 (cortex_a9_neon_vld1_1_2_regs): Likewise.
6666 (cortex_a9_neon_vld1_3_4_regs): Likewise.
6667 (cortex_a9_neon_vld2_2_regs_vld1_vld2_all_lanes): Likewise.
6668 (cortex_a9_neon_vld2_4_regs): Likewise.
6669 (cortex_a9_neon_vld3_vld4): Likewise.
6670 (cortex_a9_neon_vst1_1_2_regs_vst2_2_regs): Likewise.
6671 (cortex_a9_neon_vst1_3_4_regs): Likewise.
6672 (cortex_a9_neon_vst2_4_regs_vst3_vst4): Likewise.
6673 (cortex_a9_neon_vst3_vst4): Likewise.
6674 (cortex_a9_neon_vld1_vld2_lane): Likewise.
6675 (cortex_a9_neon_vld3_vld4_lane): Likewise.
6676 (cortex_a9_neon_vst1_vst2_lane): Likewise.
6677 (cortex_a9_neon_vst3_vst4_lane): Likewise.
6678 (cortex_a9_neon_vld3_vld4_all_lanes): Likewise.
6679 (cortex_a9_neon_mcr): Likewise.
6680 (cortex_a9_neon_mcr_2_mcrr): Likewise.
6681 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
6682 (cortex_a9_fps): Likewise.
6683 * config/arm/cortex-m4-fpu.md (cortex_m4_vmov_2): Update for attribute
6684 change.
6685 (cortex_m4_fmuls): Likewise.
6686 * config/arm/cortex-r4f.md (cortex_r4_mcr): Update for attribute
6687 change.
6688 (cortex_r4_mrc): Likewise.
6689 * config/arm/iterators.md: Update comment referring to neon_type.
6690 * config/arm/iwmmxt.md (iwmmxt_arm_movdi): Update for attribute change.
6691 (iwmmxt_movsi_insn): Likewise.
6692 * config/arm/marvell-pj4.md (pj4_vfp_to_core): Update for
6693 attribute change.
6694 (pj4_core_to_vfp): Likewise.
6695 * config/arm/neon-schedgen.ml (emit_insn_reservations): Update for
6696 attribute change.
6697 * config/arm/vfp11.md (vfp_fload): Update for attribute change.
6698 (vfp_fstore): Likewise.
6699 * doc/md.texi: Change references to neon_type to refer to type.
6700
6701 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6702
6703 * tree.h (DECL_BUILT_IN): Fix typo in comment.
6704
6705 2013-09-04 David Edelsohn <dje.gcc@gmail.com>
6706
6707 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Only emit
6708 lglobl if not weak.
6709
6710 2013-09-04 Easwaran Raman <eraman@google.com>
6711
6712 PR middle-end/57370
6713 * tree-ssa-reassoc.c (get_stmt_uid_with_default): New function,
6714 (build_and_add_sum): Use it.
6715 (appears_later_in_bb): Simplify code.
6716
6717 2013-09-04 Teresa Johnson <tejohnson@google.com>
6718
6719 * dumpfile.c (dump_finish): Don't close stderr/stdout.
6720
6721 2013-09-04 James Greenhalgh <james.greenhalgh@arm.com>
6722
6723 * config/aarch64/arm_neon.h (vaddvq_<su>64): Fix return types.
6724
6725 2013-09-04 Jan Hubicka <jh@suse.cz>
6726
6727 * Makefile.in (ipa-devirt.o): Add dependency on diagnostic.h
6728 * ipa-devirt.c: Include diganostic.h
6729 (odr_type_d): Add types and types_set.
6730 (hash_type_name): Work for types with vtables during LTO.
6731 (odr_hasher::remove): Fix comment; destroy types_set.
6732 (add_type_duplicate): New function,
6733 (get_odr_type): Use it.
6734 (dump_type_inheritance_graph): Dump type duplicates.
6735 * ipa.c (symtab_remove_unreachable_nodes): Build type inheritance
6736 graph.
6737 * tree.c (types_same_for_odr): Give exact answers on types with
6738 virtual tables.
6739
6740 2013-09-04 Dodji Seketeli <dodji@redhat.com>
6741
6742 * tree.h (DECL_BUILT_IN, DECL_IS_BUILTIN): Add more comments
6743 explaining their differences.
6744
6745 2013-09-04 Sandeep Kumar Singh<Sandeep.Singh2@kpitcummins.com>
6746
6747 * config/rx/rx.h: Add option -mcpu for target variants RX100 and RX200.
6748
6749 2013-09-03 Jeff Law <law@redhat.com>
6750
6751 * tree-ssa-threadedge.c (thread_across_edge): Record entire path
6752 when not threading through a joiner block. Pass joiner/no joiner
6753 state to register_jump_thread.
6754 * tree-ssa-threadupdate.c (register_jump_thread): Get joiner/no joiner
6755 state from argument rather than implying on path length.
6756 Dump the entire jump thread path into debugging dump.
6757 * tree-flow.h (register_jump_thread): Update prototype.
6758
6759 2013-08-29 Xinliang David Li <davidxl@google.com>
6760
6761 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6762 Remove a trivial gcc_assert.
6763
6764 2013-08-29 Xinliang David Li <davidxl@google.com>
6765
6766 * tree-vect-slp.c (destroy_bb_vec_info): Data ref cleanup.
6767 * tree-vect-loop.c (destroy_bb_vec_info): Ditto.
6768 * tree-vect-data-refs.c (vect_compute_data_ref_alignment):
6769 Delay base decl alignment adjustment.
6770 * tree-vectorizer.c (vect_destroy_datarefs): New function.
6771 * tree-vectorizer.h: New data structure.
6772 (set_dr_misalignment): New function.
6773 (dr_misalignment): Ditto.
6774 * tree-vect-stmts.c (vectorizable_store): Ensure alignment.
6775 (vectorizable_load): Ditto.
6776 (ensure_base_align): New function.
6777 (vectorize_loops): Add dbg_cnt support.
6778 (execute_vect_slp): Ditto.
6779 * dbgcnt.def: New debug counter.
6780 * Makefile: New dependency.
6781
6782 2013-09-03 Meador Inge <meadori@codesourcery.com>
6783
6784 Revert:
6785
6786 2013-08-30 Meador Inge <meadori@codesourcery.com>
6787
6788 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
6789
6790 2013-09-03 David Edelsohn <dje.gcc@gmail.com>
6791
6792 * config/rs6000/rs6000.h (ASM_OUTPUT_DEF_FROM_DECLS): Emit lglobl for
6793 function descriptor.
6794
6795 2013-09-03 Richard Biener <rguenther@suse.de>
6796
6797 * tree-affine.c (add_elt_to_tree): Fix association issue,
6798 avoid useless converts and make sure to always return a
6799 properly typed result.
6800
6801 2013-09-03 Richard Biener <rguenther@suse.de>
6802
6803 PR middle-end/57656
6804 * fold-const.c (negate_expr_p): Fix division case.
6805 (negate_expr): Likewise.
6806
6807 2013-09-03 Richard Biener <rguenther@suse.de>
6808
6809 PR lto/58285
6810 * tree-streamer-out.c: Include tm.h.
6811 * Makefile.in (tree-streamer-out.o): Depend on $(TM_H).
6812
6813 2013-09-03 Jan Hubicka <jh@suse.cz>
6814
6815 * tree-profile.c (tree_profiling): Cleanup CFG when done.
6816
6817 2013-09-03 Alan Modra <amodra@gmail.com>
6818
6819 * config.gcc (powerpc*-*-linux*): Add support for little-endian
6820 multilibs to big-endian target and vice versa.
6821 * config/rs6000/t-linux64: Use := assignment on all vars.
6822 (MULTILIB_EXTRA_OPTS): Remove fPIC.
6823 (MULTILIB_OSDIRNAMES): Specify using mapping from multilib_options.
6824 * config/rs6000/t-linux64le: New file.
6825 * config/rs6000/t-linux64bele: New file.
6826 * config/rs6000/t-linux64lebe: New file.
6827
6828 2013-09-02 Jan Hubicka <jh@suse.cz>
6829
6830 * ipa-inline-transform.c (inline_transform): Do not
6831 optimize_inline_calls when not optimizing.
6832
6833 2013-09-02 Jan Hubicka <jh@suse.cz>
6834
6835 * lto-symtab.c (lto_symtab_merge_symbols): Add comments; merge
6836 duplicated nodes for assembler names.
6837 * symtab.c (symtab_unregister_node): Do not attempt to unlink
6838 hard registers from assembler name hash.
6839
6840 2013-09-02 Jan Hubicka <jh@suse.cz>
6841
6842 * ipa-split.c (execute_split_functions): Split externally visible
6843 functions called once.
6844
6845 2013-09-02 Martin Jambor <mjambor@suse.cz>
6846
6847 PR ipa/58106
6848 * ipa-prop.c (ipa_edge_duplication_hook): Always put new rdesc to the
6849 linked list. When finding the correct duplicate, also consider also
6850 the caller in additon to its inlined_to node.
6851
6852 2013-09-02 James Greenhalgh <james.greenhalgh@arm.com>
6853
6854 * config/aarch64/aarch64-simd-builtins.def
6855 (dup_lane_scalar): Remove.
6856 * config/aarch64/aarch64-simd.md
6857 (aarch64_simd_dup): Add 'w->w' alternative.
6858 (aarch64_dup_lane<mode>): Allow for VALL.
6859 (aarch64_dup_lane_scalar<mode>): Remove.
6860 (aarch64_dup_lane_<vswap_width_name><mode>): New.
6861 (aarch64_get_lane_signed<mode>): Add w->w altenative.
6862 (aarch64_get_lane_unsigned<mode>): Likewise.
6863 (aarch64_get_lane<mode>): Likewise.
6864 * config/aarch64/aarch64.c (aarch64_evpc_dup): New.
6865 (aarch64_expand_vec_perm_const_1): Use aarch64_evpc_dup.
6866 * config/aarch64/iterators.md (VSWAP_WIDTH): New.
6867 (VCON): Change container of V2SF.
6868 (vswap_width_name): Likewise.
6869 * config/aarch64/arm_neon.h
6870 (__aarch64_vdup_lane_any): New.
6871 (__aarch64_vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6872 (vdup<q>_n_<psuf><8,16,32,64>): Convert to C implementation.
6873 (vdup<q>_lane<q>_<fpsu><8,16,32,64>): Likewise.
6874
6875 2013-09-02 Eric Botcazou <ebotcazou@adacore.com>
6876
6877 PR middle-end/56382
6878 * expr.c (emit_move_complex): Do not move complex FP values as parts if
6879 the source or the destination is a single hard register.
6880
6881 2013-09-02 Richard Biener <rguenther@suse.de>
6882
6883 PR middle-end/57511
6884 * tree-scalar-evolution.c (instantiate_scev_name): Allow
6885 non-linear SCEVs.
6886
6887 2013-09-02 Richard Biener <rguenther@suse.de>
6888
6889 * tree-affine.c (add_elt_to_tree): Avoid converting all pointer
6890 arithmetic to sizetype.
6891
6892 2013-09-02 Bin Cheng <bin.cheng@arm.com>
6893
6894 * tree-ssa-loop-ivopts.c (set_autoinc_for_original_candidates):
6895 Find auto-increment use both before and after candidate.
6896
6897 2013-09-02 Marek Polacek <polacek@redhat.com>
6898
6899 * Makefile.in (ubsan.o): Add $(TM_P_H) dependency.
6900
6901 2013-09-01 Jan Hubicka <jh@suse.cz>
6902
6903 * Makefile.in: Add ipa-profile.o
6904 (ipa.o, ipa-devrit.o, ipa-inline-analysis.o): Adjust dependencies.
6905 * cgraph.c (struct cgraph_propagate_frequency_data,
6906 cgraph_propagate_frequency_1, cgraph_propagate_frequency): Move to
6907 ipa-profile.c; replace cgraph_ by ipa_ prefix.
6908 * cgraph.h (cgraph_propagate_frequency): Remove.
6909 * ipa-inline-analysis.c: Include ipa-utils.h;
6910 drop duplicated cfgloop.h.
6911 (inline_update_callee_summaries): Update.
6912 * ipa-profile.c: New file.
6913 * ipa-utils.h (ipa_propagate_frequency): Declare.
6914 * ipa.c: Do not include pointer-set.h, hash-table.h, lto-streamer.h,
6915 data-streamer.h, value-prof.h.
6916 (symtab_remove_unreachable_nodes): Update profile.
6917 (struct histogram_entry, histogram, histogram_pool, histogram_hash,
6918 account_time_size, cmp_counts, dump_histogram,
6919 ipa_profile_generate_summary, ipa_profile_write_summary,
6920 ipa_profile_read_summary, ipa_profile, gate_ipa_profile,
6921 pass_data_ipa_profile, pass_ipa_profile, make_pass_ipa_profile):
6922 Move to ipa-profile.c.
6923
6924 2013-09-01 John David Anglin <danglin@gcc.gnu.org>
6925
6926 * config/pa/pa.md: Allow "const 0" operand 1 in "scc" insns.
6927
6928 2013-09-01 Jan Hubicka <jh@suse.cz>
6929
6930 * common.opt (fdevirtualize-speculatively): New function.
6931 * invoke.texi (fdevirtualize-speculatively): Document.
6932 * ipa-devirt.c: Include ipa-inline.h
6933 (likely_target_p): New function.
6934 (ipa_devirt): New function.
6935 (gate_ipa_devirt): New function.
6936 (pass_data_ipa_devirt): New static var.
6937 (pass_ipa_devirt): Likewise.
6938 (make_pass_ipa_devirt): New function.
6939 * opts.c (default_options): Add OPT_fdevirtualize_speculatively.
6940 (common_handle_option): Disable devirtualization when
6941 value range profiling is available.
6942 * passes.def (pass_ipa_devirt): Add.
6943 * timever.def (TV_IPA_DEVIRT): New timevar.
6944 * tree-pass.h (make_pass_ipa_devirt):
6945
6946 2013-09-01 Iain Sandoe <iain@codesourcery.com>
6947
6948 * config/darwin.h (LINK_COMMAND_SPEC_A): Revise sanitizer specs to
6949 include sanitize(undefined).
6950
6951 2013-08-31 Diego Novillo <dnovillo@google.com>
6952
6953 * Makefile.in (TREE_CORE_H): Define.
6954 (TREE_H): Use.
6955 (GTFILES): Add tree-core.h.
6956 * builtins.c (built_in_class_names): Use BUILT_IN_LAST to
6957 size the array.
6958 * tree-core.h: New file.
6959 Move all data structures, enum, typedefs, global
6960 declarations and constants from ...
6961 * tree.h: ... here.
6962
6963 2013-08-31 Jan Hubicka <jh@suse.cz>
6964
6965 * bulitins.c (expand_builtin): Do not early exit for gcov
6966 instrumented functions.
6967
6968 2013-08-31 Marek Polacek <polacek@redhat.com>
6969
6970 * ubsan.c: Include tm_p.h.
6971
6972 2013-08-31 Jan Hubicka <jh@suse.cz>
6973
6974 * gimple-streamer-in.c (input_gimple_stmt): Silence parameter unused
6975 warning.
6976
6977 * cgraph.c (cgraph_get_body): Update call of lto_input_function_body.
6978 * gimple-streamer-in.c (input_gimple_stmt): Move sanity check to ...
6979 * tree-cfg.c (verify_gimple_label): ... here.
6980 * ipa-utils.c: Include lto-streamer.h, ipa-inline.h
6981 (ipa_merge_profiles): New function.
6982 * lto-streamer-in.c (lto_read_body): Take node instead of fn_decl.
6983 (lto_input_function_body): Likewise.
6984 * ipa-utils.h (ipa_merge_profiles): Declare.
6985 * lto-streamer.h (lto_input_function_body): Update prototype.
6986 (emit_label_in_global_context_p): Remove.
6987 * lto-symtab.c: Include ipa-utils.h
6988 (lto_cgraph_replace_node): Use ipa_merge_profiles.
6989
6990 2013-08-31 Jan Hubicka <jh@suse.cz>
6991
6992 * cgraph.c (cgraph_speculative_call_info): Fix ref lookup
6993
6994 2013-08-31 Jan Hubicka <jh@suse.cz>
6995
6996 * basic-block.h (apply_scale): Make scale parmeter gcov_type.
6997
6998 2013-08-31 Uros Bizjak <ubizjak@gmail.com>
6999
7000 * config/alpha/alpha.c (alpha_emit_conditional_move): Update
7001 "cmp" RTX before signed_comparison_operator check to account
7002 for "code" changes.
7003
7004 2013-08-30 Jan Hubicka <jh@suse.cz>
7005
7006 * ipa-prop.c (ipa_set_jf_known_type): Check that we add only records.
7007 (detect_type_change_1): Rename to ...
7008 (detect_type_change): ... this one; early return on non-polymorphic
7009 types.
7010 (detect_type_change_ssa): Add comp_type parameter; update
7011 use of detect_type_change.
7012 (compute_complex_assign_jump_func): Add param_type parameter;
7013 update use of detect_type_change_ssa.
7014 (compute_complex_ancestor_jump_func): Likewise.
7015 (ipa_get_callee_param_type): New function.
7016 (ipa_compute_jump_functions_for_edge): Compute parameter type;
7017 update calls to the jump function computation functions.
7018
7019 2013-08-30 Teresa Johnson <tejohnson@google.com>
7020 Steven Bosscher <steven@gcc.gnu.org>
7021
7022 * cfgrtl.c (fixup_new_cold_bb): New routine.
7023 (commit_edge_insertions): Invoke fixup_partitions.
7024 (find_partition_fixes): New routine.
7025 (fixup_partitions): Ditto.
7026 (verify_hot_cold_block_grouping): Update comments.
7027 (rtl_verify_edges): Invoke find_partition_fixes.
7028 (rtl_verify_bb_pointers): Update comments.
7029 (rtl_verify_bb_layout): Ditto.
7030 * basic-block.h (probably_never_executed_edge_p): Declare.
7031 (fixup_partitions): Ditto.
7032 * cfgcleanup.c (try_optimize_cfg): Invoke fixup_partitions.
7033 * bb-reorder.c (sanitize_hot_paths): New function.
7034 (find_rarely_executed_basic_blocks_and_crossing_edges): Invoke
7035 sanitize_hot_paths.
7036 * predict.c (probably_never_executed_edge_p): New routine.
7037 * cfg.c (check_bb_profile): Add partition insanity warnings.
7038
7039 2013-08-30 Meador Inge <meadori@codesourcery.com>
7040
7041 * tree-vrp.c (check_array_ref): Bail out on zero-length arrays.
7042
7043 2013-08-30 Marek Polacek <polacek@redhat.com>
7044
7045 * Makefile.in (ubsan.o): Add.
7046 (c-family/c-ubsan.o): Add.
7047 (builtins.o): Add ubsan.h dependency.
7048 * ubsan.h: New file.
7049 * ubsan.c: New file.
7050 * common.opt: Add -fsanitize=undefined option.
7051 (flag_sanitize): Add variable.
7052 (fsanitize=): Add option. Add Driver.
7053 (fsanitize=thread): Remove option.
7054 (fsanitize=address): Likewise.
7055 (static-libubsan): New option.
7056 * doc/invoke.texi: Document the new flag and -static-libubsan.
7057 * sanitizer.def (DEF_SANITIZER_BUILTIN): Define.
7058 (BUILT_IN_UBSAN_HANDLE_BUILTIN_UNREACHABLE): Define.
7059 * builtin-attrs.def (ATTR_COLD): Define.
7060 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7061 * builtins.def (BUILT_IN_UBSAN_HANDLE_DIVREM_OVERFLOW,
7062 BUILT_IN_UBSAN_HANDLE_SHIFT_OUT_OF_BOUNDS): Define.
7063 * flag-types.h (sanitize_code): New enum.
7064 * opts.c (common_handle_option): Parse command line arguments
7065 of -fsanitize=. Add -fsanitize=unreachable option.
7066 * varasm.c (get_variable_section): Adjust.
7067 (assemble_noswitch_variable): Likewise.
7068 (assemble_variable): Likewise.
7069 (output_constant_def_contents): Likewise.
7070 (categorize_decl_for_section): Likewise.
7071 (place_block_symbol): Likewise.
7072 (output_object_block): Likewise.
7073 * builtins.def: Likewise.
7074 * toplev.c (compile_file): Likewise.
7075 (process_options): Likewise.
7076 * cppbuiltin.c: Likewise.
7077 * tsan.c (tsan_pass): Likewise.
7078 (tsan_gate): Likewise.
7079 (tsan_gate_O0): Likewise.
7080 * cfgexpand.c (partition_stack_vars): Likewise.
7081 (expand_stack_vars): Likewise.
7082 (defer_stack_allocation): Likewise.
7083 (expand_used_vars): Likewise.
7084 * cfgcleanup.c (old_insns_match_p): Likewise.
7085 * asan.c (asan_finish_file): Likewise.
7086 (asan_instrument): Likewise.
7087 (gate_asan): Likewise.
7088 (initialize_sanitizer_builtins): Build BT_FN_VOID_PTR_PTR_PTR.
7089 (ATTR_COLD_NOTHROW_LEAF_LIST): Define.
7090 (asan_global_struct): Use pointer_sized_int_node instead
7091 calling build_nonstandard_integer_type.
7092 (initialize_sanitizer_builtins): Likewise.
7093 (asan_finish_file): Likewise.
7094 * gcc.c: Document %{%:function(args):X}.
7095 (static_spec_functions): Add sanitize.
7096 (handle_spec_function): Add retval_nonnull argument and if non-NULL,
7097 store funcval != NULL there.
7098 (do_spec_1): Adjust handle_spec_function caller.
7099 (handle_braces): Allow %:function(args) as condition.
7100 (sanitize_spec_function): New function.
7101 (ADD_STATIC_LIBUBSAN_LIBS): Define.
7102 (LIBUBSAN_SPEC): Likewise.
7103 (LIBUBSAN_EARLY_SPEC): Likewise.
7104 (SANITIZER_SPEC): Handle libubsan.
7105 (SANITIZER_EARLY_SPEC): Likewise.
7106 * config/darwin.h (LINK_COMMAND_SPEC_A): Use %:sanitize(address)
7107 instead of fsanitize=address.
7108 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Use %:sanitize(address)
7109 instead of fsanitize=address*.
7110 * builtins.c: Include ubsan.h.
7111 (fold_builtin_0): Instrument __builtin_unreachable.
7112 * config/rs6000/rs6000.h (FRAME_GROWS_DOWNWARD): Use flag_sanitize
7113 instead of flag_asan.
7114 * tree.h (enum tree_index): Add TI_POINTER_SIZED_TYPE.
7115 (pointer_sized_int_node): Define.
7116 * tree.c (build_common_tree_nodes): Initialize pointer_sized_int_node.
7117
7118 2013-08-30 Mike Stump <mikestump@comcast.net>
7119
7120 * doc/install.texi (Prerequisites): Note regression in Tcl 8.6
7121 with RE patterns.
7122
7123 2013-08-29 Jan Hubicka <jh@suse.cz>
7124
7125 * cgraph.c (cgraph_function_body_availability): Handle weakref
7126 correctly.
7127 * passes.def: Remove pass_fixup_cfg.
7128 * ipa-inline.c (ipa_inline): When not optimizing, do not inline;
7129 track when we need to remove functions.
7130 (gate_ipa_inline): Execute inlining always; add comment why.
7131 (pass_data_ipa_inline): Remove TODO_remove_functions.
7132 * ipa-inline-analysis.c (inline_generate_summary): When not optimizing
7133 do not produce summaries.
7134 * symtab.c (change_decl_assembler_name): Handle renaming of weakrefs.
7135 (symtab_nonoverwritable_alias): Assert we are not called on weakref.
7136 * varpool.c (cgraph_variable_initializer_availability): Fix weakrefs,
7137 constant pool and vtable.
7138
7139 2013-08-30 Tejas Belagod <tejas.belagod@arm.com>
7140
7141 * config/aarch64/arm_neon.h (__AARCH64_UINT64_C, __AARCH64_INT64_C):
7142 New arm_neon.h's internal macros to specify 64-bit constants.
7143 Avoid using stdint.h's macros.
7144
7145 2013-08-30 Joern Rennecke <joern.rennecke@embecosm.com>
7146
7147 * recog.c (verify_changes): Verify that changes[i].old is non-zero
7148 before applying REG_P.
7149
7150 2013-08-30 Jakub Jelinek <jakub@redhat.com>
7151
7152 PR tree-optimization/58277
7153 * tree-ssa-strlen.c (strlen_enter_block): If do_invalidate gave up
7154 after seeing too many stmts with vdef in between dombb and current
7155 bb, invalidate everything.
7156
7157 2013-08-30 Richard Biener <rguenther@suse.de>
7158
7159 * fold-const.c (fold_single_bit_test): Fix overflow test.
7160
7161 2013-08-30 Eric Botcazou <ebotcazou@adacore.com>
7162
7163 * function.c (assign_parm_setup_reg): For a parameter passed by pointer
7164 and which can live in a register, always retrieve the value on entry.
7165 * var-tracking.c (add_stores): Treat the copy on entry for a parameter
7166 passed by invisible reference specially.
7167 (emit_notes_in_bb) <MO_VAL_USE>: Emit notes before the instruction.
7168 (vt_add_function_parameter): Correctly deal with a parameter passed by
7169 invisible reference.
7170
7171 2013-08-30 Jan Hubicka <jh@suse.cz>
7172
7173 * tree.c (set_call_expr_flags): Fix handling of TM_PURE.
7174
7175 2013-08-30 Richard Biener <rguenther@suse.de>
7176
7177 PR tree-optimization/58228
7178 * tree-vect-data-refs.c (vect_analyze_data_ref_access): Do not
7179 allow invariant loads in nested loop vectorization.
7180
7181 2013-08-30 Richard Biener <rguenther@suse.de>
7182
7183 PR tree-optimization/58223
7184 * tree-loop-distribution.c (has_anti_dependence): Rename to ...
7185 (has_anti_or_output_dependence): ... this and adjust to also
7186 look for output dependences.
7187 (mark_nodes_having_upstream_mem_writes): Adjust.
7188 (rdg_flag_uses): Likewise.
7189
7190 2013-08-30 Richard Biener <rguenther@suse.de>
7191
7192 PR tree-optimization/58010
7193 * tree-vect-loop.c (vect_create_epilog_for_reduction): Remove
7194 assert that we have a loop-closed PHI.
7195
7196 2013-08-29 Jan Hubicka <jh@suse.cz>
7197
7198 * lto-symtab.c (lto_cgraph_replace_node): Free decl_in_state.
7199 * cgraph.c (cgraph_release_function_body): Free decl_in_state.
7200 * lto-section-in.c (lto_free_function_in_decl_state): New function.
7201 (lto_free_function_in_decl_state_for_node): New function.
7202
7203 2013-08-29 Xinliang David Li <davidxl@google.com>
7204
7205 * loop-unroll.c (report_unroll_peel): Minor message change.
7206 * tree-vect-loop-manip.c (vect_do_peeling_for_alignment):
7207 Emit alignment peeling message with default -fopt-info.
7208 (vect_loop_versioning): Emit loop version info message.
7209 * tree-vectorizer.c (vectorize_loops): Minor message change.
7210 (execute_vect_slp): Ditto.
7211
7212 2013-08-29 Eric Botcazou <ebotcazou@adacore.com>
7213
7214 * cgraphclones.c (cgraph_create_virtual_clone): Compute the DECL_NAME
7215 of the clone from the DECL_NAME of the original function.
7216
7217 2013-08-29 Oleg Endo <olegendo@gcc.gnu.org>
7218
7219 * passes.c (register_pass): Add overload.
7220 * tree-pass.h (register_pass): Forward declare it. Add comment.
7221
7222 2013-08-29 Jan Hubicka <jh@suse.cz>
7223
7224 * lto-streamer-out.c (hash_tree): Stream DECL_FINAL_P,
7225 DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P and TYPE_FINAL_P.
7226 * lto-streamer-in.c (unpack_ts_decl_with_vis_value_fields): Stream
7227 DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7228 (unpack_ts_type_common_value_fields): Stream TYPE_FINAL_P.
7229 * tree-streamer-out.c (pack_ts_decl_with_vis_value_fields):
7230 Add DECL_FINAL_P, DECL_CXX_CONSTRUCTOR_P and DECL_CXX_DESTRUCTOR_P.
7231 (pack_ts_type_common_value_fields): Add TYPE_FINAL_P.
7232
7233 2013-08-29 Teresa Johnson <tejohnson@google.com>
7234
7235 * dumpfile.c (dump_loc): Output column number.
7236 * dumpfile.h (OPTGROUP_OTHER): Add and enable under OPTGROUP_ALL.
7237 * doc/invoke.texi: Document optall -fopt-info flag.
7238 * profile.c (read_profile_edge_counts): Use new dump framework.
7239 (compute_branch_probabilities): Ditto.
7240 * passes.c (pass_manager::register_one_dump_file): Use OPTGROUP_OTHER
7241 when pass not in any opt group.
7242 * pass_manager.h (pass_manager::get_pass_profile): New method.
7243 * value-prof.c (check_counter): Use new dump framework.
7244 (check_ic_target): Ditto.
7245 * coverage.c (get_coverage_counts): Ditto.
7246 (coverage_init): Setup new dump framework.
7247
7248 2013-08-29 Richard Biener <rguenther@suse.de>
7249
7250 PR tree-optimization/58246
7251 * tree-ssa-dce.c (mark_aliased_reaching_defs_necessary_1): Properly
7252 handle the dominance check inside a basic-block.
7253
7254 2013-08-29 Richard Biener <rguenther@suse.de>
7255
7256 PR middle-end/57287
7257 * tree-ssa-copy.c (may_propagate_copy): Allow propagating
7258 of default defs that appear in abnormal PHI nodes.
7259
7260 2013-08-29 Richard Biener <rguenther@suse.de>
7261
7262 PR tree-optimization/57685
7263 * tree-vrp.c (register_edge_assert_for_1): Recurse only for
7264 single-use operands to avoid exponential complexity.
7265
7266 2013-08-28 Dehao Chen <dehao@google.com>
7267
7268 * ipa-inline.c (edge_badness): Fix integer underflow.
7269
7270 2013-08-28 Uros Bizjak <ubizjak@gmail.com>
7271
7272 * gtm-builtins.def (_ITM_free): Declare leaf.
7273
7274 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7275
7276 PR target/58067
7277 * config/i386/i386.md (*tls_global_dynamic_64_largepic): New insn.
7278 (*tls_local_dynamic_base_64_largepic): Likewise.
7279 (tls_global_dynamic_64_<mode>, tls_local_dynamic_base_64_<mode>):
7280 Remove predicate from call operand.
7281 * config/i386/i386.c (ix86_tls_get_addr): For -mcmodel=large -fpic
7282 return sum of pic_offset_table_rtx and UNSPEC_PLTOFF of the symbol.
7283
7284 2013-08-28 Jeff Law <law@redhat.com>
7285
7286 * tree-ssa-threadedge.c (thread_around_empty_block): Remove
7287 checks for the number of predecessors and successors allowed.
7288 * tree-ssa-threadupdate.c (mark_threaded_blocks): Ignore requests
7289 which require copying a joiner block if there is a request which
7290 is a subpath that requires no joiner block copying.
7291
7292 2013-08-28 Jan Hubicka <jh@suse.cz>
7293
7294 * lto-streamer-out.c (DFS_write_tree_body): Drop
7295 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX and BINFO_VPTR_INDEX.
7296 (hash_tree): Do not hash DECL_DEFER_OUTPUT, BINFO_INHERITANCE_CHAIN,
7297 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX, DECL_IN_TEXT_SECTION.
7298 * tree-streamer-in.c (unpack_ts_decl_common_value_fields):
7299 Do not read DECL_ERROR_ISSUED.
7300 (unpack_ts_decl_with_vis_value_fields): Do not read
7301 DECL_DEFER_OUTPUT.
7302 (lto_input_ts_binfo_tree_pointers): Do not read
7303 BINFO_INHERITANCE_CHAIN, BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX
7304 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Do not
7305 write DECL_ERROR_ISSUED..
7306 (pack_ts_decl_with_vis_value_fields): Do not write
7307 DECL_DEFER_OUTPUT.
7308 (write_ts_binfo_tree_pointers): Do not read BINFO_INHERITANCE_CHAIN,
7309 BINFO_SUBVTT_INDEX, BINFO_VPTR_INDEX.
7310 * print-tree.c (print_node): Do not print DECL_ERROR_ISSUED.
7311 * tree.h (tree_decl_common): Update comment.
7312 (DECL_ERROR_ISSUED): Remove.
7313
7314 2013-08-28 Jakub Jelinek <jakub@redhat.com>
7315
7316 PR middle-end/58257
7317 * omp-low.c (copy_var_decl): Copy over TREE_NO_WARNING flag.
7318
7319 2013-08-28 Jan Hubicka <jh@suse.cz>
7320
7321 * builtins.def (free): Declare leaf.
7322
7323 2013-08-27 David Malcolm <dmalcolm@redhat.com>
7324
7325 * gdbhooks.py: New.
7326 * configure.ac (gdbinit.in): Add import of gcc/gdbhooks.py.
7327 * configure: Regenerate.
7328
7329 2013-08-27 Martin Jambor <mjambor@suse.cz>
7330
7331 * ipa-prop.h (ipa_pass_through_data): New field type_preserved.
7332 (ipa_ancestor_jf_data): Likewise.
7333 (ipa_get_jf_pass_through_agg_preserved): Fix comment typo.
7334 (ipa_get_jf_pass_through_type_preserved): New function.
7335 (ipa_get_jf_ancestor_agg_preserved): Fix comment typo.
7336 (ipa_get_jf_ancestor_type_preserved): New function.
7337 * ipa-cp.c (ipa_get_jf_pass_through_result): Honor type_preserved flag.
7338 (ipa_get_jf_ancestor_result): Likewise.
7339 (propagate_vals_accross_pass_through): Use
7340 ipa_get_jf_pass_through_result to do all the value mappings.
7341 * ipa-prop.c (ipa_print_node_jump_functions_for_edge): Dump the
7342 type_preserved flag.
7343 (ipa_set_jf_cst_copy): New function.
7344 (ipa_set_jf_simple_pass_through): Set the type_preserved flag.
7345 (ipa_set_jf_arith_pass_through): Likewise.
7346 (ipa_set_ancestor_jf): Likewise.
7347 (compute_complex_assign_jump_func): Set type_preserved instead of
7348 punting.
7349 (ipa_compute_jump_functions_for_edge): Likewise.
7350 (combine_known_type_and_ancestor_jfs): Honor type_preserved.
7351 (update_jump_functions_after_inlining): Update type_preserved.
7352 Explicitely create jump functions when combining one with pass_through.
7353 (ipa_write_jump_function): Stream the type_preserved flags.
7354 (ipa_read_jump_function): Likewise.
7355
7356 2013-08-27 Jakub Jelinek <jakub@redhat.com>
7357 Aldy Hernandez <aldyh@redhat.com>
7358
7359 * Makefile.in (omp-low.o): Depend on $(TARGET_H).
7360 * cfgloop.h (struct loop): Add safelen, force_vect, simduid.
7361 * function.h (struct function): Add has_force_vect_loops and
7362 has_simduid_loops.
7363 * gimple-pretty-print.c (dump_gimple_omp_for): Handle GF_OMP_FOR_KIND*.
7364 * gimple.c (gimple_build_omp_critical): Add KIND argument and
7365 handle it.
7366 * gimple.def: Update CLAUSES comments.
7367 * gimple.h (enum gf_mask): Add GF_OMP_FOR_KIND_{FOR,SIMD}.
7368 (gimple_build_omp_for): Add argument to prototype.
7369 (gimple_omp_for_kind): New.
7370 (gimple_omp_for_set_kind): New.
7371 * gimplify.c (enum gimplify_omp_var_data): Add GOVD_LINEAR to
7372 GOVD_DATA_SHARE_CLASS.
7373 (enum omp_region_type): Add ORT_SIMD.
7374 (gimple_add_tmp_var): Handle ORT_SIMD.
7375 (gimplify_var_or_parm_decl): Same.
7376 (is_gimple_stmt): Same.
7377 (omp_firstprivatize_variable): Same.
7378 (omp_add_variable): Only use splay_tree_insert if lookup failed.
7379 (omp_notice_variable): Handle ORT_SIMD.
7380 (omp_is_private): Add SIMD argument and handle it as well as ORT_SIMD.
7381 (omp_check_private): Handle ORT_SIMD.
7382 (gimplify_scan_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7383 OMP_CLAUSE_SAFELEN.
7384 (gimplify_adjust_omp_clauses_1): Handle GOVD_LINEAR.
7385 Handle OMP_CLAUSE_LASTPRIVATE.
7386 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_LINEAR and
7387 OMP_CLAUSE_SAFELEN.
7388 (gimplify_omp_for): Handle OMP_SIMD and OMP_CLAUSE_LINEAR.
7389 (gimplify_expr): Handle OMP_SIMD.
7390 * internal-fn.c (expand_GOMP_SIMD_LANE): New.
7391 (expand_GOMP_SIMD_VF): New.
7392 (expand_GOMP_SIMD_LAST_LANE): New.
7393 * internal-fn.def (GOMP_SIMD_LANE): New.
7394 (GOMP_SIMD_VF): New.
7395 (GOMP_SIMD_LAST_LANE): New.
7396 * omp-low.c: Include target.h.
7397 (extract_omp_for_data): Handle OMP_SIMD, OMP_CLAUSE_LINEAR,
7398 OMP_CLAUSE_SAFELEN.
7399 (check_omp_nesting_restrictions): Same.
7400 (omp_max_vf): New.
7401 (lower_rec_simd_input_clauses): New.
7402 (lower_rec_input_clauses): Handle OMP_SIMD, GF_OMP_FOR_KIND_SIMD,
7403 OMP_CLAUSE_LINEAR.
7404 (lower_lastprivate_clauses): Handle OMP_CLAUSE_LINEAR,
7405 GF_OMP_FOR_KIND_SIMD, OMP_SIMD.
7406 (expand_omp_build_assign): New.
7407 (expand_omp_for_init_counts): New.
7408 (expand_omp_for_init_vars): New.
7409 (extract_omp_for_update_vars): New.
7410 (expand_omp_for_generic): Use expand_omp_for_{init,update}_vars
7411 and rewrite accordingly.
7412 (expand_omp_simd): New.
7413 (expand_omp_for): Use expand_omp_simd.
7414 (lower_omp_for_lastprivate): Unshare vinit when appropriate.
7415 (lower_omp_for): Do not lower the body.
7416 * tree-data-ref (get_references_in_stmt): Allow IFN_GOMP_SIMD_LANE
7417 in their own loops.
7418 * tree-flow.h (find_omp_clause): Remove prototype.
7419 * tree-if-conv.c (main_tree_if_conversion): Run if doing if conversion,
7420 forcing vectorization of the loop, or if flag_tree_vectorize.
7421 (gate_tree_if_conversion): Similarly.
7422 * tree-inline.c (remap_gimple_stmt): Pass for kind argument to
7423 gimple_build_omp_for.
7424 (copy_cfg_body): set has_force_vect_loops and has_simduid_loops.
7425 * tree-parloops (create_parallel_loop): Pass kind argument to
7426 gimple_build_omp_for.
7427 * tree-pretty-print.c (dump_omp_clause): Add cases for
7428 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_LINEAR, OMP_CLAUSE_SAFELEN,
7429 OMP_CLAUSE__SIMDUID_.
7430 (dump_generic_node): Handle OMP_SIMD.
7431 * tree-ssa-ccp.c (likely_value): Handle IFN_GOMP_SIMD*.
7432 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely_1): Do not
7433 unroll OMP_SIMD loops here.
7434 * tree-ssa-loop.c (gate_tree_vectorize): Run if has_force_vect_loops.
7435 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Handle
7436 loop->safelen.
7437 (vect_analyze_data_refs): Handle simd loops.
7438 * tree-vect-loop.c (vectorizable_live_operation): Handle
7439 IFN_GOMP_SIMD*.
7440 * tree-vect-stmts.c (vectorizable_call): Handle IFN_GOMP_SIMD_LANE.
7441 (vectorizable_store): Handle STMT_VINFO_SIMD_LANE_ACCESS_P.
7442 (vectorizable_load): Same.
7443 * tree-vectorizer.c: Include hash-table.h and tree-ssa-propagate.h.
7444 (struct simduid_to_vf): New.
7445 (simduid_to_vf::hash): New.
7446 (simduid_to-vf::equal): New.
7447 (struct simd_array_to_simduid): New.
7448 (simd_array_to_simduid::hash): New.
7449 (simd_array_to_simduid::equal): New.
7450 (adjust_simduid_builtins): New.
7451 (struct note_simd_array_uses_struct): New.
7452 (note_simd_array_uses_cb): New.
7453 (note_simd_array_uses): New.
7454 (vectorize_loops): Handle simd hints and adjust simd builtins
7455 accordingly.
7456 * tree-vectorizer.h (struct _stmt_vec_info): Add
7457 simd_lane_access_p field.
7458 (STMT_VINFO_SIMD_LANE_ACCESS_P): New macro.
7459 * tree.c (omp_clause_num_ops): Add entries for OMP_CLAUSE_LINEAR,
7460 OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_UNIFORM.
7461 (omp_clause_code_name): Same.
7462 (walk_tree_1): Handle OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN,
7463 OMP_CLAUSE__SIMDUID_, OMP_CLAUSE_LINEAR.
7464 * tree.def (OMP_SIMD): New entry.
7465 * tree.h (enum omp_clause_code): Add entries for OMP_CLAUSE_LINEAR,
7466 OMP_CLAUSE_UNIFORM, OMP_CLAUSE_SAFELEN, OMP_CLAUSE__SIMDUID_.
7467 (OMP_CLAUSE_DECL): Adjust range for new clauses.
7468 (OMP_CLAUSE_LINEAR_NO_COPYIN): New.
7469 (OMP_CLAUSE_LINEAR_NO_COPYOUT): New.
7470 (OMP_CLAUSE_LINEAR_STEP): New.
7471 (OMP_CLAUSE_SAFELEN_EXPR): New.
7472 (OMP_CLAUSE__SIMDUID__DECL): New.
7473 (find_omp_clause): New prototype.
7474
7475 2013-08-27 H.J. Lu <hongjiu.lu@intel.com>
7476
7477 * config/i386/driver-i386.c (host_detect_local_cpu): Update
7478 Haswell processor detection.
7479
7480 2013-08-27 Christian Widmer <shadow@umbrox.de>
7481
7482 PR target/57927
7483 * config/i386/driver-i386.c (host_detect_local_cpu): Add detection
7484 of Ivy Bridge and Haswell processors. Assume core-avx2 for unknown
7485 AVX2 capable processors.
7486
7487 2013-08-27 Tejas Belagod <tejas.belagod@arm.com>
7488
7489 * config/aarch64/arm_neon.h: Replace all inline asm implementations
7490 of vget_low_* with implementations in terms of other intrinsics.
7491
7492 2013-08-27 Marc Glisse <marc.glisse@inria.fr>
7493
7494 PR middle-end/57219
7495 * doc/extend.texi (__builtin_isinf_sign): Restrict the return
7496 values to -1, 0 and 1.
7497
7498 2013-08-27 Vidya Praveen <vidyapraveen@arm.com>
7499
7500 * config/aarch64/aarch64.md (unspec): Add UNSPEC_SISD_SSHL,
7501 UNSPEC_SISD_USHL, UNSPEC_USHL_2S, UNSPEC_SSHL_2S, UNSPEC_SISD_NEG.
7502 (<optab><mode>3_insn): Remove.
7503 (aarch64_ashl_sisd_or_int_<mode>3): New Pattern.
7504 (aarch64_lshr_sisd_or_int_<mode>3): Likewise.
7505 (aarch64_ashr_sisd_or_int_<mode>3): Likewise.
7506 (define_split for aarch64_lshr_sisd_or_int_di3): Likewise.
7507 (define_split for aarch64_lshr_sisd_or_int_si3): Likewise.
7508 (define_split for aarch64_ashr_sisd_or_int_di3): Likewise.
7509 (define_split for aarch64_ashr_sisd_or_int_si3): Likewise.
7510 (aarch64_sisd_ushl, aarch64_sisd_sshl): Likewise.
7511 (aarch64_ushl_2s, aarch64_sshl_2s, aarch64_sisd_neg_qi): Likewise.
7512 (ror<mode>3_insn): Likewise.
7513 * config/aarch64/predicates.md (aarch64_simd_register): New.
7514
7515 2013-08-27 Richard Biener <rguenther@suse.de>
7516
7517 PR tree-optimization/57521
7518 * tree-if-conv.c (if_convertible_bb_p): Verify that at least
7519 one edge is non-critical.
7520 (find_phi_replacement_condition): Make sure to use a non-critical
7521 edge. Cleanup and remove old bug workarounds.
7522 (bb_postdominates_preds): Remove.
7523 (if_convertible_loop_p_1): Do not compute post-dominators.
7524 (combine_blocks): Do not free post-dominators.
7525 (main_tree_if_conversion): Likewise.
7526 (pass_data_if_conversion): Add TODO_verify_ssa.
7527
7528 2013-08-27 DJ Delorie <dj@redhat.com>
7529
7530 * config/i386/djgpp.h (ASM_DECLARE_FUNCTION_NAME): New.
7531
7532 2013-08-27 Yufeng Zhang <yufeng.zhang@arm.com>
7533
7534 * function.c (assign_parm_find_data_types): Set passed_mode and
7535 nominal_mode to the TYPE_MODE of nominal_type for the built
7536 pointer type in case of the struct-pass-by-reference.
7537
7538 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7539
7540 * config/avr/avr-stdint.h (INT16_TYPE): Change default to "int".
7541 (UINT16_TYPE): Change default to "unsigned int".
7542
7543 * config/avr/avr.opt (mfract-convert-truncate): New option.
7544 * config/avr/avr.c (avr_out_fract): Unless TARGET_FRACT_CONV_TRUNC
7545 is set, round negative fractional integers according to n1169
7546 when converting to integer types.
7547
7548 2013-08-26 Jan Hubicka <jh@suse.cz>
7549
7550 * cgraph.c (cgraph_propagate_frequency): Do not assume that virtual
7551 methods can not be called indirectly when their address is not taken.
7552
7553 2013-08-26 Jan Hubicka <jh@suse.cz>
7554
7555 * gimple-fold.c (gimple_get_virt_method_for_binfo): Use
7556 ctor_for_folding.
7557
7558 2013-08-26 Jan Hubicka <jh@suse.cz>
7559
7560 * ipa.c (comdat_can_be_unshared_p_1): C++ constructors and destructors
7561 can be unshared.
7562
7563 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7564
7565 * reload.c (find_valid_class): Allow classes that do not include
7566 FIRST_PSEUDO_REGISTER - 1.
7567
7568 2013-08-26 Jan Hubicka <jh@suse.cz>
7569
7570 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Fix formatting;
7571 fix edge count/frequency when speculation failed; fix type check
7572 for the direct call.
7573
7574 2013-08-26 Jan Hubicka <jh@suse.cz>
7575
7576 * ipa-prop.c (ipa_print_node_params): Do not ICE during WPA.
7577
7578 2013-08-26 Jan Hubicka <jh@suse.cz>
7579
7580 * ipa-inline-transform.c (inline_transform): Be ready for basic block
7581 to be changed by edge redirection.
7582
7583 2013-08-26 Jan Hubicka <jh@suse.cz>
7584
7585 * cgraph.c (cgraph_speculative_call_info): Fix parameter order and
7586 formating; add sanity check.
7587 (cgraph_resolve_speculation): Add FIXME about scaling profiles.
7588 (cgraph_redirect_edge_call_stmt_to_callee): Fix ICE in debug dump.
7589 * ipa-inline.c (heap_edge_removal_hook): Reset node growth cache.
7590 (resolve_noninline_speculation): Update callee keys, too.
7591
7592 2013-08-26 Jan Hubicka <jh@suse.cz>
7593
7594 * tree.h (tree_decl_with_vis): Add cxx_constructor, cxx_destructor.
7595 (DECL_CXX_CONSTRUCTOR_P, DECL_CXX_DESTRUCTOR_P): New macros.
7596
7597 2013-08-26 Joern Rennecke <joern.rennecke@embecosm.com>
7598
7599 * config/i386/i386.c (x86_64_elf_select_section): Put ATTRIBUTE_UNUSED
7600 into proper place.
7601
7602 2013-08-26 Uros Bizjak <ubizjak@gmail.com>
7603
7604 * config/i386/i386.c (ix86_debug_options): Remove prototype.
7605 (x86_64_elf_select_section): Ditto.
7606 (ix86_handle_tm_regparm_attribute): Remove ATTRIBUTE_UNUSED on used
7607 arguments.
7608 (ix86_pass_by_reference): Ditto.
7609 (output_set_got): Ditto.
7610 (ix86_unary_operator_ok): Ditto.
7611 (ix86_expand_builtin): Ditto.
7612
7613 2013-08-23 Jan Hubicka <jh@suse.cz>
7614
7615 * cgraph.c (cgraph_turn_edge_to_speculative): Fix debug output.
7616
7617 2013-08-23 Jan Hubicka <jh@suse.cz>
7618
7619 * tree.h (TYPE_FINAL_P, DECL_FINAL_P): New macros.
7620 (tree_decl_with_vis): Add FINAL field.
7621
7622 2013-08-23 Jeff Law <law@redhat.com>
7623
7624 * tree-ssa-pre.c (do_regular_insertion): Include the expression in
7625 the debugging dump when the expression is fully redundant.
7626
7627 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7628
7629 * diagnostic.c (diagnostic_set_caret_max_width): Use pp_buffer.
7630 * gimple-pretty-print.c (gimple_dump_bb_buff): Likewise.
7631 * pretty-print.c (pp_formatted_text_data): Likewise.
7632 (pp_write_text_to_stream): Likewise.
7633 (pp_write_text_as_dot_label_to_stream): Likewise.
7634 (pp_append_r): Likewise.
7635 (pp_format): Likewise.
7636 (pp_flush): Likewise.
7637 (pp_clear_output_area): Likewise.
7638 (pp_append_text): Likewise.
7639 (pp_formatted_text): Likewise.
7640 (pp_remaining_character_count_for_line): Likewise.
7641 (pp_newline): Likewise.
7642 (pp_character): Likewise.
7643 (output_buffer::~output_buffer): Define.
7644 (pretty_printer::~pretty_printer): Destruct output buffer.
7645 * pretty-print.h (output_buffer::~output_buffer): Declare.
7646 (pretty_printer::~pretty_printer): Declare virtual.
7647
7648 2013-08-24 Marc Glisse <marc.glisse@inria.fr>
7649
7650 PR other/57324
7651 * hwint.h (HOST_WIDE_INT_UC, HOST_WIDE_INT_1U, HOST_WIDE_INT_M1,
7652 HOST_WIDE_INT_M1U): New macros.
7653 * fold-const.c (sign_bit_p, build_range_check, fold_unary_loc,
7654 fold_binary_loc, fold_ternary_loc): Use the new macros. Use an
7655 unsigned -1 for lshift.
7656 * cse.c (cse_insn): Likewise.
7657 * double-int.c (rshift_double, lshift_double): Likewise.
7658 * builtins.c (fold_builtin_bitop): Likewise.
7659 * combine.c (force_to_mode): Likewise.
7660 * tree.c (integer_pow2p, tree_log2, tree_floor_log2): Likewise.
7661 * simplify-rtx.c (simplify_const_unary_operation,
7662 simplify_const_binary_operation): Likewise.
7663 * tree-stdarg.c (va_list_counter_bump, va_list_ptr_read,
7664 check_va_list_escapes): Likewise.
7665 * rtlanal.c (nonzero_bits1): Likewise.
7666 * expmed.c (expand_smod_pow2): Likewise.
7667 * tree-ssa-structalias.c (UNKNOWN_OFFSET): Use HOST_WIDE_INT_MIN.
7668
7669 2013-08-23 Jan Hubicka <jh@suse.cz>
7670
7671 * cgraph.c (cgraph_turn_edge_to_speculative): Mark target node
7672 as having address taken.
7673
7674 2013-08-23 Jan Hubicka <jh@suse.cz>
7675
7676 * ipa-utils.h (method_class_type): Declare.
7677 * ipa-devirt.c (method_class_type): Export.
7678
7679 * cgraphunit.c (analyze_functions): Do basic devirtualization;
7680 do not walk base classes of anonymous types.
7681
7682 2013-08-23 Kaz Kojima <kkojima@gcc.gnu.org>
7683
7684 PR rtl-optimization/58220
7685 PR regression/58221
7686 * final.c (reemit_insn_block_notes): Use NEXT_INSN to
7687 handle SEQUENCE insns properly.
7688
7689 2013-08-23 Gabriel Dos Reis <gdr@integrable-solutions.net>
7690
7691 * pretty-print.h (pp_newline_and_flush): Declare. Remove macro
7692 definition.
7693 (pp_newline_and_indent): Likewise.
7694 (pp_separate_with): Likewise.
7695 * pretty-print.c (pp_newline_and_flush): Define.
7696 (pp_newline_and_indent): Likewise.
7697 (pp_separate_with): Likewise.
7698
7699 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7700
7701 PR target/58218
7702 * config/i386/x86-64.h (TARGET_SECTION_TYPE_FLAGS): Define.
7703 * config/i386/i386.c (x86_64_elf_section_type_flags): New function.
7704
7705 2013-08-23 Kirill Yukhin <kirill.yukhin@intel.com>
7706
7707 * config/i386/predicates.md (ext_sse_reg_operand): New.
7708 * config/i386/i386.md (*movti_internal): Use
7709 predicate to determine if EVEX is needed.
7710 (*movsi_internal): Ditto.
7711 (*movdf_internal): Ditto.
7712 (*movsf_internal): Ditto.
7713 * config/i386/mmx.md (*mov<mode>_internal): Ditto.
7714
7715 2013-08-23 Jakub Jelinek <jakub@redhat.com>
7716
7717 PR tree-optimization/58209
7718 * tree-tailcall.c (process_assignment): Handle POINTER_PLUS_EXPR.
7719 (find_tail_calls): Give up for pointer result types if m is non-NULL.
7720 (adjust_return_value_with_ops): For PLUS_EXPR and pointer result type
7721 emit POINTER_PLUS_EXPR.
7722 (create_tailcall_accumulator): For pointer result type accumulate in
7723 sizetype type.
7724
7725 2013-08-22 Paolo Carlini <paolo.carlini@oracle.com>
7726
7727 * configure.ac: Add backslashes missing from the last change.
7728 * configure: Regenerate.
7729
7730 2013-08-22 Jan Hubicka <jh@susue.cz>
7731
7732 * ipa.c (function_and_variable_visibility): First remember function
7733 was global and then make it local.
7734
7735 2013-08-22 Julian Brown <julian@codesourcery.com>
7736
7737 * configure.ac: Add aarch64 to list of arches which use "nop" in
7738 debug_line test.
7739 * configure: Regenerate.
7740
7741 2013-08-22 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
7742
7743 * config/s390/linux.h (TARGET_LIBC_HAS_FUNCTION): Define as
7744 gnu_libc_has_function.
7745 * config/s390/tpf.h: Likewise.
7746
7747 2013-08-22 Jan Hubicka <jh@susue.cz>
7748
7749 * timevar.c (validate_phases): Add cast.
7750
7751 2013-08-22 Jan Hubicka <jh@susue.cz>
7752
7753 * timevar.c (validate_phases): Use size_t for memory.
7754 * timevar.h (struct timevar_time_def): Use size_t for ggc_mem.
7755
7756 2013-08-22 Gabriel Dos Reis <gdr@integrable-solutions.net>
7757
7758 * pretty-print.h (output_buffer::output_buffer): Declare.
7759 (pretty_printer::pretty_printer): Likewise.
7760 (pp_construct): Remove.
7761 * pretty-print.c (output_buffer::output_buffer): Define.
7762 (pretty_printer::pretty_printer): Rename from pp_construct. Simplify.
7763 * gimple-pretty-print.c (print_gimple_stmt): Do not call pp_construct.
7764 (print_gimple_expr): Likewise.
7765 (print_gimple_seq): Likewise.
7766 (gimple_dump_bb): Likewise.
7767 * sched-vis.c (dump_value_slim): Likewise.
7768 (dump_insn_slim): Likewise.
7769 (dump_rtl_slim): Likewise.
7770 (str_pattern_slim): Likewise.
7771 * tree-mudflap.c (mf_varname_tree): Likewise.
7772 * graph.c (print_graph_cfg): Likewise.
7773 (start_graph_dump): Likewise.
7774 * tree-pretty-print.c (maybe_init_pretty_print): Likewise. Use
7775 placement-new.
7776 * diagnostic.c (diagnostic_initialize): Simplify early diagnostic
7777 pretty printer initialization.
7778 * coretypes.h (diagnostic_context): Remove superflous type alias
7779 declaration.
7780 (pretty_printer): Likewise. Declare directly as a class.
7781 (pretty_print_info): Remove declaration as class.
7782 * asan.c (asan_emit_stack_protection): Remove call to pp_construct
7783 and pp_clear_output_area.
7784 (asan_add_global): Likewise.
7785
7786 2013-08-22 Jan Hubicka <jh@suse.cz>
7787
7788 * cgraphunit.c (analyze_functions) Use update_type_inheritance_graph.
7789 * ipa-utils.h (update_type_inheritance_graph): Declare.
7790 (possible_polymorphic_call_target_p): Declare.
7791 (possible_polymorphic_call_target_p): New.
7792 * ipa-devirt.c: Update toplevel comments.
7793 (cached_polymorphic_call_targets): Move up.
7794 (odr_type_d): Move ID down.
7795 (polymorphic_type_binfo_p): Update comment.
7796 (odr_hasher::remove): Likewise;
7797 (get_odr_type): Set anonymous_namespace.
7798 (dump_odr_type): Dump it.
7799 (dump_type_inheritance_graph): Do not ICE when there are no ODR types.
7800 (maybe_record_node): Record node in cached_polymorphic_call_targets.
7801 (record_binfo): Add comment.
7802 (free_polymorphic_call_targets_hash): Do not ICE when cache is not
7803 built.
7804 (devirt_node_removal_hook): Do not iCE when cache is freed.
7805 (possible_polymorphic_call_target_p): New predicate.
7806 (update_type_inheritance_graph): New function.
7807
7808 2013-08-22 Alexander Ivchenko <alexander.ivchenko@intel.com>
7809 Maxim Kuznetsov <maxim.kuznetsov@intel.com>
7810 Sergey Lega <sergey.s.lega@intel.com>
7811 Anna Tikhonova <anna.tikhonova@intel.com>
7812 Ilya Tocar <ilya.tocar@intel.com>
7813 Andrey Turetskiy <andrey.turetskiy@intel.com>
7814 Ilya Verbin <ilya.verbin@intel.com>
7815 Kirill Yukhin <kirill.yukhin@intel.com>
7816 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
7817
7818 * common/config/i386/i386-common.c (OPTION_MASK_ISA_AVX512F_SET): New.
7819 (OPTION_MASK_ISA_AVX512CD_SET): Ditto.
7820 (OPTION_MASK_ISA_AVX512PF_SET): Ditto.
7821 (OPTION_MASK_ISA_AVX512ER_SET): Ditto.
7822 (OPTION_MASK_ISA_AVX2_UNSET): Update.
7823 (OPTION_MASK_ISA_AVX512F_UNSET): New.
7824 (OPTION_MASK_ISA_AVX512CD_UNSET): Ditto.
7825 (OPTION_MASK_ISA_AVX512PF_UNSET): Ditto.
7826 (OPTION_MASK_ISA_AVX512ER_UNSET): Ditto.
7827 (ix86_handle_option): Handle OPT_mavx512f, OPT_mavx512cd,
7828 OPT_mavx512pf, OPT_mavx512er cases.
7829 * config/i386/constraints.md (v): New constraint.
7830 (Yi, Yj): Replace SSE_REGS with ALL_SSE_REGS.
7831 * config/i386/cpuid.h (bit_AVX512F, bit_AVX512PF, bit_AVX512ER)
7832 (bit_AVX512CD): New.
7833 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
7834 AVX512F, AVX512ER, AVX512PF, AVX512CD features.
7835 * config/i386/i386-c.c (ix86_target_macros_internal):
7836 Conditionally define __AVX512F__, __AVX512ER__, __AVX512CD__,
7837 __AVX512PF__.
7838 * config/i386/i386-modes.def (VECTOR_MODES (INT, 128))
7839 (VECTOR_MODES (FLOAT, 128), INT_MODE (XI, 64)): New modes.
7840 * config/i386/i386.c (regclass_map, dbx_register_map)
7841 (dbx64_register_map, svr4_dbx_register_map): Add new SSE registers.
7842 (gate_insert_vzeroupper): Disable vzeroupper for TARGET_AVX512F.
7843 (ix86_target_string): Define -mavx512f, -mavx512er, -mavx512cd,
7844 -mavx512pf options.
7845 (ix86_option_override_internal): Define PTA_AVX512F, PTA_AVX512ER,
7846 PTA_AVX512PF, PTA_AVX512CD. Handle -mavx512f, -mavx512er, -mavx512cd,
7847 -mavx512pf options. Fix formatting.
7848 (ix86_conditional_register_usage): Squash EXT_REX_SSE_REGs for 32-bit
7849 targets. Squash EVEX_SSE_REGS if AVX512F is disabled.
7850 (ix86_valid_target_attribute_inner_p): Handle -mavx512f, -mavx512er,
7851 -mavx512cd, -mavx512pf options.
7852 (standard_sse_constant_opcode): Add vpternlogd for 512-bit modes.
7853 (print_reg, ix86_print_operand): Handle 'g' to output 512-bit operands.
7854 (ix86_preferred_output_reload_class): Replace SSE_REGS with
7855 ALL_SSE_REGS.
7856 (ix86_hard_regno_mode_ok): Support 512-bit registers.
7857 (ix86_set_reg_reg_cost): Ditto.
7858 (x86_order_regs_for_local_alloc): Ditto.
7859 (MAX_VECT_LEN): Extend to 64-byte.
7860 (ix86_spill_class): Replace SSE_REGS with ALL_SSE_REGS.
7861 * config/i386/i386.h (TARGET_AVX512F, TARGET_AVX512PF)
7862 (TARGET_AVX512ER, TARGET_AVX512CD): New.
7863 (BIGGEST_ALIGNMENT): Extend to 512-bits.
7864 (FIRST_PSEUDO_REGISTER, FIXED_REGISTERS): Add new registers.
7865 (CALL_USED_REGISTERS, REG_ALLOC_ORDER): Likewise.
7866 (VALID_AVX512F_SCALAR_MODE, VALID_AVX512F_REG_MODE): New.
7867 (SSE_REG_MODE_P): Support new modes.
7868 (FIRST_MMX_REG, FIRST_REX_INT_REG, FIRST_REX_SSE_REG): Add comments.
7869 (FIRST_EXT_REX_SSE_REG, LAST_EXT_REX_SSE_REG): New.
7870 (reg_class, REG_CLASS_NAMES): Add EVEX_SSE_REGS, ALL_SSE_REGS.
7871 (SSE_CLASS_P, MAYBE_SSE_CLASS_P): Replace SSE_REGS with ALL_SSE_REGS.
7872 (REG_CLASS_CONTENTS): Add new registers.
7873 (SSE_REGNO_P, SSE_REGNO, HARD_REGNO_RENAME_OK): Support new registers.
7874 (EXT_REX_SSE_REGNO_P): New.
7875 (HI_REGISTER_NAMES): Add new registers.
7876 * config/i386/i386.md: Define constants for new registers.
7877 (mode): Add new 512-bit modes.
7878 (prefix): Support evex prefix.
7879 (isa): Support avx512f, noavx512f, fma_avx512f.
7880 (ssemodesuffix): Add new 512-bit modes.
7881 (movxi): New.
7882 (*movxi_internal_avx512f): Ditto.
7883 (*movdi_internal): Replace constraint "x" with the new constraint "v".
7884 Support MODE_XI.
7885 (*movsi_internal): Likewise.
7886 (*movdf_internal): Likewise.
7887 (*movsf_internal): Likewise.
7888 (*fop_<mode>_comm_sse): Replace constraint "x" with new constraint "v".
7889 (<code><mode>3): Likewise.
7890 * config/i386/i386.opt (mavx512f, mavx512pf, mavx512er, mavx512cd):
7891 New.
7892 * config/i386/mmx.md (*mov<mode>_internal): Replace constraint "x"
7893 with the new constraint "v".
7894 * config/i386/sse.md (*mov<mode>_internal): Support new registers and
7895 modes.
7896 (<sse>_loadu<ssemodesuffix><avxsizesuffix>): Replace constraint "x"
7897 with the new constraint "v".
7898 (<sse2>_loaddqu<avxsizesuffix>): Likewise.
7899 (<sse2>_storedqu<avxsizesuffix>): Likewise.
7900 (*<plusminus_insn><mode>3): Likewise.
7901 (<sse>_vm<plusminus_insn><mode>3): Likewise.
7902 (*mul<mode>3): Likewise.
7903 (<sse>_vmmul<mode>3): Likewise.
7904 (<sse>_div<mode>3): Likewise.
7905 (<sse>_vmdiv<mode>3): Likewise.
7906 (<sse>_sqrt<mode>2): Likewise.
7907 (<sse>_vmsqrt<mode>2): Likewise.
7908 (*<code><mode>3_finite): Likewise.
7909 (*<code><mode>3) <smaxmin>: Likewise.
7910 (<sse>_vm<code><mode>3): Likewise.
7911 (*<code><mode>3) <any_logic>: Likewise.
7912 (*fma_fmadd_<mode>): Likewise.
7913 (*fma_fmsub_<mode>): Likewise.
7914 (*fma_fnmadd_<mode>): Likewise.
7915 (*fma_fnmsub_<mode>): Likewise.
7916 (*fma_fmaddsub_<mode>): Likewise.
7917 (*fma_fmsubadd_<mode>): Likewise.
7918 (*fmai_fmadd_<mode>): Likewise.
7919 (*fmai_fmsub_<mode>): Likewise.
7920 (*fmai_fnmadd_<mode>): Likewise.
7921 (*fmai_fnmsub_<mode>): Likewise.
7922 (sse_cvtsi2ss): Likewise.
7923 (sse_cvtsi2ssq): Likewise.
7924 (sse_cvtss2si): Likewise.
7925 (sse_cvtss2si_2): Likewise.
7926 (sse_cvtss2siq): Likewise.
7927 (sse_cvtss2siq_2): Likewise.
7928 (sse_cvttss2si): Likewise.
7929 (sse_cvtss2siq_2): Likewise.
7930 (float<sseintvecmodelower><mode>2): Likewise.
7931 (sse2_cvtsd2si_2): Likewise.
7932 (sse2_cvtsd2siq_2): Likewise.
7933 (*<plusminus_insn><mode>3): Likewise.
7934 (*<sse2_avx2>_<plusminus_insn><mode>3): Likewise.
7935 (*<sse4_1_avx2>_mul<mode>3): Likewise.
7936 (ashr<mode>3): Likewise.
7937 (<shift_insn><mode>3): Likewise.
7938 (avx2_<code><mode>3): Likewise.
7939 (*avx2_<code><mode>3): Likewise.
7940 (*andnot<mode>3): Likewise.
7941 (*<code><mode>3) <any_logic>: Likewise.
7942 (abs<mode>2): Likewise.
7943 (avx2_permvar<mode>): Likewise.
7944 (avx2_perm<mode>_1): Likewise.
7945 (*avx_vpermilp<mode>): Likewise.
7946 (avx_vpermilvar<mode>3): Likewise.
7947 (avx2_ashrv<mode>): Likewise.
7948 (avx2_<shift_insn>v<mode>): Likewise.
7949 * doc/invoke.texi: Document -mavx512f, -mavx512pf, -mavx512er,
7950 -mavx512cd.
7951 * doc/rtl.texi: Document XImode.
7952
7953 2013-08-21 Jeff Law <law@redhat.com>
7954
7955 * tree-flow.h (register_jump_thread): Pass vector of edges
7956 instead of each important edge.
7957 * tree-ssa-threadedge.c (thread_across_edge): Build the jump
7958 thread path into a vector and pass that to register_jump_thread.
7959 * tree-ssa-threadupdate.c (register_jump_thread): Conver the
7960 passed in edge vector to the current 3-edge form.
7961
7962 Revert:
7963 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
7964
7965 * dce.c (fini_dce): Call df_analyze again just in case
7966 delete_unmarked_insns removed anything.
7967
7968 2013-08-21 Joern Rennecke <joern.rennecke@embecosm.com>
7969
7970 * reload.h (struct reg_equivs): Rename to ..
7971 (struct reg_equivs_s): .. this.
7972
7973 2013-08-20 Martin Liska <marxin.liska@gmail.com>
7974
7975 * ipa.c (ipa_profile_read_summary): Fix buffer overflow.
7976
7977 2013-08-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
7978
7979 * config/sol2-10.h (TARGET_LIBC_HAS_FUNCTION): Don't nest comment.
7980
7981 2013-08-21 Jeff Law <law@redhat.com>
7982
7983 * tree-vrp.c (simplify_stmt_for_jump_threading): Try to
7984 simplify assignments too. If the RHS collapses to a singleton
7985 range, then return the value for the range.
7986
7987 2013-08-21 Kirill Yukhin <kirill.yukhin@intel.com>
7988
7989 * config/i386/sse.md (V16): Rename to...
7990 (VMOVE): this.
7991 (mov<mode>): Update iterator name.
7992 (*mov<mode>_internal): Ditto.
7993 (push<mode>1): Ditto.
7994 (movmisalign<mode>): Ditto.
7995
7996 2013-08-20 Jan Hubicka <jh@suse.cz>
7997
7998 PR bootstrap/58186
7999 * cgraph.c (cgraph_add_edge_to_call_site_hash): Overwrite hash
8000 entry for direct edges.
8001 (cgraph_turn_edge_to_speculative): Fix setting of can_throw_external.
8002
8003 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8004
8005 Revert my last two changes, r201865 and r201864:
8006
8007 Revert r201865:
8008 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8009
8010 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8011 instances can own GC refs.
8012
8013 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8014 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8015 (gcc::context::gt_pch_nx): Likewise.
8016 (gcc::context::gt_pch_nx): Likewise.
8017 * ggc.h (gt_ggc_mx <T>): New.
8018 (gt_pch_nx_with_op <T>): New.
8019 (gt_pch_nx <T>): New.
8020 * passes.c (opt_pass::gt_ggc_mx): New.
8021 (opt_pass::gt_pch_nx): New.
8022 (opt_pass::gt_pch_nx_with_op): New.
8023 (pass_manager::gt_ggc_mx): New.
8024 (pass_manager::gt_pch_nx): New.
8025 (pass_manager::gt_pch_nx_with_op): New.
8026 (pass_manager::operator new): Use
8027 ggc_internal_cleared_alloc_stat rather than xcalloc.
8028 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8029 (pass_manager::gt_ggc_mx): New.
8030 (pass_manager::gt_pch_nx): New.
8031 (pass_manager::gt_pch_nx_with_op): New.
8032 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8033 (opt_pass::operator new): New.
8034 (opt_pass::gt_ggc_mx): New.
8035 (opt_pass::gt_pch_nx): New.
8036 (opt_pass::gt_pch_nx_with_op): New.
8037
8038 Revert r201864:
8039 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8040
8041 * Makefile.in (GTFILES): Add context.h.
8042 * context.c (gcc::context::operator new): New.
8043 (gcc::context::gt_ggc_mx): New.
8044 (gcc::context::gt_pch_nx): New.
8045 (gcc::context::gt_pch_nx): New.
8046 * context.h (gcc::context): Add GTY((user)) marking.
8047 (gcc::context::operator new): New.
8048 (gcc::context::gt_ggc_mx): New.
8049 (gcc::context::gt_pch_nx): New.
8050 (gcc::context::gt_pch_nx): New.
8051 (g): Add GTY marking.
8052 (gt_ggc_mx (gcc::context *)): New.
8053 (gt_pch_nx (gcc::context *)): New.
8054 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8055 void *cookie)): New.
8056 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8057
8058 2013-08-20 Alexey Makhalov <makhaloff@gmail.com>
8059
8060 * dce.c (fini_dce): Call df_analyze again just in case
8061 delete_unmarked_insns removed anything.
8062
8063 2013-08-20 Teresa Johnson <tejohnson@google.com>
8064
8065 PR rtl-optimizations/57451
8066 * final.c (reemit_insn_block_notes): Prevent lexical blocks
8067 from crossing split section boundaries.
8068
8069 2013-08-20 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
8070
8071 * config/arm/linux-elf.h (MULTILIB_DEFAULTS): Remove definition.
8072 * config/arm/t-linux-eabi (MULTILIB_OPTIONS): Document association
8073 with MULTLIB_DEFAULTS.
8074
8075 2013-08-20 Nick Clifton <nickc@redhat.com>
8076
8077 * target.def (narrow_volatile_bitfield): Note that the default
8078 value is false, not !TARGET_STRICT_ALIGN.
8079 * doc/tm.texi: Regenerate.
8080
8081 2013-08-20 Pavel Chupin <pavel.v.chupin@intel.com>
8082
8083 Fix LIB_SPEC for systems without libpthread.
8084
8085 * config/gnu-user.h: Introduce GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC.
8086 * config/arm/linux-eabi.h: Use GNU_USER_TARGET_NO_PTHREADS_LIB_SPEC
8087 for Android.
8088 * config/i386/linux-common.h: Likewise.
8089 * config/mips/linux-common.h: Likewise.
8090
8091 2013-08-20 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
8092
8093 * tree-ssa-ccp.c (get_default_value): Remove redundant condition
8094 checks.
8095
8096 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8097
8098 Make opt_pass and gcc::pass_manager be GC-managed, so that pass
8099 instances can own GC refs.
8100
8101 * Makefile.in (GTFILES): Add pass_manager.h and tree-pass.h.
8102 * context.c (gcc::context::gt_ggc_mx): Traverse passes_.
8103 (gcc::context::gt_pch_nx): Likewise.
8104 (gcc::context::gt_pch_nx): Likewise.
8105 * ggc.h (gt_ggc_mx <T>): New.
8106 (gt_pch_nx_with_op <T>): New.
8107 (gt_pch_nx <T>): New.
8108 * passes.c (opt_pass::gt_ggc_mx): New.
8109 (opt_pass::gt_pch_nx): New.
8110 (opt_pass::gt_pch_nx_with_op): New.
8111 (pass_manager::gt_ggc_mx): New.
8112 (pass_manager::gt_pch_nx): New.
8113 (pass_manager::gt_pch_nx_with_op): New.
8114 (pass_manager::operator new): Use
8115 ggc_internal_cleared_alloc_stat rather than xcalloc.
8116 * pass_manager.h (class pass_manager): Add GTY((user)) marking.
8117 (pass_manager::gt_ggc_mx): New.
8118 (pass_manager::gt_pch_nx): New.
8119 (pass_manager::gt_pch_nx_with_op): New.
8120 * tree-pass.h (class opt_pass): Add GTY((user)) marking.
8121 (opt_pass::operator new): New.
8122 (opt_pass::gt_ggc_mx): New.
8123 (opt_pass::gt_pch_nx): New.
8124 (opt_pass::gt_pch_nx_with_op): New.
8125
8126 2013-08-20 David Malcolm <dmalcolm@redhat.com>
8127
8128 * Makefile.in (GTFILES): Add context.h.
8129 * context.c (gcc::context::operator new): New.
8130 (gcc::context::gt_ggc_mx): New.
8131 (gcc::context::gt_pch_nx): New.
8132 (gcc::context::gt_pch_nx): New.
8133 * context.h (gcc::context): Add GTY((user)) marking.
8134 (gcc::context::operator new): New.
8135 (gcc::context::gt_ggc_mx): New.
8136 (gcc::context::gt_pch_nx): New.
8137 (gcc::context::gt_pch_nx): New.
8138 (g): Add GTY marking.
8139 (gt_ggc_mx (gcc::context *)): New.
8140 (gt_pch_nx (gcc::context *)): New.
8141 (gt_pch_nx (gcc::context *ctxt, gt_pointer_operator op,
8142 void *cookie)): New.
8143 * gengtype.c (open_base_files) <ifiles>: Add context.h.
8144
8145 2013-08-20 Alan Modra <amodra@gmail.com>
8146
8147 PR target/57865
8148 * config/rs6000/rs6000.c (rs6000_emit_prologue): Correct ool_adjust.
8149 (rs6000_emit_epilogue): Likewise.
8150
8151 2013-08-19 Dehao Chen <dehao@google.com>
8152
8153 * value-prof.c (gimple_ic): Fix the bug of adding EH edge.
8154
8155 2013-08-19 Peter Bergner <bergner@vnet.ibm.com>
8156 Jakub Jelinek <jakub@redhat.com>
8157
8158 * builtins.def (BUILT_IN_FABSD32): New DFP ABS builtin.
8159 (BUILT_IN_FABSD64): Likewise.
8160 (BUILT_IN_FABSD128): Likewise.
8161 * builtins.c (expand_builtin): Add support for new DFP ABS builtins.
8162 (fold_builtin_1): Likewise.
8163 * config/rs6000/dfp.md (*negtd2_fpr): Handle non-overlapping
8164 destination and source operands.
8165 (*abstd2_fpr): Likewise.
8166 (*nabstd2_fpr): Likewise.
8167
8168 2013-08-19 Richard Sandiford <rdsandiford@googlemail.com>
8169
8170 * config/mips/mips.c (mips_adjust_insn_length): Add checks for
8171 JUMP_P and INSN_P.
8172
8173 2013-08-19 Aldy Hernandez <aldyh@redhat.com>
8174
8175 * doc/invoke.texi (-fcilkplus): Clarify that implementation is
8176 incomplete.
8177
8178 2013-08-19 Alexander Ivchenko <alexander.ivchenko@intel.com>
8179
8180 * target.def (TARGET_LIBC_HAS_FUNCTION): New target hook.
8181 * builtins.c (default_libc_has_function): New.
8182 (gnu_libc_has_function): Ditto.
8183 (no_c99_libc_has_function): Ditto.
8184 (expand_builtin_cexpi): Using new target hook TARGET_LIBC_HAS_FUNCTION
8185 instead of TARGET_HAS_SINCOS and TARGET_C99_FUNCTIONS.
8186 (fold_builtin_sincos): Likewise.
8187 (fold_builtin_cexp): Likewise.
8188 * builtins.def (DEF_C94_BUILTIN): Likewise.
8189 (DEF_C99_BUILTIN): Likewise.
8190 (DEF_C99_C90RES_BUILTIN): Likewise.
8191 (DEF_C99_COMPL_BUILTIN): New define. Change all complex c99 builtin
8192 definitions to using this define.
8193 * config/darwin-protos.h (darwin_libc_has_function): New.
8194 * config/darwin.c (darwin_libc_has_function): Ditto.
8195 * config/alpha/linux.h: Remove TARGET_C99_FUNCTIONS and
8196 TARGET_HAS_SINCOS. Redefine TARGET_LIBC_HAS_FUNCTION.
8197 * config/darwin.h: Ditto.
8198 * config/elfos.h: Ditto.
8199 * config/freebsd.h: Ditto.
8200 * config/i386/cygming.h: Ditto.
8201 * config/i386/djgpp.h: Ditto.
8202 * config/i386/i386-interix.h: Ditto.
8203 * config/microblaze/microblaze.h: Ditto.
8204 * config/mmix/mmix.h: Ditto.
8205 * config/gnu-user.h: Ditto.
8206 * config/ia64/hpux.h: Ditto.
8207 * config/pa/pa-hpux.h: Ditto.
8208 * config/pdp11/pdp11.h: Ditto.
8209 * config/picochip/picochip.h: Ditto.
8210 * config/linux.h: Ditto.
8211 * config/netbsd.h: Ditto.
8212 * config/openbsd.h: Ditto.
8213 * config/rs6000/aix43.h: Ditto.
8214 * config/rs6000/aix51.h: Ditto.
8215 * config/rs6000/aix52.h: Ditto.
8216 * config/rs6000/aix53.h: Ditto.
8217 * config/rs6000/aix61.h: Ditto.
8218 * config/rs6000/darwin.h: Ditto.
8219 * config/rs6000/linux.h: Ditto.
8220 * config/rs6000/linux64.h: Ditto.
8221 * config/s390/tpf.h: Ditto.
8222 * config/sol2-10.h: Ditto.
8223 * config/sol2.h: Ditto.
8224 * config/vms/vms.h: Ditto.
8225 * config/vxworks.h: Ditto.
8226 * config/linux-android.c (linux_android_libc_has_function):
8227 New linux-specific implementation of TARGET_LIBC_HAS_FUNCTION.
8228 * config/linux-protos.h (linux_android_libc_has_function):
8229 New declaration.
8230 * config/i386/i386.c (ix86_libc_has_function): New.
8231 * config/i386/i386-protos.h
8232 (ix86_libc_has_function): New declaration.
8233 * config/i386/i386.md
8234 ("isinfxf2"): Change condition for TARGET_LIBC_HAS_FUNCTION.
8235 ("isinf<mode>2): Likewise.
8236 * convert.c (convert_to_integer): Using new target hook
8237 TARGET_LIBC_HAS_FUNCTION istead of TARGET_HAS_SINCOS and
8238 TARGET_C99_FUNCTIONS.
8239 * fortran/f95-lang.c (gfc_init_builtin_functions): Ditto.
8240 * tree-ssa-math-opts.c (execute_cse_sincos): Ditto.
8241 * coretypes.h (function_class): New enum for different
8242 classes of functions.
8243 * defaults.h: Remove TARGET_C99_FUNCTIONS and TARGET_HAS_SINCOS.
8244 * doc/tm.texi.in (TARGET_C99_FUNCTIONS): Remove documentation.
8245 (TARGET_HAS_SINCOS): Likewise.
8246 (TARGET_LIBC_HAS_FUNCTION): New.
8247 * doc/tm.texi: Regenerated.
8248 * targhooks.h (default_libc_has_function): New declaration.
8249 (no_c99_libc_has_function): Ditto.
8250 (gnu_libc_has_function): Ditto.
8251 * system.h: Add the poisoning of TARGET_C99_FUNCTIONS
8252 and TARGET_HAS_SINCOS.
8253
8254 2013-08-18 Jan Hubicka <jh@suse.cz>
8255
8256 * Makeifle-in (ipa-devirt.o): New.
8257 (GTFILES): Add ipa-utils.h and ipa-devirt.c
8258 * cgraphunit.c (decide_is_symbol_needed): Do not care about virtuals.
8259 (analyze_functions): Look into possible targets of polymorphic call.
8260 * dumpfile.c (dump_files): Add type-inheritance dump.
8261 * dumpfile.h (TDI_inheritance): New.
8262 * ipa-devirt.c: New file.
8263 * ipa-utils.h (odr_type_d): Forward declare.
8264 (odr_type): New type.
8265 (build_type_inheritance_graph): Declare.
8266 (possible_polymorphic_call_targets): Declare and introduce inline
8267 variant when only edge is pased.
8268 (dump_possible_polymorphic_call_targets): Likewise.
8269 * timevar.def (TV_IPA_INHERITANCE, TV_IPA_VIRTUAL_CALL): New.
8270 * tree.c (type_in_anonymous_namespace_p): Break out from ...
8271 (types_same_for_odr): ... here.
8272 * tree.h (type_in_anonymous_namespace_p): Declare.
8273
8274 2013-08-18 Jakub Jelinek <jakub@redhat.com>
8275
8276 PR tree-optimization/58006
8277 * tree-parloops.c (take_address_of): Don't ICE if get_name
8278 returns NULL.
8279 (eliminate_local_variables_stmt): Remove clobber stmts.
8280
8281 2013-08-18 Eric Botcazou <ebotcazou@adacore.com>
8282
8283 * cgraphunit.c (handle_alias_pairs): Reset the alias flag after the
8284 error message is issued for an alias to undefined symbol.
8285
8286 2013-08-18 Jan Hubicka <jh@suse.cz>
8287
8288 * cgraph.c (cgraph_create_indirect_edge): Discover
8289 polymorphic calls and record basic info into indirect_info.
8290 * gimple-fold.c (gimple_fold_call): When doing BINFO based
8291 devirtualization, ignore objc function calls.
8292 * ipa-cp.c (initialize_node_lattices): Be ready for polymorphic
8293 call with no parm index info.
8294 * ipa-prop.c (ipa_analyze_call_uses): Likewise.
8295 * tree.c (virtual_method_call_p): New function.
8296 * tree.h (virtual_method_call_p): Declare.
8297
8298 2013-08-16 Jan Hubicka <jh@suse.cz>
8299
8300 PR middle-end/58179
8301 * tree.c (obj_type_ref_class): Do not ICE on non-method calls.
8302
8303 2013-08-16 David Edelsohn <dje.gcc@gmail.com>
8304
8305 * config/rs6000/rs6000.md (rs6000_get_timebase_ppc32): Add length
8306 attribute.
8307
8308 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8309
8310 * gengtype.c (type_for_name): Add special-case support for
8311 locating types within the "gcc::" namespace.
8312 (open_base_files): Emit a "using namespace gcc" directive.
8313
8314 2013-08-16 Michael Meissner <meissner@linux.vnet.ibm.com>
8315
8316 PR target/58160
8317 * config/rs6000/predicates.md (fusion_gpr_mem_load): Allow the
8318 memory rtx to contain ZERO_EXTEND and SIGN_EXTEND.
8319
8320 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): Pass operands
8321 array instead of each individual operand as a separate argument.
8322 (emit_fusion_gpr_load): Likewise.
8323 (expand_fusion_gpr_load): Add new function declaration.
8324
8325 * config/rs6000/rs6000.c (fusion_gpr_load_p): Change the calling
8326 signature to have the operands passed as an array, instead of as
8327 separate arguments. Allow ZERO_EXTEND to be in the memory
8328 address, and also SIGN_EXTEND if -mpower8-fusion-sign. Do not
8329 depend on the register live/dead flags when peepholes are run.
8330 (expand_fusion_gpr_load): New function to be called from the
8331 peephole2 pass, to change the register that addis sets to be the
8332 target register.
8333 (emit_fusion_gpr_load): Change the calling signature to have the
8334 operands passed as an array, instead of as separate arguments.
8335 Allow ZERO_EXTEND to be in the memory address, and also
8336 SIGN_EXTEND if -mpower8-fusion-sign.
8337
8338 * config/rs6000/rs6000.md (UNSPEC_FUSION_GPR): Delete unused
8339 unspec enumeration.
8340 (power8 fusion peephole/peephole2): Rework the fusion peepholes to
8341 adjust the register addis loads up in the peephole2 pass. Do not
8342 depend on the register live/dead state when the peephole pass is done.
8343
8344 2013-08-16 David Malcolm <dmalcolm@redhat.com>
8345
8346 * gengtype.c (create_user_defined_type): Ensure that the kind
8347 is set to TYPE_USER_STRUCT, fixing a bug seen when an incomplete
8348 declaration is seen before the GTY((user)) marking.
8349
8350 2013-08-16 Bernd Edlinger <bernd.edlinger@hotmail.de>
8351
8352 PR target/58105
8353 * config/i386/i386.c (make_resolver_func): Set DECL_UNINLINABLE.
8354
8355 2013-08-16 Jan Hubicka <jh@suse.cz>
8356
8357 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Add new
8358 arugment expected_type.
8359 (gimple_fold_call): Use it.
8360 * gimple.h (gimple_extract_devirt_binfo_from_cst): Update prototype.
8361 * ipa-cp.c (ipa_get_indirect_edge_target_1): Update.
8362 * ipa-prop.c (ipa_analyze_virtual_call_uses): Use obj_type_ref_class.
8363 (try_make_edge_direct_virtual_call): Likewise.
8364 * tree.c (obj_type_ref_class): New.
8365 * tree.h (obj_type_ref_class): Use it.
8366
8367 2013-08-16 Gabriel Dos Reis <gdr@integrable-solutions.net>
8368
8369 * sched-vis.c (rtl_slim_pp_initialized): Remove.
8370 (rtl_slim_pp): Likewise.
8371 (init_rtl_slim_pretty_print): Likewise.
8372 (dump_value_slim): Don't call it. Use local pretty printer.
8373 (dump_insn_slim): Likewise.
8374 (dump_rtl_slim): Likewise.
8375 (str_pattern_slim): Likewise.
8376 * tree-mudflap.c (mf_varname_tree): Use local pretty printer.
8377 Simplify.
8378
8379 2013-08-16 Jakub Jelinek <jakub@redhat.com>
8380
8381 PR tree-optimization/58164
8382 * gimple.c (walk_stmt_load_store_addr_ops): For visit_addr
8383 walk gimple_goto_dest of GIMPLE_GOTO.
8384
8385 PR tree-optimization/58165
8386 * tree-call-cdce.c (shrink_wrap_one_built_in_call): If
8387 bi_call must be the last stmt in a bb, don't split_block, instead
8388 use fallthru edge from it and give up if there is none.
8389 Release conds vector when returning early.
8390
8391 2013-08-14 Xinliang David Li <davidxl@google.com>
8392
8393 * config/i386/i386.c (ix86_option_override_internal):
8394 Remove unused variable and field.
8395
8396 2013-08-14 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
8397
8398 PR target/57949
8399 * doc/invoke.texi: Add documentation of mcompat-align-parm option.
8400 * config/rs6000/rs6000.opt: Add mcompat-align-parm option.
8401 * config/rs6000/rs6000.c (rs6000_function_arg_boundary): For AIX
8402 and Linux, correct BLKmode alignment when 128-bit alignment is
8403 required and compatibility flag is not set.
8404 (rs6000_gimplify_va_arg): For AIX and Linux, honor specified alignment
8405 for zero-size arguments when compatibility flag is not set.
8406
8407 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8408
8409 PR tree-optimization/58145
8410 * tree-sra.c (build_ref_for_offset): If prev_base has
8411 TREE_THIS_VOLATILE or TREE_SIDE_EFFECTS, propagate it to MEM_REF.
8412
8413 2013-08-14 Xinliang David Li <davidxl@google.com>
8414
8415 * config/i386/i386.c (ix86_option_override_internal):
8416 Fix uninitialized variable error.
8417
8418 2013-08-14 Xinliang David Li <davidxl@google.com>
8419
8420 * config/i386/i386.opt: Define two new options.
8421 * config/i386/x86-tune.def: Add arch selector field in macros.
8422 * config/i386/i386.h: Adjust macro definition.
8423 * config/i386/i386.c (ix86_option_override_internal):
8424 Refactor the code.
8425 (parse_mtune_ctrl_str): New function.
8426 (set_ix86_tune_features): New function.
8427 (ix86_function_specific_restore): Call the new helper function.
8428
8429 2013-08-14 Andrey Belevantsev <abel@ispras.ru>
8430
8431 PR rtl-optimization/57662
8432 * sel-sched.c (code_motion_process_successors): When the current insn
8433 is removed after the recursive traversal, break from the loop.
8434 Add comments and debug printouts.
8435
8436 2013-08-14 Jakub Jelinek <jakub@redhat.com>
8437 Alexandre Oliva <aoliva@redhat.com>
8438
8439 PR target/58067
8440 * config/i386/i386.c (ix86_delegitimize_address): For CM_MEDIUM_PIC
8441 and CM_LARGE_PIC ix86_cmodel fall thru into the -m32 code, handle
8442 there also UNSPEC_PLTOFF.
8443
8444 2013-08-14 Marek Polacek <polacek@redhat.com>
8445
8446 * ipa-inline-analysis.c (add_clause): Avoid shifting integer
8447 NUM_CONDITIONS bit positions.
8448
8449 2013-08-13 Cary Coutant <ccoutant@google.com>
8450
8451 * dwarf2out.c (CHECKSUM_BLOCK): New macro.
8452 (attr_checksum): Hash vector contents instead of pointer.
8453 (attr_checksum_ordered): Likewise.
8454
8455 2013-08-13 Uros Bizjak <ubizjak@gmail.com>
8456
8457 * config/i386/sse.md (*sse2_maskmovdqu): Emit addr32 prefix
8458 when Pmode != word_mode. Add length_address attribute.
8459 (sse3_monitor_<mode>): Merge from sse3_monitor and
8460 sse3_monitor64_<mode> insn patterns. Emit addr32 prefix when
8461 Pmode != word_mode. Update insn length attribute.
8462 * config/i386/i386.c (ix86_option_override_internal): Update
8463 ix86_gen_monitor selection for merged sse3_monitor insn.
8464
8465 2013-08-13 Julian Brown <julian@codesourcery.com>
8466
8467 * config/rs6000/rs6000.c (rs6000_legitimize_reload_address): Don't
8468 perform invalid legitimization on greater-than-word-size modes for
8469 TARGET_E500_DOUBLE.
8470
8471 2013-08-13 Vladimir Makarov <vmakarov@redhat.com>
8472
8473 * ira.c (setup_class_translate_array): Use aclass instead of cl
8474 for classes not fully covered by allocno classes.
8475
8476 2013-08-13 Jakub Jelinek <jakub@redhat.com>
8477
8478 PR tree-optimization/57661
8479 * tree-inline.h (struct copy_body_data): Add blocks_to_copy field.
8480 * tree-inline.c (tree_function_versioning): Initialize it.
8481 (remap_gimple_stmt): Return GIMPLE_NOP for MEM_REF lhs clobber stmts
8482 if id->blocks_to_copy and MEM_REF's SSA_NAME is defined in a block
8483 that is not being copied.
8484
8485 PR sanitizer/56417
8486 * asan.c (instrument_strlen_call): Fix typo in comment.
8487 Use char * type even for the lhs of POINTER_PLUS_EXPR.
8488
8489 2013-08-13 Steve Ellcey <sellcey@mips.com>
8490
8491 * config/mips/mips.md (prefetch): Use lw instead of ld on
8492 loongson in 32bit mode.
8493
8494 2013-08-13 Nick Clifton <nickc@redhat.com>
8495
8496 * config.gcc: (avr-linux): Allow for tmake_file not being empty.
8497
8498 2013-08-13 Jan Hubicka <jh@suse.cz>
8499
8500 * cgraph.c (cgraph_turn_edge_to_speculative): Return newly
8501 introduced edge; fix typo in sanity check.
8502 (cgraph_resolve_speculation): Export; improve diagnostic.
8503 (cgraph_redirect_edge_call_stmt_to_callee): Better diagnostic; cancel
8504 speculation at type mismatch.
8505 * cgraph.h (cgraph_turn_edge_to_speculative): Update.
8506 (cgraph_resolve_speculation): Declare.
8507 (symtab_can_be_discarded): New function.
8508 * value-prof.c (gimple_ic_transform): Remove actual transform code.
8509 * ipa-inline-transform.c (speculation_removed): New global var.
8510 (clone_inlined_nodes): See if speculation can be removed.
8511 (inline_call): If speculations was removed, we growths may not match.
8512 * ipa-inline.c (can_inline_edge_p): Add DISREGARD_LIMITS parameter.
8513 (speculation_useful_p): New function.
8514 (resolve_noninline_speculation): New function.
8515 (inline_small_functions): Resolve useless speculations.
8516 * ipa-inline.h (speculation_useful_p): Declare
8517 * ipa.c (can_replace_by_local_alias): Simplify.
8518 (ipa_profile): Produce speculative calls in non-lto, too;
8519 add simple cost model; produce local aliases.
8520
8521 2013-08-13 David Malcolm <dmalcolm@redhat.com>
8522
8523 * config/i386/t-i386 (i386.o): Rename stray PIPELINE_H to
8524 PASS_MANAGER_H.
8525
8526 2013-08-12 Paolo Carlini <paolo.carlini@oracle.com>
8527
8528 * config/i386/i386.c (ix86_function_versions): Use error + inform.
8529
8530 2013-08-12 Uros Bizjak <ubizjak@gmail.com>
8531
8532 * config/i386/i386.md (floatunssi<mode>2 expand): Use MODEF mode
8533 iterator instead of X87MODEF.
8534
8535 2013-08-12 Perez Read <netfirewall@gmail.com>
8536
8537 PR target/58132
8538 * config/i386/i386.md (*movabs<mode>_1): Add <ptrsize> PTR before
8539 operand 0 for intel asm alternative.
8540 (*movabs<mode>_2): Ditto for operand 1.
8541
8542 2013-08-12 James Greenhalgh <james.greenhalgh@arm.com>
8543
8544 * config/aarch64/arm_none.h
8545 (vdup<bhsd>_lane_<su><8,16,32,64>): Fix macro call.
8546
8547 2013-08-12 Nick Clifton <nickc@redhat.com>
8548
8549 * config.gcc (m32r-linux): Allow for tmake_file not being empty.
8550
8551 2013-08-12 Yuri Rumyantsev <ysrumyan@gmail.com>
8552
8553 * config/i386/i386.md (floatunssi<mode>2 expand): Add new
8554 expand for QI/HImode operand to produce more effictive code for
8555 unsigned char(short) --> float(double) conversion.
8556
8557 2013-08-12 Alexander Monakov <amonakov@ispras.ru>
8558
8559 * doc/invoke.texi: Mention that -ftls-model does not force the final
8560 model.
8561
8562 2013-08-12 Marek Polacek <polacek@redhat.com>
8563 Marc Glisse <marc.glisse@inria.fr>
8564
8565 PR tree-optimization/57980
8566 * tree-tailcall.c (process_assignment): Call build_minus_one_cst
8567 when creating -1 constant.
8568
8569 2013-08-10 Jan Hubicka <jh@suse.cz>
8570
8571 Workaround binutils PR14342.
8572 * tree-profile.c (init_ic_make_global_vars): Add LTO path.
8573 (gimple_init_edge_profiler): Likewise.
8574 (gimple_gen_ic_func_profiler): Likewise.
8575
8576 2013-08-09 Jan Hubicka <jh@suse.cz>
8577
8578 * cgraph.c (cgraph_create_edge_1): Clear speculative flag.
8579
8580 2013-08-09 Xinliang David Li <davidxl@google.com>
8581
8582 * config/i386/stringop.def: New file.
8583 * config/i386/stringop.opt: New file.
8584 * config/i386/i386-opts.h: Include stringopt.def.
8585 * config/i386/i386.opt: Include stringopt.opt.
8586 * config/i386/i386.c (ix86_option_override_internal):
8587 Override default size based stringop inline strategies with options.
8588 * config/i386/i386.c (ix86_parse_stringop_strategy_string):
8589 New function.
8590
8591 2013-08-09 Jan Hubicka <jh@suse.cz>
8592
8593 * ipa-ref.c (ipa_clear_stmts_in_references): Clear lto_stmt_uid, too.
8594
8595 2013-08-09 Jan Hubicka <jh@suse.cz>
8596
8597 * cgraph.c (cgraph_resolve_speculation): Cut frequency to
8598 CGRAPH_FREQ_MAX.
8599 (dump_cgraph_node): Dump profile-id.
8600 * cgraph.h (cgraph_indirect_call_info): Add common_target_id
8601 and common_target_probability.
8602 * lto-cgraph.c (lto_output_edge): Stream common targets.
8603 (lto_output_node): Stream profile ids.
8604 (input_node): Stream profile ids.
8605 (input_edge): Stream common targets.
8606 * lto-streamer-in.c (fixup_call_stmt_edges_1): Fix formatting.
8607 * ipa.c: Include value-prof.h
8608 (ipa_profile_generate_summary): Turn indirect call statement histograms
8609 into common targets.
8610 (ipa_profile): Turn common targets into speculative edges.
8611
8612 2013-08-09 Jan Hubicka <jh@suse.cz>
8613
8614 * cgraph.h (cgraph_node): Add profile_id.
8615 * value-prof.c (cgraph_node_map): Turn into pointer_map.
8616 (init_node_map): Rewrite to handle hashes increas of incremental IDs.
8617 (del_node_map): Update.
8618 (find_func_by_funcdef_no): Replace by ...
8619 (find_func_by_profile_id): ... this one.
8620 (gimple_ic_transform): Do not remove useful histograms when
8621 speculation is not done; dump info when indirect call removal
8622 can happen at LTO.
8623 * value-prof.h (find_func_by_profile_id, gimple_ic): Declare.
8624 * gcov-io.h (__gcov_indirect_call_profiler): Replace by ...
8625 (__gcov_indirect_call_profiler_v2): .. this one.
8626 * profile.h (init_node_map): Update.
8627 * coverage.c (coverage_compute_profile_id): New function.
8628 * coverage.h (coverage_compute_profile_id): Declare.
8629 * tree-profile.c (init_ic_make_global_vars): Make
8630 __gcov_indirect_call_callee and __gcov_indirect_call_counters global.
8631 (gimple_init_edge_profiler): Update prototype of
8632 __gcov_indirect_call_profiler.
8633 (gimple_gen_ic_func_profiler): Simplify.
8634 (tree_profiling): Use init_node_map
8635
8636 2013-08-09 Jan Hubicka <jh@suse.cz>
8637
8638 * cgraphbuild.c (cgraph_rebuild_references): Rebuild only
8639 non-speculative refs.
8640 * cgraph.c (cgraph_update_edge_in_call_site_hash): New function.
8641 (cgraph_add_edge_to_call_site_hash): Deal with speculative calls.
8642 (cgraph_set_call_stmt): Likewise.
8643 (cgraph_create_edge_1): Fix release checking compilatoin;
8644 clear lto_stmt_uid.
8645 (cgraph_free_edge): Free indirect info.
8646 (cgraph_turn_edge_to_speculative): New function.
8647 (cgraph_speculative_call_info): New function.
8648 (cgraph_make_edge_direct): Return direct edge; handle speculation.
8649 (cgraph_redirect_edge_call_stmt_to_callee): Expand speculative edges.
8650 (dump_cgraph_node): Dump speculation.
8651 (verify_edge_count_and_frequency): Accept speculative edges.
8652 (verify_edge_corresponds_to_fndecl): Handle partitioned cgraph.
8653 (verify_cgraph_node): Handle speculation.
8654 * cgraph.h (cgraph_edge): Add SPECULATIVE flag.
8655 (cgraph_set_call_stmt): Update prototype.
8656 (cgraph_make_edge_direct): Update prototype.
8657 (cgraph_speculative_call_info): Declare.
8658 * ipa-cp.c (ipcp_discover_new_direct_edges): Be ready for edge
8659 to change; update call of ipa_find_references.
8660 * ipa-ref.c (ipa_record_reference): Fix return value; clear
8661 lto_stmt_uid and speculative flags.
8662 (ipa_dump_references): Dump speculation.
8663 (ipa_clone_references): Clone speculative flag.
8664 (ipa_clone_referring): Likewise.
8665 (ipa_clone_ref): New function.
8666 (ipa_find_reference): Look into lto_stmt_uids
8667 (ipa_clear_stmts_in_references): Do not clear speculative calls.
8668 * ipa-ref.h (ipa_ref): Add lto_stmt_uid and speculative flags.
8669 (ipa_find_reference): Update declaration.
8670 (ipa_clone_ref): Declare.
8671 * lto-cgraph.c (lto_output_edge): Make lto_stmt_uids start from 0;
8672 stream speculative flag.
8673 (lto_output_ref): Stream statements uids and speculation.
8674 (input_ref): Likewise.
8675 (input_edge): Stream speuclation.
8676 * cgraphclones.c (cgraph_clone_edge): Clone speculation.
8677 (cgraph_set_call_stmt_including_clones): Handle speculation.
8678 * ipa-inline.c (heap_edge_removal_hook): New function.
8679 (inline_small_functions): Register it.
8680 * lto-streamer-in.c (fixup_call_stmt_edges_1): Bounds checking;
8681 also initialize refs.
8682 * ipa-prop.c (ipa_make_edge_direct_to_target): Be ready for
8683 edge to change.
8684 (try_make_edge_direct_simple_call): Likewise.
8685 (try_make_edge_direct_simple_call): Likewise.
8686 (update_indirect_edges_after_inlining): Likewise.
8687 (remove_described_reference): Look proper lto_stmt_uid.
8688 (propagate_controlled_uses): Likewise.
8689 (propagate_controlled_uses): Liekwise.
8690 * tree-inline.c (copy_bb): Copy speculative edges.
8691 (redirect_all_calls): New function.
8692 (copy_cfg_body): Do redirection after loop info is updated.
8693 (delete_unreachable_blocks_update_callgraph): Updadte speculation.
8694
8695 2013-08-09 Jan Hubicka <jh@suse.cz>
8696
8697 * lto-streamer-out.c (output_function): Renumber PHIs.
8698 * lto-streamer-in.c (input_function): Likewise.
8699
8700 2013-08-09 James Greenhalgh <james.greenhalgh@arm.com>
8701
8702 * config/aarch64/aarch64-simd-builtins.def (get_lane_signed): Remove.
8703 (get_lane_unsigned): Likewise.
8704 (dup_lane_scalar): Likewise.
8705 (get_lane): enable for VALL.
8706 * config/aarch64/aarch64-simd.md
8707 (aarch64_dup_lane_scalar<mode>): Remove.
8708 (aarch64_get_lane_signed<mode>): Likewise.
8709 (aarch64_get_lane_unsigned<mode>): Likewise.
8710 (aarch64_get_lane_extend<GPI:mode><VDQQH:mode>): New.
8711 (aarch64_get_lane_zero_extendsi<mode>): Likewise.
8712 (aarch64_get_lane<mode>): Enable for all vector modes.
8713 (aarch64_get_lanedi): Remove misleading constraints.
8714 * config/aarch64/arm_neon.h
8715 (__aarch64_vget_lane_any): Define.
8716 (__aarch64_vget<q>_lane_<fpsu><8,16,32,64>): Likewise.
8717 (vget<q>_lane_<fpsu><8,16,32,64>): Use __aarch64_vget_lane macros.
8718 (vdup<bhsd>_lane_<su><8,16,32,64>): Likewise.
8719 * config/aarch64/iterators.md (VDQQH): New.
8720 (VDQQHS): Likewise.
8721 (vwcore): Likewise.
8722
8723 2013-08-09 Eric Botcazou <ebotcazou@adacore.com>
8724
8725 * configure.ac: Add GAS check for LEON instructions on SPARC.
8726 * configure: Regenerate.
8727 * config.in: Likewise.
8728 * config.gcc (with_cpu): Remove sparc-leon*-* and deal with LEON in the
8729 sparc*-*-* block.
8730 * config/sparc/sparc.opt (LEON, LEON3): New masks.
8731 * config/sparc/sparc.h (ASM_CPU32_DEFAULT_SPEC): Set to AS_LEON_FLAG
8732 for LEON or LEON3.
8733 (ASM_CPU_SPEC): Pass AS_LEON_FLAG if -mcpu=leon or -mcpu=leon3.
8734 (AS_LEON_FLAG): New macro.
8735 * config/sparc/sparc.c (sparc_option_override): Set MASK_LEON for leon
8736 and MASK_LEON3 for leon3 and unset them if HAVE_AS_LEON is not defined.
8737 Deal with LEON and LEON3 for the memory model.
8738 * config/sparc/sync.md (atomic_compare_and_swap<mode>): Enable if LEON3
8739 (atomic_compare_and_swap<mode>_1): Likewise.
8740 (*atomic_compare_and_swap<mode>_1): Likewise.
8741
8742 2013-08-09 Zhenqiang Chen <zhenqiang.chen@linaro.org>
8743
8744 * config/arm/neon.md (vcond): Fix floating-point vector
8745 comparisons against 0.
8746
8747 2013-08-08 Vladimir Makarov <vmakarov@redhat.com>
8748
8749 * lra-constraints.c (emit_spill_move): Remove assert.
8750 (process_alt_operands): Add more debugging
8751 output. Increase reject for spilling into memory. Decrease
8752 reject for reloading scratch.
8753 (split_reg): Use HARD_REGNO_CALLER_SAVE_MODE.
8754
8755 2013-08-08 Steve Ellcey <sellcey@mips.com>
8756
8757 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add nan2008.
8758 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Make mips16 and
8759 micromips incompatible. Add nan2008.
8760 (MULTILIB_DIRNAMES): Add nan2008.
8761 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8762 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Make mips16
8763 and micromips incompatible. Add nan2008.
8764 (MULTILIB_DIRNAMES): Add nan2008.
8765 (MULTILIB_EXCEPTIONS): Remove mips16/micromips entry.
8766
8767 2013-08-08 Richard Sandiford <rdsandiford@googlemail.com>
8768
8769 PR rtl-optimization/58079
8770 * combine.c (combine_simplify_rtx): Avoid using SUBST if
8771 simplify_comparison has widened a comparison with an integer.
8772
8773 2013-08-08 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
8774
8775 * config/arm/neon.md (movmisalign<mode>): Disable when we
8776 don't allow unaligned accesses.
8777 (*movmisalign<mode>_neon_store): Likewise.
8778 (*movmisalign<mode>_neon_load): Likewise.
8779 (*movmisalign<mode>_neon_store): Likewise.
8780 (*movmisalign<mode>_neon_load): Likewise.
8781
8782 2013-08-08 Jan Hubicka <jh@suse.cz>
8783
8784 * cgraphbuild.c (build_cgraph_edges): Do not walk into debugs.
8785 (make_pass_rebuild_cgraph_edges): Also clear references.
8786 * cgraph.c (verify_cgraph_node): Add basic ipa-ref verifier.
8787 * ipa-inline-transform.c (inline_transform): Remove all references
8788 after inlining.
8789 * cgraphunit.c (expand_function): Remove all references after
8790 expansion.
8791 * ipa-ref.c (ipa_ref_has_aliases_p): Fix formatting.
8792 (ipa_find_reference): Rewrite to iterator.
8793 (remove_stmt_references): Likewise.
8794 (ipa_clear_stmts_in_references): New function.
8795 * ipa-ref.h (ipa_clear_stmts_in_references): Declare.
8796 * cgraphclones.c (cgraph_materialize_all_clones): Remove or
8797 clear references.
8798 * ipa-split.c (split_function): Remove references in split function.
8799
8800 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
8801
8802 PR target/57431
8803 * config/arm/arm/neon.md (neon_vld1_dupdi): New expand pattern.
8804 (neon_vld1_dup<mode> VD iterator): Iterate over VD not VDX.
8805
8806 2013-08-08 Richard Earnshaw <rearnsha@arm.com>
8807
8808 PR target/56979
8809 * config/arm/arm.c (aapcs_vfp_allocate): Decompose the argument if the
8810 suggested mode for the assignment isn't compatible with the
8811 registers required.
8812
8813 2013-08-08 Bernd Edlinger <bernd.edlinger@hotmail.de>
8814
8815 PR target/58065
8816 * config/arm/arm.h (MALLOC_ABI_ALIGNMENT): Define.
8817
8818 2013-08-07 Xinliang David Li <davidxl@google.com>
8819
8820 * config/i386/i386.opt: New option -mtune-ctrl=.
8821 * config/i386/x86-tune.def: New file.
8822 * config/i386/i386.h: include x86-tune.def.
8823 * config/i386/i386.c (ix86_option_override_internal):
8824 Parsing -mtune-ctrl= option and set tune features.
8825
8826 2013-08-07 Oleg Endo <olegendo@gcc.gnu.org>
8827
8828 PR other/12081
8829 * config/rs6000/rs6000.c (gen_2arg_fn_t): Remove typedef.
8830 (rs6000_emit_swdiv, rs6000_emit_swrsqrt): Don't cast result of GEN_FCN
8831 to gen_2arg_fn_t.
8832
8833 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8834
8835 * rtl.h (update_alignments): Declare.
8836 * final.c (grow_label_align): New function extracted from...
8837 (shorten_branches): ...here. Call it.
8838 (update_alignments): New function.
8839 * reorg.c (sibling_labels): New variable.
8840 (get_label_before): Add SIBLING parameter. If it is non-zero, push
8841 the new label along with it onto the sibling_labels vector.
8842 (fill_simple_delay_slots): Adjust call to get_label_before.
8843 (fill_slots_from_thread): Likewise.
8844 (relax_delay_slots): Likewise.
8845 (make_return_insns): Likewise.
8846 (dbr_schedule): Invoke update_alignment on the sibling_labels vector.
8847
8848 2013-08-07 Eric Botcazou <ebotcazou@adacore.com>
8849
8850 * diagnostic.c (diagnostic_classify_diagnostic): Accept zero index and
8851 document its semantics.
8852 (diagnostic_report_diagnostic): Adjust accordingly.
8853
8854 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8855
8856 * config/sparc/sparc.c (insert_pass_work_around_errata): Move into...
8857 (sparc_option_override): ...and port to new C++ pass API.
8858 * config/sparc/t-sparc (sparc.o): Add dep on CONTEXT_H
8859
8860 2013-08-07 Peter Bergner <bergner@vnet.ibm.com>
8861
8862 * config/rs6000/rs6000.c (htm_expand_builtin) <case 0>: Remove.
8863
8864 2013-08-06 Caroline Tice <cmtice@google.com>
8865
8866 * gcc.c (VTABLE_VERIFICATION_SPEC): New definition.
8867 (LINK_COMMAND_SPEC): Add VTABLE_VERIFICATION_SPEC.
8868 * tree-pass.h: Add pass_vtable_verify.
8869 * varasm.c (assemble_variable): Add code to properly set the comdat
8870 section and name for the .vtable_map_vars section.
8871 (assemble_vtyv_preinit_initializer): New function.
8872 (default_sectin_type_flags): Make sure .vtable_map_vars section has
8873 LINK_ONCE flag.
8874 * output.h: Add function decl for assemble_vtv_preinit_initializer.
8875 * vtable-verify.c: New file.
8876 * vtable-verify.h: New file.
8877 * flag-types.h (enum vtv_priority): Defintions for flag_vtable_verify
8878 initialiation levels.
8879 * timevar.def (TV_VTABLE_VERIFICATION): New definition.
8880 * passes.def: Insert pass_vtable_verify.
8881 * aclocal.m4: Reorder includes.
8882 * doc/invoke.texi: Document the -fvtable-verify=, -fvtv-debug, and
8883 -fvtv-counts options.
8884 * config/gnu-user.h (GNU_USER_TARGET_STARTFILE_SPEC): Add vtv_start*.o,
8885 as appropriate, if -fvtable-verify=... is used.
8886 (GNU_USER_TARGET_ENDFILE_SPEC): Add vtv_end*.o as appropriate, if
8887 -fvtable-verify=... is used.
8888 * Makefile.in (OBJS): Add vtable-verify.o to list.
8889 (vtable-verify.o): Add new build rule.
8890 (GTFILES): Add vtable-verify.c to list.
8891 * common.opt (fvtable-verify=): New flag.
8892 (vtv_priority): Values for fvtable-verify= flag.
8893 (fvtv-counts): New flag.
8894 (fvtv-debug): New flag.
8895 * tree.h (save_vtable_map_decl): New extern function decl.
8896
8897 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8898
8899 * config/rl78/rl78.c (rl78_devirt_pass): Convert from a struct to...
8900 (pass_rl78_devirt): ...new subclass of rtl_opt_pass along with...
8901 (pass_data_rl78_devirt): ...new pass_data instance and...
8902 (make_pass_rl78_devirt): ...new function.
8903 (rl78_asm_file_start): Port pass registration to new C++ API.
8904
8905 2013-08-07 David Malcolm <dmalcolm@redhat.com>
8906
8907 * coretypes.h (rtl_opt_pass): Add.
8908 (gcc::context): Add.
8909 * config/epiphany/epiphany.c (pass_mode_switch_use): New.
8910 (epiphany_init): Port to new C++ pass API.
8911 (epiphany_optimize_mode_switching): Likewise.
8912 * pass_manager.h (pass_manager::get_pass_split_all_insns): New.
8913 (pass_manager::get_pass_mode_switching): New.
8914 (pass_manager::get_pass_peephole2): New.
8915 * mode-switching.c (pass_mode_switching): Add clone method.
8916 * recog.c (pass_peephole2): Add clone method.
8917 (pass_split_all_insns): Add clone method.
8918
8919 2013-08-06 David Malcolm <dmalcolm@redhat.com>
8920
8921 * config/mips/mips.c (insert_pass_mips_machine_reorg2): Move into...
8922 (mips_option_override): ...here, porting to new C++ API for passes.
8923
8924 2013-08-06 Jan Hubicka <jh@suse.cz>
8925
8926 * cgraph.c (cgraph_get_body): New function based on lto.c
8927 implementation.
8928 * cgraph.h (cgraph_get_body): Declare.
8929 * cgraphclones.c (cgraph_create_virtual_clone): Commonize WPA and
8930 LTO paths.
8931 * cgraphunit.c (expand_function): Get body prior expanding.
8932 * ipa.c (function_and_variable_visibility): Use gimple_has_body_p test.
8933 * lto-cgraph.c (lto_output_node): Do not stream bodies we don't
8934 really need.
8935 * passes.c (do_per_function_toporder): Get body.
8936 * tree-inline.c (expand_call_inline): Get body prior inlining it.
8937 * tree-ssa-structalias.c (ipa_pta_execute): Get body; skip clones.
8938
8939 2013-08-06 Martin Jambor <mjambor@suse.cz>
8940
8941 PR fortran/57987
8942 * cgraphunit.c (cgraph_finalize_function): Assert that nested function
8943 is not re-finalized. Rename second parameter to no_collect.
8944
8945 2013-08-06 Martin Jambor <mjambor@suse.cz>
8946
8947 PR middle-end/58041
8948 * gimple-ssa-strength-reduction.c (replace_ref): Make sure built
8949 MEM_REF has proper alignment information.
8950
8951 2013-08-05 Oleg Endo <olegendo@gcc.gnu.org>
8952
8953 PR other/12081
8954 * recog.h (rtx (*insn_gen_fn) (rtx, ...)): Replace typedef with new
8955 class insn_gen_fn.
8956 * expr.c (move_by_pieces_1, store_by_pieces_2): Replace argument
8957 rtx (*) (rtx, ...) with insn_gen_fn.
8958 * genoutput.c (output_insn_data): Cast gen_? function pointers to
8959 insn_gen_fn::stored_funcptr. Add initializer braces.
8960
8961 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8962
8963 Rewrite how instances of passes are cloned to remove assumptions
8964 about their sizes (thus allowing pass subclasses to have
8965 additional data fields, albeit non-GC-managed ones at this point).
8966
8967 * passes.c (make_pass_instance): Now that passes have clone
8968 methods, rewrite this function to eliminate XNEW and memcpy
8969 calls that used hardcoded sizes. Since this function no longer
8970 creates pass instances, rename it to...
8971 (add_pass_instance): ...this. Document the old way that passes were
8972 numbered and flagged, and rework this function to continue using it.
8973 (next_pass_1): Add an initial_pass argument for use by
8974 add_pass_instance.
8975 (position_pass): When adding multiple instances of a pass, use
8976 the pass's clone method, rather than relying on the XNEW/memcpy
8977 within the former make_pass_instance (now add_pass_instance).
8978 (pass_manager::pass_manager): When invoking next_pass_1, also supply
8979 the initial instance of the current pass within the pass manager.
8980
8981 2013-08-05 David Malcolm <dmalcolm@redhat.com>
8982
8983 This is the automated part of the conversion of passes from C
8984 structs to C++ classes.
8985
8986 Patch autogenerated by refactor_passes.py from
8987 https://github.com/davidmalcolm/gcc-refactoring-scripts
8988 revision 03fe39476a4c4ea450b49e087cfa817b5f92021e
8989
8990 * asan.c (pass_asan): Convert from a global struct to a subclass of
8991 gimple_opt_pass along with...
8992 (pass_data_asan): ...new pass_data instance and...
8993 (make_pass_asan): ...new function.
8994 (pass_asan_O0): Convert from a global struct to a subclass of
8995 gimple_opt_pass along with...
8996 (pass_data_asan_O0): ...new pass_data instance and...
8997 (make_pass_asan_O0): ...new function.
8998 * auto-inc-dec.c (pass_inc_dec): Convert from a global struct to a
8999 subclass of rtl_opt_pass along with...
9000 (pass_data_inc_dec): ...new pass_data instance and...
9001 (make_pass_inc_dec): ...new function.
9002 * bb-reorder.c (pass_reorder_blocks): Convert from a global struct to
9003 a subclass of rtl_opt_pass along with...
9004 (pass_data_reorder_blocks): ...new pass_data instance and...
9005 (make_pass_reorder_blocks): ...new function.
9006 (pass_duplicate_computed_gotos): Convert from a global struct to a
9007 subclass of rtl_opt_pass along with...
9008 (pass_data_duplicate_computed_gotos): ...new pass_data instance and...
9009 (make_pass_duplicate_computed_gotos): ...new function.
9010 (pass_partition_blocks): Convert from a global struct to a subclass of
9011 rtl_opt_pass along with...
9012 (pass_data_partition_blocks): ...new pass_data instance and...
9013 (make_pass_partition_blocks): ...new function.
9014 * bt-load.c (pass_branch_target_load_optimize1): Convert from a global
9015 struct to a subclass of rtl_opt_pass along with...
9016 (pass_data_branch_target_load_optimize1): ...new pass_data instance
9017 and...
9018 (make_pass_branch_target_load_optimize1): ...new function.
9019 (pass_branch_target_load_optimize2): Convert from a global struct to a
9020 subclass of rtl_opt_pass along with...
9021 (pass_data_branch_target_load_optimize2): ...new pass_data instance
9022 and...
9023 (make_pass_branch_target_load_optimize2): ...new function.
9024 * cfgcleanup.c (pass_jump): Convert from a global struct to a subclass
9025 of rtl_opt_pass along with...
9026 (pass_data_jump): ...new pass_data instance and...
9027 (make_pass_jump): ...new function.
9028 (pass_jump2): Convert from a global struct to a subclass of
9029 rtl_opt_pass along with...
9030 (pass_data_jump2): ...new pass_data instance and...
9031 (make_pass_jump2): ...new function.
9032 * cfgexpand.c (pass_expand): Convert from a global struct to a
9033 subclass of rtl_opt_pass along with...
9034 (pass_data_expand): ...new pass_data instance and...
9035 (make_pass_expand): ...new function.
9036 * cfgrtl.c (pass_free_cfg): Convert from a global struct to a subclass
9037 of rtl_opt_pass along with...
9038 (pass_data_free_cfg): ...new pass_data instance and...
9039 (make_pass_free_cfg): ...new function.
9040 (pass_into_cfg_layout_mode): Convert from a global struct to a
9041 subclass of rtl_opt_pass along with...
9042 (pass_data_into_cfg_layout_mode): ...new pass_data instance and...
9043 (make_pass_into_cfg_layout_mode): ...new function.
9044 (pass_outof_cfg_layout_mode): Convert from a global struct to a
9045 subclass of rtl_opt_pass along with...
9046 (pass_data_outof_cfg_layout_mode): ...new pass_data instance and...
9047 (make_pass_outof_cfg_layout_mode): ...new function.
9048 * cgraphbuild.c (pass_build_cgraph_edges): Convert from a global
9049 struct to a subclass of gimple_opt_pass along with...
9050 (pass_data_build_cgraph_edges): ...new pass_data instance and...
9051 (make_pass_build_cgraph_edges): ...new function.
9052 (pass_rebuild_cgraph_edges): Convert from a global struct to a
9053 subclass of gimple_opt_pass along with...
9054 (pass_data_rebuild_cgraph_edges): ...new pass_data instance and...
9055 (make_pass_rebuild_cgraph_edges): ...new function.
9056 (pass_remove_cgraph_callee_edges): Convert from a global struct to a
9057 subclass of gimple_opt_pass along with...
9058 (pass_data_remove_cgraph_callee_edges): ...new pass_data instance
9059 and...
9060 (make_pass_remove_cgraph_callee_edges): ...new function.
9061 * combine-stack-adj.c (pass_stack_adjustments): Convert from a global
9062 struct to a subclass of rtl_opt_pass along with...
9063 (pass_data_stack_adjustments): ...new pass_data instance and...
9064 (make_pass_stack_adjustments): ...new function.
9065 * combine.c (pass_combine): Convert from a global struct to a subclass
9066 of rtl_opt_pass along with...
9067 (pass_data_combine): ...new pass_data instance and...
9068 (make_pass_combine): ...new function.
9069 * compare-elim.c (pass_compare_elim_after_reload): Convert from a
9070 global struct to a subclass of rtl_opt_pass along with...
9071 (pass_data_compare_elim_after_reload): ...new pass_data instance
9072 and...
9073 (make_pass_compare_elim_after_reload): ...new function.
9074 * cprop.c (pass_rtl_cprop): Convert from a global struct to a subclass
9075 of rtl_opt_pass along with...
9076 (pass_data_rtl_cprop): ...new pass_data instance and...
9077 (make_pass_rtl_cprop): ...new function.
9078 * cse.c (pass_cse): Convert from a global struct to a subclass of
9079 rtl_opt_pass along with...
9080 (pass_data_cse): ...new pass_data instance and...
9081 (make_pass_cse): ...new function.
9082 (pass_cse2): Convert from a global struct to a subclass of
9083 rtl_opt_pass along with...
9084 (pass_data_cse2): ...new pass_data instance and...
9085 (make_pass_cse2): ...new function.
9086 (pass_cse_after_global_opts): Convert from a global struct to a
9087 subclass of rtl_opt_pass along with...
9088 (pass_data_cse_after_global_opts): ...new pass_data instance and...
9089 (make_pass_cse_after_global_opts): ...new function.
9090 * dce.c (pass_ud_rtl_dce): Convert from a global struct to a subclass
9091 of rtl_opt_pass along with...
9092 (pass_data_ud_rtl_dce): ...new pass_data instance and...
9093 (make_pass_ud_rtl_dce): ...new function.
9094 (pass_fast_rtl_dce): Convert from a global struct to a subclass of
9095 rtl_opt_pass along with...
9096 (pass_data_fast_rtl_dce): ...new pass_data instance and...
9097 (make_pass_fast_rtl_dce): ...new function.
9098 * df-core.c (pass_df_initialize_opt): Convert from a global struct to
9099 a subclass of rtl_opt_pass along with...
9100 (pass_data_df_initialize_opt): ...new pass_data instance and...
9101 (make_pass_df_initialize_opt): ...new function.
9102 (pass_df_initialize_no_opt): Convert from a global struct to a
9103 subclass of rtl_opt_pass along with...
9104 (pass_data_df_initialize_no_opt): ...new pass_data instance and...
9105 (make_pass_df_initialize_no_opt): ...new function.
9106 (pass_df_finish): Convert from a global struct to a subclass of
9107 rtl_opt_pass along with...
9108 (pass_data_df_finish): ...new pass_data instance and...
9109 (make_pass_df_finish): ...new function.
9110 * dse.c (pass_rtl_dse1): Convert from a global struct to a subclass of
9111 rtl_opt_pass along with...
9112 (pass_data_rtl_dse1): ...new pass_data instance and...
9113 (make_pass_rtl_dse1): ...new function.
9114 (pass_rtl_dse2): Convert from a global struct to a subclass of
9115 rtl_opt_pass along with...
9116 (pass_data_rtl_dse2): ...new pass_data instance and...
9117 (make_pass_rtl_dse2): ...new function.
9118 * dwarf2cfi.c (pass_dwarf2_frame): Convert from a global struct to a
9119 subclass of rtl_opt_pass along with...
9120 (pass_data_dwarf2_frame): ...new pass_data instance and...
9121 (make_pass_dwarf2_frame): ...new function.
9122 * except.c (pass_set_nothrow_function_flags): Convert from a global
9123 struct to a subclass of rtl_opt_pass along with...
9124 (pass_data_set_nothrow_function_flags): ...new pass_data instance
9125 and...
9126 (make_pass_set_nothrow_function_flags): ...new function.
9127 (pass_convert_to_eh_region_ranges): Convert from a global struct to a
9128 subclass of rtl_opt_pass along with...
9129 (pass_data_convert_to_eh_region_ranges): ...new pass_data instance
9130 and...
9131 (make_pass_convert_to_eh_region_ranges): ...new function.
9132 * final.c (pass_compute_alignments): Convert from a global struct to a
9133 subclass of rtl_opt_pass along with...
9134 (pass_data_compute_alignments): ...new pass_data instance and...
9135 (make_pass_compute_alignments): ...new function.
9136 (pass_final): Convert from a global struct to a subclass of
9137 rtl_opt_pass along with...
9138 (pass_data_final): ...new pass_data instance and...
9139 (make_pass_final): ...new function.
9140 (pass_shorten_branches): Convert from a global struct to a subclass of
9141 rtl_opt_pass along with...
9142 (pass_data_shorten_branches): ...new pass_data instance and...
9143 (make_pass_shorten_branches): ...new function.
9144 (pass_clean_state): Convert from a global struct to a subclass of
9145 rtl_opt_pass along with...
9146 (pass_data_clean_state): ...new pass_data instance and...
9147 (make_pass_clean_state): ...new function.
9148 * function.c (pass_instantiate_virtual_regs): Convert from a global
9149 struct to a subclass of rtl_opt_pass along with...
9150 (pass_data_instantiate_virtual_regs): ...new pass_data instance and...
9151 (make_pass_instantiate_virtual_regs): ...new function.
9152 (pass_leaf_regs): Convert from a global struct to a subclass of
9153 rtl_opt_pass along with...
9154 (pass_data_leaf_regs): ...new pass_data instance and...
9155 (make_pass_leaf_regs): ...new function.
9156 (pass_thread_prologue_and_epilogue): Convert from a global struct to a
9157 subclass of rtl_opt_pass along with...
9158 (pass_data_thread_prologue_and_epilogue): ...new pass_data instance
9159 and...
9160 (make_pass_thread_prologue_and_epilogue): ...new function.
9161 (pass_match_asm_constraints): Convert from a global struct to a
9162 subclass of rtl_opt_pass along with...
9163 (pass_data_match_asm_constraints): ...new pass_data instance and...
9164 (make_pass_match_asm_constraints): ...new function.
9165 * fwprop.c (pass_rtl_fwprop): Convert from a global struct to a
9166 subclass of rtl_opt_pass along with...
9167 (pass_data_rtl_fwprop): ...new pass_data instance and...
9168 (make_pass_rtl_fwprop): ...new function.
9169 (pass_rtl_fwprop_addr): Convert from a global struct to a subclass of
9170 rtl_opt_pass along with...
9171 (pass_data_rtl_fwprop_addr): ...new pass_data instance and...
9172 (make_pass_rtl_fwprop_addr): ...new function.
9173 * gcse.c (pass_rtl_pre): Convert from a global struct to a subclass of
9174 rtl_opt_pass along with...
9175 (pass_data_rtl_pre): ...new pass_data instance and...
9176 (make_pass_rtl_pre): ...new function.
9177 (pass_rtl_hoist): Convert from a global struct to a subclass of
9178 rtl_opt_pass along with...
9179 (pass_data_rtl_hoist): ...new pass_data instance and...
9180 (make_pass_rtl_hoist): ...new function.
9181 * gimple-low.c (pass_lower_cf): Convert from a global struct to a
9182 subclass of gimple_opt_pass along with...
9183 (pass_data_lower_cf): ...new pass_data instance and...
9184 (make_pass_lower_cf): ...new function.
9185 * gimple-ssa-strength-reduction.c (pass_strength_reduction): Convert
9186 from a global struct to a subclass of gimple_opt_pass along with...
9187 (pass_data_strength_reduction): ...new pass_data instance and...
9188 (make_pass_strength_reduction): ...new function.
9189 * ifcvt.c (pass_rtl_ifcvt): Convert from a global struct to a subclass
9190 of rtl_opt_pass along with...
9191 (pass_data_rtl_ifcvt): ...new pass_data instance and...
9192 (make_pass_rtl_ifcvt): ...new function.
9193 (pass_if_after_combine): Convert from a global struct to a subclass of
9194 rtl_opt_pass along with...
9195 (pass_data_if_after_combine): ...new pass_data instance and...
9196 (make_pass_if_after_combine): ...new function.
9197 (pass_if_after_reload): Convert from a global struct to a subclass of
9198 rtl_opt_pass along with...
9199 (pass_data_if_after_reload): ...new pass_data instance and...
9200 (make_pass_if_after_reload): ...new function.
9201 * init-regs.c (pass_initialize_regs): Convert from a global struct to
9202 a subclass of rtl_opt_pass along with...
9203 (pass_data_initialize_regs): ...new pass_data instance and...
9204 (make_pass_initialize_regs): ...new function.
9205 * ipa-cp.c (pass_ipa_cp): Convert from a global struct to a subclass
9206 of ipa_opt_pass_d along with...
9207 (pass_data_ipa_cp): ...new pass_data instance and...
9208 (make_pass_ipa_cp): ...new function.
9209 * ipa-inline-analysis.c (pass_inline_parameters): Convert from a
9210 global struct to a subclass of gimple_opt_pass along with...
9211 (pass_data_inline_parameters): ...new pass_data instance and...
9212 (make_pass_inline_parameters): ...new function.
9213 * ipa-inline.c (pass_early_inline): Convert from a global struct to a
9214 subclass of gimple_opt_pass along with...
9215 (pass_data_early_inline): ...new pass_data instance and...
9216 (make_pass_early_inline): ...new function.
9217 (pass_ipa_inline): Convert from a global struct to a subclass of
9218 ipa_opt_pass_d along with...
9219 (pass_data_ipa_inline): ...new pass_data instance and...
9220 (make_pass_ipa_inline): ...new function.
9221 * ipa-pure-const.c (pass_local_pure_const): Convert from a global
9222 struct to a subclass of gimple_opt_pass along with...
9223 (pass_data_local_pure_const): ...new pass_data instance and...
9224 (make_pass_local_pure_const): ...new function.
9225 (pass_ipa_pure_const): Convert from a global struct to a subclass of
9226 ipa_opt_pass_d along with...
9227 (pass_data_ipa_pure_const): ...new pass_data instance and...
9228 (make_pass_ipa_pure_const): ...new function.
9229 * ipa-reference.c (pass_ipa_reference): Convert from a global struct
9230 to a subclass of ipa_opt_pass_d along with...
9231 (pass_data_ipa_reference): ...new pass_data instance and...
9232 (make_pass_ipa_reference): ...new function.
9233 * ipa-split.c (pass_split_functions): Convert from a global struct to
9234 a subclass of gimple_opt_pass along with...
9235 (pass_data_split_functions): ...new pass_data instance and...
9236 (make_pass_split_functions): ...new function.
9237 (pass_feedback_split_functions): Convert from a global struct to a
9238 subclass of gimple_opt_pass along with...
9239 (pass_data_feedback_split_functions): ...new pass_data instance and...
9240 (make_pass_feedback_split_functions): ...new function.
9241 * ipa.c (pass_ipa_function_and_variable_visibility): Convert from a
9242 global struct to a subclass of simple_ipa_opt_pass along with...
9243 (pass_data_ipa_function_and_variable_visibility): ...new pass_data
9244 instance and...
9245 (make_pass_ipa_function_and_variable_visibility): ...new function.
9246 (pass_ipa_free_inline_summary): Convert from a global struct to a
9247 subclass of simple_ipa_opt_pass along with...
9248 (pass_data_ipa_free_inline_summary): ...new pass_data instance and...
9249 (make_pass_ipa_free_inline_summary): ...new function.
9250 (pass_ipa_whole_program_visibility): Convert from a global struct to a
9251 subclass of ipa_opt_pass_d along with...
9252 (pass_data_ipa_whole_program_visibility): ...new pass_data instance
9253 and...
9254 (make_pass_ipa_whole_program_visibility): ...new function.
9255 (pass_ipa_profile): Convert from a global struct to a subclass of
9256 ipa_opt_pass_d along with...
9257 (pass_data_ipa_profile): ...new pass_data instance and...
9258 (make_pass_ipa_profile): ...new function.
9259 (pass_ipa_cdtor_merge): Convert from a global struct to a subclass of
9260 ipa_opt_pass_d along with...
9261 (pass_data_ipa_cdtor_merge): ...new pass_data instance and...
9262 (make_pass_ipa_cdtor_merge): ...new function.
9263 * ira.c (pass_ira): Convert from a global struct to a subclass of
9264 rtl_opt_pass along with...
9265 (pass_data_ira): ...new pass_data instance and...
9266 (make_pass_ira): ...new function.
9267 (pass_reload): Convert from a global struct to a subclass of
9268 rtl_opt_pass along with...
9269 (pass_data_reload): ...new pass_data instance and...
9270 (make_pass_reload): ...new function.
9271 * jump.c (pass_cleanup_barriers): Convert from a global struct to a
9272 subclass of rtl_opt_pass along with...
9273 (pass_data_cleanup_barriers): ...new pass_data instance and...
9274 (make_pass_cleanup_barriers): ...new function.
9275 * loop-init.c (pass_loop2): Convert from a global struct to a subclass
9276 of rtl_opt_pass along with...
9277 (pass_data_loop2): ...new pass_data instance and...
9278 (make_pass_loop2): ...new function.
9279 (pass_rtl_loop_init): Convert from a global struct to a subclass of
9280 rtl_opt_pass along with...
9281 (pass_data_rtl_loop_init): ...new pass_data instance and...
9282 (make_pass_rtl_loop_init): ...new function.
9283 (pass_rtl_loop_done): Convert from a global struct to a subclass of
9284 rtl_opt_pass along with...
9285 (pass_data_rtl_loop_done): ...new pass_data instance and...
9286 (make_pass_rtl_loop_done): ...new function.
9287 (pass_rtl_move_loop_invariants): Convert from a global struct to a
9288 subclass of rtl_opt_pass along with...
9289 (pass_data_rtl_move_loop_invariants): ...new pass_data instance and...
9290 (make_pass_rtl_move_loop_invariants): ...new function.
9291 (pass_rtl_unswitch): Convert from a global struct to a subclass of
9292 rtl_opt_pass along with...
9293 (pass_data_rtl_unswitch): ...new pass_data instance and...
9294 (make_pass_rtl_unswitch): ...new function.
9295 (pass_rtl_unroll_and_peel_loops): Convert from a global struct to a
9296 subclass of rtl_opt_pass along with...
9297 (pass_data_rtl_unroll_and_peel_loops): ...new pass_data instance
9298 and...
9299 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9300 (pass_rtl_doloop): Convert from a global struct to a subclass of
9301 rtl_opt_pass along with...
9302 (pass_data_rtl_doloop): ...new pass_data instance and...
9303 (make_pass_rtl_doloop): ...new function.
9304 * lower-subreg.c (pass_lower_subreg): Convert from a global struct to
9305 a subclass of rtl_opt_pass along with...
9306 (pass_data_lower_subreg): ...new pass_data instance and...
9307 (make_pass_lower_subreg): ...new function.
9308 (pass_lower_subreg2): Convert from a global struct to a subclass of
9309 rtl_opt_pass along with...
9310 (pass_data_lower_subreg2): ...new pass_data instance and...
9311 (make_pass_lower_subreg2): ...new function.
9312 * lto-streamer-out.c (pass_ipa_lto_gimple_out): Convert from a global
9313 struct to a subclass of ipa_opt_pass_d along with...
9314 (pass_data_ipa_lto_gimple_out): ...new pass_data instance and...
9315 (make_pass_ipa_lto_gimple_out): ...new function.
9316 (pass_ipa_lto_finish_out): Convert from a global struct to a subclass
9317 of ipa_opt_pass_d along with...
9318 (pass_data_ipa_lto_finish_out): ...new pass_data instance and...
9319 (make_pass_ipa_lto_finish_out): ...new function.
9320 * mode-switching.c (pass_mode_switching): Convert from a global struct
9321 to a subclass of rtl_opt_pass along with...
9322 (pass_data_mode_switching): ...new pass_data instance and...
9323 (make_pass_mode_switching): ...new function.
9324 * modulo-sched.c (pass_sms): Convert from a global struct to a
9325 subclass of rtl_opt_pass along with...
9326 (pass_data_sms): ...new pass_data instance and...
9327 (make_pass_sms): ...new function.
9328 * omp-low.c (pass_expand_omp): Convert from a global struct to a
9329 subclass of gimple_opt_pass along with...
9330 (pass_data_expand_omp): ...new pass_data instance and...
9331 (make_pass_expand_omp): ...new function.
9332 (pass_lower_omp): Convert from a global struct to a subclass of
9333 gimple_opt_pass along with...
9334 (pass_data_lower_omp): ...new pass_data instance and...
9335 (make_pass_lower_omp): ...new function.
9336 (pass_diagnose_omp_blocks): Convert from a global struct to a subclass
9337 of gimple_opt_pass along with...
9338 (pass_data_diagnose_omp_blocks): ...new pass_data instance and...
9339 (make_pass_diagnose_omp_blocks): ...new function.
9340 * passes.c (pass_early_local_passes): Convert from a global struct to
9341 a subclass of simple_ipa_opt_pass along with...
9342 (pass_data_early_local_passes): ...new pass_data instance and...
9343 (make_pass_early_local_passes): ...new function.
9344 (pass_all_early_optimizations): Convert from a global struct to a
9345 subclass of gimple_opt_pass along with...
9346 (pass_data_all_early_optimizations): ...new pass_data instance and...
9347 (make_pass_all_early_optimizations): ...new function.
9348 (pass_all_optimizations): Convert from a global struct to a subclass
9349 of gimple_opt_pass along with...
9350 (pass_data_all_optimizations): ...new pass_data instance and...
9351 (make_pass_all_optimizations): ...new function.
9352 (pass_all_optimizations_g): Convert from a global struct to a subclass
9353 of gimple_opt_pass along with...
9354 (pass_data_all_optimizations_g): ...new pass_data instance and...
9355 (make_pass_all_optimizations_g): ...new function.
9356 (pass_rest_of_compilation): Convert from a global struct to a subclass
9357 of rtl_opt_pass along with...
9358 (pass_data_rest_of_compilation): ...new pass_data instance and...
9359 (make_pass_rest_of_compilation): ...new function.
9360 (pass_postreload): Convert from a global struct to a subclass of
9361 rtl_opt_pass along with...
9362 (pass_data_postreload): ...new pass_data instance and...
9363 (make_pass_postreload): ...new function.
9364 * postreload-gcse.c (pass_gcse2): Convert from a global struct to a
9365 subclass of rtl_opt_pass along with...
9366 (pass_data_gcse2): ...new pass_data instance and...
9367 (make_pass_gcse2): ...new function.
9368 * postreload.c (pass_postreload_cse): Convert from a global struct to
9369 a subclass of rtl_opt_pass along with...
9370 (pass_data_postreload_cse): ...new pass_data instance and...
9371 (make_pass_postreload_cse): ...new function.
9372 * predict.c (pass_profile): Convert from a global struct to a subclass
9373 of gimple_opt_pass along with...
9374 (pass_data_profile): ...new pass_data instance and...
9375 (make_pass_profile): ...new function.
9376 (pass_strip_predict_hints): Convert from a global struct to a subclass
9377 of gimple_opt_pass along with...
9378 (pass_data_strip_predict_hints): ...new pass_data instance and...
9379 (make_pass_strip_predict_hints): ...new function.
9380 * recog.c (pass_peephole2): Convert from a global struct to a subclass
9381 of rtl_opt_pass along with...
9382 (pass_data_peephole2): ...new pass_data instance and...
9383 (make_pass_peephole2): ...new function.
9384 (pass_split_all_insns): Convert from a global struct to a subclass of
9385 rtl_opt_pass along with...
9386 (pass_data_split_all_insns): ...new pass_data instance and...
9387 (make_pass_split_all_insns): ...new function.
9388 (pass_split_after_reload): Convert from a global struct to a subclass
9389 of rtl_opt_pass along with...
9390 (pass_data_split_after_reload): ...new pass_data instance and...
9391 (make_pass_split_after_reload): ...new function.
9392 (pass_split_before_regstack): Convert from a global struct to a
9393 subclass of rtl_opt_pass along with...
9394 (pass_data_split_before_regstack): ...new pass_data instance and...
9395 (make_pass_split_before_regstack): ...new function.
9396 (pass_split_before_sched2): Convert from a global struct to a subclass
9397 of rtl_opt_pass along with...
9398 (pass_data_split_before_sched2): ...new pass_data instance and...
9399 (make_pass_split_before_sched2): ...new function.
9400 (pass_split_for_shorten_branches): Convert from a global struct to a
9401 subclass of rtl_opt_pass along with...
9402 (pass_data_split_for_shorten_branches): ...new pass_data instance
9403 and...
9404 (make_pass_split_for_shorten_branches): ...new function.
9405 * ree.c (pass_ree): Convert from a global struct to a subclass of
9406 rtl_opt_pass along with...
9407 (pass_data_ree): ...new pass_data instance and...
9408 (make_pass_ree): ...new function.
9409 * reg-stack.c (pass_stack_regs): Convert from a global struct to a
9410 subclass of rtl_opt_pass along with...
9411 (pass_data_stack_regs): ...new pass_data instance and...
9412 (make_pass_stack_regs): ...new function.
9413 (pass_stack_regs_run): Convert from a global struct to a subclass of
9414 rtl_opt_pass along with...
9415 (pass_data_stack_regs_run): ...new pass_data instance and...
9416 (make_pass_stack_regs_run): ...new function.
9417 * regcprop.c (pass_cprop_hardreg): Convert from a global struct to a
9418 subclass of rtl_opt_pass along with...
9419 (pass_data_cprop_hardreg): ...new pass_data instance and...
9420 (make_pass_cprop_hardreg): ...new function.
9421 * reginfo.c (pass_reginfo_init): Convert from a global struct to a
9422 subclass of rtl_opt_pass along with...
9423 (pass_data_reginfo_init): ...new pass_data instance and...
9424 (make_pass_reginfo_init): ...new function.
9425 * regmove.c (pass_regmove): Convert from a global struct to a subclass
9426 of rtl_opt_pass along with...
9427 (pass_data_regmove): ...new pass_data instance and...
9428 (make_pass_regmove): ...new function.
9429 * regrename.c (pass_regrename): Convert from a global struct to a
9430 subclass of rtl_opt_pass along with...
9431 (pass_data_regrename): ...new pass_data instance and...
9432 (make_pass_regrename): ...new function.
9433 * reorg.c (pass_delay_slots): Convert from a global struct to a
9434 subclass of rtl_opt_pass along with...
9435 (pass_data_delay_slots): ...new pass_data instance and...
9436 (make_pass_delay_slots): ...new function.
9437 (pass_machine_reorg): Convert from a global struct to a subclass of
9438 rtl_opt_pass along with...
9439 (pass_data_machine_reorg): ...new pass_data instance and...
9440 (make_pass_machine_reorg): ...new function.
9441 * sched-rgn.c (pass_sched): Convert from a global struct to a subclass
9442 of rtl_opt_pass along with...
9443 (pass_data_sched): ...new pass_data instance and...
9444 (make_pass_sched): ...new function.
9445 (pass_sched2): Convert from a global struct to a subclass of
9446 rtl_opt_pass along with...
9447 (pass_data_sched2): ...new pass_data instance and...
9448 (make_pass_sched2): ...new function.
9449 * stack-ptr-mod.c (pass_stack_ptr_mod): Convert from a global struct
9450 to a subclass of rtl_opt_pass along with...
9451 (pass_data_stack_ptr_mod): ...new pass_data instance and...
9452 (make_pass_stack_ptr_mod): ...new function.
9453 * store-motion.c (pass_rtl_store_motion): Convert from a global struct
9454 to a subclass of rtl_opt_pass along with...
9455 (pass_data_rtl_store_motion): ...new pass_data instance and...
9456 (make_pass_rtl_store_motion): ...new function.
9457 * tracer.c (pass_tracer): Convert from a global struct to a subclass
9458 of gimple_opt_pass along with...
9459 (pass_data_tracer): ...new pass_data instance and...
9460 (make_pass_tracer): ...new function.
9461 * trans-mem.c (pass_diagnose_tm_blocks): Convert from a global struct
9462 to a subclass of gimple_opt_pass along with...
9463 (pass_data_diagnose_tm_blocks): ...new pass_data instance and...
9464 (make_pass_diagnose_tm_blocks): ...new function.
9465 (pass_lower_tm): Convert from a global struct to a subclass of
9466 gimple_opt_pass along with...
9467 (pass_data_lower_tm): ...new pass_data instance and...
9468 (make_pass_lower_tm): ...new function.
9469 (pass_tm_init): Convert from a global struct to a subclass of
9470 gimple_opt_pass along with...
9471 (pass_data_tm_init): ...new pass_data instance and...
9472 (make_pass_tm_init): ...new function.
9473 (pass_tm_mark): Convert from a global struct to a subclass of
9474 gimple_opt_pass along with...
9475 (pass_data_tm_mark): ...new pass_data instance and...
9476 (make_pass_tm_mark): ...new function.
9477 (pass_tm_edges): Convert from a global struct to a subclass of
9478 gimple_opt_pass along with...
9479 (pass_data_tm_edges): ...new pass_data instance and...
9480 (make_pass_tm_edges): ...new function.
9481 (pass_tm_memopt): Convert from a global struct to a subclass of
9482 gimple_opt_pass along with...
9483 (pass_data_tm_memopt): ...new pass_data instance and...
9484 (make_pass_tm_memopt): ...new function.
9485 (pass_ipa_tm): Convert from a global struct to a subclass of
9486 simple_ipa_opt_pass along with...
9487 (pass_data_ipa_tm): ...new pass_data instance and...
9488 (make_pass_ipa_tm): ...new function.
9489 * tree-call-cdce.c (pass_call_cdce): Convert from a global struct to a
9490 subclass of gimple_opt_pass along with...
9491 (pass_data_call_cdce): ...new pass_data instance and...
9492 (make_pass_call_cdce): ...new function.
9493 * tree-cfg.c (pass_build_cfg): Convert from a global struct to a
9494 subclass of gimple_opt_pass along with...
9495 (pass_data_build_cfg): ...new pass_data instance and...
9496 (make_pass_build_cfg): ...new function.
9497 (pass_split_crit_edges): Convert from a global struct to a subclass of
9498 gimple_opt_pass along with...
9499 (pass_data_split_crit_edges): ...new pass_data instance and...
9500 (make_pass_split_crit_edges): ...new function.
9501 (pass_warn_function_return): Convert from a global struct to a
9502 subclass of gimple_opt_pass along with...
9503 (pass_data_warn_function_return): ...new pass_data instance and...
9504 (make_pass_warn_function_return): ...new function.
9505 (pass_warn_function_noreturn): Convert from a global struct to a
9506 subclass of gimple_opt_pass along with...
9507 (pass_data_warn_function_noreturn): ...new pass_data instance and...
9508 (make_pass_warn_function_noreturn): ...new function.
9509 (pass_warn_unused_result): Convert from a global struct to a subclass
9510 of gimple_opt_pass along with...
9511 (pass_data_warn_unused_result): ...new pass_data instance and...
9512 (make_pass_warn_unused_result): ...new function.
9513 * tree-cfgcleanup.c (pass_merge_phi): Convert from a global struct to
9514 a subclass of gimple_opt_pass along with...
9515 (pass_data_merge_phi): ...new pass_data instance and...
9516 (make_pass_merge_phi): ...new function.
9517 * tree-complex.c (pass_lower_complex): Convert from a global struct to
9518 a subclass of gimple_opt_pass along with...
9519 (pass_data_lower_complex): ...new pass_data instance and...
9520 (make_pass_lower_complex): ...new function.
9521 (pass_lower_complex_O0): Convert from a global struct to a subclass of
9522 gimple_opt_pass along with...
9523 (pass_data_lower_complex_O0): ...new pass_data instance and...
9524 (make_pass_lower_complex_O0): ...new function.
9525 * tree-eh.c (pass_lower_eh): Convert from a global struct to a
9526 subclass of gimple_opt_pass along with...
9527 (pass_data_lower_eh): ...new pass_data instance and...
9528 (make_pass_lower_eh): ...new function.
9529 (pass_refactor_eh): Convert from a global struct to a subclass of
9530 gimple_opt_pass along with...
9531 (pass_data_refactor_eh): ...new pass_data instance and...
9532 (make_pass_refactor_eh): ...new function.
9533 (pass_lower_resx): Convert from a global struct to a subclass of
9534 gimple_opt_pass along with...
9535 (pass_data_lower_resx): ...new pass_data instance and...
9536 (make_pass_lower_resx): ...new function.
9537 (pass_lower_eh_dispatch): Convert from a global struct to a subclass
9538 of gimple_opt_pass along with...
9539 (pass_data_lower_eh_dispatch): ...new pass_data instance and...
9540 (make_pass_lower_eh_dispatch): ...new function.
9541 (pass_cleanup_eh): Convert from a global struct to a subclass of
9542 gimple_opt_pass along with...
9543 (pass_data_cleanup_eh): ...new pass_data instance and...
9544 (make_pass_cleanup_eh): ...new function.
9545 * tree-emutls.c (pass_ipa_lower_emutls): Convert from a global struct
9546 to a subclass of simple_ipa_opt_pass along with...
9547 (pass_data_ipa_lower_emutls): ...new pass_data instance and...
9548 (make_pass_ipa_lower_emutls): ...new function.
9549 * tree-if-conv.c (pass_if_conversion): Convert from a global struct to
9550 a subclass of gimple_opt_pass along with...
9551 (pass_data_if_conversion): ...new pass_data instance and...
9552 (make_pass_if_conversion): ...new function.
9553 * tree-into-ssa.c (pass_build_ssa): Convert from a global struct to a
9554 subclass of gimple_opt_pass along with...
9555 (pass_data_build_ssa): ...new pass_data instance and...
9556 (make_pass_build_ssa): ...new function.
9557 * tree-loop-distribution.c (pass_loop_distribution): Convert from a
9558 global struct to a subclass of gimple_opt_pass along with...
9559 (pass_data_loop_distribution): ...new pass_data instance and...
9560 (make_pass_loop_distribution): ...new function.
9561 * tree-mudflap.c (pass_mudflap_1): Convert from a global struct to a
9562 subclass of gimple_opt_pass along with...
9563 (pass_data_mudflap_1): ...new pass_data instance and...
9564 (make_pass_mudflap_1): ...new function.
9565 (pass_mudflap_2): Convert from a global struct to a subclass of
9566 gimple_opt_pass along with...
9567 (pass_data_mudflap_2): ...new pass_data instance and...
9568 (make_pass_mudflap_2): ...new function.
9569 * tree-nomudflap.c (pass_mudflap_1): Convert from a global struct to a
9570 subclass of gimple_opt_pass along with...
9571 (pass_data_mudflap_1): ...new pass_data instance and...
9572 (make_pass_mudflap_1): ...new function.
9573 (pass_mudflap_2): Convert from a global struct to a subclass of
9574 gimple_opt_pass along with...
9575 (pass_data_mudflap_2): ...new pass_data instance and...
9576 (make_pass_mudflap_2): ...new function.
9577 * tree-nrv.c (pass_nrv): Convert from a global struct to a subclass of
9578 gimple_opt_pass along with...
9579 (pass_data_nrv): ...new pass_data instance and...
9580 (make_pass_nrv): ...new function.
9581 (pass_return_slot): Convert from a global struct to a subclass of
9582 gimple_opt_pass along with...
9583 (pass_data_return_slot): ...new pass_data instance and...
9584 (make_pass_return_slot): ...new function.
9585 * tree-object-size.c (pass_object_sizes): Convert from a global struct
9586 to a subclass of gimple_opt_pass along with...
9587 (pass_data_object_sizes): ...new pass_data instance and...
9588 (make_pass_object_sizes): ...new function.
9589 * tree-optimize.c (pass_cleanup_cfg_post_optimizing): Convert from a
9590 global struct to a subclass of gimple_opt_pass along with...
9591 (pass_data_cleanup_cfg_post_optimizing): ...new pass_data instance
9592 and...
9593 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9594 (pass_fixup_cfg): Convert from a global struct to a subclass of
9595 gimple_opt_pass along with...
9596 (pass_data_fixup_cfg): ...new pass_data instance and...
9597 (make_pass_fixup_cfg): ...new function.
9598 * tree-pass.h (pass_mudflap_1): Replace declaration with that of...
9599 (make_pass_mudflap_1): ...new function.
9600 (pass_mudflap_2): Replace declaration with that of...
9601 (make_pass_mudflap_2): ...new function.
9602 (pass_asan): Replace declaration with that of...
9603 (make_pass_asan): ...new function.
9604 (pass_asan_O0): Replace declaration with that of...
9605 (make_pass_asan_O0): ...new function.
9606 (pass_tsan): Replace declaration with that of...
9607 (make_pass_tsan): ...new function.
9608 (pass_tsan_O0): Replace declaration with that of...
9609 (make_pass_tsan_O0): ...new function.
9610 (pass_lower_cf): Replace declaration with that of...
9611 (make_pass_lower_cf): ...new function.
9612 (pass_refactor_eh): Replace declaration with that of...
9613 (make_pass_refactor_eh): ...new function.
9614 (pass_lower_eh): Replace declaration with that of...
9615 (make_pass_lower_eh): ...new function.
9616 (pass_lower_eh_dispatch): Replace declaration with that of...
9617 (make_pass_lower_eh_dispatch): ...new function.
9618 (pass_lower_resx): Replace declaration with that of...
9619 (make_pass_lower_resx): ...new function.
9620 (pass_build_cfg): Replace declaration with that of...
9621 (make_pass_build_cfg): ...new function.
9622 (pass_early_tree_profile): Replace declaration with that of...
9623 (make_pass_early_tree_profile): ...new function.
9624 (pass_cleanup_eh): Replace declaration with that of...
9625 (make_pass_cleanup_eh): ...new function.
9626 (pass_sra): Replace declaration with that of...
9627 (make_pass_sra): ...new function.
9628 (pass_sra_early): Replace declaration with that of...
9629 (make_pass_sra_early): ...new function.
9630 (pass_early_ipa_sra): Replace declaration with that of...
9631 (make_pass_early_ipa_sra): ...new function.
9632 (pass_tail_recursion): Replace declaration with that of...
9633 (make_pass_tail_recursion): ...new function.
9634 (pass_tail_calls): Replace declaration with that of...
9635 (make_pass_tail_calls): ...new function.
9636 (pass_tree_loop): Replace declaration with that of...
9637 (make_pass_tree_loop): ...new function.
9638 (pass_tree_loop_init): Replace declaration with that of...
9639 (make_pass_tree_loop_init): ...new function.
9640 (pass_lim): Replace declaration with that of...
9641 (make_pass_lim): ...new function.
9642 (pass_tree_unswitch): Replace declaration with that of...
9643 (make_pass_tree_unswitch): ...new function.
9644 (pass_predcom): Replace declaration with that of...
9645 (make_pass_predcom): ...new function.
9646 (pass_iv_canon): Replace declaration with that of...
9647 (make_pass_iv_canon): ...new function.
9648 (pass_scev_cprop): Replace declaration with that of...
9649 (make_pass_scev_cprop): ...new function.
9650 (pass_empty_loop): Replace declaration with that of...
9651 (make_pass_empty_loop): ...new function.
9652 (pass_record_bounds): Replace declaration with that of...
9653 (make_pass_record_bounds): ...new function.
9654 (pass_graphite): Replace declaration with that of...
9655 (make_pass_graphite): ...new function.
9656 (pass_graphite_transforms): Replace declaration with that of...
9657 (make_pass_graphite_transforms): ...new function.
9658 (pass_if_conversion): Replace declaration with that of...
9659 (make_pass_if_conversion): ...new function.
9660 (pass_loop_distribution): Replace declaration with that of...
9661 (make_pass_loop_distribution): ...new function.
9662 (pass_vectorize): Replace declaration with that of...
9663 (make_pass_vectorize): ...new function.
9664 (pass_slp_vectorize): Replace declaration with that of...
9665 (make_pass_slp_vectorize): ...new function.
9666 (pass_complete_unroll): Replace declaration with that of...
9667 (make_pass_complete_unroll): ...new function.
9668 (pass_complete_unrolli): Replace declaration with that of...
9669 (make_pass_complete_unrolli): ...new function.
9670 (pass_parallelize_loops): Replace declaration with that of...
9671 (make_pass_parallelize_loops): ...new function.
9672 (pass_loop_prefetch): Replace declaration with that of...
9673 (make_pass_loop_prefetch): ...new function.
9674 (pass_iv_optimize): Replace declaration with that of...
9675 (make_pass_iv_optimize): ...new function.
9676 (pass_tree_loop_done): Replace declaration with that of...
9677 (make_pass_tree_loop_done): ...new function.
9678 (pass_ch): Replace declaration with that of...
9679 (make_pass_ch): ...new function.
9680 (pass_ccp): Replace declaration with that of...
9681 (make_pass_ccp): ...new function.
9682 (pass_phi_only_cprop): Replace declaration with that of...
9683 (make_pass_phi_only_cprop): ...new function.
9684 (pass_build_ssa): Replace declaration with that of...
9685 (make_pass_build_ssa): ...new function.
9686 (pass_build_alias): Replace declaration with that of...
9687 (make_pass_build_alias): ...new function.
9688 (pass_build_ealias): Replace declaration with that of...
9689 (make_pass_build_ealias): ...new function.
9690 (pass_dominator): Replace declaration with that of...
9691 (make_pass_dominator): ...new function.
9692 (pass_dce): Replace declaration with that of...
9693 (make_pass_dce): ...new function.
9694 (pass_dce_loop): Replace declaration with that of...
9695 (make_pass_dce_loop): ...new function.
9696 (pass_cd_dce): Replace declaration with that of...
9697 (make_pass_cd_dce): ...new function.
9698 (pass_call_cdce): Replace declaration with that of...
9699 (make_pass_call_cdce): ...new function.
9700 (pass_merge_phi): Replace declaration with that of...
9701 (make_pass_merge_phi): ...new function.
9702 (pass_split_crit_edges): Replace declaration with that of...
9703 (make_pass_split_crit_edges): ...new function.
9704 (pass_pre): Replace declaration with that of...
9705 (make_pass_pre): ...new function.
9706 (pass_profile): Replace declaration with that of...
9707 (make_pass_profile): ...new function.
9708 (pass_strip_predict_hints): Replace declaration with that of...
9709 (make_pass_strip_predict_hints): ...new function.
9710 (pass_lower_complex_O0): Replace declaration with that of...
9711 (make_pass_lower_complex_O0): ...new function.
9712 (pass_lower_complex): Replace declaration with that of...
9713 (make_pass_lower_complex): ...new function.
9714 (pass_lower_vector): Replace declaration with that of...
9715 (make_pass_lower_vector): ...new function.
9716 (pass_lower_vector_ssa): Replace declaration with that of...
9717 (make_pass_lower_vector_ssa): ...new function.
9718 (pass_lower_omp): Replace declaration with that of...
9719 (make_pass_lower_omp): ...new function.
9720 (pass_diagnose_omp_blocks): Replace declaration with that of...
9721 (make_pass_diagnose_omp_blocks): ...new function.
9722 (pass_expand_omp): Replace declaration with that of...
9723 (make_pass_expand_omp): ...new function.
9724 (pass_expand_omp_ssa): Replace declaration with that of...
9725 (make_pass_expand_omp_ssa): ...new function.
9726 (pass_object_sizes): Replace declaration with that of...
9727 (make_pass_object_sizes): ...new function.
9728 (pass_strlen): Replace declaration with that of...
9729 (make_pass_strlen): ...new function.
9730 (pass_fold_builtins): Replace declaration with that of...
9731 (make_pass_fold_builtins): ...new function.
9732 (pass_stdarg): Replace declaration with that of...
9733 (make_pass_stdarg): ...new function.
9734 (pass_early_warn_uninitialized): Replace declaration with that of...
9735 (make_pass_early_warn_uninitialized): ...new function.
9736 (pass_late_warn_uninitialized): Replace declaration with that of...
9737 (make_pass_late_warn_uninitialized): ...new function.
9738 (pass_cse_reciprocals): Replace declaration with that of...
9739 (make_pass_cse_reciprocals): ...new function.
9740 (pass_cse_sincos): Replace declaration with that of...
9741 (make_pass_cse_sincos): ...new function.
9742 (pass_optimize_bswap): Replace declaration with that of...
9743 (make_pass_optimize_bswap): ...new function.
9744 (pass_optimize_widening_mul): Replace declaration with that of...
9745 (make_pass_optimize_widening_mul): ...new function.
9746 (pass_warn_function_return): Replace declaration with that of...
9747 (make_pass_warn_function_return): ...new function.
9748 (pass_warn_function_noreturn): Replace declaration with that of...
9749 (make_pass_warn_function_noreturn): ...new function.
9750 (pass_cselim): Replace declaration with that of...
9751 (make_pass_cselim): ...new function.
9752 (pass_phiopt): Replace declaration with that of...
9753 (make_pass_phiopt): ...new function.
9754 (pass_forwprop): Replace declaration with that of...
9755 (make_pass_forwprop): ...new function.
9756 (pass_phiprop): Replace declaration with that of...
9757 (make_pass_phiprop): ...new function.
9758 (pass_tree_ifcombine): Replace declaration with that of...
9759 (make_pass_tree_ifcombine): ...new function.
9760 (pass_dse): Replace declaration with that of...
9761 (make_pass_dse): ...new function.
9762 (pass_nrv): Replace declaration with that of...
9763 (make_pass_nrv): ...new function.
9764 (pass_rename_ssa_copies): Replace declaration with that of...
9765 (make_pass_rename_ssa_copies): ...new function.
9766 (pass_sink_code): Replace declaration with that of...
9767 (make_pass_sink_code): ...new function.
9768 (pass_fre): Replace declaration with that of...
9769 (make_pass_fre): ...new function.
9770 (pass_check_data_deps): Replace declaration with that of...
9771 (make_pass_check_data_deps): ...new function.
9772 (pass_copy_prop): Replace declaration with that of...
9773 (make_pass_copy_prop): ...new function.
9774 (pass_vrp): Replace declaration with that of...
9775 (make_pass_vrp): ...new function.
9776 (pass_uncprop): Replace declaration with that of...
9777 (make_pass_uncprop): ...new function.
9778 (pass_return_slot): Replace declaration with that of...
9779 (make_pass_return_slot): ...new function.
9780 (pass_reassoc): Replace declaration with that of...
9781 (make_pass_reassoc): ...new function.
9782 (pass_rebuild_cgraph_edges): Replace declaration with that of...
9783 (make_pass_rebuild_cgraph_edges): ...new function.
9784 (pass_remove_cgraph_callee_edges): Replace declaration with that of...
9785 (make_pass_remove_cgraph_callee_edges): ...new function.
9786 (pass_build_cgraph_edges): Replace declaration with that of...
9787 (make_pass_build_cgraph_edges): ...new function.
9788 (pass_local_pure_const): Replace declaration with that of...
9789 (make_pass_local_pure_const): ...new function.
9790 (pass_tracer): Replace declaration with that of...
9791 (make_pass_tracer): ...new function.
9792 (pass_warn_unused_result): Replace declaration with that of...
9793 (make_pass_warn_unused_result): ...new function.
9794 (pass_diagnose_tm_blocks): Replace declaration with that of...
9795 (make_pass_diagnose_tm_blocks): ...new function.
9796 (pass_lower_tm): Replace declaration with that of...
9797 (make_pass_lower_tm): ...new function.
9798 (pass_tm_init): Replace declaration with that of...
9799 (make_pass_tm_init): ...new function.
9800 (pass_tm_mark): Replace declaration with that of...
9801 (make_pass_tm_mark): ...new function.
9802 (pass_tm_memopt): Replace declaration with that of...
9803 (make_pass_tm_memopt): ...new function.
9804 (pass_tm_edges): Replace declaration with that of...
9805 (make_pass_tm_edges): ...new function.
9806 (pass_split_functions): Replace declaration with that of...
9807 (make_pass_split_functions): ...new function.
9808 (pass_feedback_split_functions): Replace declaration with that of...
9809 (make_pass_feedback_split_functions): ...new function.
9810 (pass_strength_reduction): Replace declaration with that of...
9811 (make_pass_strength_reduction): ...new function.
9812 (pass_ipa_lower_emutls): Replace declaration with that of...
9813 (make_pass_ipa_lower_emutls): ...new function.
9814 (pass_ipa_function_and_variable_visibility): Replace declaration with
9815 that of...
9816 (make_pass_ipa_function_and_variable_visibility): ...new function.
9817 (pass_ipa_tree_profile): Replace declaration with that of...
9818 (make_pass_ipa_tree_profile): ...new function.
9819 (pass_early_local_passes): Replace declaration with that of...
9820 (make_pass_early_local_passes): ...new function.
9821 (pass_ipa_whole_program_visibility): Replace declaration with that
9822 of...
9823 (make_pass_ipa_whole_program_visibility): ...new function.
9824 (pass_ipa_lto_gimple_out): Replace declaration with that of...
9825 (make_pass_ipa_lto_gimple_out): ...new function.
9826 (pass_ipa_increase_alignment): Replace declaration with that of...
9827 (make_pass_ipa_increase_alignment): ...new function.
9828 (pass_ipa_inline): Replace declaration with that of...
9829 (make_pass_ipa_inline): ...new function.
9830 (pass_ipa_free_lang_data): Replace declaration with that of...
9831 (make_pass_ipa_free_lang_data): ...new function.
9832 (pass_ipa_free_inline_summary): Replace declaration with that of...
9833 (make_pass_ipa_free_inline_summary): ...new function.
9834 (pass_ipa_cp): Replace declaration with that of...
9835 (make_pass_ipa_cp): ...new function.
9836 (pass_ipa_reference): Replace declaration with that of...
9837 (make_pass_ipa_reference): ...new function.
9838 (pass_ipa_pure_const): Replace declaration with that of...
9839 (make_pass_ipa_pure_const): ...new function.
9840 (pass_ipa_pta): Replace declaration with that of...
9841 (make_pass_ipa_pta): ...new function.
9842 (pass_ipa_lto_finish_out): Replace declaration with that of...
9843 (make_pass_ipa_lto_finish_out): ...new function.
9844 (pass_ipa_tm): Replace declaration with that of...
9845 (make_pass_ipa_tm): ...new function.
9846 (pass_ipa_profile): Replace declaration with that of...
9847 (make_pass_ipa_profile): ...new function.
9848 (pass_ipa_cdtor_merge): Replace declaration with that of...
9849 (make_pass_ipa_cdtor_merge): ...new function.
9850 (pass_cleanup_cfg_post_optimizing): Replace declaration with that
9851 of...
9852 (make_pass_cleanup_cfg_post_optimizing): ...new function.
9853 (pass_init_datastructures): Replace declaration with that of...
9854 (make_pass_init_datastructures): ...new function.
9855 (pass_fixup_cfg): Replace declaration with that of...
9856 (make_pass_fixup_cfg): ...new function.
9857 (pass_expand): Replace declaration with that of...
9858 (make_pass_expand): ...new function.
9859 (pass_instantiate_virtual_regs): Replace declaration with that of...
9860 (make_pass_instantiate_virtual_regs): ...new function.
9861 (pass_rtl_fwprop): Replace declaration with that of...
9862 (make_pass_rtl_fwprop): ...new function.
9863 (pass_rtl_fwprop_addr): Replace declaration with that of...
9864 (make_pass_rtl_fwprop_addr): ...new function.
9865 (pass_jump): Replace declaration with that of...
9866 (make_pass_jump): ...new function.
9867 (pass_jump2): Replace declaration with that of...
9868 (make_pass_jump2): ...new function.
9869 (pass_lower_subreg): Replace declaration with that of...
9870 (make_pass_lower_subreg): ...new function.
9871 (pass_cse): Replace declaration with that of...
9872 (make_pass_cse): ...new function.
9873 (pass_fast_rtl_dce): Replace declaration with that of...
9874 (make_pass_fast_rtl_dce): ...new function.
9875 (pass_ud_rtl_dce): Replace declaration with that of...
9876 (make_pass_ud_rtl_dce): ...new function.
9877 (pass_rtl_dce): Replace declaration with that of...
9878 (make_pass_rtl_dce): ...new function.
9879 (pass_rtl_dse1): Replace declaration with that of...
9880 (make_pass_rtl_dse1): ...new function.
9881 (pass_rtl_dse2): Replace declaration with that of...
9882 (make_pass_rtl_dse2): ...new function.
9883 (pass_rtl_dse3): Replace declaration with that of...
9884 (make_pass_rtl_dse3): ...new function.
9885 (pass_rtl_cprop): Replace declaration with that of...
9886 (make_pass_rtl_cprop): ...new function.
9887 (pass_rtl_pre): Replace declaration with that of...
9888 (make_pass_rtl_pre): ...new function.
9889 (pass_rtl_hoist): Replace declaration with that of...
9890 (make_pass_rtl_hoist): ...new function.
9891 (pass_rtl_store_motion): Replace declaration with that of...
9892 (make_pass_rtl_store_motion): ...new function.
9893 (pass_cse_after_global_opts): Replace declaration with that of...
9894 (make_pass_cse_after_global_opts): ...new function.
9895 (pass_rtl_ifcvt): Replace declaration with that of...
9896 (make_pass_rtl_ifcvt): ...new function.
9897 (pass_into_cfg_layout_mode): Replace declaration with that of...
9898 (make_pass_into_cfg_layout_mode): ...new function.
9899 (pass_outof_cfg_layout_mode): Replace declaration with that of...
9900 (make_pass_outof_cfg_layout_mode): ...new function.
9901 (pass_loop2): Replace declaration with that of...
9902 (make_pass_loop2): ...new function.
9903 (pass_rtl_loop_init): Replace declaration with that of...
9904 (make_pass_rtl_loop_init): ...new function.
9905 (pass_rtl_move_loop_invariants): Replace declaration with that of...
9906 (make_pass_rtl_move_loop_invariants): ...new function.
9907 (pass_rtl_unswitch): Replace declaration with that of...
9908 (make_pass_rtl_unswitch): ...new function.
9909 (pass_rtl_unroll_and_peel_loops): Replace declaration with that of...
9910 (make_pass_rtl_unroll_and_peel_loops): ...new function.
9911 (pass_rtl_doloop): Replace declaration with that of...
9912 (make_pass_rtl_doloop): ...new function.
9913 (pass_rtl_loop_done): Replace declaration with that of...
9914 (make_pass_rtl_loop_done): ...new function.
9915 (pass_web): Replace declaration with that of...
9916 (make_pass_web): ...new function.
9917 (pass_cse2): Replace declaration with that of...
9918 (make_pass_cse2): ...new function.
9919 (pass_df_initialize_opt): Replace declaration with that of...
9920 (make_pass_df_initialize_opt): ...new function.
9921 (pass_df_initialize_no_opt): Replace declaration with that of...
9922 (make_pass_df_initialize_no_opt): ...new function.
9923 (pass_reginfo_init): Replace declaration with that of...
9924 (make_pass_reginfo_init): ...new function.
9925 (pass_inc_dec): Replace declaration with that of...
9926 (make_pass_inc_dec): ...new function.
9927 (pass_stack_ptr_mod): Replace declaration with that of...
9928 (make_pass_stack_ptr_mod): ...new function.
9929 (pass_initialize_regs): Replace declaration with that of...
9930 (make_pass_initialize_regs): ...new function.
9931 (pass_combine): Replace declaration with that of...
9932 (make_pass_combine): ...new function.
9933 (pass_if_after_combine): Replace declaration with that of...
9934 (make_pass_if_after_combine): ...new function.
9935 (pass_ree): Replace declaration with that of...
9936 (make_pass_ree): ...new function.
9937 (pass_partition_blocks): Replace declaration with that of...
9938 (make_pass_partition_blocks): ...new function.
9939 (pass_match_asm_constraints): Replace declaration with that of...
9940 (make_pass_match_asm_constraints): ...new function.
9941 (pass_regmove): Replace declaration with that of...
9942 (make_pass_regmove): ...new function.
9943 (pass_split_all_insns): Replace declaration with that of...
9944 (make_pass_split_all_insns): ...new function.
9945 (pass_fast_rtl_byte_dce): Replace declaration with that of...
9946 (make_pass_fast_rtl_byte_dce): ...new function.
9947 (pass_lower_subreg2): Replace declaration with that of...
9948 (make_pass_lower_subreg2): ...new function.
9949 (pass_mode_switching): Replace declaration with that of...
9950 (make_pass_mode_switching): ...new function.
9951 (pass_sms): Replace declaration with that of...
9952 (make_pass_sms): ...new function.
9953 (pass_sched): Replace declaration with that of...
9954 (make_pass_sched): ...new function.
9955 (pass_ira): Replace declaration with that of...
9956 (make_pass_ira): ...new function.
9957 (pass_reload): Replace declaration with that of...
9958 (make_pass_reload): ...new function.
9959 (pass_clean_state): Replace declaration with that of...
9960 (make_pass_clean_state): ...new function.
9961 (pass_branch_prob): Replace declaration with that of...
9962 (make_pass_branch_prob): ...new function.
9963 (pass_value_profile_transformations): Replace declaration with that
9964 of...
9965 (make_pass_value_profile_transformations): ...new function.
9966 (pass_postreload_cse): Replace declaration with that of...
9967 (make_pass_postreload_cse): ...new function.
9968 (pass_gcse2): Replace declaration with that of...
9969 (make_pass_gcse2): ...new function.
9970 (pass_split_after_reload): Replace declaration with that of...
9971 (make_pass_split_after_reload): ...new function.
9972 (pass_branch_target_load_optimize1): Replace declaration with that
9973 of...
9974 (make_pass_branch_target_load_optimize1): ...new function.
9975 (pass_thread_prologue_and_epilogue): Replace declaration with that
9976 of...
9977 (make_pass_thread_prologue_and_epilogue): ...new function.
9978 (pass_stack_adjustments): Replace declaration with that of...
9979 (make_pass_stack_adjustments): ...new function.
9980 (pass_peephole2): Replace declaration with that of...
9981 (make_pass_peephole2): ...new function.
9982 (pass_if_after_reload): Replace declaration with that of...
9983 (make_pass_if_after_reload): ...new function.
9984 (pass_regrename): Replace declaration with that of...
9985 (make_pass_regrename): ...new function.
9986 (pass_cprop_hardreg): Replace declaration with that of...
9987 (make_pass_cprop_hardreg): ...new function.
9988 (pass_reorder_blocks): Replace declaration with that of...
9989 (make_pass_reorder_blocks): ...new function.
9990 (pass_branch_target_load_optimize2): Replace declaration with that
9991 of...
9992 (make_pass_branch_target_load_optimize2): ...new function.
9993 (pass_leaf_regs): Replace declaration with that of...
9994 (make_pass_leaf_regs): ...new function.
9995 (pass_split_before_sched2): Replace declaration with that of...
9996 (make_pass_split_before_sched2): ...new function.
9997 (pass_compare_elim_after_reload): Replace declaration with that of...
9998 (make_pass_compare_elim_after_reload): ...new function.
9999 (pass_sched2): Replace declaration with that of...
10000 (make_pass_sched2): ...new function.
10001 (pass_stack_regs): Replace declaration with that of...
10002 (make_pass_stack_regs): ...new function.
10003 (pass_stack_regs_run): Replace declaration with that of...
10004 (make_pass_stack_regs_run): ...new function.
10005 (pass_df_finish): Replace declaration with that of...
10006 (make_pass_df_finish): ...new function.
10007 (pass_compute_alignments): Replace declaration with that of...
10008 (make_pass_compute_alignments): ...new function.
10009 (pass_duplicate_computed_gotos): Replace declaration with that of...
10010 (make_pass_duplicate_computed_gotos): ...new function.
10011 (pass_variable_tracking): Replace declaration with that of...
10012 (make_pass_variable_tracking): ...new function.
10013 (pass_free_cfg): Replace declaration with that of...
10014 (make_pass_free_cfg): ...new function.
10015 (pass_machine_reorg): Replace declaration with that of...
10016 (make_pass_machine_reorg): ...new function.
10017 (pass_cleanup_barriers): Replace declaration with that of...
10018 (make_pass_cleanup_barriers): ...new function.
10019 (pass_delay_slots): Replace declaration with that of...
10020 (make_pass_delay_slots): ...new function.
10021 (pass_split_for_shorten_branches): Replace declaration with that of...
10022 (make_pass_split_for_shorten_branches): ...new function.
10023 (pass_split_before_regstack): Replace declaration with that of...
10024 (make_pass_split_before_regstack): ...new function.
10025 (pass_convert_to_eh_region_ranges): Replace declaration with that
10026 of...
10027 (make_pass_convert_to_eh_region_ranges): ...new function.
10028 (pass_shorten_branches): Replace declaration with that of...
10029 (make_pass_shorten_branches): ...new function.
10030 (pass_set_nothrow_function_flags): Replace declaration with that of...
10031 (make_pass_set_nothrow_function_flags): ...new function.
10032 (pass_dwarf2_frame): Replace declaration with that of...
10033 (make_pass_dwarf2_frame): ...new function.
10034 (pass_final): Replace declaration with that of...
10035 (make_pass_final): ...new function.
10036 (pass_rtl_seqabstr): Replace declaration with that of...
10037 (make_pass_rtl_seqabstr): ...new function.
10038 (pass_release_ssa_names): Replace declaration with that of...
10039 (make_pass_release_ssa_names): ...new function.
10040 (pass_early_inline): Replace declaration with that of...
10041 (make_pass_early_inline): ...new function.
10042 (pass_inline_parameters): Replace declaration with that of...
10043 (make_pass_inline_parameters): ...new function.
10044 (pass_update_address_taken): Replace declaration with that of...
10045 (make_pass_update_address_taken): ...new function.
10046 (pass_convert_switch): Replace declaration with that of...
10047 (make_pass_convert_switch): ...new function.
10048 * tree-profile.c (pass_ipa_tree_profile): Convert from a global struct
10049 to a subclass of simple_ipa_opt_pass along with...
10050 (pass_data_ipa_tree_profile): ...new pass_data instance and...
10051 (make_pass_ipa_tree_profile): ...new function.
10052 * tree-sra.c (pass_sra_early): Convert from a global struct to a
10053 subclass of gimple_opt_pass along with...
10054 (pass_data_sra_early): ...new pass_data instance and...
10055 (make_pass_sra_early): ...new function.
10056 (pass_sra): Convert from a global struct to a subclass of
10057 gimple_opt_pass along with...
10058 (pass_data_sra): ...new pass_data instance and...
10059 (make_pass_sra): ...new function.
10060 (pass_early_ipa_sra): Convert from a global struct to a subclass of
10061 gimple_opt_pass along with...
10062 (pass_data_early_ipa_sra): ...new pass_data instance and...
10063 (make_pass_early_ipa_sra): ...new function.
10064 * tree-ssa-ccp.c (pass_ccp): Convert from a global struct to a
10065 subclass of gimple_opt_pass along with...
10066 (pass_data_ccp): ...new pass_data instance and...
10067 (make_pass_ccp): ...new function.
10068 (pass_fold_builtins): Convert from a global struct to a subclass of
10069 gimple_opt_pass along with...
10070 (pass_data_fold_builtins): ...new pass_data instance and...
10071 (make_pass_fold_builtins): ...new function.
10072 * tree-ssa-copy.c (pass_copy_prop): Convert from a global struct to a
10073 subclass of gimple_opt_pass along with...
10074 (pass_data_copy_prop): ...new pass_data instance and...
10075 (make_pass_copy_prop): ...new function.
10076 * tree-ssa-copyrename.c (pass_rename_ssa_copies): Convert from a
10077 global struct to a subclass of gimple_opt_pass along with...
10078 (pass_data_rename_ssa_copies): ...new pass_data instance and...
10079 (make_pass_rename_ssa_copies): ...new function.
10080 * tree-ssa-dce.c (pass_dce): Convert from a global struct to a
10081 subclass of gimple_opt_pass along with...
10082 (pass_data_dce): ...new pass_data instance and...
10083 (make_pass_dce): ...new function.
10084 (pass_dce_loop): Convert from a global struct to a subclass of
10085 gimple_opt_pass along with...
10086 (pass_data_dce_loop): ...new pass_data instance and...
10087 (make_pass_dce_loop): ...new function.
10088 (pass_cd_dce): Convert from a global struct to a subclass of
10089 gimple_opt_pass along with...
10090 (pass_data_cd_dce): ...new pass_data instance and...
10091 (make_pass_cd_dce): ...new function.
10092 * tree-ssa-dom.c (pass_dominator): Convert from a global struct to a
10093 subclass of gimple_opt_pass along with...
10094 (pass_data_dominator): ...new pass_data instance and...
10095 (make_pass_dominator): ...new function.
10096 (pass_phi_only_cprop): Convert from a global struct to a subclass of
10097 gimple_opt_pass along with...
10098 (pass_data_phi_only_cprop): ...new pass_data instance and...
10099 (make_pass_phi_only_cprop): ...new function.
10100 * tree-ssa-dse.c (pass_dse): Convert from a global struct to a
10101 subclass of gimple_opt_pass along with...
10102 (pass_data_dse): ...new pass_data instance and...
10103 (make_pass_dse): ...new function.
10104 * tree-ssa-forwprop.c (pass_forwprop): Convert from a global struct to
10105 a subclass of gimple_opt_pass along with...
10106 (pass_data_forwprop): ...new pass_data instance and...
10107 (make_pass_forwprop): ...new function.
10108 * tree-ssa-ifcombine.c (pass_tree_ifcombine): Convert from a global
10109 struct to a subclass of gimple_opt_pass along with...
10110 (pass_data_tree_ifcombine): ...new pass_data instance and...
10111 (make_pass_tree_ifcombine): ...new function.
10112 * tree-ssa-loop-ch.c (pass_ch): Convert from a global struct to a
10113 subclass of gimple_opt_pass along with...
10114 (pass_data_ch): ...new pass_data instance and...
10115 (make_pass_ch): ...new function.
10116 * tree-ssa-loop.c (pass_tree_loop): Convert from a global struct to a
10117 subclass of gimple_opt_pass along with...
10118 (pass_data_tree_loop): ...new pass_data instance and...
10119 (make_pass_tree_loop): ...new function.
10120 (pass_tree_loop_init): Convert from a global struct to a subclass of
10121 gimple_opt_pass along with...
10122 (pass_data_tree_loop_init): ...new pass_data instance and...
10123 (make_pass_tree_loop_init): ...new function.
10124 (pass_lim): Convert from a global struct to a subclass of
10125 gimple_opt_pass along with...
10126 (pass_data_lim): ...new pass_data instance and...
10127 (make_pass_lim): ...new function.
10128 (pass_tree_unswitch): Convert from a global struct to a subclass of
10129 gimple_opt_pass along with...
10130 (pass_data_tree_unswitch): ...new pass_data instance and...
10131 (make_pass_tree_unswitch): ...new function.
10132 (pass_predcom): Convert from a global struct to a subclass of
10133 gimple_opt_pass along with...
10134 (pass_data_predcom): ...new pass_data instance and...
10135 (make_pass_predcom): ...new function.
10136 (pass_vectorize): Convert from a global struct to a subclass of
10137 gimple_opt_pass along with...
10138 (pass_data_vectorize): ...new pass_data instance and...
10139 (make_pass_vectorize): ...new function.
10140 (pass_graphite): Convert from a global struct to a subclass of
10141 gimple_opt_pass along with...
10142 (pass_data_graphite): ...new pass_data instance and...
10143 (make_pass_graphite): ...new function.
10144 (pass_graphite_transforms): Convert from a global struct to a subclass
10145 of gimple_opt_pass along with...
10146 (pass_data_graphite_transforms): ...new pass_data instance and...
10147 (make_pass_graphite_transforms): ...new function.
10148 (pass_check_data_deps): Convert from a global struct to a subclass of
10149 gimple_opt_pass along with...
10150 (pass_data_check_data_deps): ...new pass_data instance and...
10151 (make_pass_check_data_deps): ...new function.
10152 (pass_iv_canon): Convert from a global struct to a subclass of
10153 gimple_opt_pass along with...
10154 (pass_data_iv_canon): ...new pass_data instance and...
10155 (make_pass_iv_canon): ...new function.
10156 (pass_scev_cprop): Convert from a global struct to a subclass of
10157 gimple_opt_pass along with...
10158 (pass_data_scev_cprop): ...new pass_data instance and...
10159 (make_pass_scev_cprop): ...new function.
10160 (pass_record_bounds): Convert from a global struct to a subclass of
10161 gimple_opt_pass along with...
10162 (pass_data_record_bounds): ...new pass_data instance and...
10163 (make_pass_record_bounds): ...new function.
10164 (pass_complete_unroll): Convert from a global struct to a subclass of
10165 gimple_opt_pass along with...
10166 (pass_data_complete_unroll): ...new pass_data instance and...
10167 (make_pass_complete_unroll): ...new function.
10168 (pass_complete_unrolli): Convert from a global struct to a subclass of
10169 gimple_opt_pass along with...
10170 (pass_data_complete_unrolli): ...new pass_data instance and...
10171 (make_pass_complete_unrolli): ...new function.
10172 (pass_parallelize_loops): Convert from a global struct to a subclass
10173 of gimple_opt_pass along with...
10174 (pass_data_parallelize_loops): ...new pass_data instance and...
10175 (make_pass_parallelize_loops): ...new function.
10176 (pass_loop_prefetch): Convert from a global struct to a subclass of
10177 gimple_opt_pass along with...
10178 (pass_data_loop_prefetch): ...new pass_data instance and...
10179 (make_pass_loop_prefetch): ...new function.
10180 (pass_iv_optimize): Convert from a global struct to a subclass of
10181 gimple_opt_pass along with...
10182 (pass_data_iv_optimize): ...new pass_data instance and...
10183 (make_pass_iv_optimize): ...new function.
10184 (pass_tree_loop_done): Convert from a global struct to a subclass of
10185 gimple_opt_pass along with...
10186 (pass_data_tree_loop_done): ...new pass_data instance and...
10187 (make_pass_tree_loop_done): ...new function.
10188 * tree-ssa-math-opts.c (pass_cse_reciprocals): Convert from a global
10189 struct to a subclass of gimple_opt_pass along with...
10190 (pass_data_cse_reciprocals): ...new pass_data instance and...
10191 (make_pass_cse_reciprocals): ...new function.
10192 (pass_cse_sincos): Convert from a global struct to a subclass of
10193 gimple_opt_pass along with...
10194 (pass_data_cse_sincos): ...new pass_data instance and...
10195 (make_pass_cse_sincos): ...new function.
10196 (pass_optimize_bswap): Convert from a global struct to a subclass of
10197 gimple_opt_pass along with...
10198 (pass_data_optimize_bswap): ...new pass_data instance and...
10199 (make_pass_optimize_bswap): ...new function.
10200 (pass_optimize_widening_mul): Convert from a global struct to a
10201 subclass of gimple_opt_pass along with...
10202 (pass_data_optimize_widening_mul): ...new pass_data instance and...
10203 (make_pass_optimize_widening_mul): ...new function.
10204 * tree-ssa-phiopt.c (pass_phiopt): Convert from a global struct to a
10205 subclass of gimple_opt_pass along with...
10206 (pass_data_phiopt): ...new pass_data instance and...
10207 (make_pass_phiopt): ...new function.
10208 (pass_cselim): Convert from a global struct to a subclass of
10209 gimple_opt_pass along with...
10210 (pass_data_cselim): ...new pass_data instance and...
10211 (make_pass_cselim): ...new function.
10212 * tree-ssa-phiprop.c (pass_phiprop): Convert from a global struct to a
10213 subclass of gimple_opt_pass along with...
10214 (pass_data_phiprop): ...new pass_data instance and...
10215 (make_pass_phiprop): ...new function.
10216 * tree-ssa-pre.c (pass_pre): Convert from a global struct to a
10217 subclass of gimple_opt_pass along with...
10218 (pass_data_pre): ...new pass_data instance and...
10219 (make_pass_pre): ...new function.
10220 (pass_fre): Convert from a global struct to a subclass of
10221 gimple_opt_pass along with...
10222 (pass_data_fre): ...new pass_data instance and...
10223 (make_pass_fre): ...new function.
10224 * tree-ssa-reassoc.c (pass_reassoc): Convert from a global struct to a
10225 subclass of gimple_opt_pass along with...
10226 (pass_data_reassoc): ...new pass_data instance and...
10227 (make_pass_reassoc): ...new function.
10228 * tree-ssa-sink.c (pass_sink_code): Convert from a global struct to a
10229 subclass of gimple_opt_pass along with...
10230 (pass_data_sink_code): ...new pass_data instance and...
10231 (make_pass_sink_code): ...new function.
10232 * tree-ssa-strlen.c (pass_strlen): Convert from a global struct to a
10233 subclass of gimple_opt_pass along with...
10234 (pass_data_strlen): ...new pass_data instance and...
10235 (make_pass_strlen): ...new function.
10236 * tree-ssa-structalias.c (pass_build_alias): Convert from a global
10237 struct to a subclass of gimple_opt_pass along with...
10238 (pass_data_build_alias): ...new pass_data instance and...
10239 (make_pass_build_alias): ...new function.
10240 (pass_build_ealias): Convert from a global struct to a subclass of
10241 gimple_opt_pass along with...
10242 (pass_data_build_ealias): ...new pass_data instance and...
10243 (make_pass_build_ealias): ...new function.
10244 (pass_ipa_pta): Convert from a global struct to a subclass of
10245 simple_ipa_opt_pass along with...
10246 (pass_data_ipa_pta): ...new pass_data instance and...
10247 (make_pass_ipa_pta): ...new function.
10248 * tree-ssa-uncprop.c (pass_uncprop): Convert from a global struct to a
10249 subclass of gimple_opt_pass along with...
10250 (pass_data_uncprop): ...new pass_data instance and...
10251 (make_pass_uncprop): ...new function.
10252 * tree-ssa-uninit.c (pass_late_warn_uninitialized): Convert from a
10253 global struct to a subclass of gimple_opt_pass along with...
10254 (pass_data_late_warn_uninitialized): ...new pass_data instance and...
10255 (make_pass_late_warn_uninitialized): ...new function.
10256 * tree-ssa.c (pass_init_datastructures): Convert from a global struct
10257 to a subclass of gimple_opt_pass along with...
10258 (pass_data_init_datastructures): ...new pass_data instance and...
10259 (make_pass_init_datastructures): ...new function.
10260 (pass_early_warn_uninitialized): Convert from a global struct to a
10261 subclass of gimple_opt_pass along with...
10262 (pass_data_early_warn_uninitialized): ...new pass_data instance and...
10263 (make_pass_early_warn_uninitialized): ...new function.
10264 (pass_update_address_taken): Convert from a global struct to a
10265 subclass of gimple_opt_pass along with...
10266 (pass_data_update_address_taken): ...new pass_data instance and...
10267 (make_pass_update_address_taken): ...new function.
10268 * tree-ssanames.c (pass_release_ssa_names): Convert from a global
10269 struct to a subclass of gimple_opt_pass along with...
10270 (pass_data_release_ssa_names): ...new pass_data instance and...
10271 (make_pass_release_ssa_names): ...new function.
10272 * tree-stdarg.c (pass_stdarg): Convert from a global struct to a
10273 subclass of gimple_opt_pass along with...
10274 (pass_data_stdarg): ...new pass_data instance and...
10275 (make_pass_stdarg): ...new function.
10276 * tree-switch-conversion.c (pass_convert_switch): Convert from a
10277 global struct to a subclass of gimple_opt_pass along with...
10278 (pass_data_convert_switch): ...new pass_data instance and...
10279 (make_pass_convert_switch): ...new function.
10280 * tree-tailcall.c (pass_tail_recursion): Convert from a global struct
10281 to a subclass of gimple_opt_pass along with...
10282 (pass_data_tail_recursion): ...new pass_data instance and...
10283 (make_pass_tail_recursion): ...new function.
10284 (pass_tail_calls): Convert from a global struct to a subclass of
10285 gimple_opt_pass along with...
10286 (pass_data_tail_calls): ...new pass_data instance and...
10287 (make_pass_tail_calls): ...new function.
10288 * tree-vect-generic.c (pass_lower_vector): Convert from a global
10289 struct to a subclass of gimple_opt_pass along with...
10290 (pass_data_lower_vector): ...new pass_data instance and...
10291 (make_pass_lower_vector): ...new function.
10292 (pass_lower_vector_ssa): Convert from a global struct to a subclass of
10293 gimple_opt_pass along with...
10294 (pass_data_lower_vector_ssa): ...new pass_data instance and...
10295 (make_pass_lower_vector_ssa): ...new function.
10296 * tree-vectorizer.c (pass_slp_vectorize): Convert from a global struct
10297 to a subclass of gimple_opt_pass along with...
10298 (pass_data_slp_vectorize): ...new pass_data instance and...
10299 (make_pass_slp_vectorize): ...new function.
10300 (pass_ipa_increase_alignment): Convert from a global struct to a
10301 subclass of simple_ipa_opt_pass along with...
10302 (pass_data_ipa_increase_alignment): ...new pass_data instance and...
10303 (make_pass_ipa_increase_alignment): ...new function.
10304 * tree-vrp.c (pass_vrp): Convert from a global struct to a subclass of
10305 gimple_opt_pass along with...
10306 (pass_data_vrp): ...new pass_data instance and...
10307 (make_pass_vrp): ...new function.
10308 * tree.c (pass_ipa_free_lang_data): Convert from a global struct to a
10309 subclass of simple_ipa_opt_pass along with...
10310 (pass_data_ipa_free_lang_data): ...new pass_data instance and...
10311 (make_pass_ipa_free_lang_data): ...new function.
10312 * tsan.c (pass_tsan): Convert from a global struct to a subclass of
10313 gimple_opt_pass along with...
10314 (pass_data_tsan): ...new pass_data instance and...
10315 (make_pass_tsan): ...new function.
10316 (pass_tsan_O0): Convert from a global struct to a subclass of
10317 gimple_opt_pass along with...
10318 (pass_data_tsan_O0): ...new pass_data instance and...
10319 (make_pass_tsan_O0): ...new function.
10320 * var-tracking.c (pass_variable_tracking): Convert from a global
10321 struct to a subclass of rtl_opt_pass along with...
10322 (pass_data_variable_tracking): ...new pass_data instance and...
10323 (make_pass_variable_tracking): ...new function.
10324 * web.c (pass_web): Convert from a global struct to a subclass of
10325 rtl_opt_pass along with...
10326 (pass_data_web): ...new pass_data instance and...
10327 (make_pass_web): ...new function.
10328 * config/epiphany/epiphany.h (pass_mode_switch_use): Replace
10329 declaration with that of...
10330 (make_pass_mode_switch_use): ...new function.
10331 (pass_resolve_sw_modes): Replace declaration with that of...
10332 (make_pass_resolve_sw_modes): ...new function.
10333 * config/epiphany/mode-switch-use.c (pass_mode_switch_use): Convert
10334 from a global struct to a subclass of rtl_opt_pass along with...
10335 (pass_data_mode_switch_use): ...new pass_data instance and...
10336 (make_pass_mode_switch_use): ...new function.
10337 * config/epiphany/resolve-sw-modes.c (pass_resolve_sw_modes): Convert
10338 from a global struct to a subclass of rtl_opt_pass along with...
10339 (pass_data_resolve_sw_modes): ...new pass_data instance and...
10340 (make_pass_resolve_sw_modes): ...new function.
10341 * config/i386/i386.c (pass_insert_vzeroupper): Convert from a global
10342 struct to a subclass of rtl_opt_pass along with...
10343 (pass_data_insert_vzeroupper): ...new pass_data instance and...
10344 (make_pass_insert_vzeroupper): ...new function.
10345 * config/sparc/sparc.c (pass_work_around_errata): Convert from a
10346 global struct to a subclass of rtl_opt_pass along with...
10347 (pass_data_work_around_errata): ...new pass_data instance and...
10348 (make_pass_work_around_errata): ...new function.
10349 * config/mips/mips.c (pass_mips_machine_reorg2): Convert from a global
10350 struct to a subclass of rtl_opt_pass along with...
10351 (pass_data_mips_machine_reorg2): ...new pass_data instance and...
10352 (make_pass_mips_machine_reorg2): ...new function.
10353
10354 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10355
10356 * passes.c (pass_manager::operator new): New.
10357
10358 2013-08-05 David Malcolm <dmalcolm@redhat.com>
10359
10360 Handwritten part of conversion of passes to C++ classes.
10361
10362 * Makefile.in (PASS_MANAGER_H): Add dep on pass-instances.def.
10363 (toplev.o): Add dep on PASS_MANAGER_H.
10364 * cgraphunit.c (cgraph_process_new_functions): Rework invocation
10365 of early local pases to reflect this moving from a global to a
10366 member of gcc::pass_manager.
10367 (cgraph_add_new_function): Likewise.
10368 * lto-cgraph.c (lto_output_node): Update for conversion of
10369 struct ipa_opt_pass_d to a C++ subclass of opt_pass.
10370 * passes.c (opt_pass::clone): New.
10371 (opt_pass::gate): New.
10372 (opt_pass::execute): New.
10373 (opt_pass::opt_pass): New.
10374 (pass_manager::execute_early_local_passes): New.
10375 (pass_manager::execute_pass_mode_switching): new.
10376 (finish_optimization_passes): Convert to...
10377 (pass_manager::finish_optimization_passes): ...this.
10378 (finish_optimization_passes): Update for conversion of passes to
10379 C++ classes.
10380 (register_dump_files_1): Use has_gate since we cannot portably
10381 check a vtable entry against NULL.
10382 (dump_one_pass): Likewise.
10383 (ipa_write_summaries_2): Likewise.
10384 (ipa_write_optimization_summaries_1): Likewise.
10385 (ipa_read_summaries_1): Likewise.
10386 (ipa_read_optimization_summaries_1): Likewise.
10387 (execute_ipa_stmt_fixups): Likewise.
10388 (pass_manager::pass_manager): Rewrite pass-creation, invoking
10389 pass-creation functions rather than wiring up globals, and
10390 storing the results in fields of pass_manager generated using
10391 pass-instances.def.
10392 (pass_manager::dump_profile_report): Update for conversion of
10393 passes to C++ classes.
10394 (pass_manager::execute_ipa_summary_passes): Likewise.
10395 (execute_one_ipa_transform_pass): Likewise.
10396 (execute_one_pass): Use has_gate and has_execute since we cannot
10397 portably check a vtable entry against NULL.
10398 * pass_manager.h (pass_manager::finish_optimization_passes): New.
10399 (pass_manager): Use pass-instances.def to add fields for the
10400 various pass instances.
10401 * toplev.c (finalize): Update for move of
10402 finish_optimization_passes to a method of gcc::pass_manager.
10403 * toplev.h (finish_optimization_passes): Move to method of class
10404 pass_manager.
10405 * tree-pass.h (struct pass_data): New.
10406 (opt_pass): Convert to C++ class, make it a subclass of pass_data.
10407 (opt_pass::gate): Convert to virtual function.
10408 (opt_pass::~opt_pass): New.
10409 (opt_pass::clone): New.
10410 (opt_pass::execute): Convert to virtual function.
10411 (opt_pass::opt_pass): New.
10412 (opt_pass::ctxt_): new.
10413 (gimple_opt_pass): Convert to subclass of opt_pass.
10414 (gimple_opt_pass::gimple_opt_pass): New.
10415 (rtl_opt_pass): Convert to subclass of opt_pass.
10416 (rtl_opt_pass::rtl_opt_pass): New.
10417 (ipa_opt_pass_d): Convert to subclass of opt_pass.
10418 (ipa_opt_pass_d::ipa_opt_pass_d): New.
10419 (simple_ipa_opt_pass): Convert to subclass of opt_pass.
10420 (simple_ipa_opt_pass::simple_ipa_opt_pass): New.
10421 * config/i386/i386.c (rest_of_handle_insert_vzeroupper): Rework
10422 invocation of pass_mode_switching to reflect this moving from a
10423 global to a member of gcc::pass_manager.
10424 (ix86_option_override): Rework how pass_insert_vzeroupper is
10425 added to the pass_manager to reflect autogenerated changes.
10426 * config/i386/t-i386 (i386.o) Add deps on CONTEXT_H and PASS_MANAGER_H.
10427
10428 2013-08-05 Richard Earnshaw <rearnsha@arm.com>
10429
10430 PR rtl-optimization/57708
10431 * recog.c (peep2_find_free_register): Validate all regs in a
10432 multi-reg mode.
10433
10434 2013-08-05 Jan Hubicka <jh@suse.cz>
10435
10436 PR lto/57602
10437 * cgraph.c (verify_cgraph_node): Accept local flags from other
10438 partitions.
10439 * ipa.c (symtab_remove_unreachable_nodes): Do not clear local flag.
10440 (function_and_variable_visibility): Likewise.
10441 * trans-mem.c (ipa_tm_create_version): TM versions are not local.
10442
10443 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10444
10445 * graph.c (init_graph_slim_pretty_print): Remove.
10446 (print_graph_cfg): Do not call it. Use local pretty printer.
10447 (start_graph_dump): Likewise.
10448
10449 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10450
10451 * gimple-pretty-print.c (buffer): Remove.
10452 (initialized): Likewise.
10453 (maybe_init_pretty_print): Likewise.
10454 (print_gimple_stmt): Do not call it. Use non-static local
10455 pretty_printer variable.
10456 (print_gimple_expr): Likewise.
10457 (print_gimple_seq): Likewise.
10458 (gimple_dump_bb): Likewise.
10459
10460 2013-08-05 Gabriel Dos Reis <gdr@integrable-solutions.net>
10461
10462 * asan.c (asan_pp): Remove.
10463 (asan_pp_initialized): Likewise.
10464 (asan_pp_initialize): Likewise.
10465 (asan_pp_string): Take a pretty_printer parameter. Adjust callers.
10466 (asan_emit_stack_protection): Tidy. Use local pretty printer.
10467 (asan_add_global): Likewise.
10468
10469 2013-08-04 Gabriel Dos Reis <gdr@integrable-solutions.net>
10470
10471 * pretty-print.h (pp_base): Remove. Adjust dependent macros.
10472 * diagnostic.h (diagnostic_flush_buffer): Adjust.
10473 * pretty-print.c (pp_formatted_text_data): Likewise.
10474 (pp_indent): Rename from pp_base_indent.
10475 (pp_format): Rename from pp_base_format.
10476 (pp_output_formatted_text): Rename from pp_base_output_formatted_text.
10477 (pp_format_verbatim): Rename from pp_base_format_verbatim.
10478 (pp_flush): Rename from pp_base_flush.
10479 (pp_set_line_maximum_length): Rename from
10480 pp_base_set_line_maximum_length.
10481 (pp_clear_output_area): Rename from pp_base_clear_output_area.
10482 (pp_set_prefix): Rename from pp_base_set_prefix.
10483 (pp_destroy_prefix): Rename from pp_base_destroy_prefix.
10484 (pp_emit_prefix): Rename from pp_base_emit_prefix.
10485 (pp_append_text): Rename from pp_base_append_text.
10486 (pp_formatted_text): Rename from pp_base_formatted_text.
10487 (pp_last_position_in_text): Rename from pp_base_last_position_in_text.
10488 (pp_remaining_character_count_for_line): Rename from
10489 pp_base_remaining_character_count_for_line.
10490 (pp_newline): Rename from pp_base_newline.
10491 (pp_character): Rename from pp_base_character.
10492 (pp_string): Rename from pp_base_string.
10493 (pp_maybe_space): Rename from pp_base_maybe_space.
10494 * asan.c (asan_pp_string): Adjust.
10495 (asan_emit_stack_protection): Likewise.
10496 (asan_add_global): Likewise.
10497 * sched-vis.c (str_pattern_slim): Adjust pretty printer function call.
10498 * tree-mudflap.c (mf_varname_tree): Likewise.
10499 * tree-pretty-print.c (pp_tree_identifier): Rename from
10500 pp_base_tree_identifier.
10501 * tree-pretty-print.h (pp_tree_identifier): Remove macro definition.
10502 Declare as function.
10503
10504 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10505
10506 * pretty-print.h (pp_bar_bar): New.
10507 (pp_ampersand_ampersand): Likewise.
10508 (pp_less_equal): Likewise.
10509 (pp_greater_equal): Likewise.
10510 * gimple-pretty-print.c (dump_ternary_rhs): Use specialized pretty
10511 printer functions instead of pp_string or operators and punctuators.
10512 (dump_gimple_call): Likewise.
10513 (dump_gimple_omp_for): Likewise.
10514 (dump_gimple_transaction): Likewise.
10515 (dump_gimple_phi): Likewise.
10516 (pp_gimple_stmt_1): Likewise.
10517 * sched-vis.c (print_insn): Likewise.
10518 * tree-mudflap.c (mf_varname_tree): Likewise.
10519 * tree-pretty-print.c (dump_block_node): Likewise.
10520 (dump_generic_node): Likewise.
10521
10522 2013-08-02 Jan Hubicka <jh@suse.cz>
10523
10524 * lto-cgraph.c (compute_ltrans_boundary): Add abstract origins into
10525 boundaries.
10526 * lto-streamer-out.c (tree_is_indexable): Results decls and
10527 parm decls are not indexable.
10528 (DFS_write_tree_body): Do not follow args and results.
10529 (hash_tree): Likewise.
10530 (output_functions): Rearrange so struct function is needed
10531 only when real body is output; be able to also ouptut abstract
10532 functions; output DECL_ARGUMENTS and DECL_RESULT.
10533 (lto_output): When not in WPA, ale store abstract functions.
10534 (write_symbol): Do not care about RESULT_DECL.
10535 (output_symbol_p): Handle correctly sbtract decls.
10536 * lto-streamer-in.c (input_function): Rearrange so struct
10537 function can be NULL at entry; allow streaming of
10538 functions w/o body; store DECL_ARGUMENTS and DECL_RESULT.
10539 * ipa.c (symtab_remove_unreachable_nodes): Silence confused
10540 sanity check during LTO.
10541 * tree-streamer-out.c (write_ts_decl_non_common_tree_pointers): Skip
10542 RESULT_DECl and DECL_ARGUMENTS.
10543 * tree-streamer-in.c (lto_input_ts_decl_non_common_tree_pointers):
10544 Likewise.
10545
10546 2013-08-03 Gabriel Dos Reis <gdr@integrable-solutions.net>
10547
10548 * pretty-print.h (pp_underscore): New.
10549 (pp_comma): Tidy.
10550 * gimple-pretty-print.c (dump_unary_rhs): Use specialized pretty
10551 printer functions instead of pp_character.
10552 (dump_binary_rhs): Likewise.
10553 (dump_ternary_rhs): Likewise.
10554 (dump_gimple_call_args): Likewise.
10555 (pp_points_to_solution): Likewise.
10556 (dump_gimple_call): Likewise.
10557 (dump_gimple_switch): Likewise.
10558 (dump_gimple_cond): Likewise.
10559 (dump_gimple_bind): Likewise.
10560 (dump_gimple_try): Likewise.
10561 (dump_gimple_omp_for): Likewise.
10562 (dump_gimple_omp_continue): Likewise.
10563 (dump_gimple_omp_single): Likewise.
10564 (dump_gimple_omp_sections): Likewise.
10565 (dump_gimple_omp_block): Likewise.
10566 (dump_gimple_omp_critical): Likewise.
10567 (dump_gimple_transaction): Likewise.
10568 (dump_gimple_asm): Likewise.
10569 (dump_gimple_phi): Likewise.
10570 (dump_gimple_omp_parallel): Likewise.
10571 (dump_gimple_omp_task): Likewise.
10572 (dump_gimple_omp_atomic_load): Likewise.
10573 (dump_gimple_omp_atomic_store): Likewise.
10574 (dump_gimple_mem_ops): Likewise.
10575 (pp_gimple_stmt_1): Likewise.
10576 (pp_cfg_jump): Likewise.
10577 (dump_implicit_edges): Likewise.
10578 (gimple_dump_bb_for_graph): Likewise.
10579 * graph.c (draw_cfg_node): Likewise.
10580 * langhooks.c (lhd_print_error_function): Likewise.
10581 * sched-vis.c (print_exp): Likewise.
10582 (print_value): Likewise.
10583 (print_pattern): Likewise.
10584 (print_insn): Likewise.
10585 (rtl_dump_bb_for_graph): Likewise.
10586 * tree-pretty-print.c (dump_function_declaration): Likewise.
10587 (dump_array_domain): Likewise.
10588 (dump_omp_clause): Likewise.
10589 (dump_location): Likewise.
10590 (dump_generic_node): Likewise.
10591 (print_struct_decl): Likewise.
10592 * diagnostic.c (diagnostic_show_locus): Use pp_space.
10593
10594 2013-08-03 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10595
10596 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Update
10597 candidate table when replacing a candidate statement.
10598 (replace_rhs_if_not_dup): Likewise.
10599 (replace_one_candidate): Likewise.
10600
10601 2013-08-02 Jan Hubicka <jh@suse.cz>
10602 Martin Liska <marxin.liska@gmail.com>
10603
10604 * cgraphunit.c (add_new_function): Fix logic when adding from
10605 late IPA pass.
10606 (assemble_thunk): Rename to ...
10607 (expand_thunk); .. this one; export; get it working with
10608 general functions; make produced gimple valid.
10609 * cgraph.h (expand_thunk): Declare.
10610
10611 2013-08-02 Jan Hubicka <jh@suse.cz>
10612
10613 * ipa-cp.c (gather_context_independent_values): Use
10614 ipa_get_param_move_cost.
10615 (get_replacement_map): Remove PARAM; move parameter folding
10616 into tree-inline.c
10617 (create_specialized_node): Update.
10618 * ipa-prop.c (ipa_populate_param_decls): Do not look for origins;
10619 assert that we have gimple body; update move_cost.
10620 (count_formal_params): Assert that we have gimple body.
10621 (ipa_dump_param): New function.
10622 (ipa_alloc_node_params): Break out from ...
10623 (ipa_initialize_node_params): ... here.
10624 (ipa_get_vector_of_formal_parms): ICE when used in WPA.
10625 (ipa_write_node_info): Stream move costs.
10626 (ipa_read_node_info): Read move costs.
10627 (ipa_update_after_lto_read): Do not recompute node params.
10628 * ipa-prop.h (ipa_param_descriptor): Add move_cost.
10629 (ipa_get_param): Check we are not in WPA.
10630 (ipa_get_param_move_cost): New.
10631 * tree-inline.c (tree_function_versioning): Fold replacement as needed.
10632 * ipa-inline-analysis.c (inline_node_duplication_hook): Expect only
10633 parm numbers to be present.
10634
10635 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10636
10637 PR rtl-optimization/58048
10638 * lra-constraints.c (process_alt_operands): Don't check asm
10639 operand on register.
10640
10641 2013-08-02 Eric Botcazou <ebotcazou@adacore.com>
10642
10643 * config/sparc/sparc.c (sparc_emit_membar_for_model) <SMM_TSO>: Add
10644 the implied StoreLoad barrier for atomic operations if before.
10645
10646 2013-08-02 Jan Hubicka <jh@suse.cz>
10647 Martin Liska <marxin.liska@gmail.com>
10648
10649 * cgraph.c (cgraph_function_body_availability): Do not check
10650 cgraph flags.
10651 * cgraph.h (symtab_for_node_and_aliases, symtab_nonoverwritable_alias,
10652 symtab_node_availability): Declare.
10653 * ipa.c (can_replace_by_local_alias): New.
10654 (function_and_variable_visibility): Use it.
10655 * symtab.c (symtab_for_node_and_aliases,
10656 symtab_nonoverwritable_alias_1, symtab_nonoverwritable_alias): New.
10657
10658 2013-08-02 Vladimir Makarov <vmakarov@redhat.com>
10659
10660 PR rtl-optimization/57963
10661 * lra-constraints.c (reverse_equiv_p, contains_reloaded_insn_p): New.
10662 (lra_constraints): Use them.
10663
10664 2013-08-02 Sofiane Naci <sofiane.naci@arm.com>
10665
10666 * config/arm/types.md (define_attr "type"): Add "load_acq"
10667 and "store_rel".
10668 * config/arm/cortex-a53.md (cortex_a53_load1): Update for attribute
10669 changes.
10670 (cortex_a53_store1): Likewise.
10671
10672 2013-08-01 Jan Hubicka <jh@suse.cz>
10673
10674 * ipa.c (symtab_remove_unreachable_nodes): Nodes in other
10675 partitions are not needed.
10676
10677 2013-08-01 Uros Bizjak <ubizjak@gmail.com>
10678
10679 * config/i386/i386.h (MAYBE_NON_Q_CLASS_P): New.
10680 * config/i386/i386.c (ix86_secondary_reload): Use INTEGER_CLASS_P and
10681 MAYBE_NON_Q_CLASS_P where appropriate.
10682
10683 2013-08-01 Jan Hubicka <jh@suse.cz>
10684
10685 * cgraph.h (release_function_body): Declare.
10686 * tree.c (free_lang_data_in_decl): Free, parameters and return values
10687 of unused delcarations.
10688
10689 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10690
10691 * config/arm/arm.md (minmax_arithsi_non_canon): Emit canonical
10692 RTL form when subtracting a constant.
10693
10694 2013-08-01 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
10695
10696 * config/arm/arm.md (peepholes for eq (reg1) (reg2/imm)):
10697 Generate canonical plus rtx with negated immediate instead of minus
10698 where appropriate.
10699 * config/arm/arm.c (thumb2_reorg): Handle ADCS <Rd>, <Rn> case.
10700
10701 2013-08-01 Jan Hubicka <jh@suse.cz>
10702
10703 * cgraph.c (cgraph_release_function_body): Use used_as_abstract_origin.
10704 (cgraph_release_function_body): Likewise.
10705 (cgraph_can_remove_if_no_direct_calls_p): Likewise.
10706 * cgraph.h (cgrpah_node): Rename abstract_and_needed
10707 to used_as_abstract_origin.
10708 * tree-inline-transfrom.c (can_remove_node_now_p_1): Do not remove
10709 symbols used as abstract origins.
10710 * cgraphunit.c (analyze_functions): Update.
10711 * ipa.c (symtab_remove_unreachable_nodes): Recompute
10712 used_as_abstract_origin.
10713 * tree-inline.c (tree_function_versioning): Update
10714 used_as_abstract_origin; be ready for DECL_RESULT and
10715 DECL_ARGUMENTS to be NULL.
10716
10717 * lto-symtab.c (lto_symtab_merge_symbols): Merge duplicated nodes
10718 for abstract functions.
10719 * cgraph.h (symtab_real_symbol_p): Abstract declarations are not
10720 real symbols.
10721
10722 2013-08-01 Jan Hubicka <jh@suse.cz>
10723
10724 * profile.c (compute_value_histograms): Fix thinko.
10725
10726 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10727
10728 * config.gcc (aarch64*-*-*): Add aarch-common.o to extra_objs. Add
10729 aarch-common-protos.h to extra_headers.
10730 (aarch64*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10731 * config/aarch64/aarch64.md: Include "../arm/cortex-a53.md".
10732 * config/aarch64/t-aarch64 (aarch-common.o): Define.
10733
10734 2013-08-01 Sofiane Naci <sofiane.naci@arm.com>
10735
10736 * config/aarch64/aarch64.md (define_attr "type"): Delete.
10737 Include "../arm/types.md". Define "type" attribute for all patterns.
10738 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update for
10739 attribute changes.
10740
10741 2013-07-31 Michael Meissner <meissner@linux.vnet.ibm.com>
10742
10743 * config/rs6000/predicates.md (fusion_gpr_addis): New predicates
10744 to support power8 load fusion.
10745 (fusion_gpr_mem_load): Likewise.
10746
10747 * config/rs6000/rs6000-modes.def (PTImode): Update a comment.
10748
10749 * config/rs6000/rs6000-protos.h (fusion_gpr_load_p): New
10750 declarations for power8 load fusion.
10751 (emit_fusion_gpr_load): Likewise.
10752
10753 * config/rs6000/rs6000.c (rs6000_option_override_internal): If
10754 tuning for power8, turn on fusion mode by default. Turn on sign
10755 extending fusion mode if normal fusion mode is on, and we are at
10756 -O2 or -O3.
10757 (fusion_gpr_load_p): New function, return true if we can fuse an
10758 addis instruction with a dependent load to a GPR.
10759 (emit_fusion_gpr_load): Emit the instructions for power8 load
10760 fusion to GPRs.
10761
10762 * config/rs6000/vsx.md (VSX_M2): New iterator for fusion peepholes.
10763 (VSX load fusion peepholes): New peepholes to fuse together an
10764 addi instruction with a VSX load instruction.
10765
10766 * config/rs6000/rs6000.md (GPR load fusion peepholes): New
10767 peepholes to fuse an addis instruction with a load to a GPR base
10768 register. If we are supporting sign extending fusions, convert
10769 sign extending loads to zero extending loads and add an explicit
10770 sign extension.
10771
10772 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10773
10774 * config.gcc (arm*-*-*): Add aarch-common.o to extra_objs. Add
10775 aarch-common-protos.h to extra_headers.
10776 (arm*-*-*): Add arm/aarch-common-protos.h to tm_p_file.
10777 * config/arm/arm.c (arm_early_load_addr_dep): Move from here to ...
10778 (arm_early_store_addr_dep): Likewise.
10779 (arm_no_early_alu_shift_dep): Likewise.
10780 (arm_no_early_alu_shift_value_dep): Likewise.
10781 (arm_no_early_mul_dep): Likewise.
10782 (arm_no_early_store_addr_dep): Likewise.
10783 (arm_mac_accumulator_is_mul_result): Likewise.
10784 (arm_mac_accumulator_is_result): Likewise.
10785 * config/arm/aarch-common.c: ... here. New file.
10786 * config/arm/arm-protos.h (arm_early_load_addr_dep): Move from
10787 here to ...
10788 (arm_early_store_addr_dep): Likewise.
10789 (arm_no_early_alu_shift_dep): Likewise.
10790 (arm_no_early_alu_shift_value_dep): Likewise.
10791 (arm_no_early_mul_dep): Likewise.
10792 (arm_no_early_store_addr_dep): Likewise.
10793 (arm_mac_accumulator_is_mul_result): Likewise.
10794 (arm_mac_accumulator_is_result): Likewise.
10795 * config/arm/aarch-common-protos.h: ... here. New file.
10796 * config/arm/t-arm (aarch-common.o): Define.
10797
10798 2013-07-31 Sofiane Naci <sofiane.naci@arm.com>
10799
10800 * config/arm/arm.md: Include new file "types.md".
10801 (define_attr "type"): Move from here to ...
10802 (define_attr "mul32"): Likewise.
10803 (define_attr "mul64"): Likewise.
10804 * config/arm/types.md: ... here. New file.
10805
10806 2013-07-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
10807
10808 * config.gcc (*-*-rtems*): Use __cxa_atexit by default.
10809 * config/rs6000/rtems.h (TARGET_LIBGCC_SDATA_SECTION): Define.
10810
10811 2013-07-31 Jan-Benedict Glaw <jbglaw@lug-owl.de>
10812
10813 * gen-pass-instances.awk: Fix offset of substr().
10814
10815 2013-07-31 David Malcolm <dmalcolm@redhat.com>
10816
10817 * Makefile.in (pass-instances.def): New.
10818 (passes.o): Replace dependency on passes.def with one on
10819 pass-instances.def
10820
10821 * gen-pass-instances.awk: New.
10822
10823 * passes.c (pass_manager::pass_manager): Use pass-instances.def
10824 rather than passes.def, updating local definition of NEXT_PASS
10825 macro to add an extra NUM parameter (currently unused).
10826
10827 2013-07-30 David Malcolm <dmalcolm@redhat.com>
10828
10829 * Makefile.in (PASS_MANAGER_H): New.
10830 (lto-cgraph.o): Depend on CONTEXT_H and PASS_MANAGER_H.
10831 (passes.o): Likewise.
10832 (statistics.o): Likewise.
10833 (cgraphunit.o): Likewise.
10834 (context.o): Depend on PASS_MANAGER_H.
10835
10836 * pass_manager.h: New.
10837
10838 * cgraphunit.c (cgraph_add_new_function): Update for moves
10839 of globals to fields of pass_manager.
10840 (analyze_function): Likewise.
10841 (expand_function): Likewise.
10842 (ipa_passes): Likewise.
10843 (compile): Likewise.
10844
10845 * context.c (context::context): New.
10846 * context.h (context::context): New.
10847 (context::get_passes): New.
10848 (context::passes_): New.
10849
10850 * lto-cgraph.c (input_node): Update for moves of globals to
10851 fields of pass_manager.
10852
10853 * passes.c (all_passes): Remove, in favor of a field of the
10854 same name within the new class pass_manager.
10855 (all_small_ipa_passes): Likewise.
10856 (all_lowering_passes): Likewise.
10857 (all_regular_ipa_passes): Likewise.
10858 (all_late_ipa_passes): Likewise.
10859 (all_lto_gen_passes): Likewise.
10860 (passes_by_id): Likewise.
10861 (passes_by_id_size): Likewise.
10862 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10863 the new class pass_manager.
10864 (set_pass_for_id): Convert to...
10865 (pass_manager::set_pass_for_id): ...method.
10866 (get_pass_for_id): Convert to...
10867 (pass_manager::get_pass_for_id): ...method.
10868 (register_one_dump_file): Move body of implementation into...
10869 (pass_manager::register_one_dump_file): ...here.
10870 (register_dump_files_1): Convert to...
10871 (pass_manager::register_dump_files_1): ...method.
10872 (register_dump_files): Convert to...
10873 (pass_manager::register_dump_files): ...method.
10874 (create_pass_tab): Update for moves of globals to fields of
10875 pass_manager.
10876 (dump_passes): Move body of implementation into...
10877 (pass_manager::dump_passes): ...here.
10878 (register_pass): Move body of implementation into...
10879 (pass_manager::register_pass): ...here.
10880 (init_optimization_passes): Convert into...
10881 (pass_manager::pass_manager): ...constructor for new
10882 pass_manager class, and initialize the pass_lists array.
10883 (check_profile_consistency): Update for moves of globals to
10884 fields of pass_manager.
10885 (dump_profile_report): Move body of implementation into...
10886 (pass_manager::dump_profile_report): ...here.
10887 (ipa_write_summaries_1): Update for moves of pass lists from
10888 being globals to fields of pass_manager.
10889 (ipa_write_optimization_summaries): Likewise.
10890 (ipa_read_summaries): Likewise.
10891 (ipa_read_optimization_summaries): Likewise.
10892 (execute_all_ipa_stmt_fixups): Likewise.
10893
10894 * statistics.c (statistics_fini): Update for moves of globals to
10895 fields of pass_manager.
10896
10897 * toplev.c (general_init): Replace call to
10898 init_optimization_passes with construction of the pass_manager
10899 instance.
10900
10901 * tree-pass.h (all_passes): Remove, in favor of a field of the
10902 same name within the new class pass_manager.
10903 (all_small_ipa_passes): Likewise.
10904 (all_lowering_passes): Likewise.
10905 (all_regular_ipa_passes): Likewise.
10906 (all_lto_gen_passes): Likewise.
10907 (all_late_ipa_passes): Likewise.
10908 (passes_by_id): Likewise.
10909 (passes_by_id_size): Likewise.
10910 (gcc_pass_lists): Remove, in favor of "pass_lists" field within
10911 the new class pass_manager.
10912 (get_pass_for_id): Remove.
10913
10914 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10915
10916 * config.gcc (arm): Require 64-bit host-wide-int for all ARM target
10917 configs.
10918
10919 2013-07-30 Richard Earnshaw <rearnsha@arm.com>
10920
10921 * arm.md (mulhi3): New expand pattern.
10922
10923 2013-07-30 Jan Hubicka <jh@suse.cz>
10924 Martin Liska <marxin.liska@gmail.com>
10925
10926 * profile.c (compute_value_histograms): Do not ICE when
10927 there is mismatch only on some counters.
10928
10929 2013-07-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
10930
10931 PR rtl-optimization/57637
10932 * function.c (move_insn_for_shrink_wrap): Also check the
10933 GEN set of the LIVE problem for the liveness analysis
10934 if it exists, otherwise give up.
10935
10936 2013-07-29 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
10937
10938 PR tree-optimization/57993
10939 * gimple-ssa-strength-reduction.c (replace_mult_candidate): Record
10940 replaced statement in the candidate table.
10941 (phi_add_costs): Return infinite cost when the hidden basis does
10942 not dominate all phis on which the candidate is dependent.
10943 (replace_one_candidate): Record replaced statement in the
10944 candidate table.
10945
10946 2013-07-29 Joern Rennecke <joern.rennecke@embecosm.com>
10947
10948 * config/epiphany/epiphany.md (*isub_i+2): New peephole.
10949 (ashlv2si3): New expander.
10950 (*ashlv2si3_i): New define_insn_and_split.
10951 * predicates.md (float_operation): Allow patterns with three
10952 basic sub-patterns.
10953
10954 PR rtl-optimization/58021
10955 * mode-switching.c (create_pre_exit): Always split off preceding
10956 insns if we are not at the basic block head.
10957
10958 2013-07-29 Maciej W. Rozycki <macro@codesourcery.com>
10959
10960 * config/mips/linux.h (GLIBC_DYNAMIC_LINKER): Handle `-mnan=2008'.
10961 (UCLIBC_DYNAMIC_LINKER): New macro.
10962 * config/mips/linux64.h (GLIBC_DYNAMIC_LINKER32): Handle
10963 `-mnan=2008'.
10964 (GLIBC_DYNAMIC_LINKER64, GLIBC_DYNAMIC_LINKERN32): Likewise.
10965 (UCLIBC_DYNAMIC_LINKER32): Undefine macro first. Handle
10966 `-mnan=2008'.
10967 (UCLIBC_DYNAMIC_LINKER64): Redefine macro.
10968 (UCLIBC_DYNAMIC_LINKERN32): Likewise.
10969 * config/mips/mips-modes.def: Remove RESET_FLOAT_FORMAT calls
10970 for SF and DF modes. Use ieee_quad_format for TF mode.
10971 * config/mips/mips-opts.h (mips_ieee_754_setting): New enum.
10972 * config/mips/mips.c (mips_file_start): Output a `.nan' directive.
10973 (mips_option_override): Handle `-mnan=legacy'.
10974 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Handle
10975 `-mabs=2008' and `-mnan=2008'.
10976 (OPTION_DEFAULT_SPECS): Add "nan" default.
10977 (ASM_SPEC): Handle `-mnan='.
10978 [!HAVE_AS_NAN] (HAVE_AS_NAN): New macro.
10979 * config/mips/mips.md (abs<mode>2): Handle `-mabs=2008', update
10980 comment accordingly.
10981 (neg<mode>2): Likewise.
10982 * config/mips/mips.opt (mabs, mnan): New options.
10983 * doc/install.texi (Configuration): Document `--with-nan=' option.
10984 * doc/invoke.texi (Option Summary): List MIPS `-mabs=' and
10985 `-mnan=' options.
10986 (MIPS Options): Document them.
10987 * config.gcc <mips*-*-*>: Handle `--with-nan='.
10988 * configure.ac <mips*-*-*>: Check for GAS `-mnan=2008' support.
10989 * configure: Regenerate.
10990 * config.in: Regenerate.
10991
10992 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10993
10994 * config/i386/i386.md (float post-reload splitters): Do not check
10995 for subregs of SSE registers.
10996
10997 2013-07-29 Uros Bizjak <ubizjak@gmail.com>
10998 H.J. Lu <hongjiu.lu@intel.com>
10999
11000 PR target/57954
11001 PR target/57988
11002 * config/i386/i386.md (post-reload splitter
11003 to avoid partial SSE reg dependency stalls): New pattern.
11004
11005 2013-07-29 Dominik Vogt <vogt@linux.vnet.ibm.com>
11006
11007 * config/s390/s390.md ("movcc"): Swap load and store instructions.
11008
11009 2013-07-27 Joern Rennecke <joern.rennecke@embecosm.com>
11010
11011 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
11012 Also reserve space for saving UNKNOWN_REGNUM for leaf functions.
11013
11014 2013-07-26 Cary Coutant <ccoutant@google.com>
11015
11016 * dwarf2out.c (die_checksum_ordered): Don't include template
11017 instantiations in signature.
11018 (is_template_parameter): New function.
11019 (is_template_instantiation): New function.
11020 (generate_skeleton_bottom_up): Don't include template instantiations
11021 in type unit DIE.
11022 (generate_skeleton): Likewise.
11023 (break_out_comdat_types): Move recursive call to break out nested
11024 types earlier.
11025 (prune_unused_types_mark_generic_parms_dies): Call
11026 is_template_parameter.
11027
11028 2013-07-26 Ian Bolton <ian.bolton@arm.com>
11029
11030 * config/aarch64/aarch64.md (neg<mode>2): Offer alternative that
11031 uses vector registers.
11032 * config/aarch64/iterators.md: Add attributes rtn and vas.
11033
11034 2013-07-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11035 Richard Earnshaw <richard.earnshaw@arm.com>
11036
11037 * combine.c (simplify_comparison): Re-canonicalize operands
11038 where appropriate.
11039 * config/arm/arm.md (movcond_addsi): New splitter.
11040
11041 2013-07-25 Sterling Augustine <saugustine@google.com>
11042
11043 * dwarf2out.c (size_of_pubnames): Move code to...
11044 (include_pubname_in_output): ...here. New.
11045 (want_pubnames): Rearrange.
11046 (output_pubnames): Call include_pubname_in_output. Move assertion.
11047
11048 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11049
11050 * doc/extend.texi: Fix return types for __builtin_ia32_cmp*s builtins.
11051
11052 2013-07-25 Cameron McInally <cameron.mcinally@nyu.edu>
11053
11054 PR target/38836
11055 * doc/extend.texi: Remove obsolete builtins. Fix
11056 typo for __builtin_ia32_loadss and __builtin_ia32_cmpnltss.
11057
11058 2013-07-25 Jan Hubicka <jh@suse.cz>
11059
11060 * cgraph.c (release_function_body): Break out from ...
11061 (cgraph_release_function_body): ... this one; also release DECL_RESULT
11062 and DECL_ARGUMENTS.
11063 * ipa-cp.c (get_replacement_map): Add parm_num argument; do not set
11064 old_tree in the map.
11065 (create_specialized_node): Update.
11066 * lto-cgraph.c (output_node_opt_summary): Do not translate old_tree
11067 into index.
11068 * cgraphclones.c (cgraph_create_virtual_clone): Do not copy
11069 DECL_ARGUMENTS, DECL_INITIAL and DECL_RESULT.
11070 * ipa-prop.c (ipa_populate_param_decls): Look for origin of clones.
11071 * tree-inline.c (initialize_cfun): Initialize DECL_ARGUMENTS and
11072 DECL_RESULT.
11073
11074 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11075
11076 * config/arm/arm.md (arm_addsi3, addsi3_carryin_<optab>,
11077 addsi3_carryin_alt2_<optab>): Correct output template.
11078
11079 2013-07-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11080
11081 * config/arm/arm-fixed.md (ssmulsa3, usmulusa3):
11082 Adjust for arm_restrict_it.
11083 Remove trailing whitespace.
11084
11085 2013-07-25  Mark Kettenis  <kettenis@openbsd.org>
11086
11087 * config/pa/pa.c (pa_trampoline_init): Emit __enable_execute_stack
11088 libcall if HAVE_ENABLE_EXECUTE_STACK is defined.
11089
11090 * config.gcc (hppa-*-openbsd*): Don't set tmake_file.
11091
11092 2013-07-25 Vladimir Makarov <vmakarov@redhat.com>
11093
11094 PR rtl-optimization/57960
11095 * lra-constraints.c (process_alt_operands): Use the right mode
11096 when checking strict_low.
11097
11098 2013-07-25 Jan Hubicka <jh@suse.cz>
11099
11100 * lto-symtab.c (lto_cgraph_replace_node): Release function body.
11101 * cgraph.c (cgraph_remove_node): Do not release function body
11102 when in cgraph streaming.
11103 * ipa.c (process_references, symtab_remove_unreachable_nodes): Objects
11104 in other partitions are not considered reachable; fix handling of
11105 clones.
11106
11107 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11108
11109 * config/arm/arm.md (*sibcall_insn): Remove unnecessary space.
11110
11111 2013-07-25 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
11112
11113 PR target/19599
11114 PR target/57731
11115 PR target/57837
11116 * config/arm/arm.md ("*sibcall_insn): Replace use of
11117 Ss with US. Adjust output for v5 and v4t.
11118 (*sibcall_value_insn): Likewise and loosen predicate on operand0.
11119
11120 * config/arm/constraints.md ("Ss"): Rename to US.
11121
11122 2013-07-25 Terry Guo <terry.guo@arm.com>
11123
11124 * config/arm/arm.c (thumb1_size_rtx_costs): Assign proper cost for
11125 shift_add/shift_sub0/shift_sub1 RTXs.
11126
11127 2013-07-24 Bill Schmidt <wschmidt@linux.ibm.com>
11128 Anton Blanchard <anton@au1.ibm.com>
11129
11130 * config/rs6000/altivec.md (altivec_vpkpx): Handle little endian.
11131 (altivec_vpks<VI_char>ss): Likewise.
11132 (altivec_vpks<VI_char>us): Likewise.
11133 (altivec_vpku<VI_char>us): Likewise.
11134 (altivec_vpku<VI_char>um): Likewise.
11135
11136 2013-07-24 David Malcolm <dmalcolm@redhat.com>
11137
11138 Introduce context class.
11139
11140 * Makefile.in (CONTEXT_H): New.
11141 (OBJS): Add context.o.
11142 (toplev.o): Add CONTEXT_H to dependencies.
11143 (context.o): New.
11144
11145 * toplev.c (general_init): Create the singleton gcc::context instance.
11146
11147 * context.c: New.
11148
11149 * context.h: New.
11150
11151 2013-07-24 Joern Rennecke <joern.rennecke@embecosm.com>
11152
11153 PR rtl-optimization/57968
11154 * mode-switching.c (create_pre_exit): Allow instructions that
11155 don't set a return register to need a non-exit mode.
11156
11157 2013-07-24 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11158 Anton Blanchard <anton@au1.ibm.com>
11159
11160 * config/rs6000/vector.md (vec_realign_load_<mode>): Reorder input
11161 operands to vperm for little endian.
11162 * config/rs6000/rs6000.c (rs6000_expand_builtin): Use lvsr instead
11163 of lvsl to create the control mask for a vperm for little endian.
11164
11165 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11166 Anton Blanchard <anton@au1.ibm.com>
11167
11168 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Reverse
11169 two operands for little-endian.
11170
11171 2013-07-23 Steve Ellcey <sellcey@mips.com>
11172
11173 * config/mips/mips.c (mips_case_values_threshold): New.
11174 (TARGET_CASE_VALUES_THRESHOLD): Define.
11175
11176 2013-07-23 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
11177 Anton Blanchard <anton@au1.ibm.com>
11178
11179 * config/rs6000/rs6000.c (altivec_expand_vec_perm_const): Correct
11180 selection of field for vector splat in little endian mode.
11181
11182 2013-07-23 Michael Meissner <meissner@linux.vnet.ibm.com>
11183
11184 * config/rs6000/vector.md (xor<mode>3): Move 128-bit boolean
11185 expanders to rs6000.md.
11186 (ior<mode>3): Likewise.
11187 (and<mode>3): Likewise.
11188 (one_cmpl<mode>2): Likewise.
11189 (nor<mode>3): Likewise.
11190 (andc<mode>3): Likewise.
11191 (eqv<mode>3): Likewise.
11192 (nand<mode>3): Likewise.
11193 (orc<mode>3): Likewise.
11194
11195 * config/rs6000/rs6000-protos.h (rs6000_split_logical): New
11196 declaration.
11197
11198 * config/rs6000/rs6000.c (rs6000_split_logical_inner): Add support
11199 to split multi-word logical operations.
11200 (rs6000_split_logical_di): Likewise.
11201 (rs6000_split_logical): Likewise.
11202
11203 * config/rs6000/vsx.md (VSX_L2): Delete, no longer used.
11204 (vsx_and<mode>3_32bit): Move 128-bit logical insns to rs6000.md,
11205 and allow TImode operations in 32-bit.
11206 (vsx_and<mode>3_64bit): Likewise.
11207 (vsx_ior<mode>3_32bit): Likewise.
11208 (vsx_ior<mode>3_64bit): Likewise.
11209 (vsx_xor<mode>3_32bit): Likewise.
11210 (vsx_xor<mode>3_64bit): Likewise.
11211 (vsx_one_cmpl<mode>2_32bit): Likewise.
11212 (vsx_one_cmpl<mode>2_64bit): Likewise.
11213 (vsx_nor<mode>3_32bit): Likewise.
11214 (vsx_nor<mode>3_64bit): Likewise.
11215 (vsx_andc<mode>3_32bit): Likewise.
11216 (vsx_andc<mode>3_64bit): Likewise.
11217 (vsx_eqv<mode>3_32bit): Likewise.
11218 (vsx_eqv<mode>3_64bit): Likewise.
11219 (vsx_nand<mode>3_32bit): Likewise.
11220 (vsx_nand<mode>3_64bit): Likewise.
11221 (vsx_orc<mode>3_32bit): Likewise.
11222 (vsx_orc<mode>3_64bit): Likewise.
11223
11224 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Always allow vector
11225 logical types in GPRs.
11226
11227 * config/rs6000/altivec.md (altivec_and<mode>3): Move 128-bit
11228 logical insns to rs6000.md, and allow TImode operations in
11229 32-bit.
11230 (altivec_ior<mode>3): Likewise.
11231 (altivec_xor<mode>3): Likewise.
11232 (altivec_one_cmpl<mode>2): Likewise.
11233 (altivec_nor<mode>3): Likewise.
11234 (altivec_andc<mode>3): Likewise.
11235
11236 * config/rs6000/rs6000.md (BOOL_128): New mode iterators and mode
11237 attributes for moving the 128-bit logical operations into
11238 rs6000.md.
11239 (BOOL_REGS_OUTPUT): Likewise.
11240 (BOOL_REGS_OP1): Likewise.
11241 (BOOL_REGS_OP2): Likewise.
11242 (BOOL_REGS_UNARY): Likewise.
11243 (BOOL_REGS_AND_CR0): Likewise.
11244 (one_cmpl<mode>2): Add support for DI logical operations on
11245 32-bit, splitting the operations to 32-bit.
11246 (anddi3): Likewise.
11247 (iordi3): Likewise.
11248 (xordi3): Likewise.
11249 (and<mode>3, 128-bit types): Rewrite 2013-06-06 logical operator
11250 changes to combine the 32/64-bit code, allow logical operations on
11251 TI mode in 32-bit, and to use similar match_operator patterns like
11252 scalar mode uses. Combine the Altivec and VSX code for logical
11253 operations, and move it here.
11254 (ior<mode>3, 128-bit types): Likewise.
11255 (xor<mode>3, 128-bit types): Likewise.
11256 (one_cmpl<mode>3, 128-bit types): Likewise.
11257 (nor<mode>3, 128-bit types): Likewise.
11258 (andc<mode>3, 128-bit types): Likewise.
11259 (eqv<mode>3, 128-bit types): Likewise.
11260 (nand<mode>3, 128-bit types): Likewise.
11261 (orc<mode>3, 128-bit types): Likewise.
11262 (and<mode>3_internal): Likewise.
11263 (bool<mode>3_internal): Likewise.
11264 (boolc<mode>3_internal1): Likewise.
11265 (boolc<mode>3_internal2): Likewise.
11266 (boolcc<mode>3_internal1): Likewise.
11267 (boolcc<mode>3_internal2): Likewise.
11268 (eqv<mode>3_internal1): Likewise.
11269 (eqv<mode>3_internal2): Likewise.
11270 (one_cmpl1<mode>3_internal): Likewise.
11271
11272 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11273
11274 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11275 Rename flag_stack_usage to flag_stack_usage_info.
11276
11277 2013-07-23 David Holsgrove <david.holsgrove@xilinx.com>
11278
11279 * config/microblaze/sync.md: New file.
11280 * config/microblaze/microblaze.md: Include sync.md
11281 * config/microblaze/microblaze.c: Add print_operand 'y'.
11282 * config/microblaze/constraints.md: Add memory_contraint
11283 'Q' which is a single register.
11284
11285 2013-07-23 Eric Botcazou <ebotcazou@adacore.com>
11286
11287 * doc/invoke.texi (SPARC Options): Document new leon3 processor value.
11288
11289 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11290
11291 * reload.c (find_reloads): Exit loop once we find this operand
11292 cannot be reloaded somehow for this alternative.
11293
11294 * reload.c (find_reloads): Exit loop once we find a hard register.
11295
11296 * rtlanal.c (computed_jump_p): Exit loop once we find label
11297 reference is used.
11298
11299 * i386.c (ix86_pad_returns): Exit loop after setting replace.
11300
11301 * cfgloopmanip.c (remove_path): Exit loop after setting
11302 irred_invalidated.
11303
11304 * gensupport.c (subst_dup): Avoid loop if code is not
11305 MATCH_DUP nor MATCH_OP_DUP.
11306
11307 2013-07-23 Nicklas Bo Jensen <nbjensen@gmail.com>
11308
11309 * doc/md.texi (Machine-Specific Peephole Optimizers): Fix a typo.
11310
11311 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11312
11313 * config/aarch64/aarch64.c (aarch64_hard_regno_mode_ok): Also return
11314 true for SP_REGNUM if mode == ptr_mode.
11315 * config/aarch64/aarch64.h (ADDITIONAL_REGISTER_NAMES): Add "wsp"
11316 with value R0_REGNUM + 31.
11317
11318 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11319
11320 * config/aarch64/aarch64.c (aarch64_pad_arg_upward): In big-endian,
11321 pad pointer-typed argument downward.
11322
11323 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11324
11325 * config/aarch64/aarch64.h (TARGET_CPU_CPP_BUILTINS): Define _ILP32
11326 and __ILP32__ when the ILP32 model is in use.
11327
11328 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11329
11330 * config/aarch64/aarch64.c (POINTER_BYTES): New define.
11331 (aarch64_load_symref_appropriately): In the case of
11332 SYMBOL_SMALL_ABSOLUTE, use the mode of 'dest' instead of Pmode
11333 to generate new rtx; likewise to the case of SYMBOL_SMALL_GOT.
11334 (aarch64_expand_mov_immediate): In the case of SYMBOL_FORCE_TO_MEM,
11335 change to pass 'ptr_mode' to force_const_mem and zero-extend 'mem'
11336 if 'mode' doesn't equal to 'ptr_mode'.
11337 (aarch64_output_mi_thunk): Add an assertion on the alignment of
11338 'vcall_offset'; change to call aarch64_emit_move differently depending
11339 on whether 'Pmode' equals to 'ptr_mode' or not; use 'POINTER_BYTES'
11340 to calculate the upper bound of 'vcall_offset'.
11341 (aarch64_cannot_force_const_mem): Change to also return true if
11342 mode != ptr_mode.
11343 (aarch64_legitimize_reload_address): In the case of large
11344 displacements, add new local variable 'xmode' and an assertion
11345 based on it; change to use 'xmode' to generate the new rtx and
11346 reload.
11347 (aarch64_asm_trampoline_template): Change to generate the template
11348 differently depending on TARGET_ILP32 or not; change to use
11349 'POINTER_BYTES' in the argument passed to assemble_aligned_integer.
11350 (aarch64_trampoline_size): Removed.
11351 (aarch64_trampoline_init): Add new local constant 'tramp_code_sz'
11352 and replace immediate literals with it. Change to use 'ptr_mode'
11353 instead of 'DImode' and call convert_memory_address if the mode
11354 of 'fnaddr' doesn't equal to 'ptr_mode'.
11355 (aarch64_elf_asm_constructor): Change to use assemble_aligned_integer
11356 to output symbol.
11357 (aarch64_elf_asm_destructor): Likewise.
11358 * config/aarch64/aarch64.h (TRAMPOLINE_SIZE): Change to be dependent
11359 on TARGET_ILP32 instead of aarch64_trampoline_size.
11360 * config/aarch64/aarch64.md (movsi_aarch64): Add new alternatives
11361 of 'mov' between WSP and W registers as well as 'adr' and 'adrp'.
11362 (loadwb_pair<GPI:mode>_<PTR:mode>): Rename to ...
11363 (loadwb_pair<GPI:mode>_<P:mode>): ... this. Replace PTR with P.
11364 (storewb_pair<GPI:mode>_<PTR:mode>): Likewise; rename to ...
11365 (storewb_pair<GPI:mode>_<P:mode>): ... this.
11366 (add_losym): Change to 'define_expand' and call gen_add_losym_<mode>
11367 depending on the value of 'mode'.
11368 (add_losym_<mode>): New.
11369 (ldr_got_small_<mode>): New, based on ldr_got_small.
11370 (ldr_got_small): Remove.
11371 (ldr_got_small_sidi): New.
11372 * config/aarch64/iterators.md (P): New.
11373 (PTR): Change to 'ptr_mode' in the condition.
11374
11375 2013-07-23 Yufeng Zhang <yufeng.zhang@arm.com>
11376
11377 * config.gcc (aarch64*-*-*): Support --with-abi.
11378 (aarch64*-*-elf): Support --with-multilib-list.
11379 (aarch64*-*-linux*): Likewise.
11380 (supported_defaults): Add abi to aarch64*-*-*.
11381 * configure.ac: Mention AArch64 for --with-multilib-list.
11382 * configure: Re-generated.
11383 * config/aarch64/biarchilp32.h: New file.
11384 * config/aarch64/biarchlp64.h: New file.
11385 * config/aarch64/aarch64-elf.h (ENDIAN_SPEC): New define.
11386 (ABI_SPEC): Ditto.
11387 (MULTILIB_DEFAULTS): Ditto.
11388 (DRIVER_SELF_SPECS): Ditto.
11389 (ASM_SPEC): Update to also substitute -mabi.
11390 * config/aarch64/aarch64-elf-raw.h (LINK_SPEC): Add linker script
11391 file whose name depends on -mabi= and -mbig-endian.
11392 * config/aarch64/aarch64.h (LONG_TYPE_SIZE): Change to depend on
11393 TARGET_ILP32.
11394 (POINTER_SIZE): New define.
11395 (POINTERS_EXTEND_UNSIGNED): Ditto.
11396 (enum aarch64_abi_type): New enumeration tag.
11397 (AARCH64_ABI_LP64, AARCH64_ABI_ILP32): New enumerators.
11398 (AARCH64_ABI_DEFAULT): Define to AARCH64_ABI_LP64 if undefined.
11399 (TARGET_ILP32): New define.
11400 * config/aarch64/aarch64.opt (mabi): New.
11401 (aarch64_abi): New.
11402 (ilp32, lp64): New values for -mabi.
11403 * config/aarch64/t-aarch64 (comma): New define.
11404 (MULTILIB_OPTIONS): Ditto.
11405 (MULTILIB_DIRNAMES): Ditto.
11406 * config/aarch64/t-aarch64-linux (MULTIARCH_DIRNAME): New define.
11407 * doc/invoke.texi: Document -mabi for AArch64.
11408
11409 2013-07-23 Georg-Johann Lay <avr@gjlay.de>
11410
11411 * config/avr/avr.md: Explain asm print modifier 'r' for REG.
11412
11413 2013-07-22 Bill Schmidt <wschmidt@vnet.linux.ibm.com>
11414 Anton Blanchard <anton@au1.ibm.com>
11415
11416 * config/rs6000/rs6000.c (rs6000_expand_vector_init): Fix
11417 endianness when selecting field to splat.
11418
11419 2013-07-22 Eric Christopher <echristo@gmail.com>
11420
11421 * dwarf2out.c (die_odr_checksum): New function to use
11422 CHECKSUM_ macros and ULEB128 for DIE tag.
11423 (generate_type_signature): Use.
11424
11425 2013-07-22 Eric Botcazou <ebotcazou@adacore.com>
11426
11427 * config.gcc (sparc*-*-*): Accept leon3 processor.
11428 (sparc-leon*-*): Merge with sparc*-*-* and add leon3 support.
11429 * doc/invoke.texi (SPARC Options): Adjust -mfix-ut699 entry.
11430 * config/sparc/sparc-opts.h (enum processor_type): Add PROCESSOR_LEON3.
11431 * config/sparc/sparc.opt (enum processor_type): Add leon3.
11432 (mfix-ut699): Adjust comment.
11433 * config/sparc/sparc.h (TARGET_CPU_leon3): New define.
11434 (CPP_CPU32_DEFAULT_SPEC): Add leon3 support.
11435 (CPP_CPU_SPEC): Likewise.
11436 (ASM_CPU_SPEC): Likewise.
11437 * config/sparc/sparc.c (leon3_cost): New constant.
11438 (sparc_option_override): Add leon3 support.
11439 (mem_ref): New function.
11440 (sparc_gate_work_around_errata): Return true if -mfix-ut699 is enabled.
11441 (sparc_do_work_around_errata): Look into the instruction in the delay
11442 slot and adjust accordingly. Add fix for the data cache nullify issues
11443 of the UT699. Change insertion position for the NOP.
11444 * config/sparc/leon.md (leon_fpalu, leon_fpmds, write_buf): Delete.
11445 (leon3_load): New reservation.
11446 (leon_store): Bump latency to 2.
11447 (grfpu): New automaton.
11448 (grfpu_alu): New unit.
11449 (grfpu_ds): Likewise.
11450 (leon_fp_alu): Adjust.
11451 (leon_fp_mult): Delete.
11452 (leon_fp_div): Split into leon_fp_divs and leon_fp_divd.
11453 (leon_fp_sqrt): Split into leon_fp_sqrts and leon_fp_sqrtd.
11454 * config/sparc/sparc.md (cpu): Add leon3.
11455 * config/sparc/sync.md (atomic_exchangesi): Disable if -mfix-ut699.
11456 (swapsi): Likewise.
11457 (atomic_test_and_set): Likewise.
11458 (ldstub): Likewise.
11459
11460 2013-07-22 Jürgen Urban <JuergenUrban@gmx.de>
11461
11462 * config.gcc (mips*-*-*): Add --with-fpu support. Make single the
11463 default for R5900 targets.
11464 * config/mips/mips.h (OPTION_DEFAULT_SPECS): Handle --with-fpu.
11465 (ISA_HAS_LDC1_SDC1): Set to false for TARGET_MIPS5900.
11466 * config/mips/mips.c (mips_option_override): Report an error for
11467 -march=r5900 -mhard-float -mdouble-float. Use spu_single_format
11468 for -march=r5900 -mhard-float.
11469
11470 2013-07-22 Po-Chun Chang <pchang9@cs.wisc.edu>
11471
11472 * df-problems.c (can_move_insns_across): Exit loop once we
11473 find a non-fixed, non-global register.
11474
11475 * ipa-pure-const.c (propagate_nothrow): Exit loop after
11476 setting can_throw.
11477
11478 * omega.c (omega_eliminate_red): Break after setting red_found.
11479 (omega_problem_has_red_equations): Similarly after setting found.
11480 (omega_query_variable): Similarly after setting coupled.
11481
11482 2013-07-22 Marek Polacek <polacek@redhat.com>
11483
11484 * gimplify.c: Don't include gimple.h twice.
11485
11486 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11487
11488 * config/arm/constraints.md (Pd): Allow TARGET_THUMB
11489 instead of TARGET_THUMB1.
11490 (Pz): New constraint.
11491 * config/arm/arm.md (arm_addsi3): Add alternatives for 16-bit
11492 encodings.
11493 (compare_negsi_si): Likewise.
11494 (compare_addsi2_op0): Likewise.
11495 (compare_addsi2_op1): Likewise.
11496 (addsi3_carryin_<optab>): Likewise.
11497 (addsi3_carryin_alt2_<optab>): Likewise.
11498 (addsi3_carryin_shift_<optab>): Disable cond_exec variant
11499 for arm_restrict_it.
11500 (subsi3_carryin): Likewise.
11501 (arm_subsi3_insn): Add alternatives for 16-bit encoding.
11502 (minmax_arithsi): Disable for arm_restrict_it.
11503 (minmax_arithsi_non_canon): Adjust for arm_restrict_it.
11504 (satsi_<SAT:code>): Disable cond_exec variant for arm_restrict_it.
11505 (satsi_<SAT:code>_shift): Likewise.
11506 (arm_shiftsi3): Add alternative for 16-bit encoding.
11507 (arm32_movhf): Disable for arm_restrict_it.
11508 (arm_cmpdi_unsigned): Add alternatives for 16-bit encoding.
11509 (arm_movtas_ze): Disable cond_exec variant for arm_restrict_it.
11510
11511 2013-07-22 Sofiane Naci <sofiane.naci@arm.com>
11512
11513 * config/arm/arm.md (attribute "insn"): Delete.
11514 (attribute "type"): Add "mov_imm", "mov_reg", "mov_shift",
11515 "mov_shift_reg", "mvn_imm", "mvn_reg", "mvn_shift" and "mvn_shift_reg".
11516 (not_shiftsi): Update for attribute change.
11517 (not_shiftsi_compare0): Likewise.
11518 (not_shiftsi_compare0_scratch): Likewise.
11519 (arm_one_cmplsi2): Likewise.
11520 (thumb1_one_cmplsi2): Likewise.
11521 (notsi_compare0): Likewise.
11522 (notsi_compare0_scratch): Likewise.
11523 (thumb1_movdi_insn): Likewise.
11524 (arm_movsi_insn): Likewise.
11525 (movhi_insn_arch4): Likewise.
11526 (movhi_bytes): Likewise.
11527 (arm_movqi_insn): Likewise.
11528 (thumb1_movqi_insn): Likewise.
11529 (arm32_movhf): Likewise.
11530 (thumb1_movhf): Likewise.
11531 (arm_movsf_soft_insn): Likewise.
11532 (thumb1_movsf_insn): Likewise.
11533 (thumb_movdf_insn): Likewise.
11534 (movsicc_insn): Likewise.
11535 (movsfcc_soft_insn): Likewise.
11536 (and_scc): Likewise.
11537 (cond_move): Likewise.
11538 (if_move_not): Likewise.
11539 (if_not_move): Likewise.
11540 (if_shift_move): Likewise.
11541 (if_move_shift): Likewise.
11542 (if_shift_shift): Likewise.
11543 (if_not_arith): Likewise.
11544 (if_arith_not): Likewise.
11545 (cond_move_not): Likewise.
11546 * config/arm/neon.md (neon_mov<mode>): Update for attribute change.
11547 (neon_mov<mode>): Likewise.
11548 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute change.
11549 (thumb2_movsi_vfp): Likewise.
11550 (movsf_vfp): Likewise.
11551 (thumb2_movsf_vfp): Likewise.
11552 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11553 change.
11554 (cortexa7_older_only): Likewise.
11555 (cortexa7_younger): Likewise.
11556 * config/arm/arm1020e.md (1020alu_op): Update for attribute change.
11557 (1020alu_shift_op): Likewise.
11558 (1020alu_shift_reg_op): Likewise.
11559 * config/arm/arm1026ejs.md (alu_op): Update for attribute change.
11560 (alu_shift_op): Likewise.
11561 (alu_shift_reg_op): Likewise.
11562 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute change.
11563 (11_alu_shift_op): Likewise.
11564 (11_alu_shift_reg_op): Likewise.
11565 * config/arm/arm926ejs.md (9_alu_op): Update for attribute change.
11566 (9_alu_shift_reg_op): Likewise.
11567 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11568 change.
11569 (cortex_a15_alu_shift): Likewise.
11570 (cortex_a15_alu_shift_reg): Likewise.
11571 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute change.
11572 (cortex_a5_alu_shift): Likewise.
11573 * config/arm/cortex-a53.md (cortex_a53_alu): Update for attribute
11574 change.
11575 (cortex_a53_alu_shift): Likewise.
11576 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11577 change.
11578 (cortex_a7_alu_reg): Likewise.
11579 (cortex_a7_alu_shift): Likewise.
11580 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11581 (cortex_a8_alu_shift): Likewise.
11582 (cortex_a8_alu_shift_reg): Likewise.
11583 (cortex_a8_mov): Likewise.
11584 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute change.
11585 (cortex_a9_dp_shift): Likewise.
11586 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute change.
11587 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute change.
11588 (cortex_r4_mov): Likewise.
11589 (cortex_r4_alu_shift): Likewise.
11590 (cortex_r4_alu_shift_reg): Likewise.
11591 * config/arm/fa526.md (526_alu_op): Update for attribute change.
11592 (526_alu_shift_op): Likewise.
11593 * config/arm/fa606te.md (606te_alu_op): Update for attribute change.
11594 * config/arm/fa626te.md (626te_alu_op): Update for attribute change.
11595 (626te_alu_shift_op): Likewise.
11596 * config/arm/fa726te.md (726te_shift_op): Update for attribute change.
11597 (726te_alu_op): Likewise.
11598 (726te_alu_shift_op): Likewise.
11599 (726te_alu_shift_reg_op): Likewise.
11600 * config/arm/fmp626.md (mp626_alu_op): Update for attribute change.
11601 (mp626_alu_shift_op): Likewise.
11602 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute change.
11603 (pj4_alu_e1_conds): Likewise.
11604 (pj4_alu): Likewise.
11605 (pj4_alu_conds): Likewise.
11606 (pj4_shift): Likewise.
11607 (pj4_shift_conds): Likewise.
11608 (pj4_alu_shift): Likewise.
11609 (pj4_alu_shift_conds): Likewise.
11610
11611 2013-07-22 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
11612
11613 * config/arm/predicates.md (shiftable_operator_strict_it):
11614 New predicate.
11615 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si):
11616 Disable cond_exec version for arm_restrict_it.
11617 (thumb2_smaxsi3): Convert to generate cond_exec.
11618 (thumb2_sminsi3): Likewise.
11619 (thumb32_umaxsi3): Likewise.
11620 (thumb2_uminsi3): Likewise.
11621 (thumb2_abssi2): Adjust constraints for arm_restrict_it.
11622 (thumb2_neg_abssi2): Likewise.
11623 (thumb2_mov_scc): Add alternative for 16-bit encoding.
11624 (thumb2_movsicc_insn): Adjust alternatives.
11625 (thumb2_mov_negscc): Disable for arm_restrict_it.
11626 (thumb2_mov_negscc_strict_it): New pattern.
11627 (thumb2_mov_notscc_strict_it): New pattern.
11628 (thumb2_mov_notscc): Disable for arm_restrict_it.
11629 (thumb2_ior_scc): Likewise.
11630 (thumb2_ior_scc_strict_it): New pattern.
11631 (thumb2_cond_move): Adjust for arm_restrict_it.
11632 (thumb2_cond_arith): Disable for arm_restrict_it.
11633 (thumb2_cond_arith_strict_it): New pattern.
11634 (thumb2_cond_sub): Adjust for arm_restrict_it.
11635 (thumb2_movcond): Likewise.
11636 (thumb2_extendqisi_v6): Disable cond_exec variant for arm_restrict_it.
11637 (thumb2_zero_extendhisi2_v6): Likewise.
11638 (thumb2_zero_extendqisi2_v6): Likewise.
11639 (orsi_notsi_si): Likewise.
11640 (orsi_not_shiftsi_si): Likewise.
11641
11642 2013-07-22 Georg-Johann Lay <avr@gjlay.de>
11643
11644 * config/avr/avr.c (avr_out_xload): No SBIS around LPM so that
11645 instruction sequence is 1 byte shorter.
11646
11647 2013-07-22 Uros Bizjak <ubizjak@gmail.com>
11648
11649 * config/i386/i386.md (nonlocal_goto_receiver): Delete insn if
11650 it is not needed after split.
11651
11652 2013-07-20 Iain Sandoe <iain@codesourcery.com>
11653
11654 PR target/51784
11655 * config/i386/i386.c (output_set_got) [TARGET_MACHO]: Adjust to emit a
11656 second label for nonlocal goto receivers. Don't output pic base labels
11657 unless we're producing PIC; mark that action unreachable().
11658 (ix86_save_reg): If the function contains a nonlocal label, save the
11659 PIC base reg.
11660 * config/darwin-protos.h (machopic_should_output_picbase_label): New.
11661 * config/darwin.c (emitted_pic_label_num): New GTY.
11662 (update_pic_label_number_if_needed): New.
11663 (machopic_output_function_base_name): Adjust for nonlocal receiver
11664 case.
11665 (machopic_should_output_picbase_label): New.
11666 * config/i386/i386.md (enum unspecv): UNSPECV_NLGR: New.
11667 (nonlocal_goto_receiver): New insn and split.
11668
11669 2013-07-20 James Greenhalgh <james.greenhalgh@arm.com>
11670
11671 * config/aarch64/aarch64-builtins.c
11672 (aarch64_fold_builtin): Fold abs in all modes.
11673 * config/aarch64/aarch64-simd-builtins.def
11674 (abs): Enable for all modes.
11675 * config/aarch64/arm_neon.h
11676 (vabs<q>_s<8,16,32,64): Rewrite using builtins.
11677 (vabs_f64): Add missing intrinsic.
11678
11679 2013-07-19 Ian Bolton <ian.bolton@arm.com>
11680
11681 * config/aarch64/arm_neon.h (vabs_s64): New function
11682
11683 2013-07-19 Georg-Johann Lay <avr@gjlay.de>
11684
11685 PR target/57516
11686 * config/avr/avr-fixed.md (round<mode>3_const): Turn expander to insn.
11687 * config/avr/avr.md (adjust_len): Add `round'.
11688 * config/avr/avr-protos.h (avr_out_round): New prototype.
11689 (avr_out_plus): Add `out_label' argument.
11690 * config/avr/avr.c (avr_out_plus_1): Add `out_label' argument.
11691 (avr_out_plus): Pass down `out_label' to avr_out_plus_1.
11692 Handle the case where `insn' is just a pattern.
11693 (avr_out_bitop): Handle the case where `insn' is just a pattern.
11694 (avr_out_round): New function.
11695 (avr_adjust_insn_length): Handle ADJUST_LEN_ROUND.
11696
11697 2013-07-18 David Holsgrove <david.holsgrove@xilinx.com>
11698
11699 * config/microblaze/microblaze.c (microblaze_expand_prologue):
11700 Add check for flag_stack_usage to handle -fstack-usage support
11701
11702 2013-07-18 Pat Haugen <pthaugen@us.ibm.com>
11703
11704 * config/rs6000/rs6000.c (rs6000_option_override_internal): Adjust flag
11705 interaction for new Power8 flags and VSX.
11706
11707 2013-07-18 Sriraman Tallam <tmsriram@google.com>
11708
11709 PR middle-end/57698
11710 * tree-inline.c (expand_call_inline): Emit errors during
11711 early_inlining only if optimization is not turned on.
11712
11713 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11714
11715 * passes.def: New.
11716
11717 * passes.c (init_optimization_passes): Move the construction of
11718 the pass hierarchy into a new passes.def file.
11719
11720 * Makefile.in (passes.o): Add dependency on passes.def.
11721
11722 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11723
11724 * passes.c (init_optimization_passes): Introduce macros for
11725 constructing the tree of passes (INSERT_PASSES_AFTER,
11726 PUSH_INSERT_PASSES_WITHIN, POP_INSERT_PASSES,
11727 TERMINATE_PASS_LIST).
11728
11729 2013-07-18 Vladimir Makarov <vmakarov@redhat.com>
11730 Wei Mi <wmi@google.com>
11731
11732 PR rtl-optimization/57878
11733 * lra-assigns.c (assign_by_spills): Move non_reload_pseudos to the
11734 top.
11735 (reload_pseudo_compare_func): Check nregs first for reload
11736 pseudos.
11737
11738 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11739
11740 * tree-pass.h (pass_ipa_lto_wpa_fixup): Remove redundant decl.
11741
11742 2013-07-18 Po-Chun Chang <pchang9@cs.wisc.edu>
11743
11744 * read-rtl.c (validate_const_int): Once an invalid character is
11745 seen, quit the loop.
11746
11747 * gengtype.c (write_roots): Similarly once we find the "deletable"
11748 or "if_marked" option.
11749
11750 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11751
11752 * config/arm/arm.md (attribute "insn"): Delete values "mrs", "msr",
11753 "xtab" and "sat". Move value "clz" from here to ...
11754 (attriubte "type"): ... here.
11755 (satsi_<SAT:code>): Delete "insn" attribute.
11756 (satsi_<SAT:code>_shift): Likewise.
11757 (arm_zero_extendqisi2addsi): Likewise.
11758 (arm_extendqisi2addsi): Likewise.
11759 (clzsi2): Update for attribute changes.
11760 (rbitsi2): Likewise.
11761 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Delete "insn"
11762 attribute.
11763 (arm_usatsihi): Likewise.
11764 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute change.
11765
11766 2013-07-18 Sofiane Naci <sofiane.naci@arm.com>
11767
11768 * config/arm/arm.md (attribute "type"): Rename "simple_alu_imm" to
11769 "arlo_imm". Rename "alu_reg" to "arlo_reg". Rename "simple_alu_shift"
11770 to "extend". Split "alu_shift" into "shift" and "arlo_shift". Split
11771 "alu_shift_reg" into "shift_reg" and "arlo_shift_reg". List types
11772 in alphabetical order.
11773 (attribute "core_cycles"): Update for attribute changes.
11774 (arm_addsi3): Likewise.
11775 (addsi3_compare0): Likewise.
11776 (addsi3_compare0_scratch): Likewise.
11777 (addsi3_compare_op1): Likewise.
11778 (addsi3_compare_op2): Likewise.
11779 (compare_addsi2_op0): Likewise.
11780 (compare_addsi2_op1): Likewise.
11781 (addsi3_carryin_shift_<optab>): Likewise.
11782 (subsi3_carryin_shift): Likewise.
11783 (rsbsi3_carryin_shift): Likewise.
11784 (arm_subsi3_insn): Likewise.
11785 (subsi3_compare0): Likewise.
11786 (subsi3_compare): Likewise.
11787 (arm_andsi3_insn): Likewise.
11788 (thumb1_andsi3_insn): Likewise.
11789 (andsi3_compare0): Likewise.
11790 (andsi3_compare0_scratch): Likewise.
11791 (zeroextractsi_compare0_scratch
11792 (andsi_not_shiftsi_si): Likewise.
11793 (iorsi3_insn): Likewise.
11794 (iorsi3_compare0): Likewise.
11795 (iorsi3_compare0_scratch): Likewise.
11796 (arm_xorsi3): Likewise.
11797 (thumb1_xorsi3_insn): Likewise.
11798 (xorsi3_compare0): Likewise.
11799 (xorsi3_compare0_scratch): Likewise.
11800 (satsi_<SAT:code>_shift): Likewise.
11801 (rrx): Likewise.
11802 (arm_shiftsi3): Likewise.
11803 (shiftsi3_compare0): Likewise.
11804 (not_shiftsi): Likewise.
11805 (not_shiftsi_compare0): Likewise.
11806 (not_shiftsi_compare0_scratch): Likewise.
11807 (arm_one_cmplsi2): Likewise.
11808 (thumb_one_complsi2): Likewise.
11809 (notsi_compare0): Likewise.
11810 (notsi_compare0_scratch): Likewise.
11811 (thumb1_zero_extendhisi2): Likewise.
11812 (arm_zero_extendhisi2): Likewise.
11813 (arm_zero_extendhisi2_v6): Likewise.
11814 (arm_zero_extendhisi2addsi): Likewise.
11815 (thumb1_zero_extendqisi2): Likewise.
11816 (thumb1_zero_extendqisi2_v6): Likewise.
11817 (arm_zero_extendqisi2): Likewise.
11818 (arm_zero_extendqisi2_v6): Likewise.
11819 (arm_zero_extendqisi2addsi): Likewise.
11820 (thumb1_extendhisi2): Likewise.
11821 (arm_extendhisi2): Likewise.
11822 (arm_extendhisi2_v6): Likewise.
11823 (arm_extendqisi): Likewise.
11824 (arm_extendqisi_v6): Likewise.
11825 (arm_extendqisi2addsi): Likewise.
11826 (thumb1_extendqisi2): Likewise.
11827 (thumb1_movdi_insn): Likewise.
11828 (arm_movsi_insn): Likewise.
11829 (movsi_compare0): Likewise.
11830 (movhi_insn_arch4): Likewise.
11831 (movhi_bytes): Likewise.
11832 (arm_movqi_insn): Likewise.
11833 (thumb1_movqi_insn): Likewise.
11834 (arm32_movhf): Likewise.
11835 (thumb1_movhf): Likewise.
11836 (arm_movsf_soft_insn): Likewise.
11837 (thumb1_movsf_insn): Likewise.
11838 (movdf_soft_insn): Likewise.
11839 (thumb_movdf_insn): Likewise.
11840 (arm_cmpsi_insn): Likewise.
11841 (cmpsi_shiftsi): Likewise.
11842 (cmpsi_shiftsi_swp): Likewise.
11843 (arm_cmpsi_negshiftsi_si): Likewise.
11844 (movsicc_insn): Likewise.
11845 (movsfcc_soft_insn): Likewise.
11846 (arith_shiftsi): Likewise.
11847 (arith_shiftsi_compare0
11848 (arith_shiftsi_compare0_scratch
11849 (sub_shiftsi): Likewise.
11850 (sub_shiftsi_compare0
11851 (sub_shiftsi_compare0_scratch
11852 (and_scc): Likewise.
11853 (cond_move): Likewise.
11854 (if_plus_move): Likewise.
11855 (if_move_plus): Likewise.
11856 (if_move_not): Likewise.
11857 (if_not_move): Likewise.
11858 (if_shift_move): Likewise.
11859 (if_move_shift): Likewise.
11860 (if_shift_shift): Likewise.
11861 (if_not_arith): Likewise.
11862 (if_arith_not): Likewise.
11863 (cond_move_not): Likewise.
11864 (thumb1_ashlsi3): Set type attribute.
11865 (thumb1_ashrsi3): Likewise.
11866 (thumb1_lshrsi3): Likewise.
11867 (thumb1_rotrsi3): Likewise.
11868 (shiftsi3_compare0_scratch): Likewise.
11869 * config/arm/neon.md (neon_mov<mode>): Update for attribute changes.
11870 (neon_mov<mode>): Likewise.
11871 * config/arm/thumb2.md (thumb_andsi_not_shiftsi_si): Update for
11872 attribute changes.
11873 (thumb2_movsi_insn): Likewise.
11874 (thumb2_cmpsi_neg_shiftsi): Likewise.
11875 (thumb2_extendqisi_v6): Likewise.
11876 (thumb2_zero_extendhisi2_v6): Likewise.
11877 (thumb2_zero_extendqisi2_v6): Likewise.
11878 (thumb2_shiftsi3_short): Likewise.
11879 (thumb2_addsi3_compare0_scratch): Likewise.
11880 (orsi_not_shiftsi_si): Likewise.
11881 * config/arm/vfp.md (arm_movsi_vfp): Update for attribute changes.
11882 * config/arm/arm-fixed.md (arm_ssatsihi_shift): Update for attribute
11883 changes.
11884 * config/arm/arm1020e.md (1020alu_op): Update for attribute changes.
11885 (1020alu_shift_op): Likewise.
11886 (1020alu_shift_reg_op): Likewise.
11887 * config/arm/arm1026ejs.md (alu_op): Update for attribute changes.
11888 (alu_shift_op): Likewise.
11889 (alu_shift_reg_op): Likewise.
11890 * config/arm/arm1136jfs.md (11_alu_op): Update for attribute changes.
11891 (11_alu_shift_op): Likewise.
11892 (11_alu_shift_reg_op): Likewise.
11893 * config/arm/arm926ejs.md (9_alu_op): Update for attribute changes.
11894 (9_alu_shift_reg_op): Likewise.
11895 * config/arm/cortex-a15.md (cortex_a15_alu): Update for attribute
11896 changes.
11897 (cortex_a15_alu_shift): Likewise.
11898 (cortex_a15_alu_shift_reg): Likewise.
11899 * config/arm/cortex-a5.md (cortex_a5_alu): Update for attribute
11900 changes.
11901 (cortex_a5_alu_shift): Likewise.
11902 * config/arm/cortex-a53.md (cortex_a53_alu) : Update for attribute
11903 changes.
11904 (cortex_a53_alu_shift): Likewise.
11905 * config/arm/cortex-a7.md (cortex_a7_alu_imm): Update for attribute
11906 changes.
11907 (cortex_a7_alu_reg): Likewise.
11908 (cortex_a7_alu_shift): Likewise.
11909 * config/arm/cortex-a8.md (cortex_a8_alu): Update for attribute
11910 changes.
11911 (cortex_a8_alu_shift): Likewise.
11912 (cortex_a8_alu_shift_reg): Likewise.
11913 (cortex_a8_mov): Likewise.
11914 * config/arm/cortex-a9.md (cortex_a9_dp): Update for attribute changes.
11915 (cortex_a9_dp_shift): Likewise.
11916 * config/arm/cortex-m4.md (cortex_m4_alu): Update for attribute
11917 changes.
11918 * config/arm/cortex-r4.md (cortex_r4_alu): Update for attribute
11919 changes.
11920 (cortex_r4_mov): Likewise.
11921 (cortex_r4_alu_shift): Likewise.
11922 (cortex_r4_alu_shift_reg): Likewise.
11923 * config/arm/fa526.md (526_alu_op): Update for attribute changes.
11924 (526_alu_shift_op): Likewise.
11925 * config/arm/fa606te.md (606te_alu_op): Update for attribute changes.
11926 * config/arm/fa626te.md (626te_alu_op): Update for attribute changes.
11927 (626te_alu_shift_op): Likewise.
11928 * config/arm/fa726te.md (726te_shift_op): Update for attribute changes.
11929 (726te_alu_op): Likewise.
11930 (726te_alu_shift_op): Likewise.
11931 (726te_alu_shift_reg_op): Likewise.
11932 * config/arm/fmp626.md (mp626_alu_op): Update for attribute changes.
11933 (mp626_alu_shift_op): Likewise.
11934 * config/arm/marvell-pj4.md (pj4_alu_e1): Update for attribute changes.
11935 (pj4_alu_e1_conds): Likewise.
11936 (pj4_alu): Likewise.
11937 (pj4_alu_conds): Likewise.
11938 (pj4_shift): Likewise.
11939 (pj4_shift_conds): Likewise.
11940 (pj4_alu_shift): Likewise.
11941 (pj4_alu_shift_conds): Likewise.
11942 * config/arm/arm.c (xscale_sched_adjust_cost): Update for attribute
11943 changes.
11944 (cortexa7_older_only): Likewise.
11945 (cortexa7_younger): Likewise.
11946
11947 2013-07-18 David Malcolm <dmalcolm@redhat.com>
11948
11949 * ipa-pure-const.c (generate_summary): Rename to...
11950 (pure_const_generate_summary): ... this.
11951
11952 2013-07-17 Iain Sandoe <iain@codesourcery.com>
11953
11954 * config/rs6000/darwin.h (REGISTER_NAMES): Add HTM registers.
11955
11956 2013-07-17 Yvan Roux <yvan.roux@linaro.org>
11957
11958 PR target/57909
11959 * config/arm/arm.c (gen_movmem_ldrd_strd): Fix unaligned load/store
11960 usage in HI mode.
11961
11962 2013-07-17 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
11963
11964 * config/s390/s390.c: (s390_expand_builtin): Allow -mhtm to be
11965 enabled without -march=zEC12.
11966 * config/s390/s390.h (TARGET_HTM): Do not require EC12 machine
11967 flags to be set.
11968
11969 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11970
11971 * config/mips/mips.h (ISA_HAS_FP4): Correct formatting.
11972 (ISA_HAS_FP_MADD4_MSUB4): Also enable for ISA_MIPS32R2.
11973 (ISA_HAS_NMADD4_NMSUB4): Remove the MODE argument; rewrite in
11974 terms of ISA_HAS_FP4, and also enable for ISA_MIPS32R2.
11975 (ISA_HAS_NMADD3_NMSUB3): Remove the MODE argument.
11976 * config/mips/mips.c (mips_rtx_costs) <PLUS>: Check for
11977 ISA_HAS_FP_MADD4_MSUB4 || ISA_HAS_FP_MADD3_MSUB3 rather than
11978 ISA_HAS_FP4.
11979 <MINUS, NEG>: Update according to changes to ISA_HAS_NMADD4_NMSUB4
11980 and ISA_HAS_NMADD3_NMSUB3.
11981 * config/mips/mips.md (nmadd4<mode>, nmadd3<mode>): Likewise.
11982 (nmadd4<mode>_fastmath, nmadd3<mode>_fastmath): Likewise.
11983 (nmsub4<mode>, nmsub3<mode>): Likewise.
11984 (nmsub4<mode>_fastmath, nmsub3<mode>_fastmath): Likewise.
11985
11986 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
11987
11988 * config/mips/mips.h (ISA_HAS_NMADD4_NMSUB4): Remove
11989 TARGET_MIPS5400 checking.
11990
11991 2013-07-16 Jakub Jelinek <jakub@redhat.com>
11992 Peter Bergner <bergner@vnet.ibm.com>
11993
11994 * config/rs6000/rs6000.h (FIRST_PSEUDO_REGISTERS): Mention HTM
11995 registers in the comment.
11996 (DWARF_FRAME_REGISTERS): Subtract also the 3 HTM registers.
11997 (DWARF_REG_TO_UNWIND_COLUMN): Use DWARF_FRAME_REGISTERS
11998 rather than FIRST_PSEUDO_REGISTERS.
11999
12000 2013-07-16 Peter Bergner <bergner@vnet.ibm.com>
12001
12002 * config/rs6000/rs6000.c (rs6000_option_override_internal): Do not
12003 enable extra ISA flags with TARGET_HTM.
12004
12005 2013-07-16 Maciej W. Rozycki <macro@codesourcery.com>
12006
12007 * config/mips/mips.h (ISA_HAS_MULS, ISA_HAS_MSAC, ISA_HAS_MACC):
12008 Fix comment typos.
12009
12010 2013-07-15 Cong Hou <congh@google.com>
12011
12012 * tree-vect-data-refs.c (dr_group_sort_cmp): Do not use hash function
12013 in compare function for sorting.
12014
12015 2013-07-15 Peter Bergner <bergner@vnet.ibm.com>
12016
12017 * config.gcc (powerpc*-*-*): Install htmintrin.h and htmxlintrin.h.
12018 * config/rs6000/t-rs6000 (MD_INCLUDES): Add htm.md.
12019 * config/rs6000/rs6000.opt: Add -mhtm option.
12020 * config/rs6000/rs6000-cpus.def (POWERPC_MASKS): Add OPTION_MASK_HTM.
12021 (ISA_2_7_MASKS_SERVER): Add OPTION_MASK_HTM.
12022 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
12023 __HTM__ if the HTM instructions are available.
12024 * config/rs6000/predicates.md (u3bit_cint_operand, u10bit_cint_operand,
12025 htm_spr_reg_operand): New define_predicates.
12026 * config/rs6000/rs6000.md (define_attr "type"): Add htm.
12027 (TFHAR_REGNO, TFIAR_REGNO, TEXASR_REGNO): New define_constants.
12028 Include htm.md.
12029 * config/rs6000/rs6000-builtin.def (BU_HTM_0, BU_HTM_1, BU_HTM_2,
12030 BU_HTM_3, BU_HTM_SPR0, BU_HTM_SPR1): Add support macros for defining
12031 HTM builtin functions.
12032 * config/rs6000/rs6000.c (RS6000_BUILTIN_H): New macro.
12033 (rs6000_reg_names, alt_reg_names): Add HTM SPR register names.
12034 (rs6000_init_hard_regno_mode_ok): Add support for HTM instructions.
12035 (rs6000_builtin_mask_calculate): Likewise.
12036 (rs6000_option_override_internal): Likewise.
12037 (bdesc_htm): Add new HTM builtin support.
12038 (htm_spr_num): New function.
12039 (htm_spr_regno): Likewise.
12040 (rs6000_htm_spr_icode): Likewise.
12041 (htm_expand_builtin): Likewise.
12042 (htm_init_builtins): Likewise.
12043 (rs6000_expand_builtin): Add support for HTM builtin functions.
12044 (rs6000_init_builtins): Likewise.
12045 (rs6000_invalid_builtin, rs6000_opt_mask): Add support for -mhtm
12046 option.
12047 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Add support for -mhtm.
12048 (TARGET_HTM, MASK_HTM): Define macros.
12049 (FIRST_PSEUDO_REGISTER): Adjust for new HTM SPR registers.
12050 (FIXED_REGISTERS): Likewise.
12051 (CALL_USED_REGISTERS): Likewise.
12052 (CALL_REALLY_USED_REGISTERS): Likewise.
12053 (REG_ALLOC_ORDER): Likewise.
12054 (enum reg_class): Likewise.
12055 (REG_CLASS_NAMES): Likewise.
12056 (REG_CLASS_CONTENTS): Likewise.
12057 (REGISTER_NAMES): Likewise.
12058 (ADDITIONAL_REGISTER_NAMES): Likewise.
12059 (RS6000_BTC_SPR, RS6000_BTC_VOID, RS6000_BTC_32BIT, RS6000_BTC_64BIT,
12060 RS6000_BTC_MISC_MASK, RS6000_BTM_HTM): New macros.
12061 (RS6000_BTM_COMMON): Add RS6000_BTM_HTM.
12062 * config/rs6000/htm.md: New file.
12063 * config/rs6000/htmintrin.h: New file.
12064 * config/rs6000/htmxlintrin.h: New file.
12065
12066 2013-07-15 Marcus Shawcroft <marcus.shawcroft@arm.com>
12067
12068 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12069 Define SYMBOL_TINY_GOT, update comment.
12070 * config/aarch64/aarch64.c
12071 (aarch64_load_symref_appropriately): Handle SYMBOL_TINY_GOT.
12072 (aarch64_expand_mov_immediate): Likewise.
12073 (aarch64_print_operand): Likewise.
12074 (aarch64_classify_symbol): Likewise.
12075 * config/aarch64/aarch64.md (UNSPEC_GOTTINYPIC): Define.
12076 (ldr_got_tiny): Define.
12077
12078 2013-07-13 Tobias Grosser <tobias@grosser.es>
12079
12080 PR tree-optimization/54094
12081 * graphite-clast-to-gimple.c (translate_clast_for_loop): Derive the
12082 scheduling dimension for the parallelism check from the polyhedral
12083 information in the AST.
12084 * graphite-dependences.c (carries_deps): Do not assume the schedule is
12085 in 2D + 1 form.
12086
12087 2013-07-13 Jason Merrill <jason@redhat.com>
12088
12089 * print-tree.c (debug_vec_tree): Use debug_raw.
12090 (debug_raw (vec<tree, va_gc> &)): New.
12091 (debug_raw (vec<tree, va_gc> *)): New.
12092 * tree.h: Declare them.
12093
12094 2013-07-13 Bin Cheng <bin.cheng@arm.com>
12095
12096 * ifcvt.c (ifcvt_after_combine): New static variable.
12097 (cheap_bb_rtx_cost_p): Set scale to REG_BR_PROB_BASE when optimizing
12098 for size.
12099 (if_convert): New parameter after_combine. Set ifcvt_after_combine.
12100 (rest_of_handle_if_conversion, rest_of_handle_if_after_combine,
12101 rest_of_handle_if_after_reload): Pass new argument for if_convert.
12102
12103 2013-07-12 Maciej W. Rozycki <macro@codesourcery.com>
12104
12105 * config/mips/mips.c (mips_expand_call): Remove empty statement.
12106
12107 2013-07-12 Michael Matz <matz@suse.de>
12108
12109 PR middle-end/55771
12110 * convert.c (convert_to_real): Reject non-float inner types.
12111
12112 2013-07-12 Tejas Belagod <tejas.belagod@arm.com>
12113
12114 * config/aarch64/aarch64-protos.h
12115 (aarch64_simd_immediate_valid_for_move): Remove.
12116 * config/aarch64/aarch64.c (simd_immediate_info): New member.
12117 (aarch64_simd_valid_immediate): Recognize idioms for shifting ones
12118 cases.
12119 (aarch64_output_simd_mov_immediate): Print the correct shift specifier.
12120
12121 2013-07-11 Steve Ellcey <sellcey@mips.com>
12122
12123 * config/mips/mips.c (mips_conditional_register_usage): Do not
12124 use t[0-7] registers in MIPS16 mode when optimizing for size.
12125
12126 2013-07-11 Sriraman Tallam <tmsriram@google.com>
12127
12128 * config/i386/i386.c (dispatch_function_versions): Fix array
12129 indexing of function_version_info to match actual_versions.
12130
12131 2013-07-11 Teresa Johnson <tejohnson@google.com>
12132
12133 * vec.h (struct va_gc): Move release out-of-line.
12134 (va_gc::release): Call ggc_free on released vec.
12135
12136 2013-07-11 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
12137
12138 * config/rs6000/rs6000.md (""*tls_gd_low<TLSmode:tls_abi_suffix>"):
12139 Require GOT register as additional operand in UNSPEC.
12140 ("*tls_ld_low<TLSmode:tls_abi_suffix>"): Likewise.
12141 ("*tls_got_dtprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12142 ("*tls_got_tprel_low<TLSmode:tls_abi_suffix>"): Likewise.
12143 ("*tls_gd<TLSmode:tls_abi_suffix>"): Update splitter.
12144 ("*tls_ld<TLSmode:tls_abi_suffix>"): Likewise.
12145 ("tls_got_dtprel_<TLSmode:tls_abi_suffix>"): Likewise.
12146 ("tls_got_tprel_<TLSmode:tls_abi_suffix>"): Likewise.
12147
12148 2013-07-11 Georg-Johann Lay <avr@gjlay.de>
12149
12150 PR target/57631
12151 * config/avr/avr.c (avr_set_current_function): Sanity-check signal
12152 name seen by assembler/linker rather if available.
12153
12154 2013-07-11 Andreas Schwab <schwab@suse.de>
12155
12156 * config/aarch64/aarch64-linux.h (CPP_SPEC): Define.
12157
12158 2013-07-10 Vladimir Makarov <vmakarov@redhat.com>
12159
12160 * lra-constraints.c (curr_insn_transform): Switch off optional reloads.
12161
12162 2013-07-10 Joseph Myers <joseph@codesourcery.com>
12163
12164 * doc/tm.texi.in: Move hook documentation to ....
12165 * target.def: ... here.
12166
12167 * doc/tm.texi.in (TARGET_CANONICALIZE_COMPARISON): Remove stray
12168 text on @hook line.
12169 * doc/tm.texi: Regenerate.
12170
12171 2013-07-10 Paolo Carlini <paolo.carlini@oracle.com>
12172
12173 PR c++/57869
12174 * doc/invoke.texi: Document -Wconditionally-supported.
12175
12176 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12177
12178 PR target/57844
12179 * config/avr/avr.c (avr_prologue_setup_frame): Trunk -size to mode
12180 of my_fp.
12181
12182 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12183
12184 PR target/57506
12185 * config/avr/avr-mcus.def (atmega16hva, atmega16hva2, atmega16hvb)
12186 (atmega16m1, atmega16u4, atmega32a, atmega32c1, atmega32hvb)
12187 (atmega32m1, atmega32u4, atmega32u6, atmega64c1, atmega64m1):
12188 Remove duplicate devices.
12189 * config/avr/gen-avr-mmcu-texi.c (print_mcus): Fail on duplicate MCUs.
12190 * config/avr/t-multilib: Regenerate.
12191 * config/avr/avr-tables.opt: Regenerate.
12192 * doc/avr-mmcu.texi: Regenerate.
12193
12194 2013-07-10 Georg-Johann Lay <avr@gjlay.de>
12195
12196 PR target/56987
12197 * config/avr/avr.opt (Waddr-space-convert): Fix typo.
12198
12199 2013-07-10 Graham Stott <graham.stott@btinternet.com>
12200
12201 * config/mips/mips.c (mips_rtx_costs): Very slightly increase
12202 the cost of MULT when optimizing for size.
12203
12204 2013-07-10 Jan-Benedict Glaw <jbglaw@lug-owl.de>
12205
12206 * config/cr16/cr16-protos.h: Don't include target.h.
12207
12208 2013-07-09 Joseph Myers <joseph@codesourcery.com>
12209
12210 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Only
12211 adjust register size for TDmode and TFmode for VSX registers.
12212
12213 2013-07-08 Kai Tietz <ktietz@redhat.com>
12214
12215 PR target/56892
12216 * config/i386/i386.c (TARGET_FUNCTION_ATTRIBUTE_INLINABLE_P): Define as
12217 hook_bool_const_tree_true.
12218
12219 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12220
12221 * config/s390/s390.c: Replace F*_REGNUM with FPR*_REGNUM.
12222 * config/s390/s390.h: Remove F*_REGNUM macro definitions.
12223 * config/s390/s390.md: Define FPR*_REGNUM constants.
12224 Fix FPR2_REGNUM constant (18 -> 17).
12225 ("*trunc<BFP:mode><DFP_ALL:mode>2")
12226 ("*trunc<DFP_ALL:mode><BFP:mode>2")
12227 ("trunc<BFP:mode><DFP_ALL:mode>2")
12228 ("trunc<DFP_ALL:mode><BFP:mode>2")
12229 ("*extend<BFP:mode><DFP_ALL:mode>2")
12230 ("*extend<DFP_ALL:mode><BFP:mode>2")
12231 ("extend<BFP:mode><DFP_ALL:mode>2")
12232 ("extend<DFP_ALL:mode><BFP:mode>2"): Replace FPR2_REGNUM with
12233 FPR4_REGNUM.
12234
12235 2013-07-08 Graham Stott <graham.stott@btinternet.com>
12236
12237 * Makefile.in: (c-family-warn): Define to $(STRICT_WARN)
12238
12239 2013-07-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12240
12241 * config/s390/s390.c: Rename cfun_set_fpr_bit to cfun_set_fpr_save
12242 and cfun_fpr_bit_p to cfun_fpr_save_p.
12243 (s390_frame_area, s390_register_info, s390_frame_info)
12244 (s390_emit_prologue, s390_emit_epilogue)
12245 (s390_conditional_register_usage): Use the *_REGNUM macros for FPR
12246 register numbers.
12247 * config/s390/s390.h: Define *_REGNUM macros for floating point
12248 register numbers.
12249
12250 2013-07-08 Eric Botcazou <ebotcazou@adacore.com>
12251
12252 * Makefile.in (tree-ssa-reassoc.o): Add dependency on $(PARAMS_H).
12253
12254 2013-07-08 Po-Chun Chang <pchang9@cs.wisc.edu>
12255
12256 PR rtl-optimization/57786
12257 * combine.c (distribute_notes) <case REG_DEAD>: Change all_used to bool
12258 and break out of the loop when it is set to false.
12259
12260 2013-07-08 Jakub Jelinek <jakub@redhat.com>
12261
12262 PR target/57819
12263 * simplify-rtx.c (simplify_unary_operation_1) <case ZERO_EXTEND>:
12264 Simplify (zero_extend:SI (subreg:QI (and:SI (reg:SI)
12265 (const_int 63)) 0)).
12266 * combine.c (make_extraction): Create ZERO_EXTEND or SIGN_EXTEND
12267 using simplify_gen_unary instead of gen_rtx_*_EXTEND.
12268 * config/i386/i386.md (*jcc_bt<mode>_1): New define_insn_and_split.
12269
12270 PR rtl-optimization/57829
12271 * simplify-rtx.c (simplify_binary_operation_1) <case IOR>: Ensure that
12272 mask bits outside of mode are just sign-extension from mode to HWI.
12273
12274 2013-07-08 Michael Zolotukhin <michael.v.zolotukhin@gmail.com>
12275
12276 * config/i386/i386-opts.h (enum stringop_alg): Add vector_loop.
12277 * config/i386/i386.c (expand_set_or_movmem_via_loop): Use
12278 adjust_address instead of change_address to keep info about alignment.
12279 (emit_strmov): Remove.
12280 (emit_memmov): New function.
12281 (expand_movmem_epilogue): Refactor to properly handle bigger sizes.
12282 (expand_movmem_epilogue): Likewise and return updated rtx for
12283 destination.
12284 (expand_constant_movmem_prologue): Likewise and return updated rtx for
12285 destination and source.
12286 (decide_alignment): Refactor, handle vector_loop.
12287 (ix86_expand_movmem): Likewise.
12288 (ix86_expand_setmem): Likewise.
12289 * config/i386/i386.opt (Enum): Add vector_loop to option stringop_alg.
12290
12291 2013-07-07 Uros Bizjak <ubizjak@gmail.com>
12292
12293 * config/i386/driver-i386.c (host_detect_local_cpu): Do not check
12294 signature_TM2_ebx, it interferes with signature_INTEL_ebx.
12295
12296 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12297
12298 * config/i386/sse.md (sse_movlhps): Change alternative 3
12299 of operand 2 to "m".
12300
12301 2013-07-06 Uros Bizjak <ubizjak@gmail.com>
12302
12303 PR target/57807
12304 * config/i386/sse.md (iptr): New mode attribute.
12305 (sse2_movq128): Add pointer size overrides for Intel asm dialect.
12306 (<sse>_vm<plusminus_insn><mode>3): Ditto.
12307 (<sse>_vmmul<mode>3): Ditto.
12308 (<sse>_vmdiv<mode>3): Ditto.
12309 (sse_vmrcpv4sf2): Ditto.
12310 (<sse>_vmsqrt<mode>2): Ditto.
12311 (sse_vmrsqrtv4sf2): Ditto.
12312 (<sse>_vm<code><mode>3): Ditto.
12313 (avx_vmcmp<mode>3): Ditto.
12314 (<sse>_vmmaskcmp<mode>3): Ditto.
12315 (<sse>_comi): Ditto.
12316 (<sse>_ucomi): Ditto.
12317 (*xop_vmfrcz_<mode>): Ditto.
12318 (*fmai_fmadd_<mode>): Ditto.
12319 (*fmai_fmsub_<mode>): Ditto.
12320 (*fmai_fnmadd_<mode>): Ditto.
12321 (*fmai_fnmsub_<mode>): Ditto.
12322 (*fma4i_vmfmadd_<mode>): Ditto.
12323 (*fma4i_vmfmsub_<mode>): Ditto.
12324 (*fma4i_vmfnmadd_<mode>): Ditto.
12325 (*fma4i_vmfnmsub_<mode>): Ditto.
12326 (*xop_vmfrcz_<mode>): Ditto.
12327 (sse_cvtps2pi): Ditto.
12328 (sse_cvttps2pi): Ditto.
12329 (sse_cvtss2si): Ditto.
12330 (sse_cvtss2si_2): Ditto.
12331 (sse_cvtss2siq_2): Ditto.
12332 (sse_cvttss2si): Ditto.
12333 (sse_cvttss2siq): Ditto.
12334 (sse_cvtsd2si): Ditto.
12335 (sse_cvtsd2si_2): Ditto.
12336 (sse_cvtsd2siq_2): Ditto.
12337 (sse_cvttsd2si): Ditto.
12338 (sse_cvttsd2siq): Ditto.
12339 (sse_cvtsd2ss): Ditto.
12340 (sse_cvtss2sd): Ditto.
12341 (avx2_pbroadcast<mode>): Ditto.
12342 (avx2_pbroadcast<mode>_1): Ditto.
12343 (*avx_vperm_broadcast_v4sf): Ditto.
12344
12345 (sse_movhlps): Ditto for movlp[sd]/movhp[sd] alternatives.
12346 (sse_movlhps): Ditto.
12347 (sse_storehps): Ditto.
12348 (sse_loadhps): Ditto.
12349 (sse_storelps): Ditto.
12350 (sse_loadlps): Ditto.
12351 (*vec_concatv4sf): Ditto.
12352 (*vec_interleave_highv2df): Ditto.
12353 (*vec_interleave_lowv2df): Ditto.
12354 (*vec_extractv2df_1_sse): Ditto.
12355 (*vec_extractv2df_0_sse): Ditto.
12356 (sse2_storelpd): Ditto.
12357 (sse2_loadlpd): Ditto.
12358 (sse2_movsd): Ditto.
12359 (*vec_concatv4si): Ditto.
12360 (vec_concatv2di): Ditto.
12361
12362 * config/i386/mmx.md (mmx_punpcklbw): Add pointer size overrides
12363 for Intel asm dialect.
12364 (mmx_punpcklwd): Ditto.
12365 (mmx_punpckldq): Ditto.
12366
12367 * config/i386/i386.c (ix86_print_operand) ['H']: Output 'qword ptr'
12368 for intel assembler dialect.
12369
12370 2013-07-06 Jakub Jelinek <jakub@redhat.com>
12371
12372 PR target/29776
12373 * fold-const.c (tree_call_nonnegative_warnv_p): Return true
12374 for BUILT_IN_C{LZ,LRSB}*.
12375 * tree.h (CASE_INT_FN): Add FN##IMAX case.
12376 * tree-vrp.c (extract_range_basic): Handle
12377 BUILT_IN_{FFS,PARITY,POPCOUNT,C{LZ,TZ,LRSB}}*. For
12378 BUILT_IN_CONSTANT_P if argument isn't (D) of PARM_DECL,
12379 fall thru to code calling set_value*.
12380 * builtins.c (expand_builtin): Remove *IMAX cases.
12381 (fold_builtin_bitop): For BUILT_IN_CLRSB* return NULL_TREE
12382 if width is bigger than 2*HWI.
12383
12384 2013-07-05 Vladimir Makarov <vmakarov@redhat.com>
12385
12386 PR rtl-optimization/55342
12387 * lra-int.h (lra_subreg_reload_pseudos): New.
12388 * lra.c: Add undoing optional reloads to the block diagram.
12389 (lra_subreg_reload_pseudos): New.
12390 (lra_optional_reload_pseudos): Change comments.
12391 (lra): Init and clear lra_subreg_reload_pseudos. Clear
12392 lra_optional_reload_pseudos after undo transformations.
12393 * lra-assigns.c (pseudo_prefix_title): New.
12394 (lra_setup_reg_renumber): Use it.
12395 (spill_for): Ditto. Check subreg reload pseudos too.
12396 (assign_by_spills): Consider subreg reload pseudos too.
12397 * lra-constraints.c (simplify_operand_subreg): Use
12398 lra_subreg_reload_pseudos instead of lra_optional_reload_pseudos.
12399 (curr_insn_transform): Recognize and do optional reloads.
12400 (undo_optional_reloads): New.
12401 (lra_undo_inheritance): Call undo_optional_reloads.
12402
12403 2013-07-05 Thomas Quinot <quinot@adacore.com>
12404
12405 * tree-complex.c (expand_complex_operations_1): Fix typo.
12406
12407 2013-07-04 Tejas Belagod <tejas.belagod@arm.com>
12408
12409 * config/aarch64/aarch64-protos.h (cpu_vector_cost): New.
12410 (tune_params): New member 'const vec_costs'.
12411 * config/aarch64/aarch64.c (generic_vector_cost): New.
12412 (generic_tunings): New member 'generic_vector_cost'.
12413 (aarch64_builtin_vectorization_cost): New.
12414 (aarch64_add_stmt_cost): New.
12415 (TARGET_VECTORIZE_ADD_STMT_COST): New.
12416 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST): New.
12417
12418 2013-07-03 Jakub Jelinek <jakub@redhat.com>
12419
12420 PR target/57777
12421 * config/i386/predicates.md (vsib_address_operand): Disallow
12422 SYMBOL_REF or LABEL_REF in parts.disp if TARGET_64BIT && flag_pic.
12423
12424 2013-07-03 Hans-Peter Nilsson <hp@bitrange.com>
12425
12426 PR middle-end/55030
12427 * stmt.c (expand_nl_goto_receiver): Remove almost-copy of
12428 expand_builtin_setjmp_receiver.
12429 (expand_label): Adjust, call expand_builtin_setjmp_receiver
12430 with NULL for the label parameter.
12431 * builtins.c (expand_builtin_setjmp_receiver): Don't clobber
12432 the frame-pointer. Adjust comments.
12433 [HAVE_builtin_setjmp_receiver]: Emit builtin_setjmp_receiver
12434 only if LABEL is non-NULL.
12435
12436 2013-07-03 Yufeng Zhang <yufeng.zhang@arm.com>
12437
12438 * config/aarch64/aarch64.h (enum arm_abi_type): Remove.
12439 (ARM_ABI_AAPCS64): Ditto.
12440 (arm_abi): Ditto.
12441 (ARM_DEFAULT_ABI): Ditto.
12442
12443 2013-07-03 James Greenhalgh <james.greenhalgh@arm.com>
12444
12445 * config/aarch64/aarch64-builtins.c
12446 (aarch64_simd_expand_builtin): Handle AARCH64_SIMD_STORE1.
12447 * config/aarch64/aarch64-simd-builtins.def (ld1): New.
12448 (st1): Likewise.
12449 * config/aarch64/aarch64-simd.md
12450 (aarch64_ld1<VALL:mode>): New.
12451 (aarch64_st1<VALL:mode>): Likewise.
12452 * config/aarch64/arm_neon.h
12453 (vld1<q>_<fpsu><8, 16, 32, 64>): Convert to RTL builtins.
12454
12455 2013-07-02 Sriraman Tallam <tmsriram@google.com>
12456
12457 * config/i386/i386.c (gate_insert_vzeroupper): Check if
12458 target ISA is AVX.
12459 (ix86_option_override_internal):Turn on all -mavx target flags by
12460 default as they are dependent on AVX anyway.
12461
12462 2013-07-02 Cary Coutant <ccoutant@google.com>
12463
12464 * dwarf2out.c (loc_checksum): Call hash_loc_operands for a
12465 deterministic hash.
12466 (loc_checksum_ordered): Likewise.
12467 (hash_loc_operands): Remove inline keyword.
12468
12469 2013-07-02 Jakub Jelinek <jakub@redhat.com>
12470
12471 PR tree-optimization/57741
12472 * tree-vect-loop.c (vect_is_simple_iv_evolution): Disallow
12473 non-INTEGRAL_TYPE_P non-SCALAR_FLOAT_TYPE_P SSA_NAME step_exprs,
12474 or SCALAR_FLOAT_TYPE_P SSA_NAMEs if !flag_associative_math.
12475 Allow REAL_CST step_exprs if flag_associative_math.
12476 (get_initial_def_for_induction): Handle SCALAR_FLOAT_TYPE_P step_expr.
12477
12478 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12479
12480 * config/aarch64/aarch64-simd.md (absdi2): Support abs for DI mode.
12481
12482 2013-07-02 Ian Bolton <ian.bolton@arm.com>
12483
12484 * config/aarch64/aarch64.md (*extr_insv_reg<mode>): New pattern.
12485
12486 2013-07-02 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12487
12488 * config/arm/arm.md (arm_andsi3_insn): Add alternatives for 16-bit
12489 encoding.
12490 (iorsi3_insn): Likewise.
12491 (arm_xorsi3): Likewise.
12492
12493 2013-07-01 Sofiane Naci <sofiane.naci@arm.com>
12494
12495 * arm.md (attribute "wtype"): Delete. Move attribute values from here
12496 to ...
12497 (attribute "type"): ... here, and prefix with "wmmx_".
12498 (attribute "core_cycles"): Update for attribute changes.
12499 * iwmmxt.md (tbcstv8qi): Update for attribute changes.
12500 (tbcstv4hi): Likewise.
12501 (tbcstv2si): Likewise.
12502 (iwmmxt_iordi3): Likewise.
12503 (iwmmxt_xordi3): Likewise.
12504 (iwmmxt_anddi3): Likewise.
12505 (iwmmxt_nanddi3): Likewise.
12506 (iwmmxt_arm_movdi): Likewise.
12507 (iwmmxt_movsi_insn): Likewise.
12508 (mov<mode>_internal): Likewise.
12509 (and<mode>3_iwmmxt): Likewise.
12510 (ior<mode>3_iwmmxt): Likewise.
12511 (xor<mode>3_iwmmxt): Likewise.
12512 (add<mode>3_iwmmxt): Likewise.
12513 (ssaddv8qi3): Likewise.
12514 (ssaddv4hi3): Likewise.
12515 (ssaddv2si3): Likewise.
12516 (usaddv8qi3): Likewise.
12517 (usaddv4hi3): Likewise.
12518 (usaddv2si3): Likewise.
12519 (sub<mode>3_iwmmxt): Likewise.
12520 (sssubv8qi3): Likewise.
12521 (sssubv4hi3): Likewise.
12522 (sssubv2si3): Likewise.
12523 (ussubv8qi3): Likewise.
12524 (ussubv4hi3): Likewise.
12525 (ussubv2si3): Likewise.
12526 (mulv4hi3_iwmmxt): Likewise.
12527 (smulv4hi3_highpart): Likewise.
12528 (umulv4hi3_highpart): Likewise.
12529 (iwmmxt_wmacs): Likewise.
12530 (iwmmxt_wmacsz): Likewise.
12531 (iwmmxt_wmacu): Likewise.
12532 (iwmmxt_wmacuz): Likewise.
12533 (iwmmxt_clrdi): Likewise.
12534 (iwmmxt_clrv8qi): Likewise.
12535 (iwmmxt_clr4hi): Likewise.
12536 (iwmmxt_clr2si): Likewise.
12537 (iwmmxt_uavgrndv8qi3): Likewise.
12538 (iwmmxt_uavgrndv4hi3): Likewise.
12539 (iwmmxt_uavgv8qi3): Likewise.
12540 (iwmmxt_uavgv4hi3): Likewise.
12541 (iwmmxt_tinsrb): Likewise.
12542 (iwmmxt_tinsrh): Likewise.
12543 (iwmmxt_tinsrw): Likewise.
12544 (iwmmxt_textrmub): Likewise.
12545 (iwmmxt_textrmsb): Likewise.
12546 (iwmmxt_textrmuh): Likewise.
12547 (iwmmxt_textrmsh): Likewise.
12548 (iwmmxt_textrmw): Likewise.
12549 (iwmxxt_wshufh): Likewise.
12550 (eqv8qi3): Likewise.
12551 (eqv4hi3): Likewise.
12552 (eqv2si3): Likewise.
12553 (gtuv8qi3): Likewise.
12554 (gtuv4hi3): Likewise.
12555 (gtuv2si3): Likewise.
12556 (gtv8qi3): Likewise.
12557 (gtv4hi3): Likewise.
12558 (gtv2si3): Likewise.
12559 (smax<mode>3_iwmmxt): Likewise.
12560 (umax<mode>3_iwmmxt): Likewise.
12561 (smin<mode>3_iwmmxt): Likewise.
12562 (umin<mode>3_iwmmxt): Likewise.
12563 (iwmmxt_wpackhss): Likewise.
12564 (iwmmxt_wpackwss): Likewise.
12565 (iwmmxt_wpackdss): Likewise.
12566 (iwmmxt_wpackhus): Likewise.
12567 (iwmmxt_wpackwus): Likewise.
12568 (iwmmxt_wpackdus): Likewise.
12569 (iwmmxt_wunpckihb): Likewise.
12570 (iwmmxt_wunpckihh): Likewise.
12571 (iwmmxt_wunpckihw): Likewise.
12572 (iwmmxt_wunpckilb): Likewise.
12573 (iwmmxt_wunpckilh): Likewise.
12574 (iwmmxt_wunpckilw): Likewise.
12575 (iwmmxt_wunpckehub): Likewise.
12576 (iwmmxt_wunpckehuh): Likewise.
12577 (iwmmxt_wunpckehuw): Likewise.
12578 (iwmmxt_wunpckehsb): Likewise.
12579 (iwmmxt_wunpckehsh): Likewise.
12580 (iwmmxt_wunpckehsw): Likewise.
12581 (iwmmxt_wunpckelub): Likewise.
12582 (iwmmxt_wunpckeluh): Likewise.
12583 (iwmmxt_wunpckeluw): Likewise.
12584 (iwmmxt_wunpckelsb): Likewise.
12585 (iwmmxt_wunpckelsh): Likewise.
12586 (iwmmxt_wunpckelsw): Likewise.
12587 (ror<mode>3): Likewise.
12588 (ashr<mode>3_iwmmxt): Likewise.
12589 (lshr<mode>3_iwmmxt): Likewise.
12590 (ashl<mode>3_iwmmxt): Likewise.
12591 (ror<mode>3_di): Likewise.
12592 (ashr<mode>3_di): Likewise.
12593 (lshr<mode>3_di): Likewise.
12594 (ashl<mode>3_di): Likewise.
12595 (iwmmxt_wmadds): Likewise.
12596 (iwmmxt_wmaddu): Likewise.
12597 (iwmmxt_tmia): Likewise.
12598 (iwmmxt_tmiaph): Likewise.
12599 (iwmmxt_tmiabb): Likewise.
12600 (iwmmxt_tmiatb): Likewise.
12601 (iwmmxt_tmiabt): Likewise.
12602 (iwmmxt_tmiatt): Likewise.
12603 (iwmmxt_tmovmskb): Likewise.
12604 (iwmmxt_tmovmskh): Likewise.
12605 (iwmmxt_tmovmskw): Likewise.
12606 (iwmmxt_waccb): Likewise.
12607 (iwmmxt_wacch): Likewise.
12608 (iwmmxt_waccw): Likewise.
12609 (iwmmxt_waligni): Likewise.
12610 (iwmmxt_walignr): Likewise.
12611 (iwmmxt_walignr0): Likewise.
12612 (iwmmxt_walignr1): Likewise.
12613 (iwmmxt_walignr2): Likewise.
12614 (iwmmxt_walignr3): Likewise.
12615 (iwmmxt_wsadb): Likewise.
12616 (iwmmxt_wsadh): Likewise.
12617 (iwmmxt_wsadbz): Likewise.
12618 (iwmmxt_wsadhz): Likewise.
12619 * iwmmxt2.md (iwmmxt_wabs<mode>3): Update for attribute changes.
12620 (iwmmxt_wabsdiffb): Likewise.
12621 (iwmmxt_wabsdiffh): Likewise.
12622 (iwmmxt_wabsdiffw): Likewise.
12623 (iwmmxt_waddsubhx): Likewise
12624 (iwmmxt_wsubaddhx): Likewise.
12625 (addc<mode>3): Likewise.
12626 (iwmmxt_avg4): Likewise.
12627 (iwmmxt_avg4r): Likewise.
12628 (iwmmxt_wmaddsx): Likewise.
12629 (iwmmxt_wmaddux): Likewise.
12630 (iwmmxt_wmaddsn): Likewise.
12631 (iwmmxt_wmaddun): Likewise.
12632 (iwmmxt_wmulwsm): Likewise.
12633 (iwmmxt_wmulwum): Likewise.
12634 (iwmmxt_wmulsmr): Likewise.
12635 (iwmmxt_wmulumr): Likewise.
12636 (iwmmxt_wmulwsmr): Likewise.
12637 (iwmmxt_wmulwumr): Likewise.
12638 (iwmmxt_wmulwl): Likewise.
12639 (iwmmxt_wqmulm): Likewise.
12640 (iwmmxt_wqmulwm): Likewise.
12641 (iwmmxt_wqmulmr): Likewise.
12642 (iwmmxt_wqmulwmr): Likewise.
12643 (iwmmxt_waddbhusm): Likewise.
12644 (iwmmxt_waddbhusl): Likewise.
12645 (iwmmxt_wqmiabb): Likewise.
12646 (iwmmxt_wqmiabt): Likewise.
12647 (iwmmxt_wqmiatb): Likewise.
12648 (iwmmxt_wqmiatt): Likewise.
12649 (iwmmxt_wqmiabbn): Likewise.
12650 (iwmmxt_wqmiabtn): Likewise.
12651 (iwmmxt_wqmiatbn): Likewise.
12652 (iwmmxt_wqmiattn): Likewise.
12653 (iwmmxt_wmiabb): Likewise.
12654 (iwmmxt_wmiabt): Likewise.
12655 (iwmmxt_wmiatb): Likewise.
12656 (iwmmxt_wmiatt): Likewise.
12657 (iwmmxt_wmiabbn): Likewise.
12658 (iwmmxt_wmiabtn): Likewise.
12659 (iwmmxt_wmiatbn): Likewise.
12660 (iwmmxt_wmiattn): Likewise.
12661 (iwmmxt_wmiawbb): Likewise.
12662 (iwmmxt_wmiawbt): Likewise.
12663 (iwmmxt_wmiawtb): Likewise.
12664 (iwmmxt_wmiawtt): Likewise.
12665 (iwmmxt_wmiawbbn): Likewise.
12666 (iwmmxt_wmiawbtn): Likewise.
12667 (iwmmxt_wmiawtbn): Likewise.
12668 (iwmmxt_wmiawttn): Likewise.
12669 (iwmmxt_wmerge): Likewise.
12670 (iwmmxt_tandc<mode>3): Likewise.
12671 (iwmmxt_torc<mode>3): Likewise.
12672 (iwmmxt_torvsc<mode>3): Likewise.
12673 (iwmmxt_textrc<mode>3): Likewise.
12674 * marvell-f-iwmmxt.md (wmmxt_shift): Update for attribute changes.
12675 (wmmxt_pack): Likewise.
12676 (wmmxt_mult_c1): Likewise.
12677 (wmmxt_mult_c2): Likewise.
12678 (wmmxt_alu_c1): Likewise.
12679 (wmmxt_alu_c2): Likewise.
12680 (wmmxt_alu_c3): Likewise.
12681 (wmmxt_transfer_c1): Likewise.
12682 (wmmxt_transfer_c2): Likewise.
12683 (wmmxt_transfer_c3): Likewise.
12684 (marvell_f_iwmmxt_wstr): Likewise.
12685 (marvell_f_iwmmxt_wldr): Likewise.
12686
12687 2013-06-29 Yufeng Zhang <yufeng.zhang@arm.com>
12688
12689 * config/aarch64/aarch64.c: Remove junk from the beginning of the file.
12690
12691 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12692
12693 Revert:
12694 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12695 * lra-constraints.c (need_for_split_p): Check call used hard regs
12696 living through calls.
12697
12698 * lra-constraints.c (inherit_in_ebb): Reset live_hard_regs for
12699 call used regs for call insn.
12700
12701 2013-06-28 Jakub Jelinek <jakub@redhat.com>
12702
12703 PR target/57736
12704 * config/i386/i386.c (ix86_expand_builtin): If target == NULL and
12705 mode is VOIDmode, don't create a VOIDmode pseudo to copy result into.
12706
12707 2013-06-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
12708
12709 * builtins.def: Fixed the function type of CILKPLUS_BUILTIN.
12710
12711 2013-06-28 Vladimir Makarov <vmakarov@redhat.com>
12712
12713 * lra-constraints.c (need_for_split_p): Check call used hard regs
12714 living through calls.
12715
12716 2013-06-28 Michael Meissner <meissner@linux.vnet.ibm.com>
12717
12718 PR target/57744
12719 * config/rs6000/rs6000.h (MODES_TIEABLE_P): Do not allow PTImode
12720 to tie with any other modes. Eliminate Altivec vector mode tests,
12721 since these are a subset of ALTIVEC or VSX vector modes. Simplify
12722 code, to return 0 if testing MODE2 for a condition, if we've
12723 already tested MODE1 for the same condition.
12724
12725 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12726
12727 * config/aarch64/aarch64.c (aarch64_cannot_force_const_mem): Adjust
12728 layout.
12729
12730 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12731
12732 * config/aarch64/aarch64-protos.h (aarch64_symbol_type):
12733 Update comment w.r.t SYMBOL_TINY_ABSOLUTE.
12734
12735 2013-06-28 Marcus Shawcroft <marcus.shawcroft@arm.com>
12736
12737 * config/aarch64/aarch64-protos.h (aarch64_classify_symbol_expression):
12738 Define.
12739 (aarch64_symbolic_constant_p): Remove.
12740 * config/aarch64/aarch64.c (aarch64_classify_symbol_expression): Remove
12741 static. Fix line length and white space.
12742 (aarch64_symbolic_constant_p): Remove.
12743 * config/aarch64/predicates.md (aarch64_valid_symref):
12744 Use aarch64_classify_symbol_expression.
12745
12746 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12747
12748 * config/arm/constraints.md (Ts): New constraint.
12749 * config/arm/arm.md (arm_movqi_insn): Add alternatives for
12750 16-bit encodings.
12751 (compare_scc): Use "Ts" constraint for operand 0.
12752 (ior_scc_scc): Likewise.
12753 (and_scc_scc): Likewise.
12754 (and_scc_scc_nodom): Likewise.
12755 (ior_scc_scc_cmp): Likewise for operand 7.
12756 (and_scc_scc_cmp): Likewise.
12757 * config/arm/thumb2.md (thumb2_movsi_insn):
12758 Add alternatives for 16-bit encodings.
12759 (thumb2_movhi_insn): Likewise.
12760 (thumb2_movsicc_insn): Likewise.
12761 (thumb2_and_scc): Take 'and' outside cond_exec. Use "Ts" constraint.
12762 (thumb2_negscc): Use "Ts" constraint.
12763 Move mvn instruction outside cond_exec block.
12764 * config/arm/vfp.md (thumb2_movsi_vfp): Add alternatives
12765 for 16-bit encodings.
12766
12767 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12768
12769 * config/arm/arm.md (arm_mulsi3_v6): Add alternative for 16-bit
12770 encoding.
12771 (mulsi3addsi_v6): Disable predicable variant for arm_restrict_it.
12772 (mulsi3subsi): Likewise.
12773 (mulsidi3adddi): Likewise.
12774 (mulsidi3_v6): Likewise.
12775 (umulsidi3_v6): Likewise.
12776 (umulsidi3adddi_v6): Likewise.
12777 (smulsi3_highpart_v6): Likewise.
12778 (umulsi3_highpart_v6): Likewise.
12779 (mulhisi3tb): Likewise.
12780 (mulhisi3bt): Likewise.
12781 (mulhisi3tt): Likewise.
12782 (maddhisi4): Likewise.
12783 (maddhisi4tb): Likewise.
12784 (maddhisi4tt): Likewise.
12785 (maddhidi4): Likewise.
12786 (maddhidi4tb): Likewise.
12787 (maddhidi4tt): Likewise.
12788 (zeroextractsi_compare0_scratch): Likewise.
12789 (insv_zero): Likewise.
12790 (insv_t2): Likewise.
12791 (anddi_notzesidi_di): Likewise.
12792 (anddi_notsesidi_di): Likewise.
12793 (andsi_notsi_si): Likewise.
12794 (iordi_zesidi_di): Likewise.
12795 (xordi_zesidi_di): Likewise.
12796 (andsi_iorsi3_notsi): Likewise.
12797 (smax_0): Likewise.
12798 (smax_m1): Likewise.
12799 (smin_0): Likewise.
12800 (not_shiftsi): Likewise.
12801 (unaligned_loadsi): Likewise.
12802 (unaligned_loadhis): Likewise.
12803 (unaligned_loadhiu): Likewise.
12804 (unaligned_storesi): Likewise.
12805 (unaligned_storehi): Likewise.
12806 (extv_reg): Likewise.
12807 (extzv_t2): Likewise.
12808 (divsi3): Likewise.
12809 (udivsi3): Likewise.
12810 (arm_zero_extendhisi2addsi): Likewise.
12811 (arm_zero_extendqisi2addsi): Likewise.
12812 (compareqi_eq0): Likewise.
12813 (arm_extendhisi2_v6): Likewise.
12814 (arm_extendqisi2addsi): Likewise.
12815 (arm_movt): Likewise.
12816 (thumb2_ldrd): Likewise.
12817 (thumb2_ldrd_base): Likewise.
12818 (thumb2_ldrd_base_neg): Likewise.
12819 (thumb2_strd): Likewise.
12820 (thumb2_strd_base): Likewise.
12821 (thumb2_strd_base_neg): Likewise.
12822 (arm_negsi2): Add alternative for 16-bit encoding.
12823 (arm_one_cmplsi2): Likewise.
12824
12825 2013-06-28 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
12826
12827 * config/arm/predicates.md (arm_cond_move_operator): New predicate.
12828 * config/arm/arm.md (movsfcc): Use arm_cond_move_operator predicate.
12829 (movdfcc): Likewise.
12830 * config/arm/vfp.md (*thumb2_movsf_vfp):
12831 Disable predication for arm_restrict_it.
12832 (*thumb2_movsfcc_vfp): Disable for arm_restrict_it.
12833 (*thumb2_movdfcc_vfp): Likewise.
12834 (*abssf2_vfp, *absdf2_vfp, *negsf2_vfp, *negdf2_vfp,*addsf3_vfp,
12835 *adddf3_vfp, *subsf3_vfp, *subdf3_vfpc, *divsf3_vfp,*divdf3_vfp,
12836 *mulsf3_vfp, *muldf3_vfp, *mulsf3negsf_vfp, *muldf3negdf_vfp,
12837 *mulsf3addsf_vfp, *muldf3adddf_vfp, *mulsf3subsf_vfp,
12838 *muldf3subdf_vfp, *mulsf3negsfaddsf_vfp, *fmuldf3negdfadddf_vfp,
12839 *mulsf3negsfsubsf_vfp, *muldf3negdfsubdf_vfp, *fma<SDF:mode>4,
12840 *fmsub<SDF:mode>4, *fnmsub<SDF:mode>4, *fnmadd<SDF:mode>4,
12841 *extendsfdf2_vfp, *truncdfsf2_vfp, *extendhfsf2, *truncsfhf2,
12842 *truncsisf2_vfp, *truncsidf2_vfp, fixuns_truncsfsi2, fixuns_truncdfsi2,
12843 *floatsisf2_vfp, *floatsidf2_vfp, floatunssisf2, floatunssidf2,
12844 *sqrtsf2_vfp, *sqrtdf2_vfp, *cmpsf_vfp, *cmpsf_trap_vfp, *cmpdf_vfp,
12845 *cmpdf_trap_vfp, <vrint_pattern><SDF:mode>2):
12846 Disable predication for arm_restrict_it.
12847
12848 2013-06-28 Kirill Yukhin <kirill.yukhin@intel.com>
12849
12850 * config/i386/bmiintrin.h (_bextr_u32): New.
12851 (_bextr_u64): Ditto.
12852
12853 2013-06-27 Richard Sandiford <rdsandiford@googlemail.com>
12854
12855 * config.gcc (mips*-mti-elf*, mips*-sde-elf*, mips64r5900-*-elf*)
12856 (mips64r5900el-*-elf*): Include mips/n32-elf.h.
12857 * config/mips/sde.h (LOCAL_LABEL_PREFIX, NO_DOLLAR_IN_LABEL)
12858 (LONG_DOUBLE_TYPE_SIZE, LIBGCC2_LONG_DOUBLE_TYPE_SIZE): Move to...
12859 * config/mips/n32-elf.h: ...this new file.
12860
12861 2013-06-27 Marc Glisse <marc.glisse@inria.fr>
12862
12863 PR target/57224
12864 * config/i386/i386.c (enum ix86_builtins, bdesc_args): Remove
12865 IX86_BUILTIN_CMPNGTSS and IX86_BUILTIN_CMPNGESS.
12866
12867 2013-06-27 Catherine Moore <clm@codesourcery.com>
12868
12869 * config/mips/mips-tables.opt: Regenerate.
12870 * config/mips/mips-cpus.def: Add m14ke and m14kec.
12871 * config/mips/mips.h (BASE_DRIVER_SELF_SPECS): m14ke* implies -mdspr2.
12872 * doc/invoke.texi: Add -m14kc.
12873
12874 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12875
12876 PR target/57623
12877 * config/i386/i386.md (bmi_bextr_<mode>): Swap predicates and
12878 constraints of operand 1 and 2.
12879
12880 PR target/57623
12881 * config/i386/i386.md (bmi2_bzhi_<mode>3): Swap AND arguments
12882 to match RTL canonicalization. Swap predicates and
12883 constraints of operand 1 and 2.
12884
12885 2013-06-27 Vladimir Makarov <vmakarov@redhat.com>
12886
12887 * lra-constraints.c (inherit_in_ebb): Process static hard regs too.
12888 Process OP_INOUT regs for splitting too.
12889
12890 2013-06-27 Jakub Jelinek <jakub@redhat.com>
12891
12892 * tree-vect-stmts.c (vectorizable_store): Move ptr_incr var
12893 decl before the loop, initialize to NULL.
12894 (vectorizable_load): Initialize ptr_incr to NULL.
12895
12896 2013-06-27 Martin Jambor <mjambor@suse.cz>
12897
12898 PR lto/57208
12899 * ipa-ref.h (ipa_maybe_record_reference): Declare.
12900 * ipa-ref.c (ipa_maybe_record_reference): New function.
12901 * cgraphclones.c (cgraph_create_virtual_clone): Use it.
12902 * ipa-cp.c (create_specialized_node): Record potential references from
12903 aggvals.
12904 * Makefile.in (ipa-ref.o): Add IPA_REF_H to dependencies.
12905
12906 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12907
12908 * config/aarch64/aarch64.c (aarch64_force_temporary): Add an extra
12909 parameter 'mode' of type 'enum machine_mode mode'; change to pass
12910 'mode' to force_reg.
12911 (aarch64_add_offset): Update calls to aarch64_force_temporary.
12912 (aarch64_expand_mov_immediate): Likewise.
12913
12914 2013-06-27 Yufeng Zhang <yufeng.zhang@arm.com>
12915
12916 * config/aarch64/aarch64.c (aarch64_add_offset): Change to pass
12917 'mode' to aarch64_plus_immediate and gen_rtx_PLUS.
12918
12919 2013-06-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
12920
12921 * config/s390/s390.c: Rename UNSPEC_CCU_TO_INT to
12922 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12923 (struct machine_function): Add tbegin_p.
12924 (s390_canonicalize_comparison): Fold CC mode compares to
12925 conditional jump if possible.
12926 (s390_emit_jump): Return the emitted jump.
12927 (s390_branch_condition_mask, s390_branch_condition_mnemonic):
12928 Handle CCRAWmode compares.
12929 (s390_option_override): Default to -mhtm if available.
12930 (s390_reg_clobbered_rtx): Handle floating point regs as well.
12931 (s390_regs_ever_clobbered): Use s390_regs_ever_clobbered also for
12932 FPRs instead of df_regs_ever_live_p.
12933 (s390_optimize_nonescaping_tx): New function.
12934 (s390_init_frame_layout): Extend clobbered_regs array to cover
12935 FPRs as well.
12936 (s390_emit_prologue): Call s390_optimize_nonescaping_tx.
12937 (s390_expand_tbegin): New function.
12938 (enum s390_builtin): New enum definition.
12939 (code_for_builtin): New array definition.
12940 (s390_init_builtins): New function.
12941 (s390_expand_builtin): New function.
12942 (TARGET_INIT_BUILTINS): Define.
12943 (TARGET_EXPAND_BUILTIN): Define.
12944 * common/config/s390/s390-common.c (processor_flags_table): Add PF_TX.
12945 * config/s390/predicates.md (s390_comparison): Handle CCRAWmode.
12946 (s390_alc_comparison): Likewise.
12947 * config/s390/s390-modes.def: Add CCRAWmode.
12948 * config/s390/s390.h (processor_flags): Add PF_TX.
12949 (TARGET_CPU_HTM): Define macro.
12950 (TARGET_HTM): Define macro.
12951 (TARGET_CPU_CPP_BUILTINS): Define __HTM__ for htm.
12952 * config/s390/s390.md: Rename UNSPEC_CCU_TO_INT to
12953 UNSPEC_STRCMPCC_TO_INT and UNSPEC_CCZ_TO_INT to UNSPEC_CC_TO_INT.
12954 (UNSPECV_TBEGIN, UNSPECV_TBEGINC, UNSPECV_TEND, UNSPECV_TABORT)
12955 (UNSPECV_ETND, UNSPECV_NTSTG, UNSPECV_PPA): New unspecv enum values.
12956 (TBEGIN_MASK, TBEGINC_MASK): New constants.
12957 ("*cc_to_int"): Move up.
12958 ("*mov<mode>cc", "*cjump_64", "*cjump_31"): Accept integer
12959 constants other than 0.
12960 ("*ccraw_to_int"): New insn and splitter definition.
12961 ("tbegin", "tbegin_nofloat", "tbegin_retry")
12962 ("tbegin_retry_nofloat", "tbeginc", "tend", "tabort")
12963 ("tx_assist"): New expander.
12964 ("tbegin_1", "tbegin_nofloat_1", "*tbeginc_1", "*tend_1")
12965 ("*tabort_1", "etnd", "ntstg", "*ppa"): New insn definition.
12966 * config/s390/s390.opt: Add -mhtm option.
12967 * config/s390/s390-protos.h (s390_emit_jump): Add return type.
12968 * config/s390/htmxlintrin.h: New file.
12969 * config/s390/htmintrin.h: New file.
12970 * config/s390/s390intrin.h: New file.
12971 * doc/extend.texi: Document htm builtins.
12972 * config.gcc: Add the new header files to extra_headers.
12973
12974 2013-06-26 Thomas Schwinge <thomas@codesourcery.com>
12975
12976 * config/i386/gnu.h [TARGET_LIBC_PROVIDES_SSP]
12977 (TARGET_CAN_SPLIT_STACK, TARGET_THREAD_SPLIT_STACK_OFFSET): Undefine.
12978
12979 2013-06-26 Michael Meissner <meissner@linux.vnet.ibm.com>
12980 Pat Haugen <pthaugen@us.ibm.com>
12981 Peter Bergner <bergner@vnet.ibm.com>
12982
12983 * config/rs6000/power8.md: New.
12984 * config/rs6000/rs6000-cpus.def (RS6000_CPU table): Adjust processor
12985 setting for power8 entry.
12986 * config/rs6000/t-rs6000 (MD_INCLUDES): Add power8.md.
12987 * config/rs6000/rs6000.c (is_microcoded_insn, is_cracked_insn): Adjust
12988 test for Power4/Power5 only.
12989 (insn_must_be_first_in_group, insn_must_be_last_in_group): Add Power8
12990 support.
12991 (force_new_group): Adjust comment.
12992 * config/rs6000/rs6000.md: Include power8.md.
12993
12994 2013-06-26 Greta Yorsh <Greta.Yorsh@arm.com>
12995
12996 * config/arm/arm.h (MAX_CONDITIONAL_EXECUTE): Define macro.
12997 * config/arm/arm-protos.h (arm_max_conditional_execute): New
12998 declaration.
12999 (tune_params): Update comment.
13000 * config/arm/arm.c (arm_cortex_a15_tune): Set max_cond_insns to 2.
13001 (arm_max_conditional_execute): New function.
13002 (thumb2_final_prescan_insn): Use max_insn_skipped and
13003 MAX_INSN_PER_IT_BLOCK to compute maximum instructions in a block.
13004
13005 2013-06-25 Jakub Jelinek <jakub@redhat.com>
13006
13007 PR tree-optimization/57705
13008 * tree-vect-loop.c (vect_is_simple_iv_evolution): Allow
13009 SSA_NAME step, provided that it is not defined inside the loop.
13010 (vect_analyze_scalar_cycles_1): Disallow SSA_NAME step in nested loop.
13011 (get_initial_def_for_induction): Handle SSA_NAME IV step.
13012
13013 2013-06-25 Martin Jambor <mjambor@suse.cz>
13014
13015 PR middle-end/57670
13016 * cgraph.h (cgraph_indirect_call_info): New flag member_ptr.
13017 * ipa-prop.c (ipa_print_node_jump_functions): Mark member pointer
13018 calls in the dump.
13019 (ipa_note_param_call): Initialize member_ptr flag.
13020 (ipa_analyze_indirect_call_uses): Set member_ptr flag.
13021 (ipa_make_edge_direct_to_target): Bail out if member_ptr is set.
13022 (ipa_write_indirect_edge_info): Stream member_ptr flag.
13023 (ipa_read_indirect_edge_info): Likewise.
13024
13025 2013-06-25 Richard Biener <rguenther@suse.de>
13026
13027 PR middle-end/56977
13028 * passes.c (init_optimization_passes): Move pass_fold_builtins
13029 and pass_dce earlier with -Og.
13030
13031 2013-06-25 Eric Botcazou <ebotcazou@adacore.com>
13032
13033 * expr.c (expand_expr_real_1) <ARRAY_REF>: Fix formatting glitches.
13034 <BIT_FIELD_REF>: Remove trailing TAB.
13035 * varasm.c (output_constructor_bitfield): Fix formatting glitch and
13036 remove blank line.
13037
13038 2013-06-24 Martin Jambor <mjambor@suse.cz>
13039
13040 PR tree-optimization/57358
13041 * ipa-prop.c (ipa_func_spec_opts_forbid_analysis_p): New function.
13042 (ipa_compute_jump_functions_for_edge): Bail out if it returns true.
13043 (ipa_analyze_params_uses): Generate pessimistic info when true.
13044
13045 2013-06-24 Martin Jambor <mjambor@suse.cz>
13046
13047 PR tree-optimization/57539
13048 * cgraphclones.c (cgraph_clone_node): Add parameter new_inlined_to, set
13049 global.inlined_to of the new node to it. All callers changed.
13050 * ipa-inline-transform.c (clone_inlined_nodes): New variable
13051 inlining_into, pass it to cgraph_clone_node.
13052 * ipa-prop.c (ipa_propagate_indirect_call_infos): Do not call
13053 ipa_free_edge_args_substructures.
13054 (ipa_edge_duplication_hook): Only add edges from inlined nodes to
13055 rdesc linked list. Do not assert rdesc edges have inlined caller.
13056 Assert we have found an rdesc in the rdesc list.
13057
13058 2013-06-24 Richard Biener <rguenther@suse.de>
13059
13060 * pointer-set.h (struct pointer_set_t): Move here from pointer-set.c.
13061 (pointer_set_lookup): Declare.
13062 (class pointer_map): New template class implementing a
13063 generic pointer to T map.
13064 (pointer_map<T>::pointer_map, pointer_map<T>::~pointer_map,
13065 pointer_map<T>::contains, pointer_map<T>::insert,
13066 pointer_map<T>::traverse): New functions.
13067 * pointer-set.c (struct pointer_set_t): Moved to pointer-set.h.
13068 (pointer_set_lookup): New function.
13069 (pointer_set_contains): Use pointer_set_lookup.
13070 (pointer_set_insert): Likewise.
13071 (insert_aux): Remove.
13072 (struct pointer_map_t): Embed a pointer_set_t.
13073 (pointer_map_create): Adjust.
13074 (pointer_map_destroy): Likewise.
13075 (pointer_map_contains): Likewise.
13076 (pointer_map_insert): Likewise.
13077 (pointer_map_traverse): Likewise.
13078 * tree-streamer.h (struct streamer_tree_cache_d): Use a
13079 pointer_map<unsigned> instead of a pointer_map_t.
13080 * tree-streamer.c (streamer_tree_cache_insert_1): Adjust.
13081 (streamer_tree_cache_lookup): Likewise.
13082 (streamer_tree_cache_create): Likewise.
13083 (streamer_tree_cache_delete): Likewise.
13084 * lto-streamer.h (struct lto_tree_ref_encoder): Use a
13085 pointer_map<unsigned> instead of a pointer_map_t.
13086 (lto_init_tree_ref_encoder): Adjust.
13087 (lto_destroy_tree_ref_encoder): Likewise.
13088 * lto-section-out.c (lto_output_decl_index): Likewise.
13089 (lto_record_function_out_decl_state): Likewise.
13090 * dominance.c (iterate_fix_dominators): Use pointer_map<int>.
13091
13092 2013-06-24 Richard Biener <rguenther@suse.de>
13093
13094 PR tree-optimization/57488
13095 * tree-ssa-pre.c (insert): Clear NEW sets before each iteration.
13096
13097 2013-06-24 Alan Modra <amodra@gmail.com>
13098
13099 * config/rs6000/rs6000.c (vspltis_constant): Correct for little-endian.
13100 (gen_easy_altivec_constant): Likewise.
13101 * config/rs6000/predicates.md (easy_vector_constant_add_self,
13102 easy_vector_constant_msb): Likewise.
13103
13104 2013-06-23 Jakub Jelinek <jakub@redhat.com>
13105
13106 PR target/57688
13107 * common/config/i386/i386-common.c (ix86_handle_option): For OPT_mlzcnt
13108 add missing return true.
13109
13110 2013-06-23 Oleg Endo <olegendo@gcc.gnu.org>
13111
13112 PR target/52483
13113 * config/sh/predicates.md (general_extend_operand): Invoke
13114 general_movsrc_operand for memory operands.
13115 (general_movsrc_operand): Allow reg+reg addressing, do not use
13116 general_operand for memory operands.
13117
13118 2013-06-23 Sriraman Tallam <tmsriram@google.com>
13119
13120 * config/i386/i386.c (ix86_pragma_target_parse): Restore target
13121 when current target options does not apply.
13122 * config/i386/i386-protos.h (ix86_reset_previous_fndecl): New function.
13123 * config/i386/i386.c (ix86_reset_previous_fndecl): Ditto.
13124 * config/i386/bmiintrin.h: Pass appropriate target
13125 attributes to header.
13126 * config/i386/mmintrin.h: Ditto.
13127 * config/i386/nmmintrin.h: Ditto.
13128 * config/i386/avx2intrin.h: Ditto.
13129 * config/i386/fxsrintrin.h: Ditto.
13130 * config/i386/tbmintrin.h: Ditto.
13131 * config/i386/xsaveintrin.h: Ditto.
13132 * config/i386/f16cintrin.h: Ditto.
13133 * config/i386/xtestintrin.h: Ditto.
13134 * config/i386/xsaveoptintrin.h: Ditto.
13135 * config/i386/bmi2intrin.h: Ditto.
13136 * config/i386/lzcntintrin.h: Ditto.
13137 * config/i386/smmintrin.h: Ditto.
13138 * config/i386/wmmintrin.h: Ditto.
13139 * config/i386/x86intrin.h: Remove all header include guards.
13140 * config/i386/prfchwintrin.h: Ditto.
13141 * config/i386/pmmintrin.h: Ditto.
13142 * config/i386/tmmintrin.h: Ditto.
13143 * config/i386/xmmintrin.h: Ditto.
13144 * config/i386/popcntintrin.h: Ditto.
13145 * config/i386/rdseedintrin.h: Ditto.
13146 * config/i386/ammintrin.h: Ditto.
13147 * config/i386/emmintrin.h: Ditto.
13148 * config/i386/immintrin.h: Remove all header include guards.
13149 * config/i386/fma4intrin.h: Ditto.
13150 * config/i386/lwpintrin.h: Ditto.
13151 * config/i386/xopintrin.h: Ditto.
13152 * config/i386/ia32intrin.h: Ditto.
13153 * config/i386/avxintrin.h: Ditto.
13154 * config/i386/rtmintrin.h: Ditto.
13155 * config/i386/fmaintrin.h: Ditto.
13156 * config/i386/mm3dnow.h: Ditto.
13157
13158 2013-06-22 Sriraman Tallam <tmsriram@google.com>
13159
13160 * common/config/i386/i386-common.c: Handle LZCNT.
13161
13162 2013-06-22 Andi Kleen <ak@linux.intel.com>
13163
13164 * doc/extend.texi: Use __atomic_store_n instead of
13165 __atomic_store in HLE example.
13166
13167 2013-06-22 Oleg Endo <olegendo@gcc.gnu.org>
13168
13169 * config/sh/sh.c: Remove <cstdlib> workaround.
13170
13171 2013-06-21 Andi Kleen <ak@linux.intel.com>
13172
13173 * doc/extend.texi: Dont use __atomic_clear in HLE example. Fix typo.
13174
13175 2013-06-21 Andi Kleen <ak@linux.intel.com>
13176
13177 * doc/extend.texi: Document that __atomic_clear and
13178 __atomic_test_and_set should only be used with bool.
13179
13180 2013-06-20 Jan Hubicka <jh@suse.cz>
13181
13182 * gimple-fold.c (gimple_extract_devirt_binfo_from_cst): Use
13183 types_same_for_odr.
13184 * tree.c (decls_same_for_odr): New function.
13185 (same_for_edr): New function.
13186 (types_same_for_odr): New function.
13187 (get_binfo_at_offset): Use it.
13188 * tree.h (types_same_for_odr): Declare.
13189
13190 2013-06-20 Oleg Endo <olegendo@gcc.gnu.org>
13191 Jason Merrill <jason@redhat.com>
13192
13193 * system.h: Include <cstdlib> as well as <stdlib.h>.
13194
13195 2013-06-20 Uros Bizjak <ubizjak@gmail.com>
13196
13197 PR target/57655
13198 * config/i386/i386.c (construct_container): Report error if
13199 long double is used with disabled x87 float returns.
13200
13201 2013-06-20 Jan Hubicka <jh@suse.cz>
13202
13203 * lto-cgraph.c (input_symtab): Do not set cgraph state.
13204
13205 2013-06-20 Joern Rennecke <joern.rennecke@embecosm.com>
13206
13207 PR rtl-optimization/57425
13208 PR rtl-optimization/57569
13209 * alias.c (write_dependence_p): Remove parameters mem_mode and
13210 canon_mem_addr. Add parameters x_mode, x_addr and x_canonicalized.
13211 Changed all callers.
13212 (canon_anti_dependence): Get comments and semantics in sync.
13213 Add parameter mem_canonicalized. Changed all callers.
13214 * rtl.h (canon_anti_dependence): Update prototype.
13215
13216 2013-06-20 Richard Biener <rguenther@suse.de>
13217
13218 * data-streamer-in.c (streamer_read_uhwi): Optimize single
13219 byte case, inline streamer_read_uchar and defer section
13220 overrun check.
13221
13222 2013-06-20 Richard Biener <rguenther@suse.de>
13223
13224 PR tree-optimization/57584
13225 * tree-ssa-loop-niter.c (expand_simple_operations): Avoid including
13226 SSA names into the expanded expression that take part in
13227 abnormal coalescing.
13228
13229 2013-06-19 Sharad Singhai <singhai@google.com>
13230
13231 * gcov.c (print_usage): Handle new option.
13232 (process_args): Ditto.
13233 (get_gcov_intermediate_filename): New function.
13234 (output_intermediate_file): New function.
13235 (output_gcov_file): New function
13236 (generate_results): Handle new option.
13237 (release_function): Relase demangled name.
13238 (read_graph_file): Handle demangled name.
13239 (output_lines): Ditto.
13240 * doc/gcov.texi: Document gcov intermediate format.
13241
13242 2013-06-19 Vladimir Makarov <vmakarov@redhat.com>
13243
13244 PR bootstrap/57604
13245 * lra.c (emit_add3_insn, emit_add2_insn): New functions.
13246 (lra_emit_add): Use the functions. Add comment about Y as an
13247 address segment.
13248
13249 2013-06-19 David Edelsohn <dje.gcc@gmail.com>
13250
13251 PR driver/57652
13252 * collect2.c (collect_atexit): New.
13253 (collect_exit): Delete.
13254 (main): Register collect_atexit with atexit.
13255 (collect_wait): Change collect_exit to exit.
13256 (do_wait): Same.
13257 * collect2.h (collect_exit): Delete.
13258 * tlink.c (do_tlink): Rename exit to ret. Change collect_exit to exit.
13259
13260 2013-06-19 Wei Mi <wmi@google.com>
13261
13262 PR rtl-optimization/57518
13263 * ira.c (set_paradoxical_subreg): Set pdx_subregs[regno]
13264 if regno is used in paradoxical subreg.
13265 (update_equiv_regs): Check pdx_subregs[regno] before
13266 set a reg to be equivalent with a mem.
13267
13268 2013-06-19 Matthias Klose <doko@ubuntu.com>
13269
13270 PR driver/57651
13271 * file-find.h (find_a_file): Add a mode parameter.
13272 * file-find.c (find_a_file): Likewise.
13273 * gcc-ar.c (main): Call find_a_file with R_OK for the plugin,
13274 with X_OK for the executables.
13275 * collect2.c (main): Call find_a_file with X_OK.
13276
13277 2013-06-19 Steve Ellcey <sellcey@mips.com>
13278
13279 PR target/56942
13280 * config/mips/mips.md (casesi_internal_mips16_<mode>):
13281 Use NEXT_INSN instead of next_real_insn.
13282
13283 2013-06-19 Jan Hubicka <jh@suse.cz>
13284
13285 * cgraph.h (const_value_known_p): Replace by ...
13286 (ctor_for_folding): .. this one.
13287 * cgraphunit.c (process_function_and_variable_attributes): Use it.
13288 * lto-cgraph.c (compute_ltrans_boundary): Use ctor_for_folding.
13289 * expr.c (expand_expr_real_1): Likewise.
13290 (string_constant): Likewise.
13291 * tree-ssa-loop-ivcanon.c (constant_after_peeling): Likewise.
13292 * ipa.c (process_references): Likewise.
13293 (symtab_remove_unreachable_nodes): Likewise.
13294 * ipa-inline-analysis.c (param_change_prob): Likewise.
13295 * gimple-fold.c (canonicalize_constructor_val): Likewise.
13296 (get_base_constructor): Likwise.
13297 * varpool.c (varpool_remove_node): Likewise.
13298 (varpool_remove_initializer): LIkewise.
13299 (dump_varpool_node): LIkwise.
13300 (const_value_known_p): Rewrite to ...
13301 (ctor_for_folding): ... this one.
13302
13303 2013-06-19 Jakub Jelinek <jakub@redhat.com>
13304
13305 PR driver/57651
13306 * gcc-ar.c (main): If not CROSS_DIRECTORY_STRUCTURE, look for
13307 PERSONALITY in $PATH derived prefixes.
13308
13309 2013-06-19 Jeff Law <law@redhat.com>
13310
13311 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): Fix typo
13312 in comment.
13313
13314 * tree-ssa-forwprop.c (simplify_bitwise_binary_boolean): New function.
13315 (simplify_bitwise_binary): Use it to simpify certain binary ops on
13316 booleans.
13317
13318 2013-06-19 Sofiane Naci <sofiane.naci@arm.com>
13319
13320 * config/arm/vfp.md: Move VFP instruction classification documentation
13321 to ...
13322 * config/arm/arm.md: ... here. Update instruction classification
13323 documentation.
13324
13325 2013-06-19 Richard Earnshaw <rearnsha@arm.com>
13326
13327 arm.md (split for eq(reg, 0)): Add variants for ARMv5 and Thumb2.
13328 (peepholes for eq(reg, not-0)): Ensure condition register is dead after
13329 pattern. Use more efficient sequences on ARMv5 and Thumb2.
13330
13331 2013-06-19 Steven Bosscher <steven@gcc.gnu.org>
13332
13333 PR target/57609
13334 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13335 with NEXT_INSN. Use tablejump_p to check for jump table data
13336 insns.
13337
13338 2013-06-19 Paolo Carlini <paolo.carlini@oracle.com>
13339
13340 PR c++/56544
13341 * doc/cpp.texi [Standard Predefined Macros, __cplusplus]: Document
13342 that now in C++ the value is correct per the C++ standards.
13343
13344 2013-06-19 Richard Biener <rguenther@suse.de>
13345
13346 * expr.c (expand_expr_real_1): Use SCOPE_FILE_SCOPE_P to check
13347 for global context.
13348
13349 2013-06-19 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13350
13351 Revert:
13352 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13353
13354 PR target/57609
13355 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13356 with next_active_insn.
13357
13358 2013-06-18 Sriraman Tallam <tmsriram@google.com>
13359
13360 * ipa-inline.c (inline_always_inline_functions): Pretend always_inline
13361 functions are inlined during failures to flag an error.
13362 * tree-inline.c (expand_call_inline): Allow the error to be flagged
13363 in early inline pass.
13364
13365 2013-06-18 H.J. Lu <hongjiu.lu@intel.com>
13366
13367 * config/i386/i386.c (initial_ix86_tune_features): Fix a typo
13368 in comments.
13369
13370 2013-06-18 Julian Brown <julian@codesourcery.com>
13371
13372 * config/arm/arm.c (neon_vector_mem_operand): Add strict argument.
13373 Permit virtual register pre-reload if !strict.
13374 (coproc_secondary_reload_class): Adjust for neon_vector_mem_operand
13375 change.
13376 * config/arm/arm-protos.h (neon_vector_mem_operand): Adjust
13377 prototype.
13378 * config/arm/neon.md (movmisalign<mode>): Use
13379 neon_perm_struct_or_reg_operand instead of
13380 neon_struct_or_register_operand.
13381 (*movmisalign<mode>_neon_load, *movmisalign<mode>_neon_store): Use
13382 neon_permissive_struct_operand instead of neon_struct_operand.
13383 * config/arm/constraints.md (Un, Um, Us): Adjust calls to
13384 neon_vector_mem_operand.
13385 * config/arm/predicates.md (neon_struct_operand): Adjust call to
13386 neon_vector_mem_operand.
13387 (neon_permissive_struct_operand): New.
13388 (neon_struct_or_register_operand): Rename to...
13389 (neon_perm_struct_or_reg_operand): This. Adjust call to
13390 neon_vector_mem_operand.
13391
13392 2013-06-18 Richard Biener <rguenther@suse.de>
13393
13394 * Makefile.in (LTO_STREAMER_H): Add pointer-set.h dependency.
13395 * lto-streamer.h: Include pointer-set.h.
13396 (struct lto_decl_slot): Remove.
13397 (struct lto_tree_ref_encoder): Make tree_hash_table a pointer-map.
13398 Remove next_index entry.
13399 (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13400 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13401 (lto_init_tree_ref_encoder): Adjust.
13402 (lto_destroy_tree_ref_encoder): Likewise.
13403 * lto-section-out.c (lto_hash_decl_slot_node, lto_eq_decl_slot_node,
13404 lto_hash_type_slot_node, lto_eq_type_slot_node): Remove.
13405 (lto_output_decl_index): Adjust.
13406 (lto_new_out_decl_state): Likewise.
13407 (lto_record_function_out_decl_state): Likewise.
13408 * lto-streamer-out.c (copy_function): Likewise.
13409
13410 2013-06-18 Richard Biener <rguenther@suse.de>
13411
13412 * Makefile.in (cgraphunit.o): Add $(CFGLOOP_H) dependency.
13413 * cgraphunit.c: Include cfgloop.h.
13414 (init_lowered_empty_function): Initialize the loop tree.
13415 (assemble_thunk): Insert new BBs into loops.
13416
13417 2013-06-18 Richard Biener <rguenther@suse.de>
13418
13419 * tree-streamer.h (streamer_tree_cache_create): Adjust prototype.
13420 * tree-streamer.c (streamer_tree_cache_create): Make maintaining
13421 the map from cache entry to cache index optional.
13422 (streamer_tree_cache_replace_tree): Adjust accordingly.
13423 (streamer_tree_cache_append): Likewise.
13424 (streamer_tree_cache_delete): Likewise.
13425 * lto-streamer-in.c (lto_data_in_create): Do not maintain the
13426 streamer cache map from cache entry to cache index.
13427 * lto-streamer-out.c (create_output_block): Adjust.
13428
13429 2013-06-18 Sofiane Naci <sofiane.naci@arm.com>
13430
13431 * config/arm/arm.md (attribute "insn"): Move multiplication and
13432 division attributes to...
13433 (attribute "type"): ... here. Remove mult.
13434 (attribute "mul32"): New attribute.
13435 (attribute "mul64"): Add umaal.
13436 (*arm_mulsi3): Update attributes.
13437 (*arm_mulsi3_v6): Likewise.
13438 (*thumb_mulsi3): Likewise.
13439 (*thumb_mulsi3_v6): Likewise.
13440 (*mulsi3_compare0): Likewise.
13441 (*mulsi3_compare0_v6): Likewise.
13442 (*mulsi_compare0_scratch): Likewise.
13443 (*mulsi_compare0_scratch_v6): Likewise.
13444 (*mulsi3addsi): Likewise.
13445 (*mulsi3addsi_v6): Likewise.
13446 (*mulsi3addsi_compare0): Likewise.
13447 (*mulsi3addsi_compare0_v6): Likewise.
13448 (*mulsi3addsi_compare0_scratch): Likewise.
13449 (*mulsi3addsi_compare0_scratch_v6): Likewise.
13450 (*mulsi3subsi): Likewise.
13451 (*mulsidi3adddi): Likewise.
13452 (*mulsi3addsi_v6): Likewise.
13453 (*mulsidi3adddi_v6): Likewise.
13454 (*mulsidi3_nov6): Likewise.
13455 (*mulsidi3_v6): Likewise.
13456 (*umulsidi3_nov6): Likewise.
13457 (*umulsidi3_v6): Likewise.
13458 (*umulsidi3adddi): Likewise.
13459 (*umulsidi3adddi_v6): Likewise.
13460 (*smulsi3_highpart_nov6): Likewise.
13461 (*smulsi3_highpart_v6): Likewise.
13462 (*umulsi3_highpart_nov6): Likewise.
13463 (*umulsi3_highpart_v6): Likewise.
13464 (mulhisi3): Likewise.
13465 (*mulhisi3tb): Likewise.
13466 (*mulhisi3bt): Likewise.
13467 (*mulhisi3tt): Likewise.
13468 (maddhisi4): Likewise.
13469 (*maddhisi4tb): Likewise.
13470 (*maddhisi4tt): Likewise.
13471 (maddhidi4): Likewise.
13472 (*maddhidi4tb): Likewise.
13473 (*maddhidi4tt): Likewise.
13474 (divsi3): Likewise.
13475 (udivsi3): Likewise.
13476 * config/arm/thumb2.md (thumb2_mulsi_short): Update attributes.
13477 (thumb2_mulsi_short_compare0): Likewise.
13478 (thumb2_mulsi_short_compare0_scratch): Likewise.
13479 * config/arm/arm1020e.md (1020mult1): Update attribute change.
13480 (1020mult2): Likewise.
13481 (1020mult3): Likewise.
13482 (1020mult4): Likewise.
13483 (1020mult5): Likewise.
13484 (1020mult6): Likewise.
13485 * config/arm/cortex-a15.md (cortex_a15_mult32): Update attribute
13486 change.
13487 (cortex_a15_mult64): Likewise.
13488 (cortex_a15_sdiv): Likewise.
13489 (cortex_a15_udiv): Likewise.
13490 * config/arm/arm1026ejs.md (mult1): Update attribute change.
13491 (mult2): Likewise.
13492 (mult3): Likewise.
13493 (mult4): Likewise.
13494 (mult5): Likewise.
13495 (mult6): Likewise.
13496 * config/arm/marvell-pj4.md (pj4_ir_mul): Update attribute change.
13497 (pj4_ir_div): Likewise.
13498 * config/arm/arm1136jfs.md (11_mult1): Update attribute change.
13499 (11_mult2): Likewise.
13500 (11_mult3): Likewise.
13501 (11_mult4): Likewise.
13502 (11_mult5): Likewise.
13503 (11_mult6): Likewise.
13504 (11_mult7): Likewise.
13505 * config/arm/cortex-a8.md (cortex_a8_mul): Update attribute change.
13506 (cortex_a8_mla): Likewise.
13507 (cortex_a8_mull): Likewise.
13508 (cortex_a8_smulwy): Likewise.
13509 (cortex_a8_smlald): Likewise.
13510 * config/arm/cortex-m4.md (cortex_m4_alu): Update attribute change.
13511 * config/arm/cortex-r4.md (cortex_r4_mul_4): Update attribute change.
13512 (cortex_r4_mul_3): Likewise.
13513 (cortex_r4_mla_4): Likewise.
13514 (cortex_r4_mla_3): Likewise.
13515 (cortex_r4_smlald): Likewise.
13516 (cortex_r4_mull): Likewise.
13517 (cortex_r4_sdiv): Likewise.
13518 (cortex_r4_udiv): Likewise.
13519 * config/arm/cortex-a7.md (cortex_a7_mul): Update attribute change.
13520 (cortex_a7_idiv): Likewise.
13521 * config/arm/arm926ejs.md (9_mult1): Update attribute change.
13522 (9_mult2): Likewise.
13523 (9_mult3): Likewise.
13524 (9_mult4): Likewise.
13525 (9_mult5): Likewise.
13526 (9_mult6): Likewise.
13527 * config/arm/cortex-a53.md (cortex_a53_mul): Update attribute change.
13528 (cortex_a53_sdiv): Likewise.
13529 (cortex_a53_udiv): Likewise.
13530 * config/arm/fa726te.md (726te_mult_op): Update attribute change.
13531 * config/arm/fmp626.md (mp626_mult1): Update attribute change.
13532 (mp626_mult2): Likewise.
13533 (mp626_mult3): Likewise.
13534 (mp626_mult4): Likewise.
13535 * config/arm/fa526.md (526_mult1): Update attribute change.
13536 (526_mult2): Likewise.
13537 * config/arm/arm-generic.md (mult): Update attribute change.
13538 (mult_ldsched_strongarm): Likewise.
13539 (mult_ldsched): Likewise.
13540 (multi_cycle): Likewise.
13541 * config/arm/cortex-a5.md (cortex_a5_mul): Update attribute change.
13542 * config/arm/fa606te.md (606te_mult1): Update attribute change.
13543 (606te_mult2): Likewise.
13544 (606te_mult3): Likewise.
13545 (606te_mult4): Likewise.
13546 * config/arm/cortex-a9.md (cortex_a9_mult16): Update attribute change.
13547 (cortex_a9_mac16): Likewise.
13548 (cortex_a9_multiply): Likewise.
13549 (cortex_a9_mac): Likewise.
13550 (cortex_a9_multiply_long): Likewise.
13551 * config/arm/fa626te.md (626te_mult1): Update attribute change.
13552 (626te_mult2): Likewise.
13553 (626te_mult3): Likewise.
13554 (626te_mult4): Likewise.
13555
13556 2013-06-18 Richard Biener <rguenther@suse.de>
13557
13558 PR lto/57334
13559 * lto-symtab.c (lto_symtab_merge_decls): Process nodes properly.
13560
13561 2013-06-18 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
13562
13563 PR target/57609
13564 * config/s390/s390.c (s390_chunkify_start): Replace next_real_insn
13565 with next_active_insn.
13566
13567 2013-06-18 Alan Modra <amodra@gmail.com>
13568
13569 * config/rs6000/rs6000.h (enum data_align): New.
13570 (LOCAL_ALIGNMENT, DATA_ALIGNMENT): Use rs6000_data_alignment.
13571 (DATA_ABI_ALIGNMENT): Define.
13572 (CONSTANT_ALIGNMENT): Correct comment.
13573 * config/rs6000/rs6000-protos.h (rs6000_data_alignment): Declare.
13574 * config/rs6000/rs6000.c (rs6000_data_alignment): New function.
13575
13576 2013-06-17 David Malcolm <dmalcolm@redhat.com>
13577
13578 * ggc-page.c (ggc_pch_write_object) <d>: Remove erroneous
13579 ATTRIBUTE_UNUSED marking.
13580
13581 2013-06-17 Sofiane Naci <sofiane.naci@arm.com>
13582
13583 * config/aarch64/aarch64-simd.md (aarch64_dup_lane<mode>): Add r<-w
13584 alternative and update.
13585 (aarch64_dup_lanedi): Delete.
13586 * config/aarch64/arm_neon.h (vdup<bhsd>_lane_*): Update.
13587 * config/aarch64/aarch64-simd-builtins.def: Update.
13588
13589 2013-06-17 Richard Biener <rguenther@suse.de>
13590
13591 * lto-streamer.h (enum LTO_tags): Add LTO_tree_scc.
13592 (lto_input_scc): Declare.
13593 (lto_input_tree_1): Likewise.
13594 (struct lto_stats_d): Add num_tree_bodies_output and
13595 num_pickle_refs_output.
13596 * lto-streamer-in.c (lto_read_body): Use streamer_tree_cache_get_tree.
13597 (lto_read_tree_1): Split out from ...
13598 (lto_read_tree): ... this.
13599 (lto_input_scc): New function.
13600 (lto_input_tree_1): Split out from ...
13601 (lto_input_tree): ... this. Handle LTO_tree_scc.
13602 (lto_data_in_create): Create the streamer cache without hashes.
13603 * lto-streamer-out.c (create_output_block): Create the streamer
13604 cache with hashes when not doing WPA.
13605 (lto_write_tree_1): Split out from ...
13606 (lto_write_tree): ... this.
13607 (get_symbol_initial_value): New function.
13608 (lto_output_tree_1): Split out from ...
13609 (lto_output_tree): ... this. Write trees as series of SCCs
13610 using a DFS walk via DFS_write_tree.
13611 (struct sccs, struct scc_entry): New types.
13612 (next_dfs_num, sccstack, sccstate, sccstate_obstack): New globals.
13613 (DFS_write_tree_body): New function.
13614 (DFS_write_tree): Likewise.
13615 (hash_tree): Likewise.
13616 (scc_entry_compare): Likewise.
13617 (hash_scc): Likewise.
13618 (tree_is_indexable): DEBUG_EXPR_DECLs are local entities.
13619 * tree-streamer-in.c (lto_input_ts_list_tree_pointers): Stream
13620 TREE_CHAIN as regular reference.
13621 (streamer_read_integer_cst): Remove.
13622 (streamer_get_pickled_tree): Adjust.
13623 * tree-streamer-out.c (streamer_write_chain): Disable streaming
13624 of DECL_EXTERNALs in BLOCK_VARS for now.
13625 (write_ts_list_tree_pointers): Stream TREE_CHAIN as regular
13626 reference.
13627 * tree-streamer.c (streamer_tree_cache_add_to_node_array):
13628 Add hash value argument and record that if hashes are recorded
13629 in the cache.
13630 (streamer_tree_cache_insert_1): Adjust.
13631 (streamer_tree_cache_insert): Likewise.
13632 (streamer_tree_cache_insert_at): Rename to ...
13633 (streamer_tree_cache_replace_tree): ... this and adjust.
13634 (streamer_tree_cache_append): Adjust.
13635 (record_common_node): Likewise.
13636 (streamer_tree_cache_create): Add argument whether to
13637 record hash values together with trees.
13638 (streamer_tree_cache_delete): Adjust.
13639 * tree-streamer.h (struct streamer_tree_cache_d): Add
13640 vector of hashes.
13641 (streamer_read_integer_cst): Remove.
13642 (streamer_tree_cache_insert): Adjust.
13643 (streamer_tree_cache_append): Likewise.
13644 (streamer_tree_cache_insert_at): Rename to ...
13645 (streamer_tree_cache_replace_tree): ... this and adjust.
13646 (streamer_tree_cache_create): Add argument whether to record hashes.
13647 (streamer_tree_cache_get): Rename to ...
13648 (streamer_tree_cache_get_tree): ... this.
13649 (streamer_tree_cache_get_hash): New function.
13650 * tree.c (cache_integer_cst): New function.
13651 * tree.h (cache_integer_cst): Declare.
13652 (ANON_AGGRNAME_FORMAT, ANON_AGGRNAME_P): Move here from cp/cp-tree.h.
13653 * lto-symtab.c (lto_varpool_replace_node): Only release
13654 DECL_INITIAL of non-prevailing decls.
13655 * varpool.c (varpool_remove_initializer): Do not release
13656 DECL_INITIAL when we are still in CGRAPH_LTO_STREAMING.
13657
13658 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13659
13660 * config/mips/mips.h (ISA_HAS_MUL3): Include TARGET_MIPS5900.
13661 (ISA_HAS_MULT, ISA_HAS_DMULT, ISA_HAS_DIV, ISA_HAS_DDIV): New macros.
13662 * config/mips/mips.md (mul<mode>3, mul<mode>3_internal)
13663 (mul<mode>3_r4000): Require ISA_HAS_<D>MULT.
13664 (mul<mode>3_mul3): Handle TARGET_MIPS5900.
13665 (mulsidi3_64bit_dmul): Remove redundant TARGET_64BIT test.
13666 (<su>muldi3_highpart, <su>muldi3_highpart_internal, <u>mulditi3)
13667 (<u>mulditi3_internal, <u>mulditi3_r4000): Require ISA_HAS_DMULT
13668 instead of TARGET_64BIT.
13669 (divmod<mode>4, udivmod<mode>4, <u>divmod<GPR:mode>4_hilo_<HILO:mode>):
13670 Require ISA_HAS_<D>DIV.
13671
13672 2013-06-16 Richard Sandiford <rdsandiford@googlemail.com>
13673
13674 * config.gcc (mips*-mti-linux*, mips64*-*-linux*, mipsisa64*-*-linux*)
13675 (mips*-*-linux*): Move default with_llsc setting to where other
13676 defaults are set.
13677 (mips*-*-vxworks*): Move with_arch default from with_cpu block to
13678 with_arch block.
13679 (mips64r5900-*-*, mips64r5900el-*-*, mipsr5900-*-*, mipsr5900el-*-*):
13680 Likewise. Remove default with_tune setting. Move default float
13681 setting to its own block. Handle with_llsc in the same block as above.
13682
13683 2013-06-16 Joern Rennecke <joern.rennecke@embecosm.com>
13684
13685 PR rtl-optimization/57425
13686 PR rtl-optimization/57569
13687 * alias.c (write_dependence_p): Add new parameters mem_mode,
13688 canon_mem_addr and mem_canonicalized. Change type of writep to bool.
13689 Changed all callers.
13690 (canon_anti_dependence): New function.
13691 * cse.c (check_dependence): Use canon_anti_dependence.
13692 * cselib.c (cselib_invalidate_mem): Likewise.
13693 * rtl.h (canon_anti_dependence): Declare.
13694
13695 2013-06-16 Jürgen Urban <JuergenUrban@gmx.de>
13696
13697 * config/mips/mips.h (ISA_HAS_LL_SC): Exclude TARGET_MIPS5900.
13698 * config/mips/mips.c (mips_start_ll_sc_sync_block): Output
13699 ".set mips3" for 64-bit targets.
13700
13701 2013-06-15 Dehao Chen <dehao@google.com>
13702
13703 * tree-flow.h (gimple_check_call_matching_types): Add new argument.
13704 * gimple-low.c (gimple_check_call_matching_types): Likewise.
13705 (gimple_check_call_args): Likewise.
13706 * value-prof.c (check_ic_target): Likewise.
13707 * ipa-inline.c (early_inliner): Likewise.
13708 * ipa-prop.c (update_indirect_edges_after_inlining): Likewise.
13709 * cgraph.c (cgraph_create_edge_1): Likewise.
13710 (cgraph_make_edge_direct): Likewise.
13711
13712 2013-06-14 Michael Meissner <meissner@linux.vnet.ibm.com>
13713
13714 PR target/57615
13715 * config/rs6000/rs6000.md (mov<mode>_ppc64): Call
13716 rs6000_output_move_128bit to handle emitting quad memory
13717 operations. Set attribute length to 8 bytes.
13718
13719 2013-06-14 Vidya Praveen <vidyapraveen@arm.com>
13720
13721 * config/aarch64/aarch64-simd.md (aarch64_<su>mlal_lo<mode>):
13722 New pattern.
13723 (aarch64_<su>mlal_hi<mode>, aarch64_<su>mlsl_lo<mode>): Likewise.
13724 (aarch64_<su>mlsl_hi<mode>, aarch64_<su>mlal<mode>): Likewise.
13725 (aarch64_<su>mlsl<mode>): Likewise.
13726
13727 2013-06-14 Mike Stump <mikestump@comcast.net>
13728
13729 * Makefile.in (TARGET_H): Add insn-codes.h.
13730
13731 2013-06-14 Alan Modra <amodra@gmail.com>
13732
13733 PR middle-end/57134
13734 PR middle-end/57586
13735 * expr.c (expand_expr_real_1 <normal_inner_ref>): Pass
13736 EXPAND_MEMORY and EXPAND_WRITE to recursive call. Don't use
13737 bitfield expansion when EXPAND_MEMORY.
13738 (expand_expr_real_1 <VIEW_CONVERT_EXPR>): Pass modifier likewise.
13739
13740 2013-06-13 Michael Meissner <meissner@linux.vnet.ibm.com>
13741
13742 * config/rs6000/rs6000.c (rs6000_option_override_internal): Move
13743 test for clearing quad memory on 32-bit later.
13744
13745 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
13746
13747 * fold-const.c (negate_expr_p): Handle VECTOR_CST.
13748 (fold_negate_expr): Likewise.
13749 (fold_real_zero_addition_p): Handle vectors.
13750 (fold_binary_loc) <PLUS_EXPR, MINUS_EXPR>: Likewise.
13751
13752 2013-06-14 Alan Modra <amodra@gmail.com>
13753
13754 * varasm.c (force_const_mem): Revert 2013-06-07 change.
13755
13756 2013-06-13 Jan Hubicka <jh@suse.cz>
13757
13758 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
13759 Local comdats are not externally visible.
13760 * symtab.c (dump_symtab_base): Dump externally visible.
13761 (verify_symtab_base): Verify back links in the symtab hash.
13762
13763 2013-06-13 Bin Cheng <bin.cheng@arm.com>
13764
13765 * fold-const.c (operand_equal_p): Consider NOP_EXPR and
13766 CONVERT_EXPR as equal nodes.
13767
13768 2013-06-13 Bin Cheng <bin.cheng@arm.com>
13769
13770 * rtlanal.c (noop_move_p): Check the code to be executed for COND_EXEC.
13771
13772 2013-06-13 Marc Glisse <marc.glisse@inria.fr>
13773
13774 * tree-ssa-forwprop.c (simplify_bitwise_binary, associate_plusminus):
13775 Generalize to complex and vector.
13776 * tree.c (build_all_ones_cst): New function.
13777 * tree.h (build_all_ones_cst): Declare it.
13778
13779 2013-06-13 Alan Modra <amodra@gmail.com>
13780
13781 * config/rs6000/rs6000.h (LONG_DOUBLE_LARGE_FIRST): Define.
13782 * config/rs6000/rs6000.md (signbittf2): New insn.
13783 (extenddftf2_internal): Use LONG_DOUBLE_LARGE_FIRST.
13784 (abstf2_internal, cmptf_internal2): Likewise.
13785 * config/rs6000/spe.md (spe_abstf2_cmp, spe_abstf2_tst): Likewise.
13786
13787 2013-06-12 Michael Meissner <meissner@linux.vnet.ibm.com>
13788 Pat Haugen <pthaugen@us.ibm.com>
13789 Peter Bergner <bergner@vnet.ibm.com>
13790
13791 * config/rs6000/rs6000.c (emit_load_locked): Add support for
13792 power8 byte, half-word, and quad-word atomic instructions.
13793 (emit_store_conditional): Likewise.
13794 (rs6000_expand_atomic_compare_and_swap): Likewise.
13795 (rs6000_expand_atomic_op): Likewise.
13796
13797 * config/rs6000/sync.md (larx): Add new modes for power8.
13798 (stcx): Likewise.
13799 (AINT): New mode iterator to include TImode as well as normal
13800 integer modes on power8.
13801 (fetchop_pred): Use int_reg_operand instead of gpc_reg_operand so
13802 that VSX registers are not considered. Use AINT mode iterator
13803 instead of INT1 to allow inclusion of quad word atomic operations
13804 on power8.
13805 (load_locked<mode>): Likewise.
13806 (store_conditional<mode>): Likewise.
13807 (atomic_compare_and_swap<mode>): Likewise.
13808 (atomic_exchange<mode>): Likewise.
13809 (atomic_nand<mode>): Likewise.
13810 (atomic_fetch_<fetchop_name><mode>): Likewise.
13811 (atomic_nand_fetch<mode>): Likewise.
13812 (mem_thread_fence): Use gen_loadsync_<mode> instead of enumerating
13813 each type.
13814 (ATOMIC): On power8, add QImode, HImode modes.
13815 (load_locked<QHI:mode>_si): Varients of load_locked for QI/HI
13816 modes that promote to SImode.
13817 (load_lockedti): Convert TImode arguments to PTImode, so that we
13818 get a guaranteed even/odd register pair.
13819 (load_lockedpti): Likewise.
13820 (store_conditionalti): Likewise.
13821 (store_conditionalpti): Likewise.
13822
13823 * config/rs6000/rs6000.md (QHI): New mode iterator for power8
13824 atomic load/store instructions.
13825 (HSI): Likewise.
13826
13827 2013-06-12 Richard Sandiford <rdsandiford@googlemail.com>
13828
13829 * config/mips/mips.md (extended_mips16): Include GOT and constant-pool
13830 loads.
13831 (insn_count): New attribute, with most cases extracted from...
13832 (length): ...here. Redefine most cases in terms of insn_count.
13833 (single_insn): Delete.
13834 (can_delay): Use insn_count to check for single instructions.
13835 (*mul<mode>3_r4300, mul<mode>3_r4000, *mul_acc_si, *mul_acc_si_r3900)
13836 (*msac_using_macc, *mul_sub_si, <u>mulsidi3_32bit_r4000)
13837 (<u>mulsidi3_64bit_r4000, <su>muldi3_highpart_internal)
13838 (<su>mulsi3_highpart_split, <su>muldi3_highpart_internal)
13839 (<u>mulditi3_r4000, *div<mode>3, *recip<mode>3, divmod<mode>4)
13840 (udivmod<mode>4, sqrt<mode>2, *rsqrt<mode>a, *rsqrt<mode>b)
13841 (fix_truncdfsi2_macro, fix_truncsfsi2_macro, *lea_high64)
13842 (*lea64, cprestore_<mode>, clear_hazard_<mode>, <unnamed insn>)
13843 (casesi_internal_mips16_<mode>, *tls_get_tp_<mode>_split)
13844 (tls_get_tp_mips16, *tls_get_tp_mips16_call_<mode>): Use "insn_count"
13845 rather than "length".
13846 (tls_get_tp_<mode>): Likewise. Remove redundant "no_delay" attribute.
13847 * config/mips/mips-ps-3d.md (mips_c_cond_4s, mips_cabs_cond_4s):
13848 Use "insn_count" rather than "length".
13849 * config/mips/mips-dsp.md
13850 (mips_l<SHORT:size><u>x_ext<GPR:mode>_<P:mode>)
13851 (mips_l<GPR:size>x_<P:mode>, *mips_lw<u>x_<P:mode>_ext): Remove
13852 length attributes.
13853
13854 2013-06-12 Marc Glisse <marc.glisse@inria.fr>
13855
13856 PR tree-optimization/57361
13857 * tree-ssa-dse.c (dse_possible_dead_store_p): Handle self-assignment.
13858
13859 2013-06-12 Sofiane Naci <sofiane.naci@arm.com>
13860
13861 * config/aarch64/aarch64-simd.md (aarch64_combine<mode>): Convert
13862 to split.
13863 (aarch64_simd_combine<mode>): New instruction expansion.
13864 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine): New
13865 function prototype.
13866 * config/aarch64/aarch64.c (aarch64_split_combine): New function.
13867 * config/aarch64/iterators.md (Vdbl): Add entry for DF.
13868
13869 2013-06-12 Jan Hubicka <jh@suse.cz>
13870
13871 * cgraph.c (verify_edge_corresponds_to_fndecl): Be lax about
13872 decl has when in streaming stage.
13873 * lto-symtab.c (lto_symtab_merge_symbols): Likewise.
13874 * cgraph.h (cgraph_state): Add CGRAPH_LTO_STREAMING.
13875
13876 2013-06-12 Roland Stigge <stigge@antcom.de>
13877
13878 PR target/57578
13879 * config/rs6000/t-linux (MULTIARCH_DIRNAME): Fix SPE version detection.
13880
13881 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13882
13883 PR tree-optimization/57537
13884 * tree-vect-patterns.c (vect_recog_widen_mult_pattern): If
13885 vect_handle_widen_op_by_const, convert oprnd1 to half_type1.
13886
13887 2013-06-12 Richard Biener <rguenther@suse.de>
13888
13889 * data-streamer.h (streamer_write_char_stream): CSE
13890 obs->current_pointer.
13891 * data-streamer-out.c (streamer_write_uhwi_stream): Inline
13892 streamer_write_char_stream manually and optimize the resulting loop.
13893 (streamer_write_hwi_stream): Likewise.
13894
13895 2013-06-12 Jan Hubicka <jh@suse.cz>
13896
13897 * lto-symtab.c (lto_symtab_merge_symbols): Populate symtab hashtable.
13898 * cgraph.h (varpool_create_empty_node): Declare.
13899 * lto-cgraph.c (input_node, input_varpool_node): Forcingly create
13900 duplicated nodes.
13901 * symtab.c (symtab_unregister_node): Be lax about missin entries
13902 in node hash.
13903 (symtab_get_node): Update comment.
13904 * varpool.c (varpool_create_empty_node): Break out from ...
13905 (varpool_node_for_decl): ... here.
13906 * lto-streamer.h (lto_file_decl_data): Add RESOLUTION_MAP.
13907
13908 2013-06-12 Eric Botcazou <ebotcazou@adacore.com>
13909
13910 * expr.c (expand_expr_real_1) <TARGET_MEM_REF>: Use straight-line flow.
13911 <MEM_REF>: Use 'type' instead of TREE_TYPE (exp) and tidy up the first
13912 part. Use straight-line flow at the end.
13913 <COMPONENT_REF>: Remove superfluous else.
13914 <VIEW_CONVERT_EXPR>: Use 'type' instead of TREE_TYPE (exp).
13915
13916 2013-06-12 Jakub Jelinek <jakub@redhat.com>
13917
13918 PR target/56564
13919 * varasm.c (decl_binds_to_current_def_p): Call binds_local_p
13920 target hook even for !TREE_PUBLIC decls. If no resolution info
13921 is available, return false for common and external decls.
13922
13923 2013-06-12 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
13924
13925 * config/rl78/constraints.md (U): New constraint.
13926 * config/rl78/rl78.md (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): Add
13927 valloc attribute.
13928
13929 2013-06-11 Michael Meissner <meissner@linux.vnet.ibm.com>
13930
13931 PR target/57589
13932 * config/rs6000/driver-rs6000.c (elf_platform): Make buffer static
13933 to allow returning address to AT_PLATFORM name.
13934
13935 2013-06-11 Jan Hubicka <jh@suse.cz>
13936
13937 * cgraph.c (cgraph_create_function_alias): Set weakref flag.
13938 * cgraph.h (symtab_node_base): Add weakref flag.
13939 * cgraphunit.c (cgraph_reset_node): Clear weakref flag.
13940 (handle_alias_pairs): Set weakref flag, do not set DECL_EXTERNAL.
13941 (output_weakrefs): Use weakref flag.
13942 * fold-const.c (simple_operand_p): Handle WEAK.
13943 * gimple-fold.c (can_refer_decl_in_current_unit_p): Drop weakref.
13944 * ipa.c (varpool_externally_visible_p): Drop weakref.
13945 (function_and_variable_visibility): Update comment; fix weakref
13946 sanity checks; do not clear DECL_WEAK on them.
13947 * lto-cgraph.c (lto_output_node): update.
13948 (lto_output_varpool_node): Update.
13949 (input_overwrite_node): Update.
13950 (input_node): Update.
13951 (input_varpool_node): Update.
13952 * lto-symtab.c (lto_symtab_symbol_p): Do not special case weakrefs.
13953 (lto_symtab_merge_symbols): Add sanity check.
13954 (lto_symtab_prevailing_decl): Do not special case weakrefs.
13955 * passes.c (rest_of_decl_compilation): Set static flag, too.
13956 * symtab.c (dump_symtab_base): Dump weakref.
13957 (verify_symtab_base): Sanity check weakrefs.
13958 (symtab_make_decl_local): Remove duplicated code.
13959 (symtab_alias_ultimate_target): Simplify.
13960 * varpool.c (varpool_create_variable_alias): Set weakref flag.
13961
13962 2013-06-11 Tom de Vries <tom@codesourcery.com>
13963
13964 * genautomata.c (gen_regexp_sequence): Handle els_num == -1. Handle
13965 sequence_vect == NULL.
13966
13967 2013-06-11 DJ Delorie <dj@redhat.com>
13968
13969 * config/rl78/rl78.c (TARGET_UNWIND_WORD_MODE): Define.
13970 (rl78_unwind_word_mode): New.
13971
13972 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13973
13974 * final.c (debug_prefix_maps): Make static.
13975
13976 2013-06-11 David Malcolm <dmalcolm@redhat.com>
13977
13978 * function.c (initial_trampoline): Remove stray copy.
13979
13980 2013-06-11 Sofiane Naci <sofiane.naci@arm.com>
13981
13982 * config/aarch64/aarch64-simd.md (move_lo_quad_<mode>): Update.
13983
13984 2013-06-11 Martin Jambor <mjambor@suse.cz>
13985
13986 * ipa-cp.c (ipa_get_indirect_edge_target_1): Check that param_index is
13987 within bounds at the beginning of the function.
13988
13989 2013-06-11 Alan Modra <amodra@gmail.com>
13990
13991 * varasm.c (get_section): Don't die on !DECL_P decl. Tidy error
13992 reporting.
13993 (get_named_section): Don't NULL !DECL_P decl.
13994
13995 2013-06-11 Igor Zamyatin <igor.zamyatin@intel.com>
13996
13997 * doc/invoke.texi (core-avx2): Document.
13998 (slm): Likewise.
13999 (atom): Updated with MOVBE.
14000
14001 2013-06-11 Richard Biener <rguenther@suse.de>
14002
14003 * collect2.c (main): Do not redirect ld stdout/stderr when debugging.
14004
14005 2013-06-11 Anton Blanchard <anton@samba.org>
14006
14007 * config/rs6000/rs6000.c (rs6000_adjust_atomic_subword): Calculate
14008 correct shift value in little-endian mode.
14009
14010 2013-06-11 Jakub Jelinek <jakub@redhat.com>
14011
14012 PR target/56564
14013 * varasm.c (get_variable_align): Move #endif to the right place.
14014
14015 2013-06-10 Cary Coutant <ccoutant@google.com>
14016
14017 * dwarf2out.c (hash_external_ref): Use die_symbol or signature
14018 for hash so that hash table traversal order is deterministic.
14019
14020 2013-06-10 Michael Meissner <meissner@linux.vnet.ibm.com>
14021 Pat Haugen <pthaugen@us.ibm.com>
14022 Peter Bergner <bergner@vnet.ibm.com>
14023
14024 * config/rs6000/vector.md (GPR move splitter): Do not split moves
14025 of vectors in GPRS if they are direct moves or quad word load or
14026 store moves.
14027
14028 * config/rs6000/rs6000-protos.h (rs6000_output_move_128bit): Add
14029 declaration.
14030 (direct_move_p): Likewise.
14031 (quad_load_store_p): Likewise.
14032
14033 * config/rs6000/rs6000.c (enum rs6000_reg_type): Simplify register
14034 classes into bins based on the physical register type.
14035 (reg_class_to_reg_type): Likewise.
14036 (IS_STD_REG_TYPE): Likewise.
14037 (IS_FP_VECT_REG_TYPE): Likewise.
14038 (reload_fpr_gpr): Arrays to determine what insn to use if we can
14039 use direct move instructions.
14040 (reload_gpr_vsx): Likewise.
14041 (reload_vsx_gpr): Likewise.
14042 (rs6000_init_hard_regno_mode_ok): Precalculate the register type
14043 information that is a simplification of register classes. Also
14044 precalculate direct move reload helpers.
14045 (direct_move_p): New function to return true if the operation can
14046 be done as a direct move instruciton.
14047 (quad_load_store_p): New function to return true if the operation
14048 is a quad memory operation.
14049 (rs6000_legitimize_address): If quad memory, only allow register
14050 indirect for TImode addresses.
14051 (rs6000_legitimate_address_p): Likewise.
14052 (enum reload_reg_type): Delete, replace with rs6000_reg_type.
14053 (rs6000_reload_register_type): Likewise.
14054 (register_to_reg_type): Return register type.
14055 (rs6000_secondary_reload_simple_move): New helper function for
14056 secondary reload and secondary memory needed to identify anything
14057 that is a simple move, and does not need reloading.
14058 (rs6000_secondary_reload_direct_move): New helper function for
14059 secondary reload to identify cases that can be done with several
14060 instructions via the direct move instructions.
14061 (rs6000_secondary_reload_move): New helper function for secondary
14062 reload to identify moves between register types that can be done.
14063 (rs6000_secondary_reload): Add support for quad memory operations
14064 and for direct move.
14065 (rs6000_secondary_memory_needed): Likewise.
14066 (rs6000_debug_secondary_memory_needed): Change argument names.
14067 (rs6000_output_move_128bit): New function to return the move to
14068 use for 128-bit moves, including knowing about the various
14069 limitations of quad memory operations.
14070
14071 * config/rs6000/vsx.md (vsx_mov<mode>): Add support for quad
14072 memory operations. call rs6000_output_move_128bit for the actual
14073 instruciton(s) to generate.
14074 (vsx_movti_64bit): Likewise.
14075
14076 * config/rs6000/rs6000.md (UNSPEC_P8V_FMRGOW): New unspec values.
14077 (UNSPEC_P8V_MTVSRWZ): Likewise.
14078 (UNSPEC_P8V_RELOAD_FROM_GPR): Likewise.
14079 (UNSPEC_P8V_MTVSRD): Likewise.
14080 (UNSPEC_P8V_XXPERMDI): Likewise.
14081 (UNSPEC_P8V_RELOAD_FROM_VSX): Likewise.
14082 (UNSPEC_FUSION_GPR): Likewise.
14083 (FMOVE128_GPR): New iterator for direct move.
14084 (f32_lv): New mode attribute for load/store of SFmode/SDmode values.
14085 (f32_sv): Likewise.
14086 (f32_dm): Likewise.
14087 (zero_extend<mode>di2_internal1): Add support for power8 32-bit
14088 loads and direct move instructions.
14089 (zero_extendsidi2_lfiwzx): Likewise.
14090 (extendsidi2_lfiwax): Likewise.
14091 (extendsidi2_nocell): Likewise.
14092 (floatsi<mode>2_lfiwax): Likewise.
14093 (lfiwax): Likewise.
14094 (floatunssi<mode>2_lfiwzx): Likewise.
14095 (lfiwzx): Likewise.
14096 (fix_trunc<mode>_stfiwx): Likewise.
14097 (fixuns_trunc<mode>_stfiwx): Likewise.
14098 (mov<mode>_hardfloat, 32-bit floating point): Likewise.
14099 (mov<move>_hardfloat64, 64-bit floating point): Likewise.
14100 (parity<mode>2_cmpb): Set length/type attr.
14101 (unnamed shift right patterns, mov<mode>_internal2): Change type attr
14102 for 'mr.' to fast_compare.
14103 (bpermd_<mode>): Change type attr to popcnt.
14104 (p8_fmrgow_<mode>): New insns for power8 direct move support.
14105 (p8_mtvsrwz_1): Likewise.
14106 (p8_mtvsrwz_2): Likewise.
14107 (reload_fpr_from_gpr<mode>): Likewise.
14108 (p8_mtvsrd_1): Likewise.
14109 (p8_mtvsrd_2): Likewise.
14110 (p8_xxpermdi_<mode>): Likewise.
14111 (reload_vsx_from_gpr<mode>): Likewise.
14112 (reload_vsx_from_gprsf): Likewise.
14113 (p8_mfvsrd_3_<mode>): LIkewise.
14114 (reload_gpr_from_vsx<mode>): Likewise.
14115 (reload_gpr_from_vsxsf): Likewise.
14116 (p8_mfvsrd_4_disf): Likewise.
14117 (multi-word GPR splits): Do not split direct moves or quad memory
14118 operations.
14119
14120 2013-06-10 David Malcolm <dmalcolm@redhat.com>
14121
14122 * tree-into-ssa.c (interesting_blocks): Make static.
14123
14124 2013-06-10 Jakub Jelinek <jakub@redhat.com>
14125
14126 PR target/56564
14127 * varasm.c (align_variable): Don't use DATA_ALIGNMENT or
14128 CONSTANT_ALIGNMENT if !decl_binds_to_current_def_p (decl).
14129 Use DATA_ABI_ALIGNMENT for that case instead if defined.
14130 (get_variable_align): New function.
14131 (get_variable_section, emit_bss, emit_common,
14132 assemble_variable_contents, place_block_symbol): Use
14133 get_variable_align instead of DECL_ALIGN.
14134 (assemble_noswitch_variable): Add align argument, use it
14135 instead of DECL_ALIGN.
14136 (assemble_variable): Adjust caller. Use get_variable_align
14137 instead of DECL_ALIGN.
14138 * config/i386/i386.h (DATA_ALIGNMENT): Adjust x86_data_alignment
14139 caller.
14140 (DATA_ABI_ALIGNMENT): Define.
14141 * config/i386/i386-protos.h (x86_data_alignment): Adjust prototype.
14142 * config/i386/i386.c (x86_data_alignment): Add opt argument. If
14143 opt is false, only return the psABI mandated alignment increase.
14144 * config/c6x/c6x.h (DATA_ALIGNMENT): Renamed to...
14145 (DATA_ABI_ALIGNMENT): ... this.
14146 * config/mmix/mmix.h (DATA_ALIGNMENT): Renamed to...
14147 (DATA_ABI_ALIGNMENT): ... this.
14148 * config/mmix/mmix.c (mmix_data_alignment): Adjust function comment.
14149 * config/s390/s390.h (DATA_ALIGNMENT): Renamed to...
14150 (DATA_ABI_ALIGNMENT): ... this.
14151 * doc/tm.texi.in (DATA_ABI_ALIGNMENT): Document.
14152 * doc/tm.texi: Regenerated.
14153
14154 2013-06-10 Uros Bizjak <ubizjak@gmail.com>
14155
14156 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Also use
14157 cmp_code to construct REG_EQUAL note.
14158
14159 2013-06-09 Jakub Jelinek <jakub@redhat.com>
14160
14161 PR target/57568
14162 * config/i386/i386.md (TARGET_READ_MODIFY_WRITE peepholes): Ensure
14163 that operands[2] doesn't overlap with operands[0].
14164
14165 2013-06-09 David Edelsohn <dje.gcc@gmail.com>
14166 Jan Hubicka <jh@suse.cz>
14167
14168 * config/rs6000/rs6000.c (print_operand, 'z'): Remove historical
14169 hack to mark symbols as used.
14170
14171 2013-06-08 Vladimir Makarov <vmakarov@redhat.com>
14172
14173 PR rtl-optimization/57559
14174 * lra-constraints.c (process_alt_operands): Don't discourage
14175 memory with known offset for offsetable memory constraint.
14176 * lra.c (lra_emit_add): Exchange y and z for 2-op add insn.
14177
14178 2013-06-08 Eric Botcazou <ebotcazou@adacore.com>
14179
14180 * varasm.c (struct oc_local_state): Reorder fields.
14181 (output_constructor_bitfield): Replace OUTER parameter with BIT_OFFSET
14182 and adjust accordingly.
14183 (output_constructor): Reorder initialization code and adjust call to
14184 output_constructor_bitfield.
14185
14186 2013-06-07 Jan Hubicka <jh@suse.cz>
14187
14188 * symtab.c (symtab_resolve_alias): Do not remove alias attribute.
14189
14190 2013-06-07 David Malcolm <dmalcolm@redhat.com>
14191
14192 * tree-object-size.c (unknown): Make const.
14193
14194 2013-06-07 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
14195
14196 * config/s390/s390.md (cpu_facility): Add cpu_zarch.
14197 ("*movmem_short", "*clrmem_short", "*cmpmem_short): Use cpu_zarch
14198 for last alternative in the cpu_facility attribute.
14199
14200 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14201
14202 PR target/56315
14203 * config/arm/arm.md (*xordi3_insn): Change to insn_and_split.
14204 (xordi3): Change operand 2 constraint to arm_xordi_operand.
14205 * config/arm/arm.c (const_ok_for_dimode_op): Handle XOR.
14206 * config/arm/constraints.md (Dg): New constraint.
14207 * config/arm/neon.md (xordi3_neon): Remove.
14208 (neon_veor<mode>): Generate xordi3 instead of xordi3_neon.
14209 * config/arm/predicates.md (arm_xordi_operand): New predicate.
14210
14211 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14212
14213 * config/arm/arm.md (anddi3_insn): Remove duplicate alternatives.
14214 Clean up alternatives.
14215
14216 2013-06-07 Alan Modra <amodra@gmail.com>
14217
14218 * config/rs6000/rs6000.c (setup_incoming_varargs): Round up
14219 va_list_gpr_size.
14220
14221 2013-06-07 Alan Modra <amodra@gmail.com>
14222
14223 * varasm.c (force_const_mem): Assert mode is not VOID or BLK.
14224
14225 2013-06-07 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14226
14227 * config/arm/constraints.md (Df): New constraint.
14228 * config/arm/arm.md (iordi3_insn): Use Df constraint instead of De.
14229 Correct length attribute for last two alternatives.
14230
14231 2013-06-07 Alan Modra <amodra@gmail.com>
14232
14233 * config/rs6000/rs6000.c (rs6000_option_override_internal): Don't
14234 override user -mfp-in-toc.
14235 (offsettable_ok_by_alignment): Consider just the current access
14236 rather than the whole object, unless BLKmode. Handle
14237 CONSTANT_POOL_ADDRESS_P constants that lack a decl too.
14238 (use_toc_relative_ref): Allow CONSTANT_POOL_ADDRESS_P constants
14239 for -mcmodel=medium.
14240 * config/rs6000/linux64.h (SUBSUBTARGET_OVERRIDE_OPTIONS): Don't
14241 override user -mfp-in-toc or -msum-in-toc. Default to
14242 -mno-fp-in-toc for -mcmodel=medium.
14243
14244 2013-06-06 DJ Delorie <dj@redhat.com>
14245
14246 * config/rl78/rl78.c (rl78_valid_pointer_mode): New, implements
14247 TARGET_VALID_POINTER_MODE.
14248
14249 2013-06-06 Michael Meissner <meissner@linux.vnet.ibm.com>
14250 Pat Haugen <pthaugen@us.ibm.com>
14251 Peter Bergner <bergner@vnet.ibm.com>
14252
14253 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions):
14254 Document new power8 builtins.
14255
14256 * config/rs6000/vector.md (and<mode>3): Add a clobber/scratch of a
14257 condition code register, to allow 128-bit logical operations to be
14258 done in the VSX or GPR registers.
14259 (nor<mode>3): Use the canonical form for nor.
14260 (eqv<mode>3): Add expanders for power8 xxleqv, xxlnand, xxlorc,
14261 vclz*, and vpopcnt* vector instructions.
14262 (nand<mode>3): Likewise.
14263 (orc<mode>3): Likewise.
14264 (clz<mode>2): LIkewise.
14265 (popcount<mode>2): Likewise.
14266
14267 * config/rs6000/predicates.md (int_reg_operand): Rework tests so
14268 that only the GPRs are recognized.
14269
14270 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
14271 support for new power8 builtins.
14272
14273 * config/rs6000/rs6000-builtin.def (xscvspdpn): Add new power8
14274 builtin functions.
14275 (xscvdpspn): Likewise.
14276 (vclz): Likewise.
14277 (vclzb): Likewise.
14278 (vclzh): Likewise.
14279 (vclzw): Likewise.
14280 (vclzd): Likewise.
14281 (vpopcnt): Likewise.
14282 (vpopcntb): Likewise.
14283 (vpopcnth): Likewise.
14284 (vpopcntw): Likewise.
14285 (vpopcntd): Likewise.
14286 (vgbbd): Likewise.
14287 (vmrgew): Likewise.
14288 (vmrgow): Likewise.
14289 (eqv): Likewise.
14290 (eqv_v16qi3): Likewise.
14291 (eqv_v8hi3): Likewise.
14292 (eqv_v4si3): Likewise.
14293 (eqv_v2di3): Likewise.
14294 (eqv_v4sf3): Likewise.
14295 (eqv_v2df3): Likewise.
14296 (nand): Likewise.
14297 (nand_v16qi3): Likewise.
14298 (nand_v8hi3): Likewise.
14299 (nand_v4si3): Likewise.
14300 (nand_v2di3): Likewise.
14301 (nand_v4sf3): Likewise.
14302 (nand_v2df3): Likewise.
14303 (orc): Likewise.
14304 (orc_v16qi3): Likewise.
14305 (orc_v8hi3): Likewise.
14306 (orc_v4si3): Likewise.
14307 (orc_v2di3): Likewise.
14308 (orc_v4sf3): Likewise.
14309 (orc_v2df3): Likewise.
14310
14311 * config/rs6000/rs6000.c (rs6000_option_override_internal): Only
14312 allow power8 quad mode in 64-bit.
14313 (rs6000_builtin_vectorized_function): Add support to vectorize
14314 ISA 2.07 count leading zeros, population count builtins.
14315 (rs6000_expand_vector_init): On ISA 2.07 use xscvdpspn to form
14316 V4SF vectors instead of xscvdpsp to avoid IEEE related traps.
14317 (builtin_function_type): Add vgbbd builtin function which takes an
14318 unsigned argument.
14319 (altivec_expand_vec_perm_const): Add support for new power8 merge
14320 instructions.
14321
14322 * config/rs6000/vsx.md (VSX_L2): New iterator for 128-bit types,
14323 that does not include TImdoe for use with 32-bit.
14324 (UNSPEC_VSX_CVSPDPN): Support for power8 xscvdpspn and xscvspdpn
14325 instructions.
14326 (UNSPEC_VSX_CVDPSPN): Likewise.
14327 (vsx_xscvdpspn): Likewise.
14328 (vsx_xscvspdpn): Likewise.
14329 (vsx_xscvdpspn_scalar): Likewise.
14330 (vsx_xscvspdpn_directmove): Likewise.
14331 (vsx_and<mode>3): Split logical operations into 32-bit and
14332 64-bit. Add support to do logical operations on TImode as well as
14333 VSX vector types. Allow logical operations to be done in either
14334 VSX registers or in general purpose registers in 64-bit mode. Add
14335 splitters if GPRs were used. For AND, add clobber of CCmode to
14336 allow use of ANDI on GPRs. Rewrite nor to use the canonical RTL
14337 encoding.
14338 (vsx_and<mode>3_32bit): Likewise.
14339 (vsx_and<mode>3_64bit): Likewise.
14340 (vsx_ior<mode>3): Likewise.
14341 (vsx_ior<mode>3_32bit): Likewise.
14342 (vsx_ior<mode>3_64bit): Likewise.
14343 (vsx_xor<mode>3): Likewise.
14344 (vsx_xor<mode>3_32bit): Likewise.
14345 (vsx_xor<mode>3_64bit): Likewise.
14346 (vsx_one_cmpl<mode>2): Likewise.
14347 (vsx_one_cmpl<mode>2_32bit): Likewise.
14348 (vsx_one_cmpl<mode>2_64bit): Likewise.
14349 (vsx_nor<mode>3): Likewise.
14350 (vsx_nor<mode>3_32bit): Likewise.
14351 (vsx_nor<mode>3_64bit): Likewise.
14352 (vsx_andc<mode>3): Likewise.
14353 (vsx_andc<mode>3_32bit): Likewise.
14354 (vsx_andc<mode>3_64bit): Likewise.
14355 (vsx_eqv<mode>3_32bit): Add support for power8 xxleqv, xxlnand,
14356 and xxlorc instructions.
14357 (vsx_eqv<mode>3_64bit): Likewise.
14358 (vsx_nand<mode>3_32bit): Likewise.
14359 (vsx_nand<mode>3_64bit): Likewise.
14360 (vsx_orc<mode>3_32bit): Likewise.
14361 (vsx_orc<mode>3_64bit): Likewise.
14362
14363 * config/rs6000/rs6000.h (VLOGICAL_REGNO_P): Update comment.
14364
14365 * config/rs6000/altivec.md (UNSPEC_VGBBD): Add power8 vgbbd
14366 instruction.
14367 (p8_vmrgew): Add power8 vmrgew and vmrgow instructions.
14368 (p8_vmrgow): Likewise.
14369 (altivec_and<mode>3): Add clobber of CCmode to allow AND using
14370 GPRs to be split under VSX.
14371 (p8v_clz<mode>2): Add power8 count leading zero support.
14372 (p8v_popcount<mode>2): Add power8 population count support.
14373 (p8v_vgbbd): Add power8 gather bits by bytes by doubleword
14374 support.
14375
14376 * config/rs6000/rs6000.md (eqv<mode>3): Add support for powerp eqv
14377 instruction.
14378
14379 * config/rs6000/altivec.h (vec_eqv): Add defines to export power8
14380 builtin functions.
14381 (vec_nand): Likewise.
14382 (vec_vclz): Likewise.
14383 (vec_vclzb): Likewise.
14384 (vec_vclzd): Likewise.
14385 (vec_vclzh): Likewise.
14386 (vec_vclzw): Likewise.
14387 (vec_vgbbd): Likewise.
14388 (vec_vmrgew): Likewise.
14389 (vec_vmrgow): Likewise.
14390 (vec_vpopcnt): Likewise.
14391 (vec_vpopcntb): Likewise.
14392 (vec_vpopcntd): Likewise.
14393 (vec_vpopcnth): Likewise.
14394 (vec_vpopcntw): Likewise.
14395
14396 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14397
14398 PR rtl-optimization/57468
14399 * config/i386/i386.c (inline_secondary_memory_needed): Ignore
14400 spilled pseudos.
14401
14402 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14403
14404 PR rtl-optimization/57459
14405 * lra-constraints.c (update_ebb_live_info): Fix typo for operand
14406 type when setting live regs.
14407
14408 2013-06-06 Vladimir Makarov <vmakarov@redhat.com>
14409
14410 * config/s390/s390.opt (mlra): New option.
14411 * config/s390/s390.c (s390_decompose_address): Check displacement
14412 for all registers for LRA.
14413 (s390_secondary_reload): Don't used secondary reloads for LRA.
14414 (s390_lra_p): New function.
14415 (TARGET_LRA_P): Define.
14416 * config/s390/s390.md (*movmem_short, *clrmem_short): Change value
14417 of attribute cpu_facility to zarch for the last alternative.
14418 (*cmpmem_short): Ditto.
14419
14420 2013-06-06 Eric Botcazou <ebotcazou@adacore.com>
14421
14422 * config/arm/arm.c (arm_r3_live_at_start_p): New predicate.
14423 (arm_compute_static_chain_stack_bytes): Use it. Tidy up.
14424 (arm_expand_prologue): Likewise.
14425
14426 2013-06-06 Teresa Johnson <tejohnson@google.com>
14427
14428 PR c++/53743
14429 * ifcvt.c (find_if_case_1): Replace BB_COPY_PARTITION with assert
14430 as this is now done by redirect_edge_and_branch_force.
14431 * function.c (thread_prologue_and_epilogue_insns): Insert new bb after
14432 barriers, and fix interaction with splitting.
14433 * emit-rtl.c (try_split): Copy REG_CROSSING_JUMP notes.
14434 * cfgcleanup.c (try_forward_edges): Fix early return value to properly
14435 reflect changes made in the routine.
14436 * bb-reorder.c (emit_barrier_after_bb): Move to cfgrtl.c.
14437 (fix_up_fall_thru_edges): Remove incorrect check for bb layout order
14438 since this is called in cfglayout mode, and replace partition fixup
14439 with assert as that is now done by force_nonfallthru_and_redirect.
14440 (add_reg_crossing_jump_notes): Handle the fact that some jumps may
14441 already be marked with region crossing note.
14442 (insert_section_boundary_note): Make non-static, gate on flag
14443 has_bb_partition, rewrite to also check for multiple partitions.
14444 (rest_of_handle_reorder_blocks): Remove call to
14445 insert_section_boundary_note, now done later during free_cfg.
14446 (duplicate_computed_gotos): Don't duplicate partition crossing edge.
14447 * bb-reorder.h (insert_section_boundary_note): Declare.
14448 * Makefile.in (cfgrtl.o): Depend on bb-reorder.h
14449 * cfgrtl.c (rest_of_pass_free_cfg): If partitions exist
14450 invoke insert_section_boundary_note.
14451 (try_redirect_by_replacing_jump): Remove unnecessary
14452 check for region crossing note.
14453 (fixup_partition_crossing): New function.
14454 (rtl_redirect_edge_and_branch): Fixup partition boundaries.
14455 (emit_barrier_after_bb): Move here from bb-reorder.c, handle insertion
14456 in non-cfglayout mode.
14457 (force_nonfallthru_and_redirect): Fixup partition boundaries,
14458 remove old code that tried to do this. Emit barrier correctly
14459 when we are in cfglayout mode.
14460 (last_bb_in_partition): New function.
14461 (rtl_split_edge): Correctly fixup partition boundaries.
14462 (commit_one_edge_insertion): Remove old code that tried to
14463 fixup region crossing edge since this is now handled in
14464 split_block, and set up insertion point correctly since
14465 block may now end in a jump.
14466 (verify_hot_cold_block_grouping): Guard against checking when not in
14467 linearized RTL mode.
14468 (rtl_verify_edges): Add checks for incorrect/missing REG_CROSSING_JUMP
14469 notes.
14470 (rtl_verify_flow_info_1): Move verify_hot_cold_block_grouping to
14471 rtl_verify_flow_info, so not called in cfglayout mode.
14472 (rtl_verify_flow_info): Move verify_hot_cold_block_grouping here.
14473 (fixup_reorder_chain): Remove old code that attempted to fixup region
14474 crossing note as this is now handled in force_nonfallthru_and_redirect.
14475 (duplicate_insn_chain): Don't duplicate switch section notes.
14476 (rtl_can_remove_branch_p): Remove unnecessary check for region crossing
14477 note.
14478 * basic-block.h (emit_barrier_after_bb): Declare.
14479
14480 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14481
14482 * config/arm/arm-fixed.md (add<mode>3,usadd<mode>3,ssadd<mode>3,
14483 sub<mode>3, ussub<mode>3, sssub<mode>3, arm_ssatsihi_shift,
14484 arm_usatsihi): Adjust alternatives for arm_restrict_it.
14485
14486 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14487
14488 * config/arm/arm-ldmstm.ml: Set "predicable_short_it" to "no"
14489 where appropriate.
14490 * config/arm/ldmstm.md: Regenerate.
14491
14492 2013-06-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14493
14494 * config/arm/sync.md (atomic_loaddi_1):
14495 Disable predication for arm_restrict_it.
14496 (arm_load_exclusive<mode>): Likewise.
14497 (arm_load_exclusivesi): Likewise.
14498 (arm_load_exclusivedi): Likewise.
14499 (arm_load_acquire_exclusive<mode>): Likewise.
14500 (arm_load_acquire_exclusivesi): Likewise.
14501 (arm_load_acquire_exclusivedi): Likewise.
14502 (arm_store_exclusive<mode>): Likewise.
14503 (arm_store_exclusive<mode>): Likewise.
14504 (arm_store_release_exclusivedi): Likewise.
14505 (arm_store_release_exclusive<mode>): Likewise.
14506
14507 2013-06-06 Richard Biener <rguenther@suse.de>
14508
14509 * lto-streamer.h (enum LTO_tags): Move LTO_tree_pickle_reference
14510 after LTO_null.
14511 (lto_tag_is_tree_code_p): Adjust.
14512 (lto_tag_is_gimple_code_p): Likewise.
14513 (lto_gimple_code_to_tag): Likewise.
14514 (lto_tag_to_gimple_code): Likewise.
14515 (lto_tree_code_to_tag): Likewise.
14516 (lto_tag_to_tree_code): Likewise.
14517 * data-streamer.h (streamer_write_hwi_in_range): Use
14518 uhwi streaming to stream the normalized range.
14519 (streamer_read_hwi_in_range): Likewise.
14520
14521 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14522
14523 * config/arm/arm.md (enabled_for_depr_it): New attribute.
14524 (predicable_short_it): Likewise.
14525 (predicated): Likewise.
14526 (enabled): Handle above.
14527 (define_cond_exec): Set predicated attribute to yes.
14528
14529 2013-06-05 Mike Stump <mikestump@comcast.net>
14530
14531 * gdbinit.in (__FUNCTION__): Add.
14532
14533 2013-06-05 Uros Bizjak <ubizjak@gmail.com>
14534
14535 * config/alpha/alpha.c (alpha_emit_conditional_move): Swap all
14536 GE, GT, GEU and GTU compares, modulo DImode compares with zero.
14537
14538 2013-06-05 Jan Hubicka <jh@suse.cz>
14539
14540 * varasm.c (mark_decl_referenced): Revert the removal until targets
14541 are fixed.
14542
14543 2013-06-05 David Edelsohn <dje.gcc@gmail.com>
14544
14545 * config/rs6000/rs6000.c (print_operand, 'z'): Use DECL_PRESERVE_P
14546 instead of mark_decl_referenced.
14547
14548 2013-06-05 Jan Hubicka <jh@suse.cz>
14549
14550 * cgraph.c (cgraph_remove_node): Clear forced_by_abi.
14551 (cgraph_node_cannot_be_local_p_1): Honnor symbol.forced_by_abi
14552 and symtab_used_from_object_file_p.
14553 (cgraph_make_node_local_1): Clear forced_by_abi.
14554 (cgraph_can_remove_if_no_direct_calls_and): Use forced_by_abi
14555 * cgraph.h (symtab_node_base): Add forced_by_abi.
14556 (decide_is_variable_needed): Remove.
14557 (varpool_can_remove_if_no_refs): Honnor symbol.forced_by_abi.
14558 * cgraphunit.c (cgraph_decide_is_function_needed): Rename to ..
14559 (decide_is_symbol_needed): ... this one; handle symbols in general;
14560 always analyze virtuals; honnor forced_by_abi.
14561 (cgraph_finalize_function): Update.
14562 (varpool_finalize_decl): Update.
14563 (symbol_defined_and_needed): Remove.
14564 (analyze_functions): Update.
14565 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
14566 output_refs, input_overwrite_node): Handle forced_by_abi.
14567 * ipa.c (cgraph_address_taken_from_non_vtable_p): Rename to ...
14568 (address_taken_from_non_vtable_p): ... this one.
14569 (comdat_can_be_unshared_p_1): New function.
14570 (cgraph_comdat_can_be_unshared_p): Rename to ...
14571 (comdat_can_be_unshared_p): ... this one; handle symbols in general.
14572 (varpool_externally_visible_p): Use comdat_can_be_unshared_p.
14573 (function_and_variable_visibility): Clear forced_by_abi as needed.
14574 * trans-mem.c (ipa_tm_mark_forced_by_abi_node): New functoin.
14575 (ipa_tm_create_version_alias, ipa_tm_create_version): Update.
14576 * symtab.c (dump_symtab_base): Dump forced_by_abi.
14577 * varpool.c (decide_is_variable_needed): Remove.
14578
14579 2013-06-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14580
14581 * config/arm/arm.c (MAX_INSN_PER_IT_BLOCK): New macro.
14582 (arm_option_override): Override arm_restrict_it where appropriate.
14583 (thumb2_final_prescan_insn): Use MAX_INSN_PER_IT_BLOCK.
14584 * config/arm/arm.opt (mrestrict-it): New command-line option.
14585 * doc/invoke.texi: Document -mrestrict-it.
14586
14587 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14588
14589 * tsan.c (tsan_atomic_table): Make const.
14590
14591 2013-06-05 Richard Biener <rguenther@suse.de>
14592
14593 * tree-streamer.c (streamer_tree_cache_insert_1): Update the
14594 index associated with the tree we are supposed to replace.
14595 * tree-streamer-out.c (pack_ts_base_value_fields): Output
14596 TREE_ASM_WRITTEN as zero for everything but SSA names.
14597
14598 2013-06-05 David Malcolm <dmalcolm@redhat.com>
14599
14600 * tree-ssa-structalias.c (call_stmt_vars): Make static.
14601
14602 2013-06-04 Jan Hubicka <jh@suse.cz>
14603
14604 * lto-cgraph.c (get_alias_symbol): Remove weakref sanity check.
14605 (input_node, input_varpool_node): Handle correctly external same
14606 body aliases.
14607 * ipa.c (symtab_remove_unreachable_nodes): Do not remove external
14608 nodes at ltrans stage.
14609
14610 2013-06-04 Jan Hubicka <jh@suse.cz>
14611
14612 * ipa-inline.c (update_caller_keys): Fix availability test.
14613 (update_callee_keys): Likewise.
14614 * symtab.c (symtab_alias_ultimate_target): Make availaiblity logic
14615 to follow ELF standard.
14616
14617 2013-06-04 Jürgen Urban <JuergenUrban@gmx.de>
14618
14619 * config.gcc (mipsr5900-*-elf*, mipsr5900el-*-elf*, mips64r5900-*-elf*)
14620 (mips64r5900el-*-elf*): New configurations.
14621 * config/mips/mips-cpus.def (r5900): New processor.
14622 * config/mips/mips-tables.opt: Regenerate.
14623 * config/mips/mips.c (mips_rtx_cost_data): Add an R5900 entry.
14624 (mips_issue_rate): Handle PROCESSOR_R5900.
14625 (mips_reorg_process_insns): Force reorder mode for the R5900.
14626 * config/mips/mips.h (TARGET_MIPS5900): Define.
14627 (ISA_HAS_CONDMOVE, ISA_HAS_PREFETCH, ISA_HAS_HILO_INTERLOCKS): Include
14628 TARGET_MIPS5900.
14629 (ISA_HAS_LOAD_DELAY, ISA_HAS_XFER_DELAY, ISA_HAS_FCMP_DELAY): Exclude
14630 TARGET_MIPS5900.
14631 * config/mips/mips.md (processor): Add r5900.
14632 (MOVECC): Disallow CCmode conditions for TARGET_MIPS5900.
14633
14634 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14635
14636 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Call
14637 into function to generate MOVI instruction.
14638 * config/aarch64/aarch64.c (aarch64_simd_container_mode): New function.
14639 (aarch64_preferred_simd_mode): Turn into wrapper.
14640 (aarch64_output_scalar_simd_mov_immediate): New function.
14641 * config/aarch64/aarch64-protos.h: Add prototype for above.
14642
14643 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14644
14645 * config/aarch64/aarch64.c (simd_immediate_info): Remove
14646 element_char member.
14647 (sizetochar): Return signed char.
14648 (aarch64_simd_valid_immediate): Remove elchar and other
14649 unnecessary variables.
14650 (aarch64_output_simd_mov_immediate): Take rtx instead of &rtx.
14651 Calculate element_char as required.
14652 * config/aarch64/aarch64-protos.h: Update and move prototype
14653 for aarch64_output_simd_mov_immediate.
14654 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>):
14655 Update arguments.
14656
14657 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14658
14659 * config/aarch64/aarch64.c (simd_immediate_info): Struct to hold
14660 information completed by aarch64_simd_valid_immediate.
14661 (aarch64_legitimate_constant_p): Update arguments.
14662 (aarch64_simd_valid_immediate): Work with struct rather than many
14663 pointers.
14664 (aarch64_simd_scalar_immediate_valid_for_move): Update arguments.
14665 (aarch64_simd_make_constant): Update arguments.
14666 (aarch64_output_simd_mov_immediate): Work with struct rather than
14667 many pointers. Output immediate directly rather than as operand.
14668 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14669 Update prototype.
14670 * config/aarch64/constraints.md (Dn): Update arguments.
14671
14672 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14673
14674 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): No
14675 longer static.
14676 (aarch64_simd_immediate_valid_for_move): Remove.
14677 (aarch64_simd_scalar_immediate_valid_for_move): Update call.
14678 (aarch64_simd_make_constant): Update call.
14679 (aarch64_output_simd_mov_immediate): Update call.
14680 * config/aarch64/aarch64-protos.h (aarch64_simd_valid_immediate):
14681 Add prototype.
14682 * config/aarch64/constraints.md (Dn): Update call.
14683
14684 2013-06-04 Ian Bolton <ian.bolton@arm.com>
14685
14686 * config/aarch64/aarch64.c (aarch64_simd_valid_immediate): Change
14687 return type to bool for prototype.
14688 (aarch64_legitimate_constant_p): Check for true instead of not -1.
14689 (aarch64_simd_valid_immediate): Fix up each return to return a bool.
14690 (aarch64_simd_immediate_valid_for_move): Update retval for bool.
14691
14692 2013-06-04 Catherine Moore <clm@codesourcery.com>
14693
14694 * config/mips/mips.opt (meva): New.
14695 * config/mips/mips.h (TARGET_CPU_CPP_BUILTINS): Define __mips_eva.
14696 (ASM_SPEC): Handle -meva.
14697 * doc/invoke.texi (meva): Document.
14698
14699 2013-06-04 Alan Modra <amodra@gmail.com>
14700
14701 * config/rs6000/rs6000.c (output_toc): Correct little-endian float
14702 constant output.
14703
14704 2013-06-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14705
14706 * rtl.def: Add extra fourth optional field to define_cond_exec.
14707 * gensupport.c (process_one_cond_exec): Process attributes from
14708 define_cond_exec.
14709 * doc/md.texi: Document fourth field in define_cond_exec.
14710
14711 2013-06-04 Eric Botcazou <ebotcazou@adacore.com>
14712
14713 * expmed.c (extract_bit_field_1): In the larger-than-a-word case, factor
14714 out the processing order as in store_bit_field_1.
14715
14716 2013-06-04 Jan Hubicka <jh@suse.cz>
14717
14718 PR middle-end/57500
14719 * cgraphunit.c (cgraph_process_same_body_aliases): Create
14720 non-VAR_DECL node if it does not exist yet.
14721
14722 2013-06-03 Richard Sandiford <rdsandiford@googlemail.com>
14723
14724 * config.gcc (mipsisa64sr71k-*-elf*, mipsisa64sb1-*-elf*)
14725 (mipsisa64sb1el-*-elf*, mips64-*-elf*, mips64el-*-elf*)
14726 (mips64orion-*-elf*, mips64orionel-*-elf*): Remove
14727 target_cpu_default setting.
14728
14729 2013-06-03 Teresa Johnson <tejohnson@google.com>
14730
14731 * dumpfile.c (opt_info_switch_p): Change -fopt-info
14732 default to -fopt-info=optimized instead of all.
14733 * doc/invoke.texi: Ditto.
14734 * tree-vectorizer.c (vectorize_loops): Emit loop vectorization
14735 success under MSG_OPTIMIZED_LOCATIONS, and use dump_printf_loc.
14736 (execute_vect_slp): Emit BB vectorization success under
14737 MSG_OPTIMIZED_LOCATIONS.
14738 * tree-vect-slp.c (vect_slp_transform_bb): Change
14739 MSG_OPTIMIZED_LOCATIONS to MSG_NOTE.
14740 * tree-vect-loop.c (vect_transform_loop): Ditto.
14741
14742 2013-06-03 Jason Merrill <jason@redhat.com>
14743
14744 PR c++/57415
14745 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
14746 Use TARGET_EXPR for C++.
14747
14748 2013-06-03 Jakub Jelinek <jakub@redhat.com>
14749
14750 PR rtl-optimization/57268
14751 * sched-deps.c (sched_analyze_2): Don't flush_pending_lists
14752 if DEBUG_INSN_P (insn).
14753
14754 Reapply
14755 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14756
14757 PR rtl-optimization/57268
14758 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14759 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14760
14761 2013-06-03 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14762
14763 * config/i386/i386.c (ix86_lea_outperforms): Fix formatting.
14764 (ix86_avoid_lea_for_addr): Likewise.
14765 (exact_dependency_1): Likewise.
14766 (ix86_adjust_cost): Likewise.
14767 (swap_top_of_ready_list): Fix formatting and !reload_completed check
14768 removed.
14769 (do_reorder_for_imul): Fix typo, formatting and
14770 !reload_completed check removed.
14771 (ix86_sched_reorder): Fix typo and formatting.
14772 (fold_builtin_cpu): Move M_INTEL_SLM at the end of processor types
14773 list.
14774
14775 2013-06-03 Sofiane Naci <sofiane.naci@arm.com>
14776
14777 * config/aarch64/aarch64.md (*movdi_aarch64): Define "simd" attribute.
14778
14779 2013-06-03 Eric Botcazou <ebotcazou@adacore.com>
14780
14781 * varasm.c (output_constant) <CONSTRUCTOR>: Minor formatting tweak.
14782 <STRING_CST>: Likewise.
14783 <VECTOR_CST>: Likewise.
14784
14785 2013-06-01 Janus Weil <janus@gcc.gnu.org>
14786 Mikael Morin <mikael@gcc.gnu.org>
14787
14788 * configure.ac: Add AC_HEADER_TIOCGWINSZ macro.
14789 * config.in: Regenerated.
14790 * configure: Regenerated.
14791
14792 2013-06-01 Jan Hubicka <jh@suse.cz>
14793
14794 PR middle-end/57366
14795 * cgraphunit.c (compile): When weakref is not supported,
14796 set up transparent aliases before final output pass.
14797 * varasm.c (assemble_alias): Do not try to do it here.
14798
14799 2013-06-01 Jan Hubicka <jh@suse.cz>
14800
14801 PR middle-end/57467
14802 * passes.c (for_per_function): Skip unanalyzed functions.
14803
14804 2013-06-01 Jan Hubicka <jh@suse.cz>
14805
14806 * lto-symtab.c (lto_symtab_merge_cgraph_nodes_1): Rename to ...
14807 (lto_symtab_merge_symbols_1): ... this one.
14808 (lto_symtab_merge_cgraph_nodes): Rename to ...
14809 (lto_symtab_merge_symbols): ... this one; simplify.
14810 * cgraph.c (same_body_aliases_done): Rename to ...
14811 (cpp_implicit_aliases_done): ... this one.
14812 (cgraph_create_function_alias): Update.
14813 (cgraph_same_body_alias): Update.
14814 (dump_cgraph_node): Remove alias dumping; simplify thunk dumping.
14815 (verify_edge_corresponds_to_fndecl): Simplify.
14816 * cgraph.h (symtab_node_base): Add cpp_implicit_alias, alias_target.
14817 (cgraph_node): Remove same_body_alias.
14818 (varpool_node): Remove alias_of and extra_name_alias.
14819 (same_body_aliases_done): Rename to ..
14820 (cpp_implicit_aliases_done): ... this one.
14821 (symtab_alias_ultimate_target): Add default parameter.
14822 (symtab_resolve_alias): New function.
14823 (fixup_same_cpp_alias_visibility): Declare.
14824 (cgraph_function_node): Add default parameter.
14825 (cgraph_node_asm_name): Likewise.
14826 (cgraph_function_or_thunk_node): Add default parameter; do
14827 not ICE when it is NULL.
14828 (varpool_variable_node): Likewise.
14829 * tree-emutls.c (create_emultls_var): Update.
14830 (ipa_lower_emutls): Update.
14831 * cgraphunit.c (cgraph_decide_is_function_needed): Update.
14832 (cgraph_reset_node): Reset alias info.
14833 (cgraph_finalize_function): Update.
14834 (fixup_same_cpp_alias_visibility): Move to symtab.c.
14835 (analyze_function): Simplify.
14836 (cgraph_process_same_body_aliases): Simplify.
14837 (analyze_functions): Fixup same body aliases.
14838 (handle_alias_pairs): Simplify.
14839 (assemble_thunk): Update.
14840 (assemble_thunks_and_aliases): Update.
14841 (output_weakrefs): Rewrite.
14842 * lto-cgraph.c (lto_output_node): Rewrite alias handling.
14843 (lto_output_varpool_node): Likewise.
14844 (compute_ltrans_boundary): Remve assert.
14845 (get_alias_symbol): New functoin.
14846 (input_node): Rewrite alias handling.
14847 (input_varpool_node): Likewise.
14848 * ipa-pure-const.c (propagate_pure_const): Fix formating.
14849 * ipa.c (process_references): Handle weakrefs correctly.
14850 (symtab_remove_unreachable_nodes): Likewise.
14851 * trans-mem.c (get_cg_data): Update.
14852 (ipa_tm_create_version_alias): Update.
14853 (ipa_tm_execute): Update.
14854 * symtab.c (dump_symtab_base): Dump aliases.
14855 (verify_symtab_base): Verify aliases.
14856 (symtab_node_availability): New function.
14857 (symtab_alias_ultimate_target): Simplify.
14858 (fixup_same_cpp_alias_visibility): Move here from cgraphunit.c;
14859 handle all the fixup cases.
14860 (symtab_resolve_alias): New function.
14861 * passes.c (ipa_write_summaries): Handle weakrefs.
14862 * varpool.c (varpool_analyze_node): Simplify.
14863 (assemble_aliases): Update.
14864 (varpool_create_variable_alias): Simplify.
14865 (varpool_extra_name_alias): Simplify.
14866 * lto-streamer.h (lto_symtab_merge_cgraph_nodes): Rename to...
14867 (lto_symtab_merge_symbols): ... this one.
14868
14869 2013-06-01 Dinar Temirbulatov <dinar@kugelworks.com>
14870
14871 Revert
14872 PR rtl-optimization/57268
14873 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14874 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14875
14876 2013-06-01 Tobias Burnus <burnus@net-b.de>
14877
14878 Partially reverted:
14879 2013-05-31 Tobias Burnus <burnus@net-b.de>
14880
14881 PR middle-end/57073
14882 * tree-ssa-math-opts.c (execute_cse_sincos): Move check
14883 further up.
14884
14885 2013-05-31 Dinar Temirbulatov <dinar@kugelworks.com>
14886
14887 PR rtl-optimization/57268
14888 * sched-deps.c (sched_analyze_2): Flush dependence lists if
14889 the sum of the read and write lists exceeds MAX_PENDING_LIST_LENGTH.
14890
14891 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14892
14893 * config/rs6000/predicates.md (rs6000_cbranch_operator): Accept some
14894 unordered comparison operators when -fno-trapping-math is in effect
14895 on the e500.
14896 * config/rs6000/rs6000.c (rs6000_generate_compare): Remove dead code
14897 and implement unordered comparison operators properly on the e500.
14898
14899 2013-05-31 Eric Botcazou <ebotcazou@adacore.com>
14900
14901 * simplify-rtx.c (simplify_byte_swapping_operation): Use proper macro
14902 for constant scalar integers.
14903 (simplify_relational_operation_1): Likewise.
14904
14905 2013-05-31 Segher Boessenkool <segher@kernel.crashing.org>
14906
14907 * config/rs6000/rs6000-opts.h (enum processor_type): Reorder.
14908 * config/rs6000/rs6000.md (cpu): Reorder. Split long line.
14909 Fix comment.
14910
14911 2013-05-31 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
14912 Igor Zamyatin <igor.zamyatin@intel.com>
14913
14914 Silvermont (SLM) architecture performance tuning.
14915 * config/i386/i386.h (enum ix86_tune_indices): Add
14916 X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS.
14917 (TARGET_SPLIT_MEM_OPND_FOR_FP_CONVERTS): New define.
14918
14919 * config/i386/i386.c (initial_ix86_tune_features)
14920 <X86_TUNE_SPLIT_MEM_OPND_FOR_FP_CONVERTS>: Initialize.
14921 (ix86_lea_outperforms): Handle Silvermont tuning.
14922 (ix86_avoid_lea_for_add): Add new argument to ix86_lea_outperforms
14923 call.
14924 (ix86_use_lea_for_mov): Likewise.
14925 (ix86_avoid_lea_for_addr): Likewise.
14926 (ix86_lea_for_add_ok): Likewise.
14927 (exact_dependency_1): New function.
14928 (exact_store_load_dependency): Likewise.
14929 (ix86_adjust_cost): Handle Silvermont tuning.
14930 (do_reoder_for_imul): Likewise.
14931 (swap_top_of_ready_list): New function.
14932 (ix86_sched_reorder): Changed to handle Silvermont tuning.
14933
14934 * config/i386/i386.md (peepholes that split memory operand in fp
14935 converts): New.
14936
14937 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14938
14939 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately):
14940 Remove un-necessary braces.
14941
14942 2013-05-31 Marcus Shawcroft <marcus.shawcroft@arm.com>
14943
14944 * config/aarch64/aarch64.c (aarch64_classify_symbol):
14945 Use SYMBOL_TINY_ABSOLUTE for AARCH64_CMODEL_TINY_PIC.
14946
14947 2013-05-31 Tobias Burnus <burnus@net-b.de>
14948
14949 PR middle-end/57073
14950 * tree-ssa-math-opts.c (execute_cse_sincos): Move check further up.
14951
14952 2013-05-31 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
14953
14954 PR target/56315
14955 * config/arm/arm.c (const_ok_for_dimode_op): Handle IOR.
14956 * config/arm/arm.md (*iordi3_insn): Change to insn_and_split.
14957 * config/arm/neon.md (iordi3_neon): Remove.
14958 (neon_vorr<mode>): Generate iordi3 instead of iordi3_neon.
14959 * config/arm/predicates.md (imm_for_neon_logic_operand):
14960 Move to earlier in the file.
14961 (neon_logic_op2): Likewise.
14962 (arm_iordi_operand_neon): New predicate.
14963
14964 2013-05-31 Richard Biener <rguenther@suse.de>
14965
14966 PR tree-optimization/57478
14967 PR tree-optimization/57453
14968 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Uses in PHI nodes
14969 are life as well.
14970
14971 2013-05-31 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
14972
14973 * config/rl78/rl78.md (mulqi3,mulhi3): New define_expands.
14974 (*mulqi3_rl78,*mulhi3_rl78,*mulhi3_g13): New define_insns.
14975
14976 2013-05-30 Tobias Burnus <burnus@net-b.de>
14977 Thomas Koenig <tkoenig@gcc.gnu.org>
14978
14979 PR middle-end/57073
14980 * tree-ssa-math-opts.c (execute_cse_sincos): Optimize
14981 powi (-1.0, k) to (k & 1) ? -1.0 : 1.0.
14982
14983 2013-05-30 Steven Bosscher <steven@gcc.gnu.org>
14984
14985 * rtlanal.c (tablejump_p): Expect table and label to be adjacent.
14986
14987 2013-05-30 Vladimir Makarov <vmakarov@redhat.com>
14988
14989 * target.def (register_usage_leveling_p): New hook.
14990 * targhooks.c (default_register_usage_leveling_p): New.
14991 * targhooks.h (default_register_usage_leveling_p): New prototype.
14992 * lra-assigns.c (register_usage_leveling_p): Use the hook.
14993 * doc/tm.texi.in (TARGET_REGISTER_USAGE_LEVELING_P): New hook.
14994 * doc/tm.texi: Update.
14995 * config/i386/i386.c (TARGET_REGISTER_USAGE_LEVELING_P): Define.
14996
14997 2013-05-30 Ian Bolton <ian.bolton@arm.com>
14998
14999 * config/aarch64/aarch64.md (insv<mode>): New define_expand.
15000 (*insv_reg<mode>): New define_insn.
15001
15002 2013-05-30 Joern Rennecke <joern.rennecke@embecosm.com>
15003
15004 PR rtl-optimization/57439
15005 * postreload.c (move2add_valid_value_p): Check that we have
15006 a zero subreg_regno_offset when accessing the register in
15007 the requested mode.
15008
15009 2013-05-30 Yuri Rumyantsev <yuri.s.rumyantsev@intel.com>
15010 Igor Zamyatin <igor.zamyatin@intel.com>
15011
15012 Silvermont (SLM) architecture pipeline model, tuning and
15013 insn selection.
15014 * config.gcc: Add slm config options and target.
15015
15016 * config/i386/slm.md: New.
15017
15018 * config/i386/driver-i386.c (host_detect_local_cpu): Check movbe.
15019
15020 * config/i386/i386-c.c (ix86_target_macros_internal): New case
15021 PROCESSOR_SLM.
15022 (ix86_target_macros_internal): Likewise.
15023
15024 * config/i386/i386.c (slm_cost): New cost.
15025 (m_SLM): New macro flag.
15026 (initial_ix86_tune_features): Set m_SLM.
15027 (x86_accumulate_outgoing_args): Likewise.
15028 (x86_arch_always_fancy_math_387): Likewise.
15029 (processor_target_table): Add slm cost.
15030 (cpu_names): Add slm cpu name.
15031 (x86_option_override_internal): Set SLM ISA.
15032 (ix86_issue_rate): New case PROCESSOR_SLM.
15033 (ia32_multipass_dfa_lookahead): Likewise.
15034 (fold_builtin_cpu): Add slm.
15035
15036 * config/i386/i386.h (TARGET_SLM): New target macro.
15037 (target_cpu_default): Add TARGET_CPU_DEFAULT_slm.
15038 (processor_type): Add PROCESSOR_SLM.
15039
15040 * config/i386/i386.md (cpu): Add new value "slm".
15041 (slm.md): Include slm.md.
15042
15043 2013-05-30 Bernd Schmidt <bernds@codesourcery.com>
15044 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15045
15046 * config/arm/arm-protos.h: Add and update function protos.
15047 * config/arm/arm.c (use_simple_return_p): New added.
15048 (thumb2_expand_return): Check simple_return flag.
15049 * config/arm/arm.md: Add simple_return and conditional simple_return.
15050 * config/arm/iterators.md: Add iterator for return and simple_return.
15051
15052 2013-05-30 Zhenqiang Chen <zhenqiang.chen@linaro.org>
15053
15054 * config/arm/arm.c (arm_add_cfa_adjust_cfa_note): New added.
15055 (arm_emit_multi_reg_pop): Add REG_CFA_ADJUST_CFA notes.
15056 (arm_emit_vfp_multi_reg_pop): Likewise.
15057 (thumb2_emit_ldrd_pop): Likewise.
15058 (arm_expand_epilogue): Add misc REG_CFA notes.
15059 (arm_unwind_emit): Skip REG_CFA_ADJUST_CFA and REG_CFA_RESTORE.
15060
15061 2013-05-29 Lawrence Crowl <crowl@google.com>
15062
15063 * config/arm/t-arm: Update for below.
15064
15065 * config/arm/arm.c (arm_libcall_uses_aapcs_base::libcall_htab):
15066 Change type to hash_table. Update dependent calls and types.
15067
15068 * config/i386/t-cygming: Update for below.
15069
15070 * config/i386/t-interix: Update for below.
15071
15072 * config/i386/winnt.c (i386_pe_section_type_flags::htab):
15073 Change type to hash_table. Update dependent calls and types.
15074 (i386_find_on_wrapper_list::wrappers): Likewise.
15075
15076 * config/ia64/t-ia64: Update for below.
15077
15078 * config/ia64/ia64.c (bundle_state_table):
15079 Change type to hash_table. Update dependent calls and types.
15080
15081 * config/mips/mips.c (mips_reorg_process_insns::htab):
15082 Change type to hash_table. Update dependent calls and types.
15083
15084 * config/sol2.c (solaris_comdat_htab):
15085 Change type to hash_table. Update dependent calls and types.
15086
15087 * config/t-sol2: Update for above.
15088
15089 2013-05-29 Teresa Johnson <tejohnson@google.com>
15090
15091 * passes.c (dump_passes): Use FOR_EACH_FUNCTION since
15092 functions are not yet marked as defined.
15093
15094 2013-05-29 Michael Meissner <meissner@linux.vnet.ibm.com>
15095 Pat Haugen <pthaugen@us.ibm.com>
15096 Peter Bergner <bergner@vnet.ibm.com>
15097
15098 * config/rs6000/vector.md (VEC_I): Add support for new power8 V2DI
15099 instructions.
15100 (VEC_A): Likewise.
15101 (VEC_C): Likewise.
15102 (vrotl<mode>3): Likewise.
15103 (vashl<mode>3): Likewise.
15104 (vlshr<mode>3): Likewise.
15105 (vashr<mode>3): Likewise.
15106
15107 * config/rs6000/rs6000-c.c (altivec_overloaded_builtins): Add
15108 support for power8 V2DI builtins.
15109
15110 * config/rs6000/rs6000-builtin.def (abs_v2di): Add support for
15111 power8 V2DI builtins.
15112 (vupkhsw): Likewise.
15113 (vupklsw): Likewise.
15114 (vaddudm): Likewise.
15115 (vminsd): Likewise.
15116 (vmaxsd): Likewise.
15117 (vminud): Likewise.
15118 (vmaxud): Likewise.
15119 (vpkudum): Likewise.
15120 (vpksdss): Likewise.
15121 (vpkudus): Likewise.
15122 (vpksdus): Likewise.
15123 (vrld): Likewise.
15124 (vsld): Likewise.
15125 (vsrd): Likewise.
15126 (vsrad): Likewise.
15127 (vsubudm): Likewise.
15128 (vcmpequd): Likewise.
15129 (vcmpgtsd): Likewise.
15130 (vcmpgtud): Likewise.
15131 (vcmpequd_p): Likewise.
15132 (vcmpgtsd_p): Likewise.
15133 (vcmpgtud_p): Likewise.
15134 (vupkhsw): Likewise.
15135 (vupklsw): Likewise.
15136 (vaddudm): Likewise.
15137 (vmaxsd): Likewise.
15138 (vmaxud): Likewise.
15139 (vminsd): Likewise.
15140 (vminud): Likewise.
15141 (vpksdss): Likewise.
15142 (vpksdus): Likewise.
15143 (vpkudum): Likewise.
15144 (vpkudus): Likewise.
15145 (vrld): Likewise.
15146 (vsld): Likewise.
15147 (vsrad): Likewise.
15148 (vsrd): Likewise.
15149 (vsubudm): Likewise.
15150
15151 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): Add
15152 support for power8 V2DI instructions.
15153
15154 * config/rs6000/altivec.md (UNSPEC_VPKUHUM): Add support for
15155 power8 V2DI instructions. Combine pack and unpack insns to use an
15156 iterator for each mode. Check whether a particular mode supports
15157 Altivec instructions instead of just checking TARGET_ALTIVEC.
15158 (UNSPEC_VPKUWUM): Likewise.
15159 (UNSPEC_VPKSHSS): Likewise.
15160 (UNSPEC_VPKSWSS): Likewise.
15161 (UNSPEC_VPKUHUS): Likewise.
15162 (UNSPEC_VPKSHUS): Likewise.
15163 (UNSPEC_VPKUWUS): Likewise.
15164 (UNSPEC_VPKSWUS): Likewise.
15165 (UNSPEC_VPACK_SIGN_SIGN_SAT): Likewise.
15166 (UNSPEC_VPACK_SIGN_UNS_SAT): Likewise.
15167 (UNSPEC_VPACK_UNS_UNS_SAT): Likewise.
15168 (UNSPEC_VPACK_UNS_UNS_MOD): Likewise.
15169 (UNSPEC_VUPKHSB): Likewise.
15170 (UNSPEC_VUNPACK_HI_SIGN): Likewise.
15171 (UNSPEC_VUNPACK_LO_SIGN): Likewise.
15172 (UNSPEC_VUPKHSH): Likewise.
15173 (UNSPEC_VUPKLSB): Likewise.
15174 (UNSPEC_VUPKLSH): Likewise.
15175 (VI2): Likewise.
15176 (VI_char): Likewise.
15177 (VI_scalar): Likewise.
15178 (VI_unit): Likewise.
15179 (VP): Likewise.
15180 (VP_small): Likewise.
15181 (VP_small_lc): Likewise.
15182 (VU_char): Likewise.
15183 (add<mode>3): Likewise.
15184 (altivec_vaddcuw): Likewise.
15185 (altivec_vaddu<VI_char>s): Likewise.
15186 (altivec_vadds<VI_char>s): Likewise.
15187 (sub<mode>3): Likewise.
15188 (altivec_vsubcuw): Likewise.
15189 (altivec_vsubu<VI_char>s): Likewise.
15190 (altivec_vsubs<VI_char>s): Likewise.
15191 (altivec_vavgs<VI_char>): Likewise.
15192 (altivec_vcmpbfp): Likewise.
15193 (altivec_eq<mode>): Likewise.
15194 (altivec_gt<mode>): Likewise.
15195 (altivec_gtu<mode>): Likewise.
15196 (umax<mode>3): Likewise.
15197 (smax<mode>3): Likewise.
15198 (umin<mode>3): Likewise.
15199 (smin<mode>3): Likewise.
15200 (altivec_vpkuhum): Likewise.
15201 (altivec_vpkuwum): Likewise.
15202 (altivec_vpkshss): Likewise.
15203 (altivec_vpkswss): Likewise.
15204 (altivec_vpkuhus): Likewise.
15205 (altivec_vpkshus): Likewise.
15206 (altivec_vpkuwus): Likewise.
15207 (altivec_vpkswus): Likewise.
15208 (altivec_vpks<VI_char>ss): Likewise.
15209 (altivec_vpks<VI_char>us): Likewise.
15210 (altivec_vpku<VI_char>us): Likewise.
15211 (altivec_vpku<VI_char>um): Likewise.
15212 (altivec_vrl<VI_char>): Likewise.
15213 (altivec_vsl<VI_char>): Likewise.
15214 (altivec_vsr<VI_char>): Likewise.
15215 (altivec_vsra<VI_char>): Likewise.
15216 (altivec_vsldoi_<mode>): Likewise.
15217 (altivec_vupkhsb): Likewise.
15218 (altivec_vupkhs<VU_char>): Likewise.
15219 (altivec_vupkls<VU_char>): Likewise.
15220 (altivec_vupkhsh): Likewise.
15221 (altivec_vupklsb): Likewise.
15222 (altivec_vupklsh): Likewise.
15223 (altivec_vcmpequ<VI_char>_p): Likewise.
15224 (altivec_vcmpgts<VI_char>_p): Likewise.
15225 (altivec_vcmpgtu<VI_char>_p): Likewise.
15226 (abs<mode>2): Likewise.
15227 (vec_unpacks_hi_v16qi): Likewise.
15228 (vec_unpacks_hi_v8hi): Likewise.
15229 (vec_unpacks_lo_v16qi): Likewise.
15230 (vec_unpacks_hi_<VP_small_lc>): Likewise.
15231 (vec_unpacks_lo_v8hi): Likewise.
15232 (vec_unpacks_lo_<VP_small_lc>): Likewise.
15233 (vec_pack_trunc_v8h): Likewise.
15234 (vec_pack_trunc_v4si): Likewise.
15235 (vec_pack_trunc_<mode>): Likewise.
15236
15237 * config/rs6000/altivec.h (vec_vaddudm): Add defines for power8
15238 V2DI builtins.
15239 (vec_vmaxsd): Likewise.
15240 (vec_vmaxud): Likewise.
15241 (vec_vminsd): Likewise.
15242 (vec_vminud): Likewise.
15243 (vec_vpksdss): Likewise.
15244 (vec_vpksdus): Likewise.
15245 (vec_vpkudum): Likewise.
15246 (vec_vpkudus): Likewise.
15247 (vec_vrld): Likewise.
15248 (vec_vsld): Likewise.
15249 (vec_vsrad): Likewise.
15250 (vec_vsrd): Likewise.
15251 (vec_vsubudm): Likewise.
15252 (vec_vupkhsw): Likewise.
15253 (vec_vupklsw): Likewise.
15254
15255 2013-05-29 Jan Hubicka <jh@suse.cz>
15256
15257 * cgraph.h (symtab_node_base): Add definition, alias and analyzed
15258 flags; reorder rest of fields in more consistent way.
15259 (varpool_node): Remove analyzed, finalized and alias.
15260 (cgraph_ndoe): Likewise.
15261 (symtab_alias_ultimate_target): New function.
15262 (cgraph_function_node): Move offline.
15263 (cgraph_reset_node): Declare.
15264 (cgraph_comdat_can_be_unshared_p): Remove.
15265 (varpool_remove_initializer): Declare.
15266 (varpool_first_defined_variable, varpool_next_defined_variable
15267 cgraph_first_defined_function, cgraph_next_defined_function): Update.
15268 (cgraph_function_with_gimple_body_p): Update.
15269 (varpool_all_refs_explicit_p): Update.
15270 (symtab_alias_target): New function.
15271 (cgraph_alias_aliased_node, varpool_alias_aliased_node): Rename to ...
15272 (cgraph_alias_target, varpool_alias_target): .. this one; simplify.
15273 (cgraph_function_or_thunk_node): Simplify using
15274 symtab_alias_ultimate_target.
15275 (varpool_variable_node): Likewise.
15276 * cgraph.c (cgraph_create_function_alias): Update.
15277 (cgraph_add_thunk): Update.
15278 (cgraph_remove_node): Update.
15279 (dump_cgraph_node): Do not dump removed flags.
15280 (cgraph_function_body_availability): Update.
15281 (cgraph_propagate_frequency): Update.
15282 (verify_cgraph_node): Check sanity of local flag.
15283 (cgraph_function_node): Move here from cgraph.h; revamp for
15284 cgraph_function_or_thunk_node.
15285 * lto-symtab.c (lto_varpool_replace_node): Update.
15286 (lto_symtab_resolve_can_prevail_p): Update.
15287 (lto_symtab_merge_cgraph_nodes): Update.
15288 * ipa-cp.c (determine_versionability, initialize_node_lattices,
15289 propagate_constants_accross_call, devirtualization_time_bonus,
15290 ipcp_propagate_stage): Update.
15291 * tree-emutls.c (create_emultls_var, ipa_lower_emutls): Update.
15292 * ipa-inline-transform.c (clone_inlined_nodes,
15293 preserve_function_body_p): Update.
15294 * ipa-reference.c (propagate): Update.
15295 (write_node_summary_p): Update.
15296 * toplev.c (wrapup_global_declaration_2): Update.
15297 * cgraphunit.c (cgraph_analyze_function): Rename to ...
15298 (analyze_function) ... this one.
15299 (cgraph_process_new_functions): Update.
15300 (cgraph_reset_node): Export.
15301 (cgraph_finalize_function): Update.
15302 (cgraph_add_new_function): Update.
15303 (process_function_and_variable_attributes): Update.
15304 (varpool_finalize_decl): Update.
15305 (symbol_finalized): Remove.
15306 (symbol_finalized_and_needed): Rename to ...
15307 (symbol_defined_and_needed): ... update.
15308 (cgraph_analyze_functions): Update.
15309 (handle_alias_pairs): Update.
15310 (mark_functions_to_output): Update.
15311 (assemble_thunk): Update.
15312 (output_in_order): Update.
15313 (output_weakrefs): Update.
15314 (finalize_compilation_unit): Update.
15315 * lto-cgraph.c (reachable_from_other_partition_p, lto_output_node,
15316 lto_output_varpool_node, compute_ltrans_boundary, input_overwrite_node,
15317 input_node, input_varpool_node): Update.
15318 * dbxout.c (dbxout_expand_expr): Update.
15319 * cgraphclones.c (cgraph_clone_node): Update.
15320 (cgraph_copy_node_for_versioning): Update.
15321 (cgraph_materialize_clone): Update.
15322 (cgraph_materialize_all_clones): Update.
15323 * ipa-pure-const.c (analyze_function, pure_const_write_summary,
15324 propagate_pure_const, propagate_nothrow): Update.
15325 * lto-streamer-out.c (lto_output, write_symbol): Update.
15326 * ipa-utils.c (ipa_reverse_postorder): Update.
15327 * ipa-inline.c (can_inline_edge_p): Update.
15328 (update_caller_keys, ipa_inline): Update.
15329 * dwarf2out.c (reference_to_unused,
15330 premark_types_used_by_global_vars_helper): Update.
15331 * tree-eh.c (tree_could_trap_p): Update.
15332 * ipa-split.c (consider_split, execute_split_functions): Update.
15333 * ipa.c (cgraph_non_local_node_p_1, cgraph_local_node_p,
15334 has_addr_references_p): Update; move ahead in file for better
15335 readability.
15336 (process_references): Simplify.
15337 (symtab_remove_unreachable_nodes): Update; cleanup way function/var
15338 bodies are removed.
15339 (cgraph_comdat_can_be_unshared_p): Make static.
15340 (cgraph_externally_visible_p): Update.
15341 (varpool_externally_visible_p): Update.
15342 (function_and_variable_visibility): Update.
15343 * trans-mem.c (get_cg_data, ipa_tm_mayenterirr_function,
15344 ipa_tm_mark_force_output_node): Update.
15345 * ipa-inline-analysis.c (dump_inline_summary, initialize_inline_failed,
15346 estimate_edge_devirt_benefit, inline_generate_summary,
15347 inline_write_summary): Update.
15348 * gimple-fold.c (can_refer_decl_in_current_unit_p): Update.
15349 * ipa-prop.c (ipa_compute_jump_functions): Update.
15350 (ipa_print_node_params, ipa_prop_read_section,
15351 ipa_update_after_lto_read, read_replacements_section): Update.
15352 * varasm.c (mark_decl_referenced): Update.
15353 (assemble_alias, dump_tm_clone_pairs): Update.
15354 * tree-inline.c (copy_bb): Update.
15355 (estimate_num_insns, optimize_inline_calls, tree_function_versioning):
15356 Update.
15357 * symtab.c (dump_symtab_base): Print new flags.
15358 (verify_symtab_base): Verify new flags.
15359 (symtab_alias_ultimate_target): New function.
15360 * tree-ssa-structalias.c (get_constraint_for_ssa_var,
15361 create_variable_info_for, associate_varinfo_to_alias, ipa_pta_execute):
15362 Update.
15363 * passes.c (ipa_write_summaries, ipa_write_optimization_summaries):
15364 Update.
15365 * i386.c (ix86_get_function_versions_dispatcher,
15366 ix86_generate_version_dispatcher_body): Update.
15367 (fold_builtin_cpu): Use varpool_add_new_variable.
15368 * varpool.c (varpool_remove_initializer): Break out from ...
15369 (varpool_remove_node): ... this one.
15370 (dump_varpool_node, varpool_node_for_asm,
15371 cgraph_variable_initializer_availability, varpool_analyze_node,
15372 varpool_assemble_decl, varpool_remove_unreferenced_decls,
15373 varpool_finalize_named_section_flags, varpool_create_variable_alias):
15374 Update.
15375
15376 2013-05-29 Jan Hubicka <jh@suse.cz>
15377
15378 * passes.c (init_optimization_passes): Move OMP expansion into lowering.
15379
15380 2013-05-29 Easwaran Raman <eraman@google.com>
15381
15382 PR tree-optimization/57442
15383 * tree-ssa-reassoc.c (appears_later_in_bb): Return correct value
15384 when control exits the main loop.
15385
15386 2013-05-29 Sandeep Kumar Singh <Sandeep.Singh2@kpitcummins.com>
15387
15388 * rx/rx.h (TARGET_CPU_CPP_BUILTINS): Add macros for RX100, RX200,
15389 and RX600.
15390 * rx/rx.opt: Add macro for rx100 with string rx100 and value RX100.
15391 * rx/rx-opts.h (rx_cpu_types): Add new cpu type rx100.
15392 * rx/t-rx: Add rx100 under multi library matches option for nofpu
15393 option.
15394
15395 2013-05-29 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15396
15397 PR tree-optimization/57441
15398 * gimple-ssa-strength-reduction.c (analyze_candidates_and_replace):
15399 Don't limit size of incr_vec to number of candidates.
15400
15401 2013-05-29 Steve Ellcey <sellcey@imgtec.com>
15402
15403 * config/mips/mti-linux.h (SYSROOT_SUFFIX_SPEC): Add micromips
15404 and mips16 directories.
15405 * config/mips/t-mti-linux (MULTILIB_OPTIONS): Add micromips and mips16.
15406 (MULTILIB_DIRNAMES): Ditto.
15407 (MULTILIB_EXCEPTIONS): Add new exceptions.
15408 * config/mips/t-mti-elf (MULTILIB_OPTIONS): Add micromips.
15409 (MULTILIB_DIRNAMES): Ditto.
15410 (MULTILIB_EXCEPTIONS): Add new exceptions.
15411
15412 2012-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15413 Marcus Shawcroft <marcus.shawcroft@arm.com>
15414
15415 * config/aarch64/aarch64-protos.h (aarch64_symbol_type): Define
15416 SYMBOL_TINY_ABSOLUTE.
15417 * config/aarch64/aarch64.c (aarch64_load_symref_appropriately): Handle
15418 SYMBOL_TINY_ABSOLUTE.
15419 (aarch64_expand_mov_immediate): Likewise.
15420 (aarch64_classify_symbol): Likewise.
15421 (aarch64_mov_operand_p): Remove ATTRIBUTE_UNUSED.
15422 Permit SYMBOL_TINY_ABSOLUTE.
15423 * config/aarch64/predicates.md (aarch64_mov_operand): Permit CONST.
15424
15425 2013-05-29 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15426 Marcus Shawcroft <marcus.shawcroft@arm.com>
15427
15428 * config/aarch64/aarch64.c (aarch64_classify_symbol): Remove comment.
15429 Refactor if/switch. Replace gcc_assert with if.
15430
15431 2013-05-29 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
15432
15433 * config/i386/i386.c (initial_ix86_tune_features): Enable
15434 FP Reassociation for AMD bdver1 and bdver2.
15435
15436 2013-05-29 Martin Jambor <mjambor@suse.cz>
15437
15438 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REF, REALPART_EXPR
15439 and IMAGPART_EXPR do not occur within other handled_components.
15440
15441 2013-05-29 Richard Biener <rguenther@suse.de>
15442
15443 * tree-vect-slp.c (vect_bb_slp_scalar_cost): Guard vinfo
15444 access on whether the use is in the BB we currently try to
15445 vectorize.
15446 (vect_bb_vectorization_profitable_p): Pass the BB we currently
15447 vectorize to vect_bb_slp_scalar_cost.
15448
15449 2013-05-29 Richard Biener <rguenther@suse.de>
15450
15451 * tree-vect-slp.c (vect_bb_slp_scalar_cost): New function
15452 computing scalar cost offsetted by stmts that are kept live
15453 by scalar uses.
15454 (vect_bb_vectorization_profitable_p): Use vect_bb_slp_scalar_cost
15455 for computation of scalar cost.
15456
15457 2013-05-28 Steve Ellcey <sellcey@mips.com>
15458
15459 * config/mips/mips-cpus.def (mips32r2): Change processor type.
15460
15461 2013-05-28 Balaji V. Iyer <balaji.v.iyer@intel.com>
15462
15463 * doc/extend.texi (C Extensions): Added documentation about Cilk Plus
15464 array notation built-in reduction functions.
15465 * doc/passes.texi (Passes): Added documentation about changes done
15466 for Cilk Plus.
15467 * doc/invoke.texi (C Dialect Options): Added documentation about
15468 the -fcilkplus flag.
15469 * Makefile.in (C_COMMON_OBJS): Added c-family/array-notation-common.o.
15470 (BUILTINS_DEF): Depend on cilkplus.def.
15471 * builtins.def: Include cilkplus.def. Define DEF_CILKPLUS_BUILTIN.
15472 * builtin-types.def: Define BT_FN_INT_PTR_PTR_PTR.
15473 * cilkplus.def: New file.
15474
15475 2013-05-28 Joern Rennecke <joern.rennecke@embecosm.com>
15476
15477 PR rtl-optimization/57439
15478 * postreload.c (move2add_use_add2_insn): Use gen_lowpart_common.
15479
15480 2013-05-28 Easwaran Raman <eraman@google.com>
15481
15482 PR tree-optimization/57337
15483 * tree-ssa-reassoc.c (appears_later_in_bb): New function.
15484 (find_insert_point): Correctly identify the insertion point
15485 when two statements with the same UID is compared.
15486
15487 2013-05-28 Richard Biener <rguenther@suse.de>
15488
15489 PR tree-optimization/56787
15490 * tree-vect-data-refs.c (vect_analyze_data_refs): Drop clobbers
15491 from the list of data references.
15492 * tree-vect-loop.c (vect_determine_vectorization_factor): Skip
15493 clobbers.
15494 (vect_analyze_loop_operations): Likewise.
15495 (vect_transform_loop): Remove clobbers.
15496
15497 2013-05-28 Martin Jambor <mjambor@suse.cz>
15498
15499 * tree-cfg.c (verify_expr): Verify that BIT_FIELD_REFs, IMAGPART_EXPRs
15500 and REALPART_EXPRs have scalar type.
15501
15502 2013-05-28 Richard Biener <rguenther@suse.de>
15503
15504 PR tree-optimization/57411
15505 * tree-ssa-copy.c (may_propagate_copy): Cannot propagate
15506 virtual operands.
15507 * tree-ssa-dom.c (eliminate_const_or_copy): Special-case
15508 virtual operand propagation.
15509
15510 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15511
15512 * config/sparc/sparc.c (sparc_expand_vec_perm_bmask): Use %g0 as
15513 destination register for bmasksi_vis.
15514 (vector_init_bshuffle): Likewise.
15515 * config/sparc/sparc.md (vec_perm_constv8qi): Likewise.
15516
15517 2013-05-28 Eric Botcazou <ebotcazou@adacore.com>
15518
15519 * doc/invoke.texi (SPARC Options): Document -mfix-ut699.
15520 * builtins.c (expand_builtin_mathfn) <BUILT_IN_SQRT>: Try to widen the
15521 mode if the instruction isn't available in the original mode.
15522 * config/sparc/sparc.opt (mfix-ut699): New option.
15523 * config/sparc/sparc.md (muldf3_extend): Disable if -mfix-ut699.
15524 (divdf3): Turn into expander.
15525 (divdf3_nofix): New insn.
15526 (divdf3_fix): Likewise.
15527 (divsf3): Disable if -mfix-ut699.
15528 (sqrtdf2): Turn into expander.
15529 (sqrtdf2_nofix): New insn.
15530 (sqrtdf2_fix): Likewise.
15531 (sqrtsf2): Disable if -mfix-ut699.
15532
15533 2013-05-27 Richard Biener <rguenther@suse.de>
15534
15535 PR middle-end/57412
15536 * omp-low.c (expand_omp_atomic_pipeline): Use the correct latch
15537 block for the new loop.
15538
15539 2013-05-27 Richard Biener <rguenther@suse.de>
15540
15541 PR tree-optimization/57343
15542 * tree-ssa-loop-niter.c (number_of_iterations_ne_max): Do not
15543 use multiple_of_p if not TYPE_OVERFLOW_UNDEFINED.
15544 (number_of_iterations_cond): Do not build the folded tree.
15545
15546 2013-05-27 Richard Biener <rguenther@suse.de>
15547
15548 Revert
15549 PR middle-end/57381
15550 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15551 OEP_CONSTANT_ADDRESS_OF retained.
15552
15553 PR tree-optimization/57417
15554 * tree-ssa-sccvn.c (vn_reference_fold_indirect): Fix test
15555 for unchanged base.
15556 (set_ssa_val_to): Compare addresses using
15557 get_addr_base_and_unit_offset.
15558
15559 2013-05-27 Joern Rennecke <joern.rennecke@embecosm.com>
15560
15561 PR rtl-optimization/56833
15562 * postreload.c (move2add_record_mode): New function.
15563 (move2add_record_sym_value, move2add_valid_value_p): Likewise.
15564 (move2add_use_add2_insn): Use move2add_record_sym_value.
15565 (move2add_use_add3_insn): Likewise.
15566 (reload_cse_move2add): Use move2add_valid_value_p and
15567 move2add_record_mode. Invalidate call-clobbered and REG_INC
15568 affected regs by setting reg_mode to VOIDmode.
15569 (move2add_note_store): Don't pretend the inside of a SUBREG is
15570 the actual destination. Invalidate single/leading registers by
15571 setting reg_mode to VOIDmode.
15572 Use move2add_record_sym_value, move2add_valid_value_p and
15573 move2add_record_mode.
15574
15575 2013-05-27 Richard Biener <rguenther@suse.de>
15576
15577 PR tree-optimization/57396
15578 * tree-affine.c (double_int_constant_multiple_p): Properly
15579 return false for val == 0 and div != 0.
15580
15581 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15582
15583 * config/mips/mips.h: Use #elif in preprocessor conditions.
15584
15585 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15586
15587 PR target/53916
15588 * config/mips/constraints.md (kl): New constraint.
15589 * config/mips/mips.md (divmod<mode>4, udivmod<mode>4): Delete.
15590 (divmod<mode>4_internal): Rename to divmod<mode>4. Use "kl" as the
15591 constraint for operand 0. Split after CSE for MIPS16. Emit a move
15592 from LO for MIPS16.
15593 (udivmod<mode>4_internal): Likewise udivmod<mode>4.
15594
15595 2013-05-25 Richard Sandiford <rdsandiford@googlemail.com>
15596
15597 PR target/55777
15598 * config/mips/mips.c (mips_can_inline_p): New function.
15599 (TARGET_CAN_INLINE_P): Define.
15600
15601 2013-05-25 Steven Bosscher <steven@gcc.gnu.org>
15602
15603 * sched-int.h (ds_t, dw_t): Make unsigned int.
15604 Fix documentation that describes how all the ds_t bits are used.
15605 Reserve the last bit for delayed-branch scheduling.
15606 (BITS_PER_DEP_STATUS): Move to ds_t typedef.
15607 (BITS_PER_DEP_WEAK): Fix definition and documentation.
15608 (gen_dep_weak_1): Remove prototype.
15609 * sched-deps.c (get_dep_weak_1): Make static.
15610 * target.def (speculate_insn, needs_block_p, gen_spec_check,
15611 get_insn_spec_ds, get_insn_checked_ds): Adjust hook prototypes.
15612 * doc/tm.texi: Regenerate.
15613 * config/ia64/ia64.c (ia64_needs_block_p): Update prototype.
15614
15615 2013-05-24 Steven Bosscher <steven@gcc.gnu.org>
15616
15617 PR debug/56950
15618 * haifa-sched.c (sched_extend_bb): Ignore DEBUG_INSNs.
15619
15620 2013-05-24 Nathan Sidwell <nathan@codesourcery.com>
15621 Sandra Loosemore <sandra@codesourcery.com>
15622
15623 * config.gcc (powerpc-*): Allow native for with-cpu.
15624
15625 2013-05-24 Jeff Law <law@redhat.com>
15626
15627 PR tree-optimization/57124
15628 * tree-vrp.c (simplify_cond_using_ranges): Only simplify a
15629 conversion feeding a condition if the range has an overflow
15630 if -fstrict-overflow. Add warnings for when we do make the
15631 transformation.
15632
15633 2013-05-24 Dehao Chen <dehao@google.com>
15634
15635 * tree-cfg.c (locus_discrim_map): Fix the typo.
15636 (locus_discrim_hasher): Likewise.
15637 (locus_discrim_hasher::hash): Likewise.
15638 (locus_discrim_hasher::equal): Likewise.
15639
15640 2013-05-24 Martin Jambor <mjambor@suse.cz>
15641
15642 PR tree-optimization/57294
15643 * cgraph.h (ipa_record_stmt_references): Declare.
15644 * cgraphbuild.c (ipa_record_stmt_references): New function.
15645 (build_cgraph_edges): Use ipa_record_stmt_references.
15646 (rebuild_cgraph_edges): Likewise.
15647 (cgraph_rebuild_references): Likewise.
15648 * ipa-prop.c (ipa_modify_call_arguments): Discard references
15649 associated with the old statement and build references from the
15650 newly built statements.
15651 * ipa-ref.c (ipa_remove_stmt_references): New function.
15652 * ipa-ref.h (ipa_remove_stmt_references): Declare.
15653
15654 2013-05-24 Vladimir Makarov <vmakarov@redhat.com>
15655
15656 * lra-constraints.c (emit_spill_move): Use smaller mode for
15657 mem-mem moves.
15658 (check_and_process_move): Consider mem-reg moves for secondary
15659 too.
15660 (curr_insn_transform): Don't lose insns emitted before for
15661 secondary memory moves.
15662 (inherit_in_ebb): Mark defined reg. Add usage only if it is not a
15663 reg set up in the current insn.
15664
15665 2013-05-24 Dehao Chen <dehao@google.com>
15666
15667 * tree-cfg.c (locus_descrim_hasher::hash): Change discriminator
15668 hash function.
15669 (locus_descrim_hasher::equal): Likewise.
15670 (build_gimple_cfg): New discriminator assignment algorithm.
15671 (make_edges): Likewise.
15672 (next_discriminator_for_locus): Likewise.
15673 (same_line_p): Likewise.
15674 (assign_discriminators): Likewise.
15675 (make_cond_expr_edges): Likewise.
15676 (make_gimple_switch_edges): Likewise.
15677 (make_goto_expr_edges): Likewise.
15678 (make_gimple_asm_edges): Likewise.
15679
15680 2013-05-24 Ian Bolton <ian.bolton@arm.com>
15681
15682 * config/aarch64/aarch64.c (aarch64_print_operand): Change the
15683 X format specifier to only display bottom 16 bits.
15684 * config/aarch64/aarch64.md (insv_imm<mode>): Allow any size of
15685 immediate to match for operand 2, since it will be masked.
15686
15687 2013-05-24 Richard Biener <rguenther@suse.de>
15688
15689 PR tree-optimization/57287
15690 * tree-ssa-uninit.c (compute_uninit_opnds_pos): Disregard
15691 all SSA names that occur in abnormal PHIs.
15692
15693 2013-05-24 Alexander Ivchenko <alexander.ivchenko@intel.com>
15694
15695 PR tree-ssa/57385
15696 * tree-ssa-sccvn.c (fully_constant_vn_reference_p): Check
15697 that index is not negative.
15698
15699 2013-05-24 Eric Botcazou <ebotcazou@adacore.com>
15700
15701 PR rtl-optimization/55177
15702 * simplify-rtx.c (simplify_unary_operation_1) <NOT>: Deal with BSWAP.
15703 (simplify_byte_swapping_operation): New.
15704 (simplify_binary_operation_1): Call it for AND, IOR and XOR.
15705 (simplify_relational_operation_1): Deal with BSWAP.
15706
15707 2013-05-23 Richard Henderson <rth@redhat.com>
15708
15709 PR target/56742
15710 * config/i386/i386.c (ix86_seh_fixup_eh_fallthru): New.
15711 (ix86_reorg): Call it.
15712
15713 2013-05-23 Uros Bizjak <ubizjak@gmail.com>
15714
15715 PR target/57379
15716 * config/alpha/alpha.md (unspec): Add UNSPEC_XFLT_COMPARE.
15717 * config/alpha/alpha.c (alpha_emit_xfloating_compare): Construct
15718 REG_EQUAL note as UNSPEC_XFLT_COMPARE unspec.
15719
15720 2013-05-23 Christian Bruel <christian.bruel@st.com>
15721
15722 PR debug/57351
15723 * config/arm/arm.c (arm_dwarf_register_span): Do not use dbx number.
15724
15725 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15726 Marcus Shawcroft <marcus.shawcroft@arm.com>
15727
15728 * config/aarch64/aarch64.md (*movdi_aarch64): Replace Usa with S.
15729 * config/aarch64/constraints.md (Usa): Remove.
15730 * doc/md.texi (AArch64 Usa): Remove.
15731
15732 2013-05-23 Chris Schlumberger-Socha <chris.schlumberger-socha@arm.com>
15733 Marcus Shawcroft <marcus.shawcroft@arm.com>
15734
15735 * config/aarch64/aarch64-protos.h (aarch64_mov_operand_p): Define.
15736 * config/aarch64/aarch64.c (aarch64_mov_operand_p): Define.
15737 * config/aarch64/predicates.md (aarch64_const_address): Remove.
15738 (aarch64_mov_operand): Use aarch64_mov_operand_p.
15739
15740 2013-05-23 Vidya Praveen <vidyapraveen@arm.com>
15741
15742 * config/aarch64/aarch64-simd.md (clzv4si2): Support for CLZ
15743 instruction (AdvSIMD).
15744 * config/aarch64/aarch64-builtins.c
15745 (aarch64_builtin_vectorized_function): Handler for BUILT_IN_CLZ.
15746 * config/aarch64/aarch-simd-builtins.def: Entry for CLZ.
15747
15748 2013-05-23 Martin Jambor <mjambor@suse.cz>
15749
15750 PR middle-end/57347
15751 * tree.h (contains_bitfld_component_ref_p): Declare.
15752 * tree-sra.c (contains_bitfld_comp_ref_p): Move...
15753 * tree.c (contains_bitfld_component_ref_p): ...here. Adjust its
15754 caller.
15755 * ipa-prop.c (determine_known_aggregate_parts): Check that LHS does
15756 not access a bit-field. Assert all final offsets are byte-aligned.
15757
15758 2013-05-23 Richard Biener <rguenther@suse.de>
15759
15760 PR tree-optimization/57380
15761 * tree-ssa-phiprop.c (propagate_with_phi): Do not require at
15762 least one invariant or re-used load.
15763 * passes.c (init_optimization_passes): Move pass_phiprop before
15764 pass_forwprop.
15765
15766 2013-05-23 James Greenhalgh <james.greenhalgh@arm.com>
15767
15768 * config/aarch64/aarch64-simd.md
15769 (aarch64_cm<optab>di): Add clobber of CC_REGNUM to unsplit pattern.
15770
15771 2013-05-23 Richard Biener <rguenther@suse.de>
15772
15773 PR middle-end/57381
15774 * fold-const.c (operand_equal_p): Compare FIELD_DECLs with
15775 OEP_CONSTANT_ADDRESS_OF retained.
15776
15777 2013-05-23 Jakub Jelinek <jakub@redhat.com>
15778
15779 PR middle-end/57344
15780 * expmed.c (store_split_bit_field): If op0 is a REG or SUBREG of a REG,
15781 don't lower unit. Handle unit not being always BITS_PER_WORD.
15782
15783 2013-05-23 Richard Biener <rguenther@suse.de>
15784
15785 PR rtl-optimization/57341
15786 * ira.c (validate_equiv_mem_from_store): Use anti_dependence
15787 instead of true_dependence.
15788
15789 2013-05-22 David Malcolm <dmalcolm@redhat.com>
15790
15791 * bb-reorder.c (branch_threshold): Make const.
15792 (exec_threshold): Ditto.
15793
15794 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
15795 Pat Haugen <pthaugen@us.ibm.com>
15796 Peter Bergner <bergner@vnet.ibm.com>
15797
15798 * doc/extend.texi (PowerPC AltiVec/VSX Built-in Functions): Add
15799 documentation for the power8 crypto builtins.
15800
15801 * config/rs6000/t-rs6000 (MD_INCLUDES): Add crypto.md.
15802
15803 * config/rs6000/rs6000-builtin.def (BU_P8V_AV_1): Add support
15804 macros for defining power8 builtin functions.
15805 (BU_P8V_AV_2): Likewise.
15806 (BU_P8V_AV_P): Likewise.
15807 (BU_P8V_VSX_1): Likewise.
15808 (BU_P8V_OVERLOAD_1): Likewise.
15809 (BU_P8V_OVERLOAD_2): Likewise.
15810 (BU_CRYPTO_1): Likewise.
15811 (BU_CRYPTO_2): Likewise.
15812 (BU_CRYPTO_3): Likewise.
15813 (BU_CRYPTO_OVERLOAD_1): Likewise.
15814 (BU_CRYPTO_OVERLOAD_2): Likewise.
15815 (XSCVSPDP): Fix typo, point to the correct instruction.
15816 (VCIPHER): Add power8 crypto builtins.
15817 (VCIPHERLAST): Likewise.
15818 (VNCIPHER): Likewise.
15819 (VNCIPHERLAST): Likewise.
15820 (VPMSUMB): Likewise.
15821 (VPMSUMH): Likewise.
15822 (VPMSUMW): Likewise.
15823 (VPERMXOR_V2DI): Likewise.
15824 (VPERMXOR_V4SI: Likewise.
15825 (VPERMXOR_V8HI: Likewise.
15826 (VPERMXOR_V16QI: Likewise.
15827 (VSHASIGMAW): Likewise.
15828 (VSHASIGMAD): Likewise.
15829 (VPMSUM): Likewise.
15830 (VPERMXOR): Likewise.
15831 (VSHASIGMA): Likewise.
15832
15833 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros): Define
15834 __CRYPTO__ if the crypto instructions are available.
15835 (altivec_overloaded_builtins): Add support for overloaded power8
15836 builtins.
15837
15838 * config/rs6000/rs6000.c (rs6000_expand_ternop_builtin): Add
15839 support for power8 crypto builtins.
15840 (builtin_function_type): Likewise.
15841 (altivec_init_builtins): Add support for builtins that take vector
15842 long long (V2DI) arguments.
15843
15844 * config/rs6000/crypto.md: New file, define power8 crypto
15845 instructions.
15846
15847 2013-05-22 Michael Meissner <meissner@linux.vnet.ibm.com>
15848 Pat Haugen <pthaugen@us.ibm.com>
15849 Peter Bergner <bergner@vnet.ibm.com>
15850
15851 * doc/invoke.texi (Option Summary): Add power8 options.
15852 (RS/6000 and PowerPC Options): Likewise.
15853
15854 * doc/md.texi (PowerPC and IBM RS6000 constraints): Update to use
15855 constraints.md instead of rs6000.h. Reorder w* constraints. Add
15856 wm, wn, wr documentation.
15857
15858 * config/rs6000/constraints.md (wm): New constraint for VSX
15859 registers if direct move instructions are enabled.
15860 (wn): New constraint for no registers.
15861 (wq): New constraint for quad word even GPR registers.
15862 (wr): New constraint if 64-bit instructions are enabled.
15863 (wv): New constraint if power8 vector instructions are enabled.
15864 (wQ): New constraint for quad word memory locations.
15865
15866 * config/rs6000/predicates.md (const_0_to_15_operand): New
15867 constraint for 0..15 for crypto instructions.
15868 (gpc_reg_operand): If VSX allow registers in VSX registers as well
15869 as GPR and floating point registers.
15870 (int_reg_operand): New predicate to match only GPR registers.
15871 (base_reg_operand): New predicate to match base registers.
15872 (quad_int_reg_operand): New predicate to match even GPR registers
15873 for quad memory operations.
15874 (vsx_reg_or_cint_operand): New predicate to allow vector logical
15875 operations in both GPR and VSX registers.
15876 (quad_memory_operand): New predicate for quad memory operations.
15877 (reg_or_indexed_operand): New predicate for direct move support.
15878
15879 * config/rs6000/rs6000-cpus.def (ISA_2_5_MASKS_EMBEDDED):
15880 Inherit from ISA_2_4_MASKS, not ISA_2_2_MASKS.
15881 (ISA_2_7_MASKS_SERVER): New mask for ISA 2.07 (i.e. power8).
15882 (POWERPC_MASKS): Add power8 options.
15883 (power8 cpu): Use ISA_2_7_MASKS_SERVER instead of specifying the
15884 various options.
15885
15886 * config/rs6000/rs6000-c.c (rs6000_target_modify_macros):
15887 Define _ARCH_PWR8 and __POWER8_VECTOR__ for power8.
15888
15889 * config/rs6000/rs6000.opt (-mvsx-timode): Add documentation.
15890 (-mpower8-fusion): New power8 options.
15891 (-mpower8-fusion-sign): Likewise.
15892 (-mpower8-vector): Likewise.
15893 (-mcrypto): Likewise.
15894 (-mdirect-move): Likewise.
15895 (-mquad-memory): Likewise.
15896
15897 * config/rs6000/rs6000.c (power8_cost): Initial definition for power8.
15898 (rs6000_hard_regno_mode_ok): Make PTImode only match even GPR
15899 registers.
15900 (rs6000_debug_reg_print): Print the base register class if -mdebug=reg.
15901 (rs6000_debug_vector_unit): Add p8_vector.
15902 (rs6000_debug_reg_global): If -mdebug=reg, print power8 constraint
15903 definitions. Also print fusion state.
15904 (rs6000_init_hard_regno_mode_ok): Set up power8 constraints.
15905 (rs6000_builtin_mask_calculate): Add power8 builtin support.
15906 (rs6000_option_override_internal): Add support for power8.
15907 (rs6000_common_init_builtins): Add debugging for skipped builtins
15908 if -mdebug=builtin.
15909 (rs6000_adjust_cost): Add power8 support.
15910 (rs6000_issue_rate): Likewise.
15911 (insn_must_be_first_in_group): Likewise.
15912 (insn_must_be_last_in_group): Likewise.
15913 (force_new_group): Likewise.
15914 (rs6000_register_move_cost): Likewise.
15915 (rs6000_opt_masks): Likewise.
15916
15917 * config/rs6000/rs6000.h (ASM_CPU_POWER8_SPEC): If we don't have a
15918 power8 capable assembler, default to power7 options.
15919 (TARGET_DIRECT_MOVE): Likewise.
15920 (TARGET_CRYPTO): Likewise.
15921 (TARGET_P8_VECTOR): Likewise.
15922 (VECTOR_UNIT_P8_VECTOR_P): Define power8 vector support.
15923 (VECTOR_UNIT_VSX_OR_P8_VECTOR_P): Likewise.
15924 (VECTOR_MEM_P8_VECTOR_P): Likewise.
15925 (VECTOR_MEM_VSX_OR_P8_VECTOR_P): Likewise.
15926 (VECTOR_MEM_ALTIVEC_OR_VSX_P): Likewise.
15927 (TARGET_XSCVDPSPN): Likewise.
15928 (TARGET_XSCVSPDPN): Likewsie.
15929 (TARGET_SYNC_HI_QI): Likewise.
15930 (TARGET_SYNC_TI): Likewise.
15931 (MASK_CRYPTO): Likewise.
15932 (MASK_DIRECT_MOVE): Likewise.
15933 (MASK_P8_FUSION): Likewise.
15934 (MASK_P8_VECTOR): Likewise.
15935 (REG_ALLOC_ORDER): Move fr13 to be lower in priority so that the TFmode
15936 temporary used by some of the direct move instructions to get two FP
15937 temporary registers does not force creation of a stack frame.
15938 (VLOGICAL_REGNO_P): Allow vector logical operations in GPRs.
15939 (MODES_TIEABLE_P): Move the VSX tests above the Altivec tests so
15940 that any VSX registers are tieable, even if they are also an
15941 Altivec vector mode.
15942 (r6000_reg_class_enum): Add wm, wr, wv constraints.
15943 (RS6000_BTM_P8_VECTOR): Power8 builtin support.
15944 (RS6000_BTM_CRYPTO): Likewise.
15945 (RS6000_BTM_COMMON): Likewise.
15946
15947 * config/rs6000/rs6000.md (cpu attribute): Add power8.
15948 * config/rs6000/rs6000-opts.h (PROCESSOR_POWER8): Likewise.
15949 (enum rs6000_vector): Add power8 vector support.
15950
15951 2013-05-22 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
15952
15953 PR target/19599
15954 PR target/57340
15955 * config/arm/arm.c (any_sibcall_uses_r3): Rename to ..
15956 (any_sibcall_could_use_r3): this and handle indirect calls.
15957 (arm_get_frame_offsets): Rename use of any_sibcall_uses_r3.
15958
15959 2013-05-22 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
15960
15961 * config/rs6000/rs6000.h (MALLOC_ABI_ALIGNMENT): New #define.
15962
15963 2013-05-22 Richard Biener <rguenther@suse.de>
15964
15965 PR middle-end/57349
15966 * profile.c (branch_prob): Do not split blocks that are
15967 abnormally receiving from ECF_RETURNS_TWICE functions.
15968
15969 2013-05-22 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
15970
15971 * recog.c (offsettable_address_addr_space_p): Fix calculation of
15972 address mode. Move pointer mode initialization to the same place.
15973
15974 2013-05-22 Michael Zolotukhin <michael.v.zolotukhin@intel.com>
15975
15976 * read-rtl.c (copy_rtx_for_iterators): Continue applying iterators
15977 while it has any effect.
15978
15979 2013-05-21 Easwaran Raman <eraman@google.com>
15980
15981 PR tree-optimization/57322
15982 * tree-ssa-reassoc.c (build_and_add_sum): If a BB is empty, set the
15983 UID of the statement added to the BB to be 1.
15984
15985 2013-05-21 Jakub Jelinek <jakub@redhat.com>
15986
15987 PR tree-optimization/57331
15988 * tree-vrp.c (simplify_cond_using_ranges): Don't optimize comparison
15989 of conversion from pointer type to integral type with integer.
15990
15991 2013-05-21 Martin Jambor <mjambor@suse.cz>
15992
15993 PR lto/57289
15994 * ipa-prop.c (ipa_read_node_info): Process param_used and
15995 controlled_uses in the same order as when writing.
15996
15997 2013-05-21 Magnus Granberg <baldrick@free.fr>
15998
15999 PR plugins/56754
16000 * Makefile.in (PLUGIN_HEADERS): Add $(TARGET_H).
16001
16002 2013-05-21 Richard Biener <rguenther@suse.de>
16003
16004 PR tree-optimization/57318
16005 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Do not
16006 estimate stmts with side-effects as likely eliminated.
16007
16008 2013-05-21 Richard Biener <rguenther@suse.de>
16009
16010 PR tree-optimization/57330
16011 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Properly
16012 preserve the call stmts fntype.
16013
16014 2013-05-21 Richard Biener <rguenther@suse.de>
16015
16016 PR tree-optimization/57303
16017 * tree-ssa-sink.c (statement_sink_location): Improve killing
16018 stmt detection and properly handle self-assignments.
16019
16020 2013-05-21 Christian Bruel <christian.bruel@st.com>
16021
16022 * dwarf2out.c (multiple_reg_loc_descriptor): Use dbx_reg_number for
16023 spanning registers. LEAF_REG_REMAP is supported only for contiguous
16024 registers. Set register size out of the PARALLEL loop.
16025
16026 2013-05-20 Oleg Endo <olegendo@gcc.gnu.org>
16027
16028 PR target/56547
16029 * config/sh/sh.md (fmasf4): Remove empty constraints strings.
16030 (*fmasf4, *fmasf4_media): New insns.
16031
16032 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16033
16034 * config/mips/mips.h (BASE_INSN_LENGTH, NOP_INSN_LENGTH): New macros.
16035 * config/mips/mips.c (mips_symbol_insns, mips_address_insns)
16036 (mips_const_insns, mips_split_const_insns, mips_load_store_insns)
16037 (mips_idiv_insns): Update the comments to say that the returned
16038 instruction counts are in units of BASE_INSN_LENGTH.
16039 (mips_adjust_insn_length): Multiply the mips_load_label_num_insns
16040 by BASE_INSN_LENGTH rather than 4. Add the jump separately,
16041 using 2 rather than 4 as the length of indirect MIPS16 and
16042 microMIPS jumps. Use NOP_INSN_LENGTH rather than 4 as the
16043 length of a NOP. Don't divide MIPS16 lengths by 2.
16044 (mips16_split_long_branches): Assume a branch is long if the
16045 length is greater than 4 rather than 8.
16046 * config/mips/mips.md (length): Give MIPS16 lengths directly,
16047 rather than multiplying them by 2. Multiply instruction counts
16048 by BASE_INSN_LENGTH rather than 4.
16049 (*jump_mips16, tls_get_tp_mips16_<mode>)
16050 (*tls_get_tp_mips16_call_<mode>): Divide lengths by 2.
16051
16052 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16053
16054 * config/mips/mips.md (extended_mips16): Remove branch case.
16055 (length): Remove duplicated extended_mips16 test.
16056
16057 2013-05-19 Richard Sandiford <rdsandiford@googlemail.com>
16058
16059 * config/mips/t-sde: Don't build 64-bit microMIPS multilibs.
16060
16061 2013-05-18 Richard Sandiford <rdsandiford@googlemail.com>
16062
16063 * recog.h (Recog_data): Rename to...
16064 (recog_data_d): ...this.
16065 (recog_data): Update accordingly.
16066 * recog.c (recog_data): Likewise.
16067 * reload.c (save_recog_data): Likewise.
16068 * config/picochip/picochip.c (picochip_saved_recog_data): Likewise.
16069 (picochip_save_recog_data, picochip_restore_recog_data): Likewise.
16070
16071 2013-05-17 Julian Brown <julian@codesourcery.com>
16072
16073 * gcse.c (compute_ld_motion_mems): If a non-simple MEM is
16074 found in a REG_EQUAL note, invalidate it.
16075
16076 2013-05-17 Easwaran Raman <eraman@google.com>
16077
16078 * tree-ssa-reassoc.c (find_insert_point): New function.
16079 (insert_stmt_after): Likewise.
16080 (get_def_stmt): Likewise.
16081 (ensure_ops_are_available): Likewise.
16082 (not_dominated_by): Likewise.
16083 (rewrite_expr_tree): Do not move statements beyond what is
16084 necessary. Remove call to swap_ops_for_binary_stmt...
16085 (reassociate_bb): ... and move it here.
16086 (build_and_add_sum): Assign UIDs for new statements.
16087 (linearize_expr): Likewise.
16088 (do_reassoc): Renumber gimple statement UIDs.
16089
16090 2013-05-17 Jan Hubicka <jh@suse.cz>
16091
16092 * lto-symtab.c (lto_symtab_merge_cgraph_nodes): Resolve cross module
16093 weakrefs.
16094 * cgraph.c (dump_cgraph_node): Do not ice on unresolved alias.
16095 * cgraphunit.c (handle_alias_pairs): Store target of unresolved
16096 weakrefs.
16097 (output_weakrefs): Update.
16098
16099 2013-05-17 Po-Chun Chang <pchang9@cs.wisc.edu>
16100 Martin Jambor <mjambor@suse.cz>
16101
16102 PR middle-end/57276
16103 * ipa-cp.c (cgraph_edge_brings_all_agg_vals_for_node): Break when a
16104 value that corresponds to the given aggval is found in values vector.
16105
16106 2013-05-17 Uros Bizjak <ubizjak@gmail.com>
16107
16108 * config/i386/driver-i386.c (host_detect_local_cpu): Pass mmx, 3dnow,
16109 sse, sse2, sse3, ssse3 and sse4a flags to options.
16110
16111 2013-05-17 David Malcolm <dmalcolm@redhat.com>
16112
16113 * gengtype-state.c: (s_expr_writer): New class, to handle
16114 prettifying of output layout of s-expressions.
16115 (state_writer): New class, to write out gtype.state.
16116 (state_written_type_count): Move this variable into member data of
16117 state_writer.
16118 (s_expr_writer::s_expr_writer): New code: constructor for new class
16119 (state_writer::state_writer(): ditto
16120 (s_expr_writer::write_new_line): New function
16121 (s_expr_writer::write_any_indent): ditto
16122 (s_expr_writer::begin_s_expr): ditto
16123 (s_expr_writer::end_s_expr): ditto
16124 (write_state_fileloc): convert to method of state_writer...
16125 (state_writer:: write_state_fileloc): ...and use methods of
16126 s_expr_writer to write indentation into the gtype.state output file
16127 to visually represent the hierarchical structure of the list
16128 structures
16129 (write_state_fields): ditto, renaming to...
16130 (state_writer::write_state_fields)
16131 (write_state_a_string): ditto, renaming to...
16132 (state_writer::write_state_a_string)
16133 (write_state_string_option): ditto, renaming to...
16134 (state_writer::write_state_string_option)
16135 (write_state_type_option): ditto, renaming to...
16136 (state_writer::write_state_type_option)
16137 (write_state_nested_option): ditto, renaming to...
16138 (state_writer::write_state_nested_option)
16139 (write_state_option): ditto, renaming to...
16140 (state_writer::write_state_option)
16141 (write_state_options): ditto, renaming to...
16142 (state_writer::write_state_options)
16143 (write_state_lang_bitmap): ditto, renaming to...
16144 (state_writer::write_state_lang_bitmap)
16145 (write_state_version): ditto, renaming to...
16146 (state_writer::write_state_version)
16147 (write_state_scalar_type): ditto, renaming to...
16148 (state_writer::write_state_scalar_type)
16149 (write_state_string_type): ditto, renaming to...
16150 (state_writer::write_state_string_type)
16151 (write_state_undefined_type): ditto, renaming to...
16152 (state_writer::write_state_undefined_type)
16153 (write_state_struct_union_type): ditto, renaming to...
16154 (state_writer::write_state_struct_union_type)
16155 (write_state_struct_type): ditto, renaming to...
16156 (state_writer::write_state_struct_type)
16157 (write_state_user_struct_type): ditto, renaming to...
16158 (state_writer::write_state_user_struct_type)
16159 (write_state_lang_struct_type): ditto, renaming to...
16160 (state_writer::write_state_lang_struct_type)
16161 (write_state_param_struct_type): ditto, renaming to...
16162 (state_writer::write_state_param_struct_type)
16163 (write_state_pointer_type): ditto, renaming to...
16164 (state_writer::write_state_pointer_type)
16165 (write_state_array_type): ditto, renaming to...
16166 (state_writer::write_state_array_type)
16167 (write_state_gc_used): ditto, renaming to...
16168 (state_writer::write_state_gc_used)
16169 (write_state_common_type_content): ditto, renaming to...
16170 (state_writer::write_state_common_type_content)
16171 (write_state_type): ditto, renaming to...
16172 (state_writer::write_state_type)
16173 (write_state_pair_list): ditto, renaming to...
16174 (state_writer::write_state_pair_list)
16175 (write_state_pair): ditto, renaming to...
16176 (state_writer::write_state_pair)
16177 (write_state_typedefs): ditto, renaming to...
16178 (state_writer::write_state_typedefs)
16179 (write_state_structures): ditto, renaming to...
16180 (state_writer::write_state_structures)
16181 (write_state_param_structs): ditto, renaming to...
16182 (state_writer::write_state_param_structs)
16183 (write_state_variables): ditto, renaming to...
16184 (state_writer::write_state_variables)
16185 (write_state_srcdir): ditto, renaming to...
16186 (state_writer::write_state_srcdir)
16187 (write_state_files_list): ditto, renaming to...
16188 (state_writer::write_state_files_list)
16189 (write_state_languages): ditto, renaming to...
16190 (state_writer::write_state_languages)
16191 (write_state): create a state_writer instance and use it when
16192 writing out the state file
16193
16194 2013-05-17 Mike Stump <mikestump@comcast.net>
16195
16196 PR rtl-optimization/57304
16197 * web.c (union_match_dups): Ensure that DF_REF_LOC exists before
16198 accessing DF_REF_REAL_LOC.
16199
16200 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16201
16202 PR rtl-optimization/57281
16203 PR rtl-optimization/57300
16204 * config/i386/i386.md (extendsidi2_1 dead reg splitter): Remove.
16205 (extendsidi2_1 peephole2s): Add instead 2 new peephole2s, that undo
16206 what the other splitter did if the registers are dead.
16207
16208 2013-05-17 Richard Biener <rguenther@suse.de>
16209
16210 * tree-ssa-alias.c (stmt_kills_ref_p_1): Properly compare
16211 MEM_REF offsets.
16212
16213 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16214
16215 * gcc.c (SANITIZER_SPEC): Reject -fsanitize=address -fsanitize=thread
16216 linking.
16217
16218 2013-05-17 Marek Polacek <polacek@redhat.com>
16219
16220 * tree-ssa-strlen.c (handle_char_store): Don't invalidate cached
16221 length when doing non-zero store of storing '\0' to '\0'.
16222
16223 2013-05-17 Jakub Jelinek <jakub@redhat.com>
16224
16225 * tree-vect-patterns.c (vect_recog_rotate_pattern): For
16226 vect_external_def oprnd1 with loop_vinfo, try to emit
16227 optional cast, negation and and stmts on the loop preheader
16228 edge instead of into the pattern def seq.
16229
16230 PR tree-optimization/57051
16231 * fold-const.c (const_binop) <case VEC_LSHIFT_EXPR,
16232 case VEC_RSHIFT_EXPR>: Fix BYTES_BIG_ENDIAN handling.
16233
16234 2013-05-16 Nick Clifton <nickc@redhat.com>
16235
16236 * config/rl78/rl78.c (rl78_attribute_table): Add naked.
16237 (rl78_is_naked_func): New function.
16238 (rl78_expand_prologue): Skip prologue generation for naked functions.
16239 (rl78_expand_epilogue): Skip epilogue generation for naked functions.
16240 * doc/extend.texi (naked): Add RL78 to the list of processors
16241 that supports this attribute.
16242
16243 2013-05-16 Jeff Law <law@redhat.com>
16244
16245 * Makefile.in (tree-switch-conversion.o): Depend on $(OPTABS_H).
16246
16247 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16248
16249 * config/i386/driver-i386.c (host_detect_local_cpu): Determine
16250 cache parameters using detect_caches_amd also for CYRIX,
16251 NSC and TM2 signatures.
16252
16253 2013-05-16 Uros Bizjak <ubizjak@gmail.com>
16254 Dzianis Kahanovich <mahatma@eu.by>
16255
16256 PR target/45359
16257 PR target/46396
16258 * config/i386/driver-i386.c (host_detect_local_cpu): Detect
16259 VIA/Centaur processors and determine their cache parameters
16260 using detect_caches_amd.
16261
16262 2013-05-16 Teresa Johnson <tejohnson@google.com>
16263
16264 * cfgrtl.c (verify_hot_cold_block_grouping): Return err.
16265 (rtl_verify_edges): New function.
16266 (rtl_verify_bb_insns): Ditto.
16267 (rtl_verify_bb_pointers): Ditto.
16268 (rtl_verify_bb_insn_chain): Ditto.
16269 (rtl_verify_fallthru): Ditto.
16270 (rtl_verify_bb_layout): Ditto.
16271 (rtl_verify_flow_info_1): Outline checks into new functions.
16272 (rtl_verify_flow_info): Ditto.
16273
16274 2013-05-16 Steve Ellcey <sellcey@imgtec.com>
16275
16276 * cfghooks.c (copy_bbs): Add update_dominance argument.
16277 * cfghooks.h (copy_bbs): Update prototype.
16278 * tree-cfg.c (gimple_duplicate_sese_region):
16279 Add update_dominance argument.
16280 * tree-flow.h (gimple_duplicate_sese_region): Update prototype.
16281 * tree-ssa-loop-ch.c (copy_loop_headers): Update
16282 gimple_duplicate_sese_region call.
16283 * tree-vect-loop-manip.c (slpeel_tree_duplicate_loop_to_edge_cfg):
16284 Update copy_bbs call.
16285 * cfgloopmanip.c (duplicate_loop_to_header_edge): Ditto.
16286 * trans-mem.c (ipa_uninstrument_transaction): Ditto.
16287
16288 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16289
16290 * tree-vectorizer.h (NUM_PATTERNS): Increment.
16291 * tree-vect-patterns.c (vect_vect_recog_func_ptrs): Add
16292 vect_recog_rotate_pattern.
16293 (vect_recog_rotate_pattern): New function.
16294
16295 2013-05-16 Jason Merrill <jason@redhat.com>
16296
16297 * Makefile.in (LLINKER): New variable.
16298 (mostlyclean): Remove link mutex.
16299 * configure.ac: Handle --enable-link-mutex.
16300 * lock-and-run.sh: New script.
16301
16302 2013-05-16 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16303
16304 PR target/19599
16305 * config/arm/arm.c (arm_function_ok_for_sibcall): Add check
16306 for NULL decl.
16307
16308 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16309
16310 * reorg.c (link_cc0_insns): Wrap in #ifdef HAVE_cc0.
16311
16312 2013-05-16 Greta Yorsh <Greta.Yorsh@arm.com>
16313
16314 * config/arm/arm-protos.h (gen_movmem_ldrd_strd): New declaration.
16315 * config/arm/arm.c (next_consecutive_mem): New function.
16316 (gen_movmem_ldrd_strd): Likewise.
16317 * config/arm/arm.md (movmemqi): Update condition and code.
16318 (unaligned_loaddi, unaligned_storedi): New patterns.
16319
16320 2013-05-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16321
16322 * config.gcc: Obsolete *-*-solaris2.9*.
16323 * doc/install.texi (Specific, *-*-solaris2*): Document it.
16324
16325 2013-05-16 Richard Biener <rguenther@suse.de>
16326
16327 * passes.c (init_optimization_passes): Move pass_parallelize_loops
16328 earlier, after GRAPHITE transforms and IV canonicalization.
16329
16330 2013-05-16 Jakub Jelinek <jakub@redhat.com>
16331
16332 * omp-low.c (extract_omp_for_data): For collapsed loops,
16333 if at least one of the loops is known at compile time to
16334 iterate zero times, set count to 0.
16335 (expand_omp_regimplify_p): New function.
16336 (expand_omp_for_generic): For collapsed loops, if at least
16337 one of the loops isn't known to iterate at least once,
16338 add runtime check with setting count to 0.
16339 (expand_omp_for_static_nochunk, expand_omp_for_static_chunk):
16340 For unsigned types if it isn't known at compile time that
16341 the loop will iterate at least once, add runtime check to bypass
16342 the whole loop if initial condition isn't true.
16343
16344 2013-05-16 Nathan Sidwell <nathan@codesourcery.com>
16345
16346 * varasm.c (default_use_anchors_for_symbol_p): Use decl_replaceable_p.
16347
16348 2013-05-16 Marc Glisse <marc.glisse@inria.fr>
16349
16350 PR middle-end/57286
16351 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Disable some
16352 transformations to avoid an infinite loop.
16353
16354 2013-05-16 Marek Polacek <polacek@redhat.com>
16355
16356 * tree-scalar-evolution.c (scev_const_prop): Add more dumps.
16357
16358 2013-05-15 Leif Ekblad <leif@rdos.net>
16359
16360 * config/i386/i386.c (ix86_decompose_address): Use
16361 DEFAULT_TLS_SEG_REG to access TLS segment register.
16362 * config/i386/i386.h (DEFAULT_TLS_SEG_REG): New define.
16363 * config/i386/rdos.h (DEFAULT_TLS_SEG_REG): Ditto.
16364 (TARGET_TLS_DIRECT_SEG_REFS_DEFAULT): Ditto.
16365
16366 2013-05-15 Richard Sandiford <rdsandiford@googlemail.com>
16367
16368 PR target/57260
16369 * config/mips/mips.c (mips_function_ok_for_sibcall): Don't allow
16370 sibling calls to functions that would normally be lazily bound,
16371 unless $gp is call-clobbered.
16372
16373 2013-05-15 Uros Bizjak <ubizjak@gmail.com>
16374
16375 * config/i386/i386.c (ix86_option_override_internal): Update
16376 processor_alias_table for missing PTA_PRFCHW and PTA_FXSR flags. Add
16377 PTA_POPCNT to corei7 entry. Do not enable SSE prefetch on
16378 non-SSE 3dNow! targets. Enable TARGET_PRFCHW for TARGET_3DNOW targets.
16379 * config/i386/i386.md (prefetch): Enable for TARGET_PRFCHW instead
16380 of TARGET_3DNOW.
16381 (*prefetch_3dnow): Enable for TARGET_PRFCHW only.
16382
16383 2013-05-15 Andreas Schwab <schwab@suse.de>
16384
16385 * config/m68k/m68k.md (*rotlhi3_lowpart, *rotlqi3_lowpart): Name
16386 for rotlhi3+1 and rotlqi3+1, resp. Fix reference to non-existing
16387 third operand.
16388
16389 2013-05-15 Teresa Johnson <tejohnson@google.com>
16390
16391 * loop-unroll.c (report_unroll_peel): Check decision before
16392 emitting unroll/peel message.
16393
16394 2013-05-15 Teresa Johnson <tejohnson@google.com>
16395
16396 * function.h (has_bb_partition): New rtl_data flag.
16397 (bb_reorder_complete): Ditto.
16398 * cfgcleanup.c (try_crossjump_to_edge): Check for has_bb_partition
16399 instead of flag_reorder_blocks_and_partition.
16400 * cfgrtl.c (verify_hot_cold_block_grouping): Moved from bb-reorder.c,
16401 with some enhancements.
16402 (rtl_verify_flow_info_1): Call verify_hot_cold_block_grouping.
16403 * bb-reorder.c (connect_traces): Check for has_bb_partition
16404 instead of flag_reorder_blocks_and_partition.
16405 (verify_hot_cold_block_grouping): Moved to cfgrtl.c.
16406 (reorder_basic_blocks): Set bb_reorder_complete flag, remove call to
16407 verify_hot_cold_block_grouping.
16408 (partition_hot_cold_basic_blocks): Set has_bb_partition.
16409
16410 2013-05-15 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
16411
16412 PR target/19599
16413 * config/arm/predicates.md (call_insn_operand): New predicate.
16414 * config/arm/constraints.md ("Cs", "Ss"): New constraints.
16415 * config/arm/arm.md (*call_insn, *call_value_insn): Match only
16416 if insn is not a tail call.
16417 (*sibcall_insn, *sibcall_value_insn): Adjust for tailcalling through
16418 registers.
16419 * config/arm/arm.h (enum reg_class): New caller save register class.
16420 (REG_CLASS_NAMES): Likewise.
16421 (REG_CLASS_CONTENTS): Likewise.
16422 * config/arm/arm.c (arm_function_ok_for_sibcall): Allow tailcalling
16423 without decls.
16424
16425 2013-05-15 Richard Biener <rguenther@suse.de>
16426
16427 * tree-vect-loop.c (vect_transform_loop): Use MSG_NOTE instead
16428 of MSG_OPTIMIZED_LOCATIONS.
16429 * tree-vect-slp.c (vect_make_slp_decision): Likewise.
16430 (vect_slp_transform_bb): Indicate location in MSG_OPTIMIZED_LOCATIONS
16431 message.
16432 * tree-vectorizer.c (vectorize_loops): Use MSG_NOTE instead
16433 of MSG_OPTIMIZED_LOCATIONS.
16434 (execute_vect_slp): Likewise.
16435 * tree-vect-loop-manip.c (vect_do_peeling_for_loop_bound): Likewise.
16436 (vect_create_cond_for_alias_checks): Likewise.
16437 * tree-vect-patterns.c (vect_recog_dot_prod_pattern): Likewise.
16438 (vect_recog_widen_mult_pattern): Likewise.
16439 (vect_recog_widen_sum_pattern): Likewise.
16440 (vect_recog_over_widening_pattern): Likewise.
16441 (vect_recog_widen_shift_pattern): Likewise.
16442 (vect_recog_vector_vector_shift_pattern): Likewise.
16443 (vect_recog_divmod_pattern): Likewise.
16444 (vect_recog_mixed_size_cond_pattern): Likewise.
16445 (vect_recog_bool_pattern): Likewise.
16446 (vect_pattern_recog_1): Likewise.
16447
16448 2013-05-15 Martin Jambor <mjambor@suse.cz>
16449
16450 * ipa-prop.c (ipa_make_edge_direct_to_target): Redirect calls to
16451 non-functions to builtin_unreachable.
16452 * ipa-inline-transform.c (inline_call): Do not assert estimates were
16453 correct when new direct edges were discovered.
16454
16455 2013-05-15 Martin Jambor <mjambor@suse.cz>
16456
16457 * ipa-prop.c (ipa_print_node_jump_functions): Print symbol order in
16458 header, print symbol order instead of node uid, print more information
16459 about indirect edge targets.
16460 (ipa_make_edge_direct_to_target): Print symbol order instead of node
16461 uids.
16462 (ipa_make_edge_direct_to_target): Likewise.
16463 (remove_described_reference): Likewise.
16464 (propagate_controlled_uses): Likewise.
16465 (ipa_print_node_params): Also print symbol order.
16466 (ipcp_transform_function): Print symbol order instead of node uids.
16467 * cgraph.c (cgraph_redirect_edge_call_stmt_to_callee): Likewise.
16468 (cgraph_get_create_real_symbol_node): Likewise.
16469 * ipa-cp.c (print_lattice): Likewise.
16470 (print_all_lattices): Likewise.
16471 (determine_versionability): Likewise.
16472 (initialize_node_lattices): Likewise.
16473 (estimate_local_effects): Likewise.
16474 (update_profiling_info): Likewise.
16475 (create_specialized_node): Likewise.
16476 (perhaps_add_new_callers): Likewise.
16477 (decide_about_value): Likewise.
16478 (decide_whether_version_node): Likewise.
16479 (identify_dead_nodes): Likewise.
16480 * ipa-inline-analysis.c (dump_inline_edge_summary): Likewise.
16481 (dump_inline_summary): Likewise.
16482 (estimate_node_size_and_time): Likewise.
16483 (inline_analyze_function): Likewise.
16484 * ipa-inline.c (report_inline_failed_reason): Likewise.
16485 (want_early_inline_function_p): Likewise.
16486 (edge_badness): Likewise.
16487 (update_edge_key): Likewise.
16488 (inline_small_functions): Likewise. Add dumping of order to two other
16489 dumps.
16490 * ipa-pure-const.c (pure_const_read_summary): Print symbol order
16491 instead of node uids.
16492 (propagate_pure_const): Likewise.
16493 (propagate_pure_const): Likewise.
16494 * ipa-utils.c (dump_cgraph_node_set): Likewise.
16495 * lto-cgraph.c (input_node): Explicitly specify we dump uid.
16496 * lto-symtab.c (lto_cgraph_replace_node): Print symbol order instead
16497 of node uids.
16498 * tree-pretty-print.c (dump_function_header): Likewise.
16499 * tree-sra.c (convert_callers_for_node): Dump in traditional format.
16500 Print symbol order instead of node uids.
16501
16502 2013-05-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
16503
16504 * config/s390/s390.c (s390_register_move_cost): Don't impose the
16505 FPR<->GPR move cost penalty if ldgr/lgdr can be used.
16506
16507 2013-05-15 Richard Biener <rguenther@suse.de>
16508
16509 PR tree-optimization/57275
16510 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence): Fix
16511 return value for fail to do runtime alias checks for gather loads.
16512
16513 2013-05-15 Jan Hubicka <jh@suse.cz>
16514
16515 PR lto/57038
16516 PR lto/47375
16517 * lto-symtab.c (lto_symtab_symbol_p): Add external symbol;
16518 weakrefs are not external.
16519 (lto_symtab_merge_decls): Fix thinko when dealing with
16520 non-lto_symtab decls.
16521 (lto_symtab_merge_cgraph_nodes): Use lto_symtab_symbol_p.
16522 (lto_symtab_prevailing_decl): Get int sync with lto_symtab_symbol_p.
16523 * varpool.c (dump_varpool_node): Dump more flags.
16524
16525 2013-05-15 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16526
16527 * config/i386/i386.c (processor_alias_table): Add instruction
16528 FSGSBASE for AMD bdver3 architecture.
16529
16530 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16531
16532 * tree.c (warn_deprecated_use): Print file:line using locus color.
16533 * diagnostic.c (diagnostic_report_current_module): Print file:line
16534 and file:line:column using locus color.
16535
16536 2013-05-14 Mike Stump <mikestump@comcast.net>
16537
16538 * gdbinit.in: Add __null.
16539
16540 2013-05-14 Mike Stump <mikestump@comcast.net>
16541
16542 * recog.h: Rename struct recog_data to Recog_data.
16543 * recog.c: Likewise.
16544 * reload.c (can_reload_into): Likewise.
16545 * config/picochip/picochip.c: Likewise.
16546
16547 2013-05-14 Mike Stump <mikestump@comcast.net>
16548
16549 * web.c (union_match_dups): Also check DF_REF_REAL_LOC.
16550
16551 2013-05-14 Steven Bosscher <steven@gcc.gnu.org>
16552
16553 * resource.h (struct resources): Remove unch_memory member.
16554 (CLEAR_RESOURCE): Don't clear unch_memory.
16555 * resource.c (mark_referenced_resources): Don't set it.
16556 (mark_set_resources): Likewise.
16557 (mark_target_live_regs): Don't clear it.
16558 (init_resource_info): Likewise.
16559 * reorg.c (resource_conflicts_p): Don't compare it.
16560 (redundant_insn): Don't set it.
16561
16562 * rtl.h (next_label, skip_consecutive_labels, link_cc0_insns):
16563 Remove prototypes.
16564 * emit-rtl.c (next_label): Remove unused function.
16565 (skip_consecutive_labels, link_cc0_insns): Move to ...
16566 * reorg.c (skip_consecutive_labels, link_cc0_insns): ... here, the
16567 only place where these functions are used, and make them static.
16568
16569 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16570
16571 * fold-const.c (fold_negate_expr): Handle vectors.
16572 (fold_truth_not_expr): Make it static.
16573 (fold_invert_truthvalue): New static function.
16574 (invert_truthvalue_loc): Handle vectors. Do not call
16575 fold_truth_not_expr directly.
16576 (fold_unary_loc) <BIT_NOT_EXPR>: Handle comparisons.
16577 <TRUTH_NOT_EXPR>: Do not cast to boolean.
16578 (fold_comparison): Handle vector constants.
16579 (fold_binary_loc) <TRUTH_XOR_EXPR>: Remove redundant code.
16580 (fold_ternary_loc) <VEC_COND_EXPR>: Adapt more COND_EXPR optimizations.
16581 * tree.h (fold_truth_not_expr): Remove declaration.
16582
16583 2013-05-14 James Greenhalgh <james.greenhalgh@arm.com>
16584
16585 * config/aarch64/aarch64-simd.md
16586 (aarch64_vcond_internal<mode>): Rename to...
16587 (aarch64_vcond_internal<mode><mode>): ...This, for integer modes.
16588 (aarch64_vcond_internal<VDQF_COND:mode><VDQF:mode>): ...This for
16589 float modes. Clarify all iterator modes.
16590 (vcond<mode><mode>): Use new name for vcond expanders.
16591 (vcond<v_cmp_result><mode>): Likewise.
16592 (vcondu<mode><mode>: Likewise.
16593 * config/aarch64/iterators.md (VDQF_COND): New.
16594
16595 2013-05-14 Marc Glisse <marc.glisse@inria.fr>
16596
16597 PR bootstrap/57266
16598 * fold-const.c (fold_binary_loc) <shift>: Use an unsigned
16599 variable for the shift amount. Check that we shift by non-negative
16600 amounts.
16601
16602 2013-05-14 Chung-Lin Tang <cltang@codesourcery.com>
16603
16604 PR target/42017
16605 * config/arm/arm.h (EPILOGUE_USES): Only return true
16606 for LR_REGNUM after epilogue_completed.
16607
16608 2013-05-14 Joern Rennecke <joern.rennecke@embecosm.com>
16609
16610 * config/avr/avr.c (avr_encode_section_info): Bail out if the type
16611 is error_mark_node.
16612
16613 2013-05-14 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
16614
16615 PR target/57261
16616 * configure.ac (gcc_cv_ld_as_needed): Disable before Solaris 11
16617 and Solaris 11+/x86 with gld.
16618 * configure: Regenerate.
16619
16620 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16621
16622 * expmed.c (expand_shift_1): Canonicalize rotates by
16623 constant bitsize / 2 to bitsize - 1.
16624 * simplify-rtx.c (simplify_binary_operation_1) <case ROTATE,
16625 case ROTATERT>: Likewise.
16626
16627 Revert:
16628 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16629
16630 * config/i386/i386.md (rotateinv): New code attr.
16631 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16632 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16633 roll $31, %eax, etc.
16634
16635 2013-05-14 Richard Biener <rguenther@suse.de>
16636
16637 PR middle-end/57235
16638 * tree-eh.c (sink_clobbers): Give up for successors with
16639 multiple predecessors and no virtual uses.
16640
16641 2013-05-14 Eric Botcazou <ebotcazou@adacore.com>
16642
16643 * config/sparc/sp64-elf.h (CPP_SUBTARGET_SPEC): Delete.
16644 * config/sparc/openbsd64.h (CPP_SUBTARGET_SPEC): Likewise.
16645
16646 2013-05-14 Jakub Jelinek <jakub@redhat.com>
16647
16648 PR middle-end/57251
16649 * expr.c (expand_expr_real_2) <case WIDEN_MULT_EXPR>: Handle
16650 the case when both op0 and op1 have VOIDmode.
16651
16652 2013-05-14 Kaushik Phatak <kaushik.phatak@kpitcummins.com>
16653
16654 * config/rl78/rl78.md(mulsi3_g13): Add additional 'nop' required
16655 in multiply-accumulate mode.
16656
16657 2013-05-13 Guozhi Wei <carrot@google.com>
16658
16659 * dwarf2asm.c (dw2_output_indirect_constant_1): Mark new decl STATIC.
16660
16661 2013-05-13 Kai Tietz <ktietz@redhat.com>
16662
16663 PR target/56975
16664 * config/i386/cygming.h (TARGET_PECOFF): Define as true.
16665 * config/i386/i386.h (TARGET_PECOFF): Define by default as false.
16666 (PIC_OFFSET_TABLE_REGNUM): Use TARGET_PECOFF.
16667 * config/i386/i386.c (ix86_option_override_internal): Likewise.
16668 (ix86_expand_prologue): Likewise.
16669 (ix86_expand_split_stack_prologue): Likewise.
16670 (legitimate_pic_address_disp_p): Likewise.
16671 (legitimize_pic_address): Likewise.
16672 (legitimize_tls_address): Likewise.
16673 (legitimize_pe_coff_symbol): Likewise.
16674 (output_pic_addr_const): Likewise.
16675 (construct_plt_address): Likewise.
16676 (ix86_expand_call): Likewise.
16677 (x86_output_mi_thunk): Likewise.
16678 (x86_function_profiler): Likewise.
16679
16680 2013-05-13 Sofiane Naci <sofiane.naci@arm.com>
16681
16682 * config/aarch64/aarch64-simd.md (aarch64_simd_mov<mode>): Group
16683 similar switch cases.
16684 (aarch64_simd_mov): Rename to aarch64_split_simd_mov. Update.
16685 (aarch64_simd_mov_to_<mode>low): Delete.
16686 (aarch64_simd_mov_to_<mode>high): Delete.
16687 (move_lo_quad_<mode>): Add w<-r alternative.
16688 (aarch64_simd_move_hi_quad_<mode>): Likewise.
16689 (aarch64_simd_mov_from_*): Update type attribute.
16690 * config/aarch64/aarch64.c (aarch64_split_simd_move): Refacror switch
16691 statement.
16692
16693 2013-05-13 Jan Hubicka <jh@suse.cz>
16694
16695 * mode-switching.c (optimize_mode_switching): Set correct RTL profile.
16696 * config/i386/i386.c (ix86_compute_frame_layout,
16697 ix86_expand_epilogue, emit_i387_cw_initialization,
16698 ix86_expand_vector_move_misalign, ix86_fp_comparison_strategy,
16699 ix86_local_alignment): Fix use of size/speed predicates.
16700
16701 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16702
16703 PR tree-optimization/45216
16704 PR tree-optimization/57157
16705 * tree-ssa-forwprop.c (simplify_rotate): Only recognize
16706 the (-Y) & (B - 1) variant if OP is |.
16707 * expmed.c (expand_shift_1): For rotations by const0_rtx just
16708 return shifted. Use (-op1) & (prec - 1) as other_amount
16709 instead of prec - op1.
16710
16711 2013-05-13 Martin Jambor <mjambor@suse.cz>
16712
16713 PR middle-end/42371
16714 * ipa-prop.h (IPA_UNDESCRIBED_USE): New macro.
16715 (ipa_constant_data): New type.
16716 (ipa_jump_func): Use ipa_constant_data to hold information about
16717 constant jump functions.
16718 (ipa_get_jf_constant): Adjust to jump function type changes.
16719 (ipa_get_jf_constant_rdesc): New function.
16720 (ipa_param_descriptor): New field controlled_uses.
16721 (ipa_get_controlled_uses): New function.
16722 (ipa_set_controlled_uses): Likewise.
16723 * ipa-ref.h (ipa_find_reference): Declare.
16724 * ipa-prop.c (ipa_cst_ref_desc): New type.
16725 (ipa_print_node_jump_functions_for_edge): Adjust for jump function type
16726 changes.
16727 (ipa_set_jf_constant): Likewise. Also create reference descriptions.
16728 New parameter cs. Adjust all callers.
16729 (ipa_analyze_params_uses): Detect uncontrolled and controlled uses.
16730 (remove_described_reference): New function.
16731 (jfunc_rdesc_usable): Likewise.
16732 (try_make_edge_direct_simple_call): Decrement controlled use count,
16733 attempt to remove reference if it hits zero.
16734 (combine_controlled_uses_counters): New function.
16735 (propagate_controlled_uses): Likewise.
16736 (ipa_propagate_indirect_call_infos): Call propagate_controlled_uses.
16737 (ipa_edge_duplication_hook): Duplicate reference descriptions.
16738 (ipa_print_node_params): Print described use counter.
16739 (ipa_write_jump_function): Adjust to jump function type changes.
16740 (ipa_read_jump_function): New parameter CS, pass it to
16741 ipa_set_jf_constant. Adjust caller.
16742 (ipa_write_node_info): Stream controlled use count
16743 (ipa_read_node_info): Likewise.
16744 * cgraph.c (cgraph_mark_address_taken_node): Bail out instead of
16745 asserting.
16746 * ipa-cp.c (ipcp_discover_new_direct_edges): Decrement controlled use
16747 count. Remove cloning-added reference if it reaches zero.
16748 * ipa-ref.c (ipa_find_reference): New function.
16749
16750 2013-05-13 Ganesh Gopalasubramanian <Ganesh.Gopalasubramanian@amd.com>
16751
16752 * config/i386/i386.c (processor_target_table): Modified default
16753 alignment values for AMD BD and BT architectures.
16754
16755 2013-05-13 Marc Glisse <marc.glisse@inria.fr>
16756
16757 * tree-vect-generic.c (uniform_vector_p): Move ...
16758 * tree.c (uniform_vector_p): ... here.
16759 * tree.h (uniform_vector_p): Declare it.
16760 * fold-const.c (fold_binary_loc) <shift>: Turn the second argument
16761 into a scalar.
16762
16763 2013-05-13 Jakub Jelinek <jakub@redhat.com>
16764
16765 PR tree-optimization/57230
16766 * tree-ssa-strlen.c (handle_char_store): Record length for
16767 array store from STRING_CST.
16768
16769 PR tree-optimization/57230
16770 * tree-ssa-strlen.c (handle_char_store): Add missing integer_zerop
16771 check.
16772
16773 2013-05-12 Joern Rennecke <joern.rennecke@embecosm.com>
16774
16775 * config/epiphany/epiphany.c (epiphany_init): Check size of
16776 NUM_MODES_FOR_MODE_SWITCHING.
16777 (epiphany_expand_prologue):
16778 Remove CONFIG_REGNUM initial value handling code.
16779 (epiphany_optimize_mode_switching): Handle EPIPHANY_MSW_ENTITY_CONFIG.
16780 (epiphany_mode_needed, epiphany_mode_entry_exit): Likewise.
16781 (emit_set_fp_mode, epiphany_mode_after): Likewise.
16782 (epiphany_mode_needed) <Handle EPIPHANY_MSW_ENTITY_AND>:
16783 Don't return 1 for FP_MODE_NONE.
16784 * config/epiphany/epiphany.h (NUM_MODES_FOR_MODE_SWITCHING):
16785 Add value for EPIPHANY_MSW_ENTITY_CONFIG.
16786 (EPIPHANY_MSW_ENTITY_CONFIG, EPIPHANY_MSW_ENTITY_NUM): Define.
16787 * config/epiphany/epiphany.md (save_config): New pattern.
16788
16789 2013-05-12 Uros Bizjak <ubizjak@gmail.com>
16790
16791 * config/i386/i386.md (*zero_extendsidi2): Add *x->?r alternative.
16792
16793 2013-05-10 Uros Bizjak <ubizjak@gmail.com>
16794
16795 * config/i386/i386.md (memory): Handle sseishft1.
16796 * config/i386/sse.md (*vec_extractv4si): Remove memory attribute.
16797 (*vec_extractv2di_1): Ditto.
16798
16799 2013-05-10 Vladimir Makarov <vmakarov@redhat.com>
16800
16801 * lra-assigns.c (find_hard_regno_for): Add 1 to the cost of call
16802 saved registers.
16803
16804 2013-05-10 Sebastian Huber <sebastian.huber@embedded-brains.de>
16805
16806 * config/arm/t-rtems-eabi: Remove mthumb/march=armv7 multilib.
16807 Add mthumb/march=armv7-a multilib.
16808 Add mthumb/march=armv7-r multilib.
16809 Add mthumb/march=armv7-a/mfpu=neon/mfloat-abi=hard multilib.
16810
16811 2013-05-10 Ralf Corsépius <ralf.corsepius@rtems.org>
16812
16813 * config/v850/t-rtems: Add more multilibs.
16814
16815 2013-05-10 Richard Biener <rguenther@suse.de>
16816
16817 PR tree-optimization/57214
16818 * tree-ssa-loop-ivcanon.c (propagate_constants_for_unrolling): Do
16819 not propagate from SSA names that occur in abnormal PHI nodes.
16820
16821 2013-05-10 Marc Glisse <marc.glisse@inria.fr>
16822
16823 * stor-layout.c (element_precision): New function.
16824 * machmode.h (element_precision): Declare it.
16825 * tree.c (build_minus_one_cst): New function.
16826 (element_precision): Likewise.
16827 * tree.h (build_minus_one_cst): Declare new function.
16828 (element_precision): Likewise.
16829 * fold-const.c (operand_equal_p): Use element_precision.
16830 (fold_binary_loc): Handle vector types.
16831 * convert.c (convert_to_integer): Use element_precision.
16832 * gimple.c (iterative_hash_canonical_type): Handle complex and vectors
16833 separately.
16834
16835 2013-05-10 Richard Sandiford <rdsandiford@googlemail.com>
16836
16837 * config/mips/mips-protos.h (m16_uimm3_b, m16_simm4_1, m16_nsimm4_1)
16838 (m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4, m16_simm8_1)
16839 (m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1, m16_uimm8_4)
16840 (m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16841 * config/mips/mips.c (m16_check_op, m16_uimm3_b, m16_simm4_1)
16842 (m16_nsimm4_1, m16_simm5_1, m16_nsimm5_1, m16_uimm5_4, m16_nuimm5_4)
16843 (m16_simm8_1, m16_nsimm8_1, m16_uimm8_1, m16_nuimm8_1, m16_uimm8_m1_1)
16844 (m16_uimm8_4, m16_nuimm8_4, m16_simm8_8, m16_nsimm8_8): Delete.
16845 * config/mips/constraints.md (Udb8, Usb5, Usb8, Usd8, Uub8, Uuw5)
16846 (Uuw8): New constraints.
16847 (Usb4): Move into alphabetical order.
16848 * config/mips/predicates.md (db8_operand, sb5_operand, sb8_operand)
16849 (sd8_operand, ub8_operand, uw8_operand): New predicates.
16850 * config/mips/mips.md (*xor<mode>3, *xor<mode>3_mips16): Name
16851 previously unnamed patterns.
16852 (*add<mode>3_mips16, *xor<mode>3_mips16, *<optab>si3_mips16)
16853 (*ashldi3_mips16, *ashrdi3_mips16, *lshrdi3_mips16)
16854 (*slt<u>_<GPR:mode><GPR2:mode>_mips16)
16855 (*sle<u>_<GPR:mode><GPR2:mode>_mips16): Use constraints instead
16856 of set_attr_alternative/if_then_else. Use extended_mips16 instead
16857 of specific lengths.
16858
16859 2013-05-10 Jakub Jelinek <jakub@redhat.com>
16860
16861 * config/i386/i386.md (rotateinv): New code attr.
16862 (*<rotate_insn><mode>3_1, *<rotate_insn>si3_1_zext,
16863 *<rotate_insn>qi3_1_slp): Emit rorl %eax instead of
16864 roll $31, %eax, etc.
16865
16866 PR tree-optimization/45216
16867 PR tree-optimization/57157
16868 * tree-ssa-forwprop.c (simplify_rotate): New function.
16869 (ssa_forward_propagate_and_combine): Call it.
16870
16871 2013-05-10 Richard Biener <rguenther@suse.de>
16872
16873 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do not
16874 disable peeling when we version for aliasing.
16875 (vector_alignment_reachable_p): Honor explicit user alignment.
16876 (vect_supportable_dr_alignment): Likewise.
16877 * tree-vect-loop-manip.c (vect_can_advance_ivs_p): Use
16878 STMT_VINFO_LOOP_PHI_EVOLUTION_PART instead of recomputing it.
16879 * tree-vect-loop.c (vect_transform_loop): First apply versioning,
16880 then peeling to arrange for the cost-model check to come first.
16881
16882 2013-05-10 Alan Modra <amodra@gmail.com>
16883
16884 * configure.ac (HAVE_AS_TLS): Swap powerpc64 and powerpc cases.
16885 (HAVE_LD_LARGE_TOC): Don't mention AIX in help text.
16886 * configure: Regenerate.
16887
16888 2013-05-10 Alan Modra <amodra@gmail.com>
16889
16890 PR target/55033
16891 * varasm.c (default_elf_select_section): Move !DECL_P check..
16892 (get_named_section): ..to here before calling get_section_name.
16893 Adjust assertion.
16894 (default_section_type_flags): Add DECL_P check.
16895 * config/i386/winnt.c (i386_pe_section_type_flags): Likewise.
16896 * config/rs6000/rs6000.c (rs6000_xcoff_section_type_flags): Likewise.
16897
16898 2013-05-09 Joern Rennecke <joern.rennecke@embecosm.com>
16899
16900 * config/epiphany/epiphany.c (epiphany_expand_prologue):
16901 When using gen_stack_adjust_str with a register offset, add a
16902 REG_FRAME_RELATED_EXPR note.
16903
16904 2013-05-09 Uros Bizjak <ubizjak@gmail.com>
16905
16906 * config/i386/sse.md (*vec_extractv4si_0_zext): New pattern.
16907 (*vec_extractv4si_zext_mem): Ditto.
16908 (*vec_extractv2di): Add 0->x and x->x alternatives.
16909 * config/i386/mmx.md (*vec_extractv2si_zext_mem): New pattern.
16910 * config/i386/i386.md (*zero_extendsidi2): Add *Yj->?r alternative.
16911
16912 2013-05-09 Jason Merrill <jason@redhat.com>
16913
16914 N3639 C++1y VLA support
16915 * gimplify.c (gimplify_vla_decl): Don't touch an existing
16916 DECL_VALUE_EXPR.
16917
16918 * tree.c (build_constructor_va): New.
16919 * tree.h: Declare it.
16920
16921 2013-05-09 Martin Jambor <mjambor@suse.cz>
16922
16923 PR lto/57084
16924 * gimple-fold.c (canonicalize_constructor_val): Call
16925 cgraph_get_create_real_symbol_node instead of cgraph_get_create_node.
16926
16927 2013-05-09 Jan Hubicka <jh@suse.cz>
16928 Richard Biener <rguenther@suse.de>
16929
16930 PR lto/54095
16931 * symtab.c (symtab_make_decl_local): Do not add private names.
16932
16933 2013-05-09 Jan Hubicka <jh@suse.cz>
16934
16935 PR lto/54095
16936 * symtab.c (insert_to_assembler_name_hash): Handle clones.
16937 (unlink_from_assembler_name_hash): Likewise.
16938 (symtab_prevail_in_asm_name_hash, symtab_register_node,
16939 symtab_unregister_node, symtab_initialize_asm_name_hash,
16940 change_decl_assembler_name): Update.
16941
16942 2013-05-09 Sofiane Naci <sofiane.naci@arm.com>
16943
16944 * config/aarch64/aarch64.md: New movtf split.
16945 (*movtf_aarch64): Update.
16946 (aarch64_movdi_tilow): Handle TF modes and rename to
16947 aarch64_movdi_<mode>low.
16948 (aarch64_movdi_tihigh): Handle TF modes and rename to
16949 aarch64_movdi_<mode>high
16950 (aarch64_movtihigh_di): Handle TF modes and rename to
16951 aarch64_mov<mode>high_di
16952 (aarch64_movtilow_di): Handle TF modes and rename to
16953 aarch64_mov<mode>low_di
16954 (aarch64_movtilow_tilow): Remove spurious whitespace.
16955 * config/aarch64/aarch64.c (aarch64_split_128bit_move): Handle TFmode
16956 splits.
16957 (aarch64_print_operand): Update.
16958
16959 2013-05-09 Alan Modra <amodra@gmail.com>
16960
16961 * configure.ac (HAVE_AS_TLS): Enable tests for powerpcle and
16962 powerpc64le.
16963 * configure: Regenerate.
16964
16965 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16966
16967 * config/i386/mmx.md (*vec_extract* splitters): Simplify post-reload
16968 splitter preparation statements.
16969 * config/i386/sse.md (*vec_extract* splitters): Ditto.
16970 (*avx_vperm_broadcast_<mode>): Use adjust_address instead of
16971 adjust_address_nv.
16972
16973 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
16974
16975 * gimple-ssa-strength-reduction.c (count_candidates): Change
16976 return value to int.
16977 (analyze_candidates_and_replace): Change type of length to int.
16978
16979 2013-05-08 Uros Bizjak <ubizjak@gmail.com>
16980
16981 * config/i386/sse.md (PEXTR_MODE, PEXTR_MODEx): Remove.
16982 (*vec_extract<mode>): Use VI12_128 mode iterator.
16983 (*vec_extract<mode>_mem): Ditto.
16984 (*vec_extract*_mem splitters): Merge splitters using VI_128 mode
16985 attribute.
16986
16987 2013-05-08 Diego Novillo <dnovillo@google.com>
16988
16989 PR bootstrap/54659
16990
16991 Revert:
16992 2012-08-17 Diego Novillo <dnovillo@google.com>
16993
16994 PR bootstrap/54281
16995 * configure.ac: Add libintl.h to AC_CHECK_HEADERS list.
16996 * config.in: Regenerate.
16997 * configure: Regenerate.
16998 * intl.h: Always include libintl.h if HAVE_LIBINTL_H is set.
16999
17000 2013-05-08 Jan Hubicka <jh@suse.cz>
17001
17002 PR lto/54095
17003 * cgraph.c (cgraph_make_node_local_1): Se unique_name.
17004 * cgraph.h (symtab_node_base): Add unique_name.
17005 * lto-cgraph.c (lto_output_node, lto_output_varpool_node,
17006 input_overwrite_node, input_varpool_node): Stream unique_name.
17007 * cgraphclones.c (cgraph_create_virtual_clone,
17008 cgraph_function_versioning): Set unique_name.
17009 * ipa.c (function_and_variable_visibility): Set unique_name.
17010
17011 2013-05-08 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17012
17013 * gimple-ssa-strength-reduction.c (find_phi_def): Revert former "fix."
17014 (alloc_cand_and_find_basis): Restrict conditional candidate
17015 processing to CAND_MULTs.
17016
17017 2013-05-08 Jan Hubicka <jh@suse.cz>
17018
17019 PR lto/54095
17020 lto-symtab.c (lto_symtab_symbol_p): New function.
17021 (lto_symtab_resolve_can_prevail_p, lto_symtab_resolve_symbols,
17022 lto_symtab_resolve_symbols, lto_symtab_merge_decls_2,
17023 lto_symtab_merge_decls_1, lto_symtab_merge_cgraph_nodes_1):
17024 Skip static symbols.
17025
17026 2013-05-08 Paolo Carlini <paolo.carlini@oracle.com>
17027
17028 PR tree-optimization/57200
17029 * tree-ssa-loop-niter.c (do_warn_aggressive_loop_optimizations):
17030 Only call inform if the preceding warning_at returns true.
17031
17032 2013-05-07 Han Shen <shenhan@google.com>
17033
17034 * cfgexpand.c (record_or_union_type_has_array_p): New function.
17035 (expand_used_vars): Add logic handling '-fstack-protector-strong'.
17036 * common.opt (fstack-protector-strong): New option.
17037 * doc/cpp.texi (__SSP_STRONG__): New builtin "__SSP_STRONG__".
17038 * doc/invoke.texi (Optimization Options): Document
17039 "-fstack-protector-strong".
17040 * gcc.c (LINK_SSP_SPEC): Add 'fstack-protector-strong'.
17041
17042 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17043
17044 * config/mips/mips.c (mips_machine_reorg2): Return 0.
17045
17046 2013-05-07 Vladimir Makarov <vmakarov@redhat.com>
17047
17048 * ira.c (update_equiv_regs): Add insn having equiv memory even if
17049 it is not lhs of the insn.
17050 (setup_reg_equiv): Remove insn having equiv memory which it is not
17051 lhs of the insn.
17052 * lra-constraints.c (process_address): Try to improve generation
17053 code for address base + disp.
17054 (lra_constraints): Make correct the code for checking insn setting
17055 up backward equivalence. Remove insn only if it is in the init
17056 insn list.
17057 * lra-eliminations.c (update_reg_eliminate): Change return value.
17058 (lra_eliminate): Use the result.
17059
17060 2013-05-07 Uros Bizjak <ubizjak@gmail.com>
17061
17062 * config/i386/sse.md (ssescalarnummask): New mode attribute.
17063 (PEXTR_MODE, PEXTR_MODEx): New mode iterators.
17064 (*vec_extract<mode>): Merge from *sse4_1_pextrb_memory and
17065 *sse4_1_pextrw_memory using PEXTR_MODE mode iterator. Handle
17066 register target operands.
17067 (*vec_extractv8hi_sse2): New pattern.
17068 (*vec_extractv16qi_zext): Rename from *sse4_1_pextrb_<mode>.
17069 (*vec_extractv8hi_zext): Rename from *sse2_pextrw_<mode>.
17070 (*vec_extract<mode>_mem): New insn and split pattern.
17071
17072 2013-05-07 Christophe Lyon <christophe.lyon@linaro.org>
17073
17074 * config/arm/arm.c (arm_asan_shadow_offset): New function.
17075 (TARGET_ASAN_SHADOW_OFFSET): Define.
17076 * config/arm/linux-eabi.h (ASAN_CC1_SPEC): Define.
17077 (LINUX_OR_ANDROID_CC): Add ASAN_CC1_SPEC.
17078
17079 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17080
17081 * gimple-ssa-strength-reduction.c (MAX_INCR_VEC_LEN): New constant.
17082 (incr_vec_index): Return -1 if increment not found.
17083 (create_add_on_incoming_edge): Assert if increment not found.
17084 (record_increment): Limit number of increments recorded.
17085 (all_phi_incrs_profitable): Return false if an increment not found.
17086 (replace_profitable_candidates): Don't process increments that were
17087 not recorded.
17088 (analyze_candidates_and_replace): Limit size of incr_vec.
17089
17090 2013-05-07 Richard Biener <rguenther@suse.de>
17091
17092 * calls.c (special_function_p): setjmp-like functions are leaf.
17093 * builtins.def (BUILT_IN_SETJMP): setjmp is leaf.
17094 * tree-inline.c (update_ssa_across_abnormal_edges): Remove assert.
17095
17096 2013-05-07 Sofiane Naci <sofiane.naci@arm.com>
17097
17098 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): call splitter.
17099 (aarch64_simd_mov<mode>): New expander.
17100 (aarch64_simd_mov_to_<mode>low): New instruction pattern.
17101 (aarch64_simd_mov_to_<mode>high): Likewise.
17102 (aarch64_simd_mov_from_<mode>low): Likewise.
17103 (aarch64_simd_mov_from_<mode>high): Likewise.
17104 (aarch64_dup_lane<mode>): Update.
17105 (aarch64_dup_lanedi): New instruction pattern.
17106 * config/aarch64/aarch64-protos.h (aarch64_split_simd_move): New prototype.
17107 * config/aarch64/aarch64.c (aarch64_split_simd_move): New function.
17108
17109 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17110
17111 * gimple-ssa-strength-reduction.c (lazy_create_slsr_reg): Remove.
17112 (replace_mult_candidate): Remove unnecessary argument; remove
17113 unnecessary parameter from call to introduce_cast_before_cand.
17114 (replace_unconditional_candidate): Remove unnecessary parameter
17115 from call to replace_mult_candidate.
17116 (replace_conditional_candidate): Likewise.
17117 (insert_initializers): Use make_temp_ssa_name.
17118 (introduce_cast_before_cand): Remove unnecessary argument; use
17119 make_temp_ssa_name.
17120 (replace_one_candidate): Remove unnecessary argument; remove
17121 unnecessary parameter from calls to introduce_cast_before_cand.
17122 (replace_profitable_candidates): Remove unnecessary parameters
17123 from calls to replace_one_candidate.
17124
17125 2013-05-07 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17126
17127 * gimple-ssa-strength-reduction.c (find_phi_def): Don't record a
17128 phi def as possibly hiding a basis for a CAND_ADD whose operands
17129 have been commuted in the analysis.
17130 (alloc_cand_and_find_basis): Add parms to call to find_phi_def.
17131
17132 2013-05-07 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
17133
17134 * config/aarch64/aarch64.md
17135 (cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): Restrict the
17136 shift value between 0-4.
17137
17138 2013-05-07 Richard Biener <rguenther@suse.de>
17139
17140 * double-int.h (rshift): New overload.
17141 * double-int.c (rshift): New function.
17142 * tree-ssa-sccvn.c (copy_reference_ops_from_ref): Optimize.
17143 (create_reference_ops_from_ref): Remove.
17144 (vn_reference_insert): Use shared ops for constructing the
17145 reference and copy it.
17146
17147 2013-05-07 Richard Biener <rguenther@suse.de>
17148
17149 PR middle-end/57190
17150 * tree-eh.c (sink_clobbers): Properly propagate
17151 SSA_NAME_OCCURS_IN_ABNORMAL_PHI.
17152
17153 2013-05-07 Jakub Jelinek <jakub@redhat.com>
17154
17155 PR tree-optimization/57149
17156 * tree-ssa-uninit.c (uninit_undefined_value_p): New inline.
17157 (can_skip_redundant_opnd, compute_uninit_opnds_pos,
17158 collect_phi_def_edges, execute_late_warn_uninitialized): Use
17159 uninit_undefined_value_p instead of ssa_undefined_value_p.
17160
17161 PR debug/57184
17162 * expr.c (expand_expr_addr_expr_1): Handle COMPOUND_LITERAL_EXPR
17163 for modifier == EXPAND_INITIALIZER.
17164
17165 2013-05-07 Anton Blanchard <anton@samba.org>
17166
17167 * configure.ac (HAVE_LD_LARGE_TOC): Use correct linker emulation
17168 for powerpc64 little endian.
17169 * configure: Regenerate.
17170
17171 2013-05-06 Graham Stott <grahams@btinternet.com>
17172
17173 * expmed.c (init_expmed_rtl): Remove unused fields reg_fld, plus_fld,
17174 mult_fld, sdiv_fld1, udiv_fld1, sdiv_32_fld1, smod_32_fld1,
17175 wide_mult_fld1, wide_lshr_fld1, shift_fld1, shift_mult_fld1,
17176 shift_add_fld1, shift_sub0_fld1, shift_sub1_fld1.
17177
17178 2013-05-06 Graham Stott <grahams@btinternet.com>
17179
17180 * gensupport.c (add_predicate_code): Also exclude SCRATCH from rtx
17181 codes which allow non-lvalues.
17182
17183 2013-05-06 Marc Glisse <marc.glisse@inria.fr>
17184
17185 * tree.c (integer_all_onesp) <COMPLEX_CST>: Test that both
17186 components are all 1s.
17187 (integer_minus_onep): New function.
17188 * tree.h (integer_minus_onep): Declare it.
17189 * fold-const.c (fold_binary_loc) <MULT_EXPR>: Test
17190 integer_minus_onep instead of integer_all_onesp.
17191
17192 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17193
17194 PR target/52933
17195 * config/sh/sh.md (*cmp_div0s_0, *cmp_div0s_1, *movsicc_div0s): Add
17196 variations of these patterns.
17197
17198 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17199
17200 * config/i386/i386.md (isa): Add x64_sse4 member.
17201 (enabled): Handle x64_sse4.
17202 (*movdi_internal): Add *x->?r alternative to emit pextrq $0,%xmm,%reg
17203 instruction for 64bit SSE4_1 targets. Update insn attributes.
17204 (*movsi_internal): Add *x->?r alternative to emit pextrd $0,%xmm,%reg
17205 instruction for SSE4_1 targets. Update insn attributes.
17206 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17207 with *sse4_1_pextrd and *sse4_1_pextrq having const_0 selector.
17208 (*vec_extractv2di_1): Merge with *sse4_1_pextrq having
17209 const_1 selector.
17210 (*vec_extractv4si): Rename from *sse4_1_pextrd.
17211 (*vec_extractv4si_zext): Rename from *sse4_1_pextrd_zext.
17212 (*vec_extract<ssevecmodelower>_0 splitters): Merge splitters together.
17213
17214 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17215
17216 PR target/57108
17217 * config/sh/sh.md (tstsi_t_zero_extract_eq): Use QIHISIDI mode iterator.
17218
17219 2013-05-06 Maxim Kuznetsov <maks.kuznetsov@gmail.com>
17220
17221 * final.c (do_assembler_dialects): Don't handle curly braces and
17222 vertical bar escaped by % as dialect delimiters.
17223 (output_asm_insn): Print curly braces and vertical bar if escaped
17224 by % and ASSEMBLER_DIALECT defined.
17225 * doc/tm.texi.in (ASSEMBLER_DIALECT): Document new standard escapes.
17226 * doc/tm.texi: Regenerated.
17227
17228 2013-05-06 Steven Bosscher <steven@gcc.gnu.org>
17229
17230 * config/mips/mips.c: Include tree-pass.h.
17231 (mips_reorg): Split in pre- and post-dbr_schedule parts.
17232 (mips_machine_reorg2): Move mips_reorg post-dbr_schedule parts here.
17233 (pass_mips_machine_reorg2): New machine specific pass.
17234 (insert_pass_mips_machine_reorg2): New pass plugin definition.
17235 (mips_option_override): Register the new pass.
17236 * rtl.h (cleanup_barriers): Remove prototype.
17237 (dbr_schedule): Likewise.
17238 * jump.c (cleanup_barriers): Make static.
17239 * reorg.c (dbr_schedule): Likewise.
17240
17241 2013-05-06 Richard Biener <rguenther@suse.de>
17242
17243 PR tree-optimization/57185
17244 * tree-parloops.c (add_field_for_reduction): Handle anonymous
17245 SSA names properly.
17246
17247 2013-05-06 Uros Bizjak <ubizjak@gmail.com>
17248
17249 PR target/57106
17250 * config/i386/i386.c (add_parameter_dependencies): Add dependence
17251 between "first_arg" and "insn", not "last" and "insn".
17252
17253 2013-05-06 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17254
17255 * gimple-ssa-strength-reduction.c (slsr_process_phi): Re-enable.
17256 (find_candidates_in_block): Re-enable slsr_process_phi.
17257 (create_phi_basis): Fix double counting of candidate adjustment.
17258
17259 2013-05-06 Richard Biener <rguenther@suse.de>
17260
17261 PR middle-end/57147
17262 * tree-cfg.c (gimple_purge_dead_abnormal_call_edges): If
17263 the edge is also fallthru, preserve it and just clear the
17264 abnormal flag.
17265 * tree-cfgcleanup.c (remove_fallthru_edge): If the edge is
17266 also complex, preserve that and just clear the fallthru flag.
17267 * tree-inline.c (update_ssa_across_abnormal_edges): Also
17268 update virtual operands.
17269
17270 2013-05-06 Alan Modra <amodra@gmail.com>
17271
17272 * config/rs6000/linux.h (DEFAULT_ASM_ENDIAN): Define.
17273 (LINK_OS_LINUX_EMUL): Use ENDIAN_SELECT.
17274 * config/rs6000/linux64.h (DEFAULT_ASM_ENDIAN): Define.
17275 * config/rs6000/sysv4le.h (DEFAULT_ASM_ENDIAN): Define.
17276 (LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17277 * config/rs6000/sysv4.h (DEFAULT_ASM_ENDIAN): Define as -mbig.
17278
17279 2013-05-06 Alan Modra <amodra@gmail.com>
17280
17281 * config/rs6000/sysv4.h (ENDIAN_SELECT): Define, extracted from
17282 (ASM_SPEC): ..here. Emit DEFAULT_ASM_ENDIAN too.
17283 (DEFAULT_ASM_ENDIAN): Define.
17284 (CC1_SPEC, LINK_TARGET_SPEC): Use ENDIAN_SELECT.
17285 * config/rs6000/linux64.h (ASM_SPEC32): Remove endian options.
17286 Update -K PIC clause from sysv4.h.
17287 (ASM_SPEC_COMMON): Use ENDIAN_SELECT.
17288 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Likewise.
17289
17290 2013-05-06 Alan Modra <amodra@gmail.com>
17291
17292 * config/rs6000/rs6000.md (bswapdi 2nd splitter): Don't swap words
17293 twice for little-endian.
17294 (ashrdi3_no_power, ashrdi3): Support little-endian.
17295
17296 2013-05-06 Oleg Endo <olegendo@gcc.gnu.org>
17297
17298 PR target/55303
17299 * config/sh/sh.c (sh_rtx_costs): Handle SMIN and SMAX cases.
17300 * config/sh/sh.md (*clips, uminsi3, *clipu, clipu_one): New insns and
17301 related expanders.
17302 * config/sh/iterators.md (SMIN_SMAX): New code iterator.
17303 * config/sh/predicates.md (arith_reg_or_0_or_1_operand,
17304 clips_min_const_int, clips_max_const_int, clipu_max_const_int):
17305 New predicates.
17306
17307 2013-05-05 Steven Bosscher <steven@gcc.gnu.org>
17308 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
17309
17310 * config.gcc (hppa*-*-*): Remove MASK_BIG_SWITCH from CPU default.
17311 * config/pa/pa.opt: Make mbig-switch a no-op.
17312 * config/pa/pa.h (TARGET_DEFAULT): Remove MASK_BIG_SWITCH.
17313 (CASE_VECTOR_MODE): Always return SImode.
17314 (ASM_OUTPUT_ADDR_VEC_ELT, ASM_OUTPUT_ADDR_DIFF_ELT): Remove code
17315 for the !TARGET_BIG_SWITCH case.
17316 * config/pa/pa-linux.h: Likewise.
17317 * config/pa/pa-openbsd.h: Likewise.
17318 * config/pa/pa-hpux.h: Define TARGET_DEFAULT to 0.
17319 * config/pa/pa.md (short_jump): Remove define_insn.
17320 (casesi): Remove code for the !TARGET_BIG_SWITCH case.
17321 (casesi0): Remove define_insn.
17322 (type): Remove btable_branch.
17323 (pa_combine_type): Likewise.
17324 (in_nullified_branch_delay): Likewise.
17325 (in_call_delay): Likewise.
17326 (define_delay): Likewise.
17327 (define_insn_reservation "Z3"): Likewise.
17328 (define_insn_reservation "Z4"): Likewise.
17329 * config/pa/pa.c (pa_reorg): Remove code for !TARGET_BIG_SWITCH.
17330 (pa_adjust_insn_length): Remove adjustment for btable branches.
17331 * doc/invoke.texi (HPPA Options): Delete documentation for mbig-switch
17332 and mno-big-switch
17333
17334 2013-05-05 Uros Bizjak <ubizjak@gmail.com>
17335
17336 * config/i386/sse.md (*vec_extract<ssevecmodelower>_0): Merge
17337 from sse2_stored and *sse2_storeq_rex64 using SWI48 mode iterator.
17338 Add m->r,x alternatives.
17339 (*vec_extract<ssevecmodelower>_0 splitters): Merge V2DI and V4SI
17340 splitters using SWI48x mode iterator.
17341 (*vec_extract_v2di_0_sse): Rename from *sse2_storeq. Disable for
17342 TARGET_64BIT. Add m->x alternative.
17343 (*vec_extractv4si_mem): Rename from *vec_ext_v4si_mem.
17344 Add o->x alternative. Enable for TARGET_SSE.
17345 (sse_storeq): Remove expander.
17346 (*vec_extractv2di_1): Enable for TARGET_SSE. Split alternatives
17347 with memory input operand.
17348 (*vec_extractv2di_1 splitter): New.
17349 (*vec_extractv4sf_mem): Rename from *vec_extract_v4sf_mem.
17350 * config/i386/i386.md (ssevecmodelower): New mode attribute.
17351
17352 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17353
17354 * config/rs6000/rs6000.c (INT_P): Reformat. Delete obsolete comment.
17355 (INT_LOWPART): Delete.
17356 (extract_MB): Adjust.
17357 (extract_ME): Adjust.
17358 (print_operand): Adjust.
17359
17360 2013-05-04 Segher Boessenkool <segher@kernel.crashing.org>
17361
17362 * config/rs6000/predicates.md (reg_or_add_cint_operand,
17363 reg_or_sub_cint_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17364 (reg_or_logical_cint_operand, easy_fp_constant,
17365 logical_const_operand): Delete "CONST_DOUBLE" case.
17366 * config/rs6000/rs6000.c (num_insns_constant_wide): Delete
17367 "HOST_BITS_PER_WIDE_INT == 64" test.
17368 (num_insns_constant): Ditto. Delete CONST_DOUBLE DImode/VOIDmode case.
17369 (build_mask64_2_operands): Delete "HOST_BITS_PER_WIDE_INT >= 64" test.
17370 (rs6000_emit_set_const): Delete CONST_DOUBLE case.
17371 (rs6000_emit_set_long_const): Delete "HOST_BITS_PER_WIDE_INT >= 64"
17372 test.
17373 (includes_rldic_lshift_p, includes_rldicr_lshift_p): Delete
17374 CONST_DOUBLE DImode/VOIDmode case.
17375 (INT_P, INT_LOWPART): Delete CONST_DOUBLE case.
17376 (print_operand): Delete "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17377 CONST_DOUBLE VOIDmode case.
17378 (output_toc): Delete "HOST_BITS_PER_WIDE_INT == 32" case.
17379 (rs6000_rtx_costs): Delete CONST_DOUBLE DImode/VOIDmode case.
17380 * config/rs6000/rs6000.md (iordi3, xordi3, splitter for these):
17381 Delete CONST_DOUBLE case.
17382 (splitters for mov FMOVE64 const_double): Delete
17383 "HOST_BITS_PER_WIDE_INT == 32" case. Delete
17384 "HOST_BITS_PER_WIDE_INT >= 64" test.
17385 (splitter for mov DI const_int): Delete "HOST_BITS_PER_WIDE_INT == 32"
17386 case.
17387 (mov DI const_double): Delete.
17388
17389 2013-05-04 Jakub Jelinek <jakub@redhat.com>
17390
17391 * combine.c (combine_simplify_rtx) <case SUBREG>: If nonzero_bits
17392 on op shows all bits zero in mode of a lowpart subreg, return zero.
17393
17394 2013-05-03 Michael Meissner <meissner@linux.vnet.ibm.com>
17395
17396 PR target/57150
17397 * config/rs6000/rs6000.h (HARD_REGNO_CALLER_SAVE_MODE): Use DFmode
17398 to save TFmode registers and DImode to save TImode registers for
17399 caller save operations.
17400 (HARD_REGNO_CALL_PART_CLOBBERED): TFmode and TDmode do not need to
17401 mark being partially clobbered since they only use the first
17402 double word.
17403
17404 * config/rs6000/rs6000.c (rs6000_init_hard_regno_mode_ok): TFmode
17405 and TDmode only use the upper 64-bits of each VSX register.
17406
17407 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17408
17409 * gimple-ssa-strength-reduction.c (slsr_process_phi): Disable.
17410 (find_candidates_in_block): Disable slsr_process_phi.
17411
17412 2013-05-03 Guozhi Wei <carrot@google.com>
17413
17414 * coverage.c (coverage_obj_init): Move the construction of gcov
17415 constructor to ...
17416 (build_init_ctor): ... here.
17417
17418 2013-05-03 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
17419
17420 * gimple-ssa-strength-reduction.c (cand_kind): Add CAND_PHI.
17421 (slsr_cand_d): Redefine def_phi.
17422 (stride_status, phi_adjust_status, count_phis_status): New enums.
17423 (find_phi_def): New.
17424 (find_basis_for_base_expr): New.
17425 (find_basis_for_candidate): Handle hidden bases.
17426 (alloc_cand_and_find_basis): Handle phi candidates.
17427 (slsr_process_phi): New.
17428 (create_mul_ssa_cand): Exclude phi base candidates; use integer_onep.
17429 (create_mul_imm_cand): Likewise.
17430 (create_add_ssa_cand): Exclude phi base candidates.
17431 (create_add_imm_cand): Likewise.
17432 (slsr_process_cast): Likewise.
17433 (slsr_process_copy): Likewise.
17434 (find_candidates_in_block): Handle phi candidates.
17435 (dump_candidate): Likewise.
17436 (unconditional_cands): Delete.
17437 (unconditional_cands_with_known_stride_p): Delete.
17438 (phi_dependent_cand_p): New.
17439 (cand_increment): Handle phi-dependent candidates.
17440 (replace_dependent): Delete.
17441 (replace_mult_candidate): New.
17442 (replace_unconditional_candidate): New.
17443 (incr_vec_index): Move to avoid forward reference.
17444 (create_add_on_incoming_edge): New.
17445 (create_phi_basis): New.
17446 (replace_dependents): Delete.
17447 (replace_conditional_candidate): New.
17448 (phi_add_costs): New.
17449 (replace_uncond_cands_and_profitable_phis): New.
17450 (record_increment): Handle phi adjustments.
17451 (record_phi_increments): New.
17452 (record_increments): Handle phi adjustments.
17453 (phi_incr_cost): New.
17454 (lowest_cost_path): Handle phis.
17455 (total_savings): Likewise.
17456 (analyze_increments): Likewise.
17457 (ncd_with_phi): New.
17458 (ncd_of_cand_and_phis): New.
17459 (nearest_common_dominator_for_cands): Handle phi increments.
17460 (all_phi_incrs_profitable): New.
17461 (replace_profitable_candidates): Handle phi-dependent candidates.
17462 (analyze_candidates_and_replace): Likewise.
17463
17464 2013-05-03 Teresa Johnson <tejohnson@google.com>
17465
17466 PR bootstrap/57154
17467 * sched-rgn.c (compute_dom_prob_ps): Ensure accumulated probabilities
17468 do not exceed REG_BR_PROB_BASE.
17469
17470 2013-05-03 Jeff Law <law@redhat.com>
17471
17472 PR tree-optimization/57144
17473 * tree-vrp.c (simplify_cond_using_ranges): Verify the constant
17474 operand of the condition will bit into the new type when eliminating
17475 a cast feeding a condition.
17476
17477 2013-05-03 Jakub Jelinek <jakub@redhat.com>
17478
17479 PR rtl-optimization/57130
17480 * combine.c (make_compound_operation) <case SUBREG>: Pass SET instead
17481 of COMPARE as in_code to the recursive call if needed.
17482
17483 2013-05-03 Uros Bizjak <ubizjak@gmail.com>
17484
17485 * config/i386/i386.md (isa): Add x64_sse4_noavx and x64_avx members.
17486 (enabled): Handle new members.
17487 * config/i386/sse.md (*vec_concatv2si): Merge from
17488 *vec_concatv2si_sse2 and vec_concatv2si_sse.
17489 (vec_concatv2di): Merge with *vec_concatv2di_rex64.
17490
17491 2013-05-03 Joern Rennecke <joern.rennecke@embecosm.com>
17492
17493 PR tree-optimization/57027
17494 * tree-ssa-math-opts.c (convert_mult_to_fma): When checking
17495 for fnms opportunity, check we got the prerequisite kind
17496 of tree / gimple before using accessor functions.
17497
17498 2013-05-03 Richard Biener <rguenther@suse.de>
17499
17500 * double-int.h (lshift): New overload without precision
17501 and arith argument.
17502 (operator *=, operator +=, operator -=): Move ...
17503 * double-int.c (operator *=, operator +=, operator -=): ... here
17504 and implement more efficiently.
17505 (mul_double_with_sign): Remove.
17506 (lshift_double): Adjust to take unsinged shift argument, push
17507 dispatching code to callers.
17508 (mul_double_wide_with_sign): Add early out for callers that
17509 are not interested in high parts or overflow.
17510 (lshift): New function.
17511 (lshift, rshift, alshift, arshift, llshift, lrshift): Add
17512 dispatch code here.
17513 (lrotate, rrotate): Use logical shifts.
17514 * expr.c (get_inner_reference): Use lshift.
17515 * fixed-value.c (do_fixed_divide): Likewise.
17516 * tree-dfa.c (get_ref_base_and_extent): Likewise.
17517 * tree-ssa-alias.c (indirect_ref_may_alias_decl_p): Likewise.
17518 (indirect_refs_may_alias_p): Likewise.
17519 (stmt_kills_ref_p_1): Likewise.
17520
17521 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17522
17523 * config/aarch64/aarch64-simd.md (simd_fabd): Correct the description.
17524
17525 2013-05-03 Vidya Praveen <vidyapraveen@arm.com>
17526
17527 * config/aarch64/aarch64-simd.md (*fabd_scalar<mode>3): Support
17528 scalar form of FABD instruction.
17529
17530 2013-05-02 Vladimir Makarov <vmakarov@redhat.com>
17531
17532 * lra-constraints.c (process_alt_operands): Add checking alt
17533 number to choose the best alternative.
17534
17535 2013-05-02 Richard Biener <rguenther@suse.de>
17536
17537 * tree-eh.c (cleanup_empty_eh_merge_phis): Remove rename_virts
17538 bitmap and its handling.
17539 (pass_cleanup_eh): Set todo_flags_finish to TODO_verify_ssa.
17540
17541 2013-05-02 Richard Biener <rguenther@suse.de>
17542
17543 PR middle-end/57140
17544 * tree-inline.c (copy_loops): Properly handle removed loops.
17545 (copy_cfg_body): Mark destination loops for fixup if source
17546 loops needed fixup.
17547
17548 2013-05-02 Greta Yorsh <Greta.Yorsh@arm.com>
17549
17550 PR target/56732
17551 * config/arm/arm.c (arm_expand_epilogue): Check really_return before
17552 generating simple_return for naked functions.
17553
17554 2013-05-02 Martin Jambor <mjambor@suse.cz>
17555
17556 PR middle-end/56988
17557 * ipa-prop.h (ipa_agg_replacement_value): New flag by_ref.
17558 * ipa-cp.c (ipa_get_indirect_edge_target_1): Also check that by_ref
17559 flags match.
17560 (find_aggregate_values_for_callers_subset): Fill in the by_ref flag of
17561 ipa_agg_replacement_value structures.
17562 (known_aggs_to_agg_replacement_list): Likewise.
17563 * ipa-prop.c (write_agg_replacement_chain): Stream by_ref flag.
17564 (read_agg_replacement_chain): Likewise.
17565 (ipcp_transform_function): Also check that by_ref flags match.
17566
17567 2013-05-02 Richard Biener <rguenther@suse.de>
17568
17569 * graphds.h (struct graph): Add obstack member.
17570 * graphds.c (new_graph): Initialize obstack and allocate
17571 vertices from it.
17572 (add_edge): Allocate edge from the obstack.
17573 (free_graph): Free the obstack instead of all edges and vertices.
17574
17575 2013-05-02 Teresa Johnson <tejohnson@google.com>
17576
17577 * loop-unswitch.c (unswitch_loop): Use helper routines with rounding
17578 divides.
17579 * cfg.c (update_bb_profile_for_threading): Ditto.
17580 * tree-inline.c (copy_bb): Ditto.
17581 (copy_edges_for_bb): Ditto.
17582 (initialize_cfun): Ditto.
17583 (copy_cfg_body): Ditto.
17584 (expand_call_inline): Ditto.
17585 * ipa-inline-analysis.c (estimate_edge_size_and_time): Ditto.
17586 (estimate_node_size_and_time): Ditto.
17587 (inline_merge_summary): Ditto.
17588 * cgraphclones.c (cgraph_clone_edge): Ditto.
17589 (cgraph_clone_node): Ditto.
17590 * sched-rgn.c (compute_dom_prob_ps): Ditto.
17591 (compute_trg_info): Ditto.
17592
17593 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17594
17595 * config/aarch64/aarch64.md (movsi_aarch64): Only allow to/from
17596 S reg when fp attribute set.
17597 (movdi_aarch64): Only allow to/from D reg when fp attribute set.
17598
17599 2013-05-02 Ian Bolton <ian.bolton@arm.com>
17600
17601 * config/aarch64/aarch64.md (*and_one_cmpl<mode>3_compare0):
17602 New pattern.
17603 (*and_one_cmplsi3_compare0_uxtw): Likewise.
17604 (*and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
17605 (*and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
17606
17607 2013-05-02 Richard Biener <rguenther@suse.de>
17608
17609 * tree-scalar-evolution.c (scev_info_hasher): Remove.
17610 (struct instantiate_cache_entry): New type.
17611 (struct instantiate_cache_entry_hasher): New hashtable descriptor.
17612 (struct instantiate_cache_type): New type.
17613 (set_instantiated_value, get_instantiated_value): Remove.
17614 (get_instantiated_value_entry): New function.
17615 (instantiate_scev_name): Use the new cache and adjust.
17616 (instantiate_scev_poly): Adjust.
17617 (instantiate_scev_binary): Likewise.
17618 (instantiate_array_ref): Likewise.
17619 (instantiate_scev_convert): Likewise.
17620 (instantiate_scev_not): Likewise.
17621 (instantiate_scev_3): Likewise.
17622 (instantiate_scev_2): Likewise.
17623 (instantiate_scev_r): Likewise.
17624 (instantiate_scev): Likewise.
17625 (resolve_mixers): Likewise.
17626
17627 2013-05-01 Vladimir Makarov <vmakarov@redhat.com>
17628
17629 PR target/57091
17630 * lra-constraints.c (best_small_class_operands_num): Remove.
17631 (process_alt_operands): Remove small_class_operands_num. Take
17632 small classes operands into losers and only if the operand is not
17633 matched. Modify debugging output.
17634 (curr_insn_transform): Remove best_small_class_operands_num.
17635 Print insn name.
17636
17637 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17638
17639 * config/aarch64/aarch64-builtins.c
17640 (aarch64_gimple_fold_builtin.c): Fold more modes for reduc_splus_.
17641 * config/aarch64/aarch64-simd-builtins.def
17642 (reduc_splus_): Add new modes.
17643 (reduc_uplus_): New.
17644 * config/aarch64/aarch64-simd.md (aarch64_addvv4sf): Remove.
17645 (reduc_uplus_v4sf): Likewise.
17646 (reduc_splus_v4sf): Likewise.
17647 (aarch64_addv<mode>): Likewise.
17648 (reduc_uplus_<mode>): Likewise.
17649 (reduc_splus_<mode>): Likewise.
17650 (aarch64_addvv2di): Likewise.
17651 (reduc_uplus_v2di): Likewise.
17652 (reduc_splus_v2di): Likewise.
17653 (aarch64_addvv2si): Likewise.
17654 (reduc_uplus_v2si): Likewise.
17655 (reduc_splus_v2si): Likewise.
17656 (reduc_<sur>plus_<mode>): New.
17657 (reduc_<sur>plus_v2di): Likewise.
17658 (reduc_<sur>plus_v2si): Likewise.
17659 (reduc_<sur>plus_v4sf): Likewise.
17660 (aarch64_addpv4sf): Likewise.
17661 * config/aarch64/arm_neon.h
17662 (vaddv<q>_<s,u,f><8, 16, 32, 64): Rewrite using builtins.
17663 * config/aarch64/iterators.md (unspec): Remove UNSPEC_ADDV,
17664 add UNSPEC_SADDV, UNSPEC_UADDV.
17665 (SUADDV): New.
17666 (sur): Add UNSPEC_SADDV, UNSPEC_UADDV.
17667
17668 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17669
17670 * config/aarch64/arm_neon.h
17671 (v<max,min><nm><q><v>_<sfu><8, 16, 32, 64>): Rewrite using builtins.
17672
17673 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17674
17675 * config/aarch64/aarch64-builtins
17676 (aarch64_gimple_fold_builtin): Fold reduc_<su><maxmin>_ builtins.
17677
17678 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17679
17680 * config/aarch64/aarch64-simd-builtins.def
17681 (reduc_smax_): New.
17682 (reduc_smin_): Likewise.
17683 (reduc_umax_): Likewise.
17684 (reduc_umin_): Likewise.
17685 (reduc_smax_nan_): Likewise.
17686 (reduc_smin_nan_): Likewise.
17687 (fmax): Remove.
17688 (fmin): Likewise.
17689 (smax): Update for V2SF, V4SF and V2DF modes.
17690 (smin): Likewise.
17691 (smax_nan): New.
17692 (smin_nan): Likewise.
17693 * config/aarch64/aarch64-simd.md (<maxmin><mode>3): Rename to...
17694 (<su><maxmin><mode>3): ...This, refactor.
17695 (s<maxmin><mode>3): New.
17696 (<maxmin_uns><mode>3): Likewise.
17697 (reduc_<maxmin_uns>_<mode>): Refactor.
17698 (reduc_<maxmin_uns>_v4sf): Likewise.
17699 (reduc_<maxmin_uns>_v2si): Likewise.
17700 (aarch64_<fmaxmin><mode>: Remove.
17701 * config/aarch64/arm_neon.h (vmax<q>_f<32,64>): Rewrite to use
17702 new builtin names.
17703 (vmin<q>_f<32,64>): Likewise.
17704 * config/iterators.md (unspec): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17705 (FMAXMIN): New.
17706 (su): Add mappings for smax, smin, umax, umin.
17707 (maxmin): New.
17708 (FMAXMINV): Add UNSPEC_FMAXNMV, UNSPEC_FMINNMV.
17709 (FMAXMIN): Rename as...
17710 (FMAXMIN_UNS): ...This.
17711 (maxminv): Remove.
17712 (fmaxminv): Likewise.
17713 (fmaxmin): Likewise.
17714 (maxmin_uns): New.
17715 (maxmin_uns_op): Likewise.
17716
17717 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17718
17719 * config/aarch64/arm_neon.h
17720 (vac<ge, gt><sd>_f<32, 64>): Rename to...
17721 (vca<ge, gt><sd>_f<32, 64>): ...this, reimpliment in C.
17722 (vca<ge, gt, lt, le><q>_f<32, 64>): Reimpliment in C.
17723
17724 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17725
17726 * config/aarch64/aarch64-simd.md (*aarch64_fac<optab><mode>): New.
17727 * config/aarch64/iterators.md (FAC_COMPARISONS): New.
17728
17729 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17730
17731 * config/aarch64/aarch64-simd.md
17732 (vcond<mode>_internal): Handle special cases for constant masks.
17733 (vcond<mode><mode>): Allow nonmemory_operands for outcome vectors.
17734 (vcondu<mode><mode>): Likewise.
17735 (vcond<v_cmp_result><mode>): New.
17736
17737 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17738
17739 * config/aarch64/aarch64-builtins.c (BUILTIN_VALLDI): Define.
17740 (aarch64_fold_builtin): Add folding for cm<eq,ge,gt,tst>.
17741 * config/aarch64/aarch64-simd-builtins.def
17742 (cmeq): Update to BUILTIN_VALLDI.
17743 (cmgt): Likewise.
17744 (cmge): Likewise.
17745 (cmle): Likewise.
17746 (cmlt): Likewise.
17747 * config/aarch64/arm_neon.h
17748 (vc<eq, lt, le, gt, ge, tst><z><qsd>_<fpsu><8,16,32,64>): Remap
17749 to builtins or C as appropriate.
17750
17751 2013-05-01 James Greenhalgh <james.greenhalgh@arm.com>
17752
17753 * config/aarch64/aarch64-simd-builtins.def (cmhs): Rename to...
17754 (cmgeu): ...This.
17755 (cmhi): Rename to...
17756 (cmgtu): ...This.
17757 * config/aarch64/aarch64-simd.md
17758 (simd_mode): Add SF.
17759 (aarch64_vcond_internal): Use new names for unsigned comparison insns.
17760 (aarch64_cm<optab><mode>): Rewrite to not use UNSPECs.
17761 * config/aarch64/aarch64.md (*cstore<mode>_neg): Rename to...
17762 (cstore<mode>_neg): ...This.
17763 * config/aarch64/iterators.md
17764 (VALLF): new.
17765 (unspec): Remove UNSPEC_CM<EQ, LE, LT, GE, GT, HS, HI, TST>.
17766 (COMPARISONS): New.
17767 (UCOMPARISONS): Likewise.
17768 (optab): Add missing comparisons.
17769 (n_optab): New.
17770 (cmp_1): Likewise.
17771 (cmp_2): Likewise.
17772 (CMP): Likewise.
17773 (cmp): Remove.
17774 (VCMP_S): Likewise.
17775 (VCMP_U): Likewise.
17776 (V_cmp_result): Add DF, SF modes.
17777 (v_cmp_result): Likewise.
17778 (v): Likewise.
17779 (vmtype): Likewise.
17780 * config/aarch64/predicates.md (aarch64_reg_or_fp_zero): New.
17781
17782 2013-05-01 Greta Yorsh <Greta.Yorsh@arm.com>
17783
17784 * config/arm/thumb2.md (thumb2_smaxsi3,thumb2_sminsi3): Convert
17785 define_insn to define_insn_and_split.
17786 (thumb32_umaxsi3,thumb2_uminsi3): Likewise.
17787 (thumb2_negdi2,thumb2_abssi2,thumb2_neg_abssi2): Likewise.
17788 (thumb2_mov_scc,thumb2_mov_negscc,thumb2_mov_notscc): Likewise.
17789 (thumb2_movsicc_insn,thumb2_and_scc,thumb2_ior_scc): Likewise.
17790 (thumb2_negscc): Likewise.
17791
17792 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
17793
17794 * config/arm/thumb2.md (thumb2_incscc, thumb2_decscc): Delete.
17795
17796 2013-04-30 Greta Yorsh <Greta.Yorsh@arm.com>
17797
17798 * config/arm/thumb2.md: Remove trailing whitespaces.
17799
17800 2013-04-30 Richard Sandiford <rsandifo@linux.vnet.ibm.com>
17801
17802 * explow.c (plus_constant): Pass "mode" to immed_double_int_const.
17803 Use gen_int_mode rather than GEN_INT.
17804
17805 2013-04-30 H.J. Lu <hongjiu.lu@intel.com>
17806
17807 * value-prof.c (stream_in_histogram_value): Remove the strayed
17808 debug_gimple_stmt.
17809
17810 2013-04-30 Richard Biener <rguenther@suse.de>
17811
17812 PR middle-end/57122
17813 * cfghooks.c (split_edge): Properly check for the loop latch edge.
17814
17815 2013-04-30 Richard Biener <rguenther@suse.de>
17816
17817 PR middle-end/57107
17818 * tree-eh.c (sink_clobbers): Preserve virtual SSA form.
17819
17820 2013-04-30 Andrey Belevantsev <abel@ispras.ru>
17821
17822 PR rtl-optimization/56957
17823 PR rtl-optimization/57105
17824 * sel-sched.c (move_op_orig_expr_found): Remove insn_emitted
17825 variable. Use just INSN_UID for determining whether an insn
17826 should be only disconnected from the insn stream.
17827 * sel-sched-ir.h (EXPR_WAS_CHANGED): Remove.
17828
17829 2013-04-30 Jakub Jelinek <jakub@redhat.com>
17830
17831 PR tree-optimization/57104
17832 * tsan.c (instrument_expr): Don't instrument accesses to
17833 DECL_HARD_REGISTER VAR_DECLs.
17834
17835 2013-04-30 Richard Biener <rguenther@suse.de>
17836
17837 * function.h (loops_for_fn): New inline function.
17838 (set_loops_for_fn): Likewise.
17839 * cfgloop.h (place_new_loop): Add struct function parameter.
17840 (get_loop): Likewise.
17841 (get_loops): Likewise.
17842 (number_of_loops): Likewise.
17843 (fel_next): Adjust.
17844 (fel_init): Likewise.
17845 * cfg.c (get_loop_copy): Adjust.
17846 * cfgloop.c (flow_loops_dump): Likewise.
17847 (record_loop_exits): Likewise.
17848 (verify_loop_structure): Likewise.
17849 * cfgloopanal.c (mark_irreducible_loops): Likewise.
17850 (estimate_reg_pressure_cost): Likewise.
17851 (mark_loop_exit_edges): Likewise.
17852 * cfgloopmanip.c (place_new_loop): Likewise.
17853 (add_loop): Likewise.
17854 (duplicate_loop): Likewise.
17855 * graph.c (draw_cfg_nodes): Likewise.
17856 * graphite-clast-to-gimple.c (translate_clast_user): Likewise.
17857 * graphite-sese-to-poly.c (build_scop_scattering): Likewise.
17858 (extract_affine_chrec): Likewise.
17859 (build_scop_iteration_domain): Likewise.
17860 * graphite.c (graphite_initialize): Likewise.
17861 * ira-build.c (create_loop_tree_nodes): Likewise.
17862 (more_one_region_p): Likewise.
17863 (rebuild_regno_allocno_maps): Likewise.
17864 (mark_loops_for_removal): Likewise.
17865 (mark_all_loops_for_removal): Likewise.
17866 (remove_unnecessary_regions): Likewise.
17867 (ira_build): Likewise.
17868 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
17869 * loop-init.c (fix_loop_structure): Likewise.
17870 (gate_rtl_move_loop_invariants): Likewise.
17871 (gate_rtl_unswitch): Likewise.
17872 (gate_rtl_unroll_and_peel_loops): Likewise.
17873 (rtl_doloop): Likewise.
17874 * lto-streamer-in.c (input_cfg): Likewise.
17875 * lto-streamer-out.c (output_cfg): Likewise.
17876 * modulo-sched.c (sms_schedule): Likewise.
17877 * predict.c (tree_estimate_probability): Likewise.
17878 (tree_estimate_probability_driver): Likewise.
17879 (estimate_loops): Likewise.
17880 * tree-cfg.c (fixup_loop_arrays_after_move): Likewise.
17881 (move_sese_region_to_fn): Likewise.
17882 (debug_loop_num): Likewise.
17883 * tree-chrec.c (chrec_evaluate): Likewise.
17884 (hide_evolution_in_other_loops_than_loop): Likewise.
17885 (chrec_component_in_loop_num): Likewise.
17886 (reset_evolution_in_loop): Likewise.
17887 (evolution_function_is_invariant_rec_p): Likewise.
17888 * tree-if-conv.c (main_tree_if_conversion): Likewise.
17889 * tree-inline.c (copy_loops): Likewise.
17890 (copy_cfg_body): Likewise.
17891 (tree_function_versioning): Likewise.
17892 * tree-loop-distribution.c (rdg_flag_loop_exits): Likewise.
17893 * tree-scalar-evolution.c (chrec_contains_symbols_defined_in_loop):
17894 Likewise.
17895 (add_to_evolution_1): Likewise.
17896 (scev_const_prop): Likewise.
17897 * tree-scalar-evolution.h (get_chrec_loop): Likewise.
17898 * tree-ssa-loop-ch.c (copy_loop_headers): Likewise.
17899 * tree-ssa-loop-im.c (analyze_memory_references): Likewise.
17900 (tree_ssa_lim_initialize): Likewise.
17901 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Likewise.
17902 (verify_loop_closed_ssa): Likewise.
17903 * tree-ssa-loop.c (tree_ssa_loop_init): Likewise.
17904 (tree_ssa_loop_im): Likewise.
17905 (tree_ssa_loop_unswitch): Likewise.
17906 (tree_vectorize): Likewise.
17907 (check_data_deps): Likewise.
17908 (tree_ssa_loop_ivcanon): Likewise.
17909 (tree_ssa_loop_bounds): Likewise.
17910 (tree_complete_unroll): Likewise.
17911 (tree_complete_unroll_inner): Likewise.
17912 (tree_parallelize_loops): Likewise.
17913 (tree_ssa_loop_prefetch): Likewise.
17914 (tree_ssa_loop_ivopts): Likewise.
17915 * tree-ssa.c (execute_update_addresses_taken): Liekwise.
17916 * tree-vectorizer.c (vectorize_loops): Likewise.
17917
17918 2013-04-29 Mike Frysinger <vapier@gentoo.org>
17919
17920 * config/arm/bpabi.h (EABI_LINK_SPEC): Define.
17921 (BPABI_LINK_SPEC): Use new EABI_LINK_SPEC.
17922 * config/arm/linux-eabi.h (LINK_SPEC): Replace BE8_LINK_SPEC
17923 with EABI_LINK_SPEC.
17924
17925 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17926
17927 PR target/44578
17928 * config/i386/i386.md (*zero_extendsidi2): Add "!" to m->?*y
17929 alternative.
17930
17931 2013-04-29 Vladimir Makarov <vmakarov@redhat.com>
17932
17933 PR target/57097
17934 * lra-constraints.c (process_alt_operands): Discourage a bit more
17935 using memory for pseudos. Print cost dump for alternatives.
17936 Modify cost values for conflicts with early clobbers.
17937 (curr_insn_transform): Spill pseudos reassigned to NO_REGS.
17938
17939 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
17940
17941 PR target/57098
17942 * config/i386/i386.c (ix86_expand_vec_perm): Validize constant memory.
17943
17944 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17945
17946 * config/aarch64/aarch64.md (movsi_aarch64): Support LDR/STR
17947 from/to S register.
17948 (movdi_aarch64): Support LDR/STR from/to D register.
17949
17950 2013-04-29 Ian Bolton <ian.bolton@arm.com>
17951
17952 * common/config/aarch64/aarch64-common.c: Enable REE pass at O2
17953 or higher by default.
17954
17955 2013-04-29 Richard Biener <rguenther@suse.de>
17956
17957 PR middle-end/57075
17958 * tree-inline.c (copy_edges_for_bb): Still split the bbs,
17959 even if not adding abnormal edges for calls that can make
17960 abnormal gotos.
17961
17962 2013-04-29 Richard Biener <rguenther@suse.de>
17963
17964 PR middle-end/57103
17965 * tree-cfg.c (move_stmt_op): Fix condition under which to update
17966 TREE_BLOCK.
17967 (move_stmt_r): Remove redundant checking.
17968
17969 2013-04-29 Teresa Johnson <tejohnson@google.com>
17970
17971 PR bootstrap/57077
17972 * basic-block.h (apply_scale): New function.
17973 (apply_probability): Use apply_scale.
17974 * gimple-streamer-in.c (input_bb): Ditto.
17975 * lto-streamer-in.c (input_cfg): Ditto.
17976 * lto-cgraph.c (merge_profile_summaries): Ditto.
17977 * tree-optimize.c (execute_fixup_cfg): Ditto.
17978 * tree-inline.c (copy_bb): Update comment to use apply_scale.
17979 (copy_edges_for_bb): Ditto.
17980 (copy_cfg_body): Ditto.
17981
17982 2013-04-29 Tom de Vries <tom@codesourcery.com>
17983
17984 * tree-ssa-tail-merge.c (find_same_succ_bb): Skip loop latch bbs.
17985 (replace_block_by): Don't set LOOPS_NEED_FIXUP.
17986 (tail_merge_optimize): Handle current_loops == NULL.
17987
17988 2013-04-26 Jeff Law <law@redhat.com>
17989
17990 * tree-vrp.c (range_fits_type_p): Move to earlier point in file.
17991 (simplify_cond_using_ranges): Generalize code to simplify
17992 COND_EXPRs where one argument is a constant and the other
17993 is an SSA_NAME created by an integral type conversion.
17994
17995 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
17996
17997 * config/arm/arm.md (store_minmaxsi): Use only when
17998 optimize_insn_for_size_p.
17999
18000 2013-04-29 Christian Bruel <christian.bruel@st.com>
18001
18002 PR target/57108
18003 * sh.md (tstsi_t_zero_extract_eq): Set mode for operand 0.
18004
18005 2013-04-29 Richard Biener <rguenther@suse.de>
18006
18007 PR middle-end/57089
18008 * omp-low.c (expand_omp_taskreg): If the parent function had a broken
18009 loop tree make sure to schedule a fixup for the child as well.
18010 (expand_omp_for_generic): Properly add loops.
18011 (expand_omp_for_static_nochunk): Likewise.
18012 (expand_omp_for_static_chunk): Likewise.
18013 (expand_omp_for): For the degenerate case fixup loops.
18014 (expand_omp_sections): Fix default bb placement in loops.
18015 (expand_omp_atomic_pipeline): Properly add loops.
18016
18017 2013-04-29 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18018
18019 * predict.c: Fix typo in comment above #define PROB_VERY_UNLIKELY.
18020
18021 2013-04-29 Tom de Vries <tom@codesourcery.com>
18022
18023 * tree-ssa-tail-merge.c: Update header comment.
18024
18025 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18026
18027 * config/aarch64/arm_neon.h
18028 (vcvt<sd>_f<32,64>_s<32,64>): Rewrite in C.
18029 (vcvt<q>_f<32,64>_s<32,64>): Rewrite using builtins.
18030 (vcvt_<high_>_f<32,64>_f<32,64>): Likewise.
18031 (vcvt<qsd>_<su><32,64>_f<32,64>): Likewise.
18032 (vcvta<qsd>_<su><32,64>_f<32,64>): Likewise.
18033 (vcvtm<qsd>_<su><32,64>_f<32,64>): Likewise.
18034 (vcvtn<qsd>_<su><32,64>_f<32,64>): Likewise.
18035 (vcvtp<qsd>_<su><32,64>_f<32,64>): Likewise.
18036
18037 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18038
18039 * config/aarch64/aarch64-simd.md
18040 (<optab><VDQF:mode><fcvt_target>2): New, maps to fix, fixuns.
18041 (<fix_trunc_optab><VDQF:mode><fcvt_target>2): New, maps to
18042 fix_trunc, fixuns_trunc.
18043 (ftrunc<VDQF:mode>2): New.
18044 * config/aarch64/iterators.md (optab): Add fix, fixuns.
18045 (fix_trunc_optab): New.
18046
18047 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18048
18049 * config/aarch64/aarch64-builtins.c
18050 (aarch64_builtin_vectorized_function): Vectorize over ifloorf,
18051 iceilf, lround, iroundf.
18052
18053 2013-04-29 Uros Bizjak <ubizjak@gmail.com>
18054
18055 PR target/54349
18056 * config/i386/i386.h (enum ix86_tune_indices)
18057 <X86_TUNE_INTER_UNIT_MOVES_TO_VEC, X86_TUNE_INTER_UNIT_MOVES_FROM_VEC>:
18058 New, split from X86_TUNE_INTER_UNIT_MOVES.
18059 <X86_TUNE_INTER_UNIT_MOVES>: Remove.
18060 (TARGET_INTER_UNIT_MOVES_TO_VEC): New define.
18061 (TARGET_INTER_UNIT_MOVES_FROM_VEC): Ditto.
18062 (TARGET_INTER_UNIT_MOVES): Remove.
18063 * config/i386/i386.c (initial_ix86_tune_features): Update.
18064 Disable X86_TUNE_INTER_UNIT_MOVES_FROM_VEC for m_ATHLON_K8 only.
18065 (ix86_expand_convert_uns_didf_sse): Use
18066 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18067 (ix86_expand_vector_init_one_nonzero): Ditto.
18068 (ix86_expand_vector_init_interleave): Ditto.
18069 (inline_secondary_memory_needed): Return true for moves from SSE class
18070 registers for !TARGET_INTER_UNIT_MOVES_FROM_VEC targets and for moves
18071 to SSE class registers for !TARGET_INTER_UNIT_MOVES_TO_VEC targets.
18072 * config/i386/constraints.md (Yi, Ym): Depend on
18073 TARGET_INTER_UNIT_MOVES_TO_VEC.
18074 (Yj, Yn): New constraints.
18075 * config/i386/i386.md (*movdi_internal): Change constraints of
18076 operand 1 from Yi to Yj and from Ym to Yn.
18077 (*movsi_internal): Ditto.
18078 (*movdf_internal): Ditto.
18079 (*movsf_internal): Ditto.
18080 (*float<SWI48x:mode><X87MODEF:mode>2_1): Use
18081 TARGET_INTER_UNIT_MOVES_TO_VEC instead of TARGET_INTER_UNIT_MOVES.
18082 (*float<SWI48x:mode><X87MODEF:mode>2_1 splitters): Ditto.
18083 (floatdi<X87MODEF:mode>2_i387_with_xmm): Ditto.
18084 (floatdi<X87MODEF:mode>2_i387_with_xmm splitters): Ditto.
18085 * config/i386/sse.md (movdi_to_sse): Ditto.
18086 (sse2_stored): Change constraint of operand 1 from Yi to Yj.
18087 Use TARGET_INTER_UNIT_MOVES_FROM_VEC instead of
18088 TARGET_INTER_UNIT_MOVES.
18089 (sse_storeq_rex64): Change constraint of operand 1 from Yi to Yj.
18090 (sse_storeq_rex64 splitter): Use TARGET_INTER_UNIT_MOVES_FROM_VEC
18091 instead of TARGET_INTER_UNIT_MOVES.
18092 * config/i386/mmx.md (*mov<mode>_internal): Change constraint of
18093 operand 1 from Yi to Yj and from Ym to Yn.
18094
18095 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18096
18097 * config/aarch64/aarch64-simd-builtins.def (vec_unpacks_hi_): New.
18098 (float_truncate_hi_): Likewise.
18099 (float_extend_lo_): Likewise.
18100 (float_truncate_lo_): Likewise.
18101 * config/aarch64/aarch64-simd.md (vec_unpacks_lo_v4sf): New.
18102 (aarch64_float_extend_lo_v2df): Likewise.
18103 (vec_unpacks_hi_v4sf): Likewise.
18104 (aarch64_float_truncate_lo_v2sf): Likewise.
18105 (aarch64_float_truncate_hi_v4sf): Likewise.
18106 (vec_pack_trunc_v2df): Likewise.
18107 (vec_pack_trunc_df): Likewise.
18108
18109 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18110
18111 * config/aarch64/aarch64-builtins.c
18112 (aarch64_fold_builtin): Fold float conversions.
18113 * config/aarch64/aarch64-simd-builtins.def
18114 (floatv2si, floatv4si, floatv2di): New.
18115 (floatunsv2si, floatunsv4si, floatunsv2di): Likewise.
18116 * config/aarch64/aarch64-simd.md
18117 (<optab><fcvt_target><VDQF:mode>2): New, expands to float and floatuns.
18118 * config/aarch64/iterators.md (FLOATUORS): New.
18119 (optab): Add float, floatuns.
18120 (su_optab): Likewise.
18121
18122 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18123
18124 * config/aarch64/aarch64-builtins.c
18125 (aarch64_builtin_vectorized_function): Use new names for
18126 fcvt builtins.
18127 * config/aarch64/aarch64-simd-builtins.def (fcvtzs): Split as...
18128 (lbtruncv2sf, lbtruncv4sf, lbtruncv2df): ...This.
18129 (fcvtzu): Split as...
18130 (lbtruncuv2sf, lbtruncuv4sf, lbtruncuv2df): ...This.
18131 (fcvtas): Split as...
18132 (lroundv2sf, lroundv4sf, lroundv2df, lroundsf, lrounddf): ...This.
18133 (fcvtau): Split as...
18134 (lrounduv2sf, lrounduv4sf, lrounduv2df, lroundusf, lroundudf): ...This.
18135 (fcvtps): Split as...
18136 (lceilv2sf, lceilv4sf, lceilv2df): ...This.
18137 (fcvtpu): Split as...
18138 (lceiluv2sf, lceiluv4sf, lceiluv2df, lceilusf, lceiludf): ...This.
18139 (fcvtms): Split as...
18140 (lfloorv2sf, lfloorv4sf, lfloorv2df): ...This.
18141 (fcvtmu): Split as...
18142 (lflooruv2sf, lflooruv4sf, lflooruv2df, lfloorusf, lfloorudf): ...This.
18143 (lfrintnv2sf, lfrintnv4sf, lfrintnv2df, lfrintnsf, lfrintndf): New.
18144 (lfrintnuv2sf, lfrintnuv4sf, lfrintnuv2df): Likewise.
18145 (lfrintnusf, lfrintnudf): Likewise.
18146 * config/aarch64/aarch64-simd.md
18147 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Convert to
18148 define_insn.
18149 (aarch64_fcvt<frint_suffix><su><mode>): Remove.
18150 * config/aarch64/iterators.md (FCVT): Include UNSPEC_FRINTN.
18151 (fcvt_pattern): Likewise.
18152
18153 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18154
18155 * config/aarch64/aarch64-simd.md
18156 (l<fcvt_pattern><su_optab><fcvt_target><VDQF:mode>2): Rename to...
18157 (l<fcvt_pattern><su_optab><VDQF:mode><fcvt_target>2): ... This.
18158
18159 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18160
18161 * config/aarch64/arm_neon.h (vrndq<a,m,n,p>_f<32, 64>): Rename to...
18162 (vrnd<a,m,n,p>q_f<32, 64>): ...This, implement using builtin.
18163 (vrnd<a,m,n,p>_f32): Implement using builtins.
18164 (vrnd<i,x><q>_f<32, 64>): New.
18165
18166 2013-04-29 James Greenhalgh <james.greenhalgh@arm.com>
18167
18168 * config/aarch64/aarch64-builtins.c
18169 (aarch64_builtin_vectorized_function): Fold to standard pattern names.
18170 * config/aarch64/aarch64-simd-builtins.def (frintn): New.
18171 (frintz): Rename to...
18172 (btrunc): ...this.
18173 (frintp): Rename to...
18174 (ceil): ...this.
18175 (frintm): Rename to...
18176 (floor): ...this.
18177 (frinti): Rename to...
18178 (nearbyint): ...this.
18179 (frintx): Rename to...
18180 (rint): ...this.
18181 (frinta): Rename to...
18182 (round): ...this.
18183 * config/aarch64/aarch64-simd.md
18184 (aarch64_frint<frint_suffix><mode>): Delete.
18185 (<frint_pattern><mode>2): Convert to insn.
18186 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRINTN.
18187 * config/aarch64/iterators.md (FRINT): Add UNSPEC_FRINTN.
18188 (frint_pattern): Likewise.
18189 (frint_suffix): Likewise.
18190
18191 2013-04-29 Richard Biener <rguenther@suse.de>
18192
18193 PR tree-optimization/57081
18194 * loop-init.c: Include tree-flow.h.
18195 (loop_optimizer_finalize): Free number of iteration estimates.
18196 * Makefile.in (loop-init.o): Add $(TREE_FLOW_H) dependency.
18197
18198 2013-04-29 Jakub Jelinek <jakub@redhat.com>
18199
18200 PR tree-optimization/57083
18201 * tree-vrp.c (extract_range_from_binary_expr_1): For LSHIFT_EXPR with
18202 non-singleton shift count range, zero extend low_bound for uns case.
18203
18204 * config/i386/predicates.md (general_vector_operand): New predicate.
18205 * config/i386/i386.c (const_vector_equal_evenodd_p): New function.
18206 (ix86_expand_mul_widen_evenodd): Force op1 resp. op2 into register
18207 if they aren't nonimmediate operands. If their original values
18208 satisfy const_vector_equal_evenodd_p, don't shift them.
18209 * config/i386/sse.md (mul<mode>3): Use general_vector_operand
18210 predicates. For the SSE4.1 case force operands[{1,2}] into registers
18211 if not nonimmediate_operand.
18212 (vec_widen_smult_even_v4si): Use nonimmediate_operand predicates
18213 instead of register_operand.
18214 (vec_widen_<s>mult_odd_<mode>): Use general_vector_operand predicates.
18215
18216 2013-04-28 Eric Botcazou <ebotcazou@adacore.com>
18217
18218 * stor-layout.c (finalize_size_functions): Allocate a structure and
18219 reset cfun before dumping the functions.
18220
18221 2013-04-27 Jakub Jelinek <jakub@redhat.com>
18222
18223 * config/i386/i386.c (ix86_expand_call): Make cregs_size unsigned.
18224
18225 PR target/56866
18226 * config/i386/i386.c (ix86_expand_mul_widen_evenodd): Don't
18227 use xop_pmacsdqh if uns_p.
18228 * config/i386/sse.md (xop_rotr<mode>3): Fix up computation of
18229 the immediate rotate count.
18230
18231 2013-04-26 Vladimir Makarov <vmakarov@redhat.com>
18232
18233 * rtl.h (struct rtx_def): Add comment for field jump.
18234 (LRA_SUBREG_P): New macro.
18235 * recog.c (register_operand): Check LRA_SUBREG_P.
18236 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18237 stack frame.
18238 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18239 (lra_final_code_change): Skip subreg change for operators.
18240 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18241 if there are no operand changes.
18242 * lra-constraints.c (curr_insn_set): New.
18243 (match_reload): Set LRA_SUBREG_P.
18244 (emit_spill_move): Ditto.
18245 (check_and_process_move): Use curr_insn_set. Process only single
18246 set insns. Don't initialize sec_mem_p and change_p.
18247 (simplify_operand_subreg): Use LRA_SUBREG_P.
18248 (reg_in_class_p): New function.
18249 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18250 of #ifdef. Add code to remove cycling.
18251 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18252 non-null disp. Reload inner instead of disp when base and index
18253 are null. Try to put lo_sum into register.
18254 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18255 (check_and_process_move): Move code for move cost check to
18256 simple_move_p. Remove equiv_substitution.
18257 (simple_move_p): New function.
18258 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18259 curr_insn_set. Call check_and_process_move only for single set
18260 insns. Use the new function. Move call of check_and_process_move
18261 after operand equiv substitution and address process.
18262
18263 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18264
18265 PR go/57045
18266 * tree-ssa-uninit.c (compute_uninit_opnds_pos): In functions
18267 with nonlocal goto receivers or returns twice calls, ignore
18268 unininitialized values from abnormal edges to nl goto receiver
18269 or returns twice call.
18270
18271 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18272
18273 PR tree-optimization/57051
18274 * fold-const.c (const_binop): Handle VEC_LSHIFT_EXPR
18275 and VEC_RSHIFT_EXPR if shift count is a multiple of element
18276 bitsize.
18277
18278 2013-04-26 Richard Biener <rguenther@suse.de>
18279
18280 * omp-low.c (finalize_task_copyfn): Do not drop PROP_loops.
18281 (expand_omp_taskreg): Likewise. Mark loops for fixup.
18282 * tree-cfg.c (move_block_to_fn): Remap loop fathers.
18283 (fixup_loop_arrays_after_move): New function.
18284 (move_sese_region_to_fn): Properly outline the loop tree parts
18285 of the SESE region.
18286
18287 2013-04-26 Uros Bizjak <ubizjak@gmail.com>
18288
18289 * config/i386/i386.md (type, unit): Fix long lines.
18290
18291 2013-04-26 Richard Biener <rguenther@suse.de>
18292
18293 * Makefile.in (lto-streamer-in.o): Add $(CFGLOOP_H) dependency.
18294 (lto-streamer-out.o): Likewise.
18295 * cfgloop.c (init_loops_structure): Export, add struct function
18296 argument and adjust.
18297 (flow_loops_find): Adjust.
18298 * cfgloop.h (enum loop_estimation): Add EST_LAST.
18299 (init_loops_structure): Declare.
18300 * lto-streamer-in.c: Include cfgloop.h.
18301 (input_cfg): Input the loop tree.
18302 * lto-streamer-out.c: Include cfgloop.h.
18303 (output_cfg): Output the loop tree.
18304 (output_struct_function_base): Do not drop PROP_loops.
18305
18306 2013-03-26 Richard Biener <rguenther@suse.de>
18307
18308 * tree-cfg.c (execute_build_cfg): Build the loop tree.
18309 (pass_build_cfg): Provide PROP_loops.
18310 (move_sese_region_to_fn): Remove loops that are outlined into fn
18311 for now.
18312 * tree-inline.c: Include cfgloop.h.
18313 (initialize_cfun): Do not drop PROP_loops.
18314 (copy_loops): New function.
18315 (copy_cfg_body): Copy loop structure.
18316 (tree_function_versioning): Initialize destination loop tree.
18317 * tree-ssa-loop.c (pass_tree_loop_init): Do not provide PROP_loops.
18318 (pass_parallelize_loops): Do IL verification.
18319 * loop-init.c (loop_optimizer_init): Fixup loops if required.
18320 * tree-optimize.c (execute_fixup_cfg): If we need to cleanup
18321 the CFG make sure we fixup loops as well.
18322 * tree-ssa-tail-merge.c: Include cfgloop.h.
18323 (replace_block_by): When merging loop latches mark loops for fixup.
18324 * lto-streamer-out.c (output_struct_function_base): Drop
18325 PROP_loops for now.
18326 * tree-ssa-phiopt.c: Include tree-scalar-evolution.h.
18327 (tree_ssa_cs_elim): Initialize the loop optimizer and SCEV.
18328 * ipa-split.c: Include cfgloop.h.
18329 (split_function): Add the new return block to the loop tree root.
18330 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Return
18331 whether we have removed the forwarder block.
18332 (merge_phi_nodes): If we removed a forwarder mark loops for fixup.
18333 * cfgloop.h (place_new_loop): Declare.
18334 * cfgloopmanip.c (place_new_loop): Export.
18335 * Makefile.in (asan.o): Add $(CFGLOOP_H) dependency.
18336 (tree-switch-conversion.o): Likewise.
18337 (tree-complex.o): Likewise.
18338 (tree-inline.o): Likewise.
18339 (tree-ssa-tailmerge.o): Likewise.
18340 (ipa-split.o): Likewise.
18341 (tree-ssa-phiopt.o): Add $(SCEV_H) dependency.
18342 (tree-ssa-copy.o): Likewise.
18343 * tree-switch-conversion.c: Include cfgloop.h
18344 (process_switch): If we emit a bit-test cascade, schedule loops
18345 for fixup.
18346 * tree-complex.c: Include cfgloop.h.
18347 (expand_complex_div_wide): Properly add new basic-blocks to loops.
18348 * asan.c: Include cfgloop.h.
18349 (create_cond_insert_point): Properly add new basic-blocks to
18350 loops, schedule loop fixup.
18351 * cfgloop.c (verify_loop_structure): Check that looks are not
18352 marked for fixup.
18353 * omp-low.c (expand_parallel_call): Properly add new basic-blocks
18354 to loops.
18355 (expand_omp_for_generic): Likewise.
18356 (expand_omp_sections): Likewise.
18357 (expand_omp_atomic_pipeline): Schedule loops for fixup.
18358 * tree-ssa-copy.c: Include tree-scalar-evolution.h.
18359 (fini_copy_prop): Disable DCE in substitute_and_fold if SCEV
18360 is initialized, not when loops are present.
18361 * tree-parloops.c (parallelize_loops): Remove checking here.
18362 * passes.c (init_optimization_passes): Schedule a copy-propagation
18363 pass before complete unrolling of inner loops.
18364
18365 2013-04-26 Jakub Jelinek <jakub@redhat.com>
18366
18367 * Makefile.in (toplev.o): Depend on diagnostic-color.h.
18368 * diagnostic-color.c (should_colorize): Remove _WIN32 version.
18369 (colorize_init): Add argument to _WIN32 version.
18370 * toplev.c: Include diagnostic-color.h.
18371 (process_options): Default to -fdiagnostics-color=auto if
18372 GCC_COLORS env var is in the environment.
18373 * common.opt (fdiagnostics-color=): Add Var and Init.
18374 * doc/invoke.texi (-fdiagnostics-color=): Document that if GCC_COLORS
18375 env var is in the environment, the default is auto rather than never.
18376
18377 * diagnostic.h (file_name_as_prefix): Add context argument.
18378 * diagnostic.c (file_name_as_prefix): Likewise. Colorize
18379 the string as locus.
18380 * langhooks.c (lhd_print_error_function): Adjust caller.
18381
18382 2013-04-25 Lawrence Crowl <crowl@google.com>
18383
18384 * var-tracking.c (shared_hash_def::htab):
18385 Change type to hash_table. Update dependent calls and types.
18386
18387 2013-04-25 Lawrence Crowl <crowl@google.com>
18388
18389 * Makefile.in: Update as needed below.
18390
18391 * alloc-pool.c (static hash_table <alloc_pool_hasher> alloc_pool_hash):
18392 Move declaration to after the type's method definitons.
18393
18394 * attribs.c (htab_t scoped_attributes::attribute_hash):
18395 Change type to hash_table. Update dependent calls and types.
18396
18397 * bitmap.c (htab_t bitmap_desc_hash):
18398 Change type to hash_table. Update dependent calls and types.
18399
18400 * cselib.c (htab_t cselib_hash_table):
18401 Change type to hash_table. Update dependent calls and types.
18402
18403 * data-streamer.h (struct string_slot): Move to lto-streamer.h.
18404 (hash_string_slot_node): Move implementation into lto-streamer.h
18405 struct string_slot_hasher.
18406 (eq_string_slot_node): Likewise.
18407
18408 * data-streamer-out.c: Update output_block::string_hash_table
18409 dependent calls and types.
18410
18411 * dwarf2cfi.c (htab_t trace_index):
18412 Change type to hash_table. Update dependent calls and types.
18413
18414 * dwarf2out.c (htab_t break_out_includes::cu_hash_table):
18415 Change type to hash_table. Update dependent calls and types.
18416 (htab_t copy_decls_for_unworthy_types::decl_table): Likewise.
18417 (htab_t optimize_external_refs::map): Likewise.
18418 (htab_t output_comp_unit::extern_map): Likewise.
18419 (htab_t output_comdat_type_unit::extern_map): Likewise.
18420 (htab_t output_macinfo::macinfo_htab): Likewise.
18421 (htab_t optimize_location_lists::htab): Likewise.
18422 (htab_t dwarf2out_finish::comdat_type_table): Likewise.
18423
18424 * except.c (htab_t ehspec_hash_type):
18425 Change type to hash_table. Update dependent calls and types.
18426 (assign_filter_values::ttypes): Likewise.
18427 (assign_filter_values::ehspec): Likewise.
18428 (sjlj_assign_call_site_values::ar_hash): Likewise.
18429 (convert_to_eh_region_ranges::ar_hash): Likewise.
18430
18431 * gcse.c (htab_t pre_ldst_table):
18432 Change type to hash_table. Update dependent calls and types.
18433
18434 * ggc-common.c (htab_t saving_htab):
18435 Change type to hash_table. Update dependent calls and types.
18436 (htab_t loc_hash): Likewise.
18437 (htab_t ptr_hash): Likewise.
18438 (call_count): Rename ggc_call_count.
18439 (call_alloc): Rename ggc_call_alloc.
18440 (loc_descriptor): Rename make_loc_descriptor.
18441 (add_statistics): Rename ggc_add_statistics.
18442
18443 * ggc-common.c (saving_htab):
18444 Change type to hash_table. Update dependent calls and types.
18445
18446 * gimple.h (struct gimplify_ctx): Move to gimplify-ctx.h.
18447 (push_gimplify_context): Likewise.
18448 (pop_gimplify_context): Likewise.
18449 (struct gimple_temp_hash_elt): Added.
18450 (struct gimplify_hasher): Likewise.
18451 (struct gimplify_ctx.temp_htab):
18452 Change type to hash_table. Update dependent calls and types.
18453
18454 * gimple-fold.c: Include gimplify-ctx.h.
18455
18456 * gimple-ssa-strength-reduction.c (htab_t base_cand_map):
18457 Change type to hash_table. Update dependent calls and types.
18458 (base_cand_dump_callback): Rename to ssa_base_cand_dump_callback to
18459 avoid potential global name collision.
18460
18461 * gimplify.c: Include gimplify-ctx.h.
18462 (struct gimple_temp_hash_elt): Move to gimplify-ctx.h.
18463 (htab_t gimplify_ctx::temp_htab):
18464 Update dependent calls and types for new type hash_table.
18465 (gimple_tree_hash): Move into gimplify_hasher in gimplify-ctx.h.
18466 (gimple_tree_eq): Move into gimplify_hasher in gimplify-ctx.h.
18467
18468 * gimplify-ctx.h: New.
18469 (struct gimple_temp_hash_elt): Move from gimplify.c.
18470 (class gimplify_hasher): New.
18471 (struct gimplify_ctx): Move from gimple.h.
18472 (htab_t gimplify_ctx::temp_htab):
18473 Change type to hash_table. Update dependent calls and types.
18474
18475 * graphite-clast-to-gimple.c: Include graphite-htab.h.
18476 (htab_t ivs_params::newivs_index):
18477 Change type to hash_table. Update dependent calls and types.
18478 (htab_t ivs_params::params_index): Likewise.
18479 (htab_t print_generated_program::params_index): Likewise.
18480 (htab_t gloog::newivs_index): Likewise.
18481 (htab_t gloog::params_index): Likewise.
18482
18483 * graphite.c: Include graphite-htab.h.
18484 4htab_t graphite_transform_loops::bb_pbb_mapping):
18485 Change type to hash_table. Update dependent calls and types.
18486
18487 * graphite-clast-to-gimple.h: (extern gloog) Move to graphite-htab.h.
18488 (bb_pbb_map_hash): Fold into bb_pbb_htab_type in graphite-htab.h.
18489 (eq_bb_pbb_map): Fold into bb_pbb_htab_type in graphite-htab.h.
18490
18491 * graphite-dependences.c: Include graphite-htab.h.
18492 (loop_is_parallel_p): Change hash table type of parameter.
18493
18494 * graphite-htab.h: New.
18495 (typedef hash_table <bb_pbb_hasher> bb_pbb_htab_type): New.
18496 (extern find_pbb_via_hash): Move from graphite-poly.h.
18497 (extern loop_is_parallel_p): Move from graphite-poly.h.
18498 (extern get_loop_body_pbbs): Move from graphite-poly.h.
18499
18500 * graphite-poly.h (extern find_pbb_via_hash): Move to graphite-htab.h.
18501 (extern loop_is_parallel_p): Move to graphite-htab.h.
18502 (extern get_loop_body_pbbs): Move to graphite-htab.h.
18503
18504 * haifa-sched.c (htab_t delay_htab):
18505 Change type to hash_table. Update dependent calls and types.
18506 (htab_t delay_htab_i2): Likewise.
18507
18508 * ira-color.c (htab_t allocno_hard_regs_htab):
18509 Change type to hash_table. Update dependent calls and types.
18510
18511 * ira-costs.c (htab_t cost_classes_htab):
18512 Change type to hash_table. Update dependent calls and types.
18513
18514 * loop-invariant.c (htab_t merge_identical_invariants::eq):
18515 Change type to hash_table. Update dependent calls and types.
18516
18517 * loop-iv.c (htab_t bivs):
18518 Change type to hash_table. Update dependent calls and types.
18519
18520 * loop-unroll.c (htab_t opt_info::insns_to_split):
18521 Change type to hash_table. Update dependent calls and types.
18522 (htab_t opt_info::insns_with_var_to_expand): Likewise.
18523
18524 * lto-streamer.h (struct string_slot): Move from data-streamer.h
18525 (struct string_slot_hasher): New.
18526 (htab_t output_block::string_hash_table):
18527 Change type to hash_table. Update dependent calls and types.
18528
18529 * lto-streamer-in.c (freeing_string_slot_hasher): New.
18530 (htab_t file_name_hash_table):
18531 Change type to hash_table. Update dependent calls and types.
18532
18533 * lto-streamer-out.c: Update output_block::string_hash_table dependent
18534 calls and types.
18535
18536 * lto-streamer.c (htab_t tree_htab):
18537 Change type to hash_table. Update dependent calls and types.
18538
18539 * omp-low.c: Include gimplify-ctx.h.
18540
18541 * passes.c (htab_t name_to_pass_map):
18542 Change type to hash_table. Update dependent calls and types.
18543 (pass_traverse): Rename to passes_pass_traverse.
18544
18545 * plugin.c (htab_t event_tab):
18546 Change type to hash_table. Update dependent calls and types.
18547
18548 * postreload-gcse.c (htab_t expr_table):
18549 Change type to hash_table. Update dependent calls and types.
18550 (dump_hash_table_entry): Rename dump_expr_hash_table_entry.
18551
18552 * sese.c (debug_rename_map_1): Make extern.
18553 (htab_t copy_bb_and_scalar_dependences::rename_map):
18554 Change type to hash_table. Update dependent calls and types.
18555
18556 * sese.h (extern debug_rename_map): Move to .c file.
18557
18558 * store-motion.c (htab_t store_motion_mems_table):
18559 Change type to hash_table. Update dependent calls and types.
18560
18561 * trans-mem.c (htab_t tm_new_mem_hash):
18562 Change type to hash_table. Update dependent calls and types.
18563
18564 * tree-browser.c (htab_t TB_up_ht):
18565 Change type to hash_table. Update dependent calls and types.
18566
18567 * tree-cfg.c (htab_t discriminator_per_locus):
18568 Change type to hash_table. Update dependent calls and types.
18569
18570 * tree-complex.c: Include tree-hasher.h
18571 (htab_t complex_variable_components):
18572 Change type to hash_table. Update dependent calls and types.
18573
18574 * tree-eh.c (htab_t finally_tree):
18575 Change type to hash_table. Update dependent calls and types.
18576
18577 * tree-flow.h (extern int_tree_map_hash): Moved into tree-hasher
18578 struct int_tree_hasher.
18579 (extern int_tree_map_eq): Likewise.
18580 (uid_decl_map_hash): Removed.
18581 (extern decl_tree_map_eq): Likewise.
18582
18583 * tree-hasher.h: New.
18584 (struct int_tree_hasher): New.
18585 (typedef int_tree_htab_type): New.
18586
18587 * tree-inline.c: Include gimplify-ctx.h.
18588
18589 * tree-mudflap.c: Include gimplify-ctx.h.
18590
18591 * tree-parloops.c: Include tree-hasher.h.
18592 (htab_t eliminate_local_variables_stmt::decl_address):
18593 Change type to hash_table. Update dependent calls and types.
18594 (htab_t separate_decls_in_region::decl_copies): Likewise.
18595
18596 * tree-scalar-evolution.c (htab_t resolve_mixers::cache):
18597 Change type to hash_table. Update dependent calls and types.
18598
18599 * tree-sra.c (candidates):
18600 Change type to hash_table. Update dependent calls and types.
18601
18602 * tree-ssa.c (int_tree_map_eq): Moved into struct int_tree_hasher
18603 in tree-flow.h.
18604 (int_tree_map_hash): Likewise.
18605
18606 * tree-ssa-dom.c (htab_t avail_exprs):
18607 Change type to hash_table. Update dependent calls and types.
18608
18609 * tree-ssa-live.c (var_map_base_init::tree_to_index):
18610 Change type to hash_table. Update dependent calls and types.
18611
18612 * tree-ssa-loop-ivopts.c (struct ivopts_data.inv_expr_tab):
18613 Change type to hash_table. Update dependent calls and types.
18614
18615 * tree-ssa-phiopt.c (seen_ssa_names):
18616 Change type to hash_table. Update dependent calls and types.
18617
18618 * tree-ssa-strlen.c (decl_to_stridxlist_htab):
18619 Change type to hash_table. Update dependent calls and types.
18620
18621 * tree-ssa-uncprop.c (equiv):
18622 Change type to hash_table. Update dependent calls and types.
18623
18624 2013-04-25 Jakub Jelinek <jakub@redhat.com>
18625
18626 PR rtl-optimization/57003
18627 * regcprop.c (copyprop_hardreg_forward_1): If ksvd.ignore_set_reg,
18628 call note_stores with kill_clobbered_value callback again after
18629 killing regs_invalidated_by_call.
18630
18631 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18632
18633 * config/aarch64/aarch64-simd.md
18634 (aarch64_simd_bsl<mode>_internal): Rewrite RTL to not use UNSPEC_BSL.
18635 (aarch64_simd_bsl<mode>): Likewise.
18636 * config/aarch64/iterators.md (unspec): Remove UNSPEC_BSL.
18637
18638 2013-04-25 Marek Polacek <polacek@redhat.com>
18639
18640 PR tree-optimization/57066
18641 * builtins.c (fold_builtin_logb): Return +Inf for -Inf.
18642
18643 2013-04-25 James Greenhalgh <jame.greenhalgh@arm.com>
18644
18645 * config/aarch64/aarch64-simd.md (neg<mode>2): Use VDQ iterator.
18646
18647 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18648
18649 * config/aarch64/aarch64-builtins.c
18650 (aarch64_fold_builtin): New.
18651 * config/aarch64/aarch64-protos.h (aarch64_fold_builtin): New.
18652 * config/aarch64/aarch64.c (TARGET_FOLD_BUILTIN): Define.
18653 * config/aarch64/aarch64-simd-builtins.def (abs): New.
18654 * config/aarch64/arm_neon.h
18655 (vabs<q>_<f32, 64>): Implement using __builtin_aarch64_fabs.
18656
18657 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18658 Tejas Belagod <tejas.belagod@arm.com>
18659
18660 * config/aarch64/aarch64-builtins.c
18661 (aarch64_gimple_fold_builtin): New.
18662 * config/aarch64/aarch64-protos.h (aarch64_gimple_fold_builtin): New.
18663 * config/aarch64/aarch64-simd-builtins.def (addv): New.
18664 * config/aarch64/aarch64-simd.md (addpv4sf): New.
18665 (addvv4sf): Update.
18666 * config/aarch64/aarch64.c (TARGET_GIMPLE_FOLD_BUILTIN): Define.
18667
18668 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18669
18670 * config/aarch64/aarch64.md
18671 (*cmp_swp_<optab><ALLX:mode>_shft_<GPI:mode>): New pattern.
18672
18673 2013-04-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
18674
18675 * config/aarch64/aarch64.md (*ngc<mode>): New pattern.
18676 (*ngcsi_uxtw): New pattern.
18677
18678 2013-04-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
18679 Julian Brown <julian@codesourcery.com>
18680
18681 * config/arm/arm.c (neon_builtin_type_mode): Add T_V4HF.
18682 (TB_DREG): Add T_V4HF.
18683 (v4hf_UP): New macro.
18684 (neon_itype): Add NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18685 (arm_init_neon_builtins): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18686 Handle initialisation of V4HF. Adjust initialisation of reinterpret
18687 built-ins.
18688 (arm_expand_neon_builtin): Handle NEON_FLOAT_WIDEN, NEON_FLOAT_NARROW.
18689 (arm_vector_mode_supported_p): Handle V4HF.
18690 (arm_mangle_map): Handle V4HFmode.
18691 * config/arm/arm.h (VALID_NEON_DREG_MODE): Add V4HF.
18692 * config/arm/arm_neon_builtins.def: Add entries for
18693 vcvtv4hfv4sf, vcvtv4sfv4hf.
18694 * config/arm/neon.md (neon_vcvtv4sfv4hf): New pattern.
18695 (neon_vcvtv4hfv4sf): Likewise.
18696 * config/arm/neon-gen.ml: Handle half-precision floating point
18697 features.
18698 * config/arm/neon-testgen.ml: Handle Requires_FP_bit feature.
18699 * config/arm/arm_neon.h: Regenerate.
18700 * config/arm/neon.ml (type elts): Add F16.
18701 (type vectype): Add T_float16x4, T_floatHF.
18702 (type vecmode): Add V4HF.
18703 (type features): Add Requires_FP_bit feature.
18704 (elt_width): Handle F16.
18705 (elt_class): Likewise.
18706 (elt_of_class_width): Likewise.
18707 (mode_of_elt): Refactor.
18708 (type_for_elt): Handle F16, fix error messages.
18709 (vectype_size): Handle T_float16x4.
18710 (vcvt_sh): New function.
18711 (ops): Add entries for vcvt_f16_f32, vcvt_f32_f16.
18712 (string_of_vectype): Handle T_floatHF, T_float16, T_float16x4.
18713 (string_of_mode): Handle V4HF.
18714 * doc/arm-neon-intrinsics.texi: Regenerate.
18715
18716 2013-04-25 James Greenhalgh <james.greenhalgh@arm.com>
18717
18718 * config/aarch64/aarch64.c (aarch64_print_operand): Fix asm_fprintf
18719 format specifier in 'X' case.
18720
18721 2013-04-25 Alan Modra <amodra@gmail.com>
18722
18723 PR target/57052
18724 * config/rs6000/rs6000.md (rotlsi3_internal7): Rename to
18725 rotlsi3_internal7le and condition on !BYTES_BIG_ENDIAN.
18726 (rotlsi3_internal8be): New BYTES_BIG_ENDIAN insn.
18727 Repeat for many other rotate/shift and mask patterns using subregs.
18728 Name lshiftrt insns.
18729 (ashrdisi3_noppc64): Rename to ashrdisi3_noppc64be and condition
18730 on WORDS_BIG_ENDIAN.
18731
18732 2013-04-25 Alan Modra <amodra@gmail.com>
18733
18734 * config.gcc: Support little-endian powerpc-linux targets.
18735 * config/rs6000/linux.h (LINK_OS_LINUX_EMUL): Define.
18736 (LINK_OS_LINUX_SPEC): Define.
18737 * config/rs6000/linuxspe.h (TARGET_DEFAULT):
18738 Preserve MASK_LITTLE_ENDIAN.
18739 * config/rs6000/default64.h (TARGET_DEFAULT): Likewise.
18740 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Likewise.
18741 * config/rs6000/linux64.h (OPTION_LITTLE_ENDIAN): Don't zero.
18742 (LINK_OS_LINUX_EMUL32, LINK_OS_LINUX_EMUL64): Define.
18743 (LINK_OS_LINUX_SPEC32, LINK_OS_LINUX_SPEC64): Use above.
18744 * config/rs6000/rs6000.c (output_toc): Don't use .tc for TARGET_ELF.
18745 Correct fp word order for little-endian. Don't shift toc entries
18746 smaller than a word for little-endian.
18747 * config/rs6000/rs6000.md (bswaphi2, bswapsi2 split): Comment.
18748 (bswapdi2 splits): Correct low-part subreg for little-endian.
18749 Remove wrong BYTES_BIG_ENDIAN tests, and rename vars to remove
18750 low/high where such is correct only for be.
18751 * config/rs6000/sysv4.h (SUBTARGET_OVERRIDE_OPTIONS): Allow
18752 little-endian for -mcall-aixdesc.
18753
18754 2013-04-25 Alan Modra <amodra@gmail.com>
18755
18756 * config/rs6000/rs6000.c (rs6000_secondary_reload_inner): Use
18757 replace_equiv_address_nv.
18758
18759 2013-04-25 Alan Modra <amodra@gmail.com>
18760
18761 * config/rs6000/rs6000.c (rs6000_emit_set_long_const): Tidy.
18762
18763 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18764
18765 Revert:
18766 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18767 * rtl.h (struct rtx_def): ...
18768
18769 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18770
18771 PR rtl-optimizations/57046
18772 * lra-constraints (split_reg): Set up lra_risky_transformations_p
18773 for multi-reg splits.
18774
18775 2013-04-24 H.J. Lu <hongjiu.lu@intel.com>
18776
18777 * config/i386/x86-64.h (ASM_SPEC): Support -mx32.
18778
18779 2013-04-24 Sterling Augustine <saugustine@google.com>
18780
18781 * dwarf2out.c (skeleton_debug_str_hash, add_skeleton_AT_string)
18782 (comp_dir_string, debug_str_dwo_section): New.
18783 (DEBUG_STR_DWO_SECTION): Rename to ...
18784 (DEBUG_DWO_STR_SECTION): ... this.
18785 (DEBUG_NORM_STR_SECTION): Delete.
18786 (DEBUG_STR_SECTION, DEBUG_STR_SECTION_FLAGS): Edit definitions.
18787 (DEBUG_STR_DWO_SECTION_FLAGS): New.
18788 (find_AT_string): Move most logic to ...
18789 (find_AT_string_in_table): ... here. New.
18790 (add_top_level_skeleton_die_attrs): Call comp_dir_string and
18791 add_skeleton_AT_string. Delete logic.
18792 (output_skeleton_debug_sections): Remove call to
18793 add_top_level_skeleton_die_attrs.
18794 (add_comp_dir_attribute): Move logic to comp_dir_string.
18795 (dwarf2out_init): Initialize debug_str_dwo_section.
18796 (output_indirect_string): Call find_string_form.
18797 (output_indirect_strings): Rewrite.
18798 (prune_unused_types): Empty skeleton_debug_str_hash.
18799 Call get_skeleton_type_unit and add_top_level_skeleton_die_attrs.
18800 (dwarf2out_finish): Call output_indirect_strings.
18801
18802 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
18803
18804 * doc/cpp.texi: Remove __GXX_EXPERIMENTAL_CXX1Y__.
18805
18806 2013-04-24 Vladimir Makarov <vmakarov@redhat.com>
18807
18808 * rtl.h (struct rtx_def): Add comment for field jump.
18809 (LRA_SUBREG_P): New macro.
18810 * recog.c (register_operand): Check LRA_SUBREG_P.
18811 * lra.c (lra): Add note at the end of RTL code. Align non-empty
18812 stack frame.
18813 * lra-spills.c (lra_spill): Align stack after spilling pseudos.
18814 (lra_final_code_change): Skip subreg change for operators.
18815 * lra-eliminations.c (eliminate_regs_in_insn): Make return earlier
18816 if there are no operand changes.
18817 * lra-constraints.c (curr_insn_set): New.
18818 (match_reload): Set LRA_SUBREG_P.
18819 (emit_spill_move): Ditto.
18820 (check_and_process_move): Use curr_insn_set. Process only single
18821 set insns. Don't initialize sec_mem_p and change_p.
18822 (simplify_operand_subreg): Use LRA_SUBREG_P.
18823 (reg_in_class_p): New function.
18824 (process_alt_operands): Use it. Use #if HAVE_ATTR_enabled instead
18825 of #ifdef. Add code to remove cycling.
18826 (process_address): Check EXTRA_CONSTRAINT_STR. Process even if
18827 non-null disp. Reload inner instead of disp when base and index
18828 are null. Try to put lo_sum into register.
18829 (EBB_PROBABILITY_CUTOFF): Redefine probability in percents.
18830 (check_and_process_move): Move code for move cost check to
18831 simple_move_p. Remove equiv_substitution.
18832 (simple_move_p): New function.
18833 (curr_insn_transform): Initialize sec_mem_p and change_p. Set up
18834 curr_insn_set. Call check_and_process_move only for single set
18835 insns. Use the new function. Move call of check_and_process_move
18836 after operand equiv substitution and address process.
18837
18838 2013-04-24 James Greenhalgh <james.greenhalgh@arm.com>
18839
18840 * config/aarch64/arm_neon.h (vld1<q>_lane*): Fix constraints.
18841 (vld1<q>_dup_<sufp><8, 16, 32, 64>): Likewise.
18842 (vld1<q>_<sufp><8, 16, 32, 64>): Likewise.
18843
18844 2013-04-24 Paolo Carlini <paolo.carlini@oracle.com>
18845
18846 * doc/cpp.texi: Document __GXX_EXPERIMENTAL_CXX1Y__.
18847
18848 2013-04-24 Marek Polacek <polacek@redhat.com>
18849
18850 * tree-scalar-evolution.h (analyze_scalar_evolution): Remove.
18851 * tree-scalar-evolution.c (get_exit_conditions_rec): Likewise.
18852 (select_loops_exit_conditions): Likewise.
18853 (number_of_iterations_for_all_loops): Likewise.
18854 (analyze_scalar_evolution_for_all_loop_phi_nodes): Likewise.
18855 (scev_analysis): Likewise.
18856
18857 2013-04-02 Catherine Moore <clm@codesourcery.com>
18858 Chao-ying Fu <fu@mips.com>
18859
18860 * config/mips/micromips.md (jraddiusp): New pattern.
18861 * config/mips/mips.c (mips_expand_epilogue): Use the JRADDIUSP
18862 instruction if possible.
18863
18864 2013-04-24 Alan Modra <amodra@gmail.com>
18865
18866 * config/rs6000/driver-rs6000.c (elf_dcachebsize): Fix comment pasto.
18867
18868 2013-04-24 Julian Brown <julian@codesourcery.com>
18869 Chung-Lin Tang <cltang@codesourcery.com>
18870
18871 * dwarf2out.c (gen_enumeration_type_die): Fix HOST_BITS_PER_WIDE_INT
18872 dependency behavior in enumeration type DIE generation. Add TODO note
18873 to comments about future DW_FORM_sdata/udata re-work of related code.
18874
18875 2013-04-23 Lawrence Crowl <crowl@google.com>
18876
18877 * Makefile.in: Update as needed below.
18878
18879 * hash-table.h (class hash_table):
18880 Correct many methods with parameter types compare_type to the correct
18881 value_type. (Correct code was unlikely to notice the change.)
18882 (hash_table::elements_with_deleted) New.
18883 (class hashtable::iterator): New.
18884 (hashtable::begin()): New.
18885 (hashtable::end()): New.
18886 (FOR_EACH_HASH_TABLE_ELEMENT): New.
18887
18888 * statistics.c (statistics_hashes):
18889 Change type to hash_table. Update dependent calls and types.
18890
18891 * tree-into-ssa.c (var_infos):
18892 Change type to hash_table. Update dependent calls and types.
18893
18894 * tree-ssa-coalesce.c (struct coalesce_list_d.list):
18895 Change type to hash_table. Update dependent calls and types.
18896
18897 * tree-ssa-loop-im.c (struct mem_ref.refs):
18898 Change type to hash_table. Update dependent calls and types.
18899
18900 * tree-ssa-reassoc.c (undistribute_ops_list::ctable):
18901 Change type to hash_table. Update dependent calls and types.
18902
18903 * tree-ssa-sccvn.c (vn_tables_s::nary):
18904 Change type to hash_table. Update dependent calls and types.
18905 (vn_tables_s::phis): Likewise.
18906 (vn_tables_s::references): Likewise.
18907
18908 * tree-ssa-sccvn.h (vn_nary_op_eq): Update parameter and return types.
18909 (vn_reference_eq): Update parameter and return types.
18910
18911 * tree-ssa-structalias.c (pointer_equiv_class_table):
18912 Change type to hash_table. Update dependent calls and types.
18913 (location_equiv_class_table): Likewise.
18914
18915 * tree-vect-data-refs.c: Consequential changes for making
18916 peeling a hash_table.
18917
18918 * tree-vect-loop.c (new_loop_vec_info): Dependent hash_table update.
18919 (destroy_loop_vec_info): Dependent hash_table update.
18920
18921 * tree-vectorizer.h (peeling_htab):
18922 Change type to hash_table. Update dependent calls and types.
18923
18924 2013-04-23 Shiva Chen <shiva0217@gmail.com>
18925
18926 * lra-assigns.c (find_hard_regno_for): Use lra_reg_val_equal_p
18927 to check the register content is equal or not.
18928 * lra-constraints.c (match_reload): Use lra_assign_reg_val
18929 to assign register content record.
18930 * lra-eliminations.c (update_reg_eliminate): Use
18931 lra_update_reg_val_offset to update register content offset.
18932 * lra-int.h (struct lra_reg): Add offset member.
18933 (lra_reg_val_equal_p): New static inline function.
18934 (lra_update_reg_val_offset): New static inline function.
18935 (lra_assign_reg_val): New static inline function.
18936 * lra.c (lra_create_new_reg): Use lra_assign_reg_val
18937 to assign register content record.
18938 (initialize_lra_reg_info_element): Initial offset to zero.
18939
18940 2013-04-23 Catherine Moore <clm@codesourcery.com>
18941
18942 * config/mips/mips.md (*movhi_internal, *movqi_internal): New
18943 operands. Record compression.
18944
18945 2013-04-23 Xinliang David Li <davidxl@google.com>
18946
18947 * cfghhooks.c (dump_bb_for_graph): Support 'slim' graph dump.
18948
18949 2013-04-23 Richard Biener <rguenther@suse.de>
18950
18951 PR middle-end/57036
18952 * tree-inline.c (copy_edges_for_bb): Add can_make_abnormal_goto
18953 parameter, only add abnormal goto edges from the copied body
18954 if the call could perform abnormal gotos.
18955 (copy_cfg_body): Adjust.
18956
18957 2013-04-23 Sofiane Naci <sofiane.naci@arm.com>
18958
18959 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add simd attribute.
18960
18961 2013-04-23 Andreas Schwab <schwab@linux-m68k.org>
18962
18963 * coretypes.h (gimple_stmt_iterator): Add struct to make
18964 compatible with C.
18965
18966 2013-04-23 Richard Biener <rguenther@suse.de>
18967
18968 PR tree-optimization/57026
18969 * tree-vrp.c (simplify_conversion_using_ranges): Do not propagate
18970 from SSA names occuring in abnormal PHI nodes.
18971
18972 2013-04-22 Andi Kleen <ak@linux.intel.com>
18973
18974 * lto/lto.c (print_lto_report_1): Fix LTO report names.
18975
18976 2013-04-22 Andi Kleen <ak@linux.intel.com>
18977
18978 * lto/lto.c (print_lto_report_1): Declare early.
18979 (read_cgraph_and_symbols): Call print_lto_report_1 early.
18980
18981 2013-04-22 Andi Kleen <ak@linux.intel.com>
18982
18983 * common.opt (-flto-report-wpa): Add.
18984 * doc/invoke.texi (-flto-report-wpa): Add.
18985 * lto/lto.c (do_whole_program_analysis): Check for lto-report-wpa.
18986 (lto_main): dito.
18987
18988 2013-04-22 Xinliang David Li <davidxl@google.com>
18989
18990 * graph.c (draw_cfg_node_succ_edges): Add branch probility as label.
18991 * cfghhooks.c (dump_bb_for_graph): Dump profile count and frquency.
18992 * Makefile.in: New dependency
18993
18994 David Daney <ddaney.cavm@gmail.com>
18995
18996 * configure.ac (gcc_cv_as_micromips_support): Use the
18997 --fatal-warnings option.
18998 * configure: Regenerate.
18999
19000 2013-04-22 Marek Polacek <polacek@redhat.com>
19001
19002 PR sanitizer/56990
19003 * tsan.c (instrument_expr): Don't instrument expression
19004 in case its size is zero.
19005
19006 2013-04-22 Uros Bizjak <ubizjak@gmail.com>
19007
19008 PR target/57032
19009 Revert:
19010 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
19011
19012 * config/alpha/alpha.c (TARGET_LRA_P): New define.
19013
19014 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19015
19016 * coretypes.h (gimple_stmt_iterator_d): Forward declare.
19017 (gimple_stmt_iterator): New typedef.
19018 * gimple.h (gimple_stmt_iterator): Rename to...
19019 (gimple_stmt_iterator_d): ... This.
19020 * doc/tm.texi.in (TARGET_FOLD_BUILTIN): Detail restriction that
19021 trees be valid for GIMPLE and GENERIC.
19022 (TARGET_GIMPLE_FOLD_BUILTIN): New.
19023 * gimple-fold.c (gimple_fold_call): Call target hook
19024 gimple_fold_builtin.
19025 * hooks.c (hook_bool_gsiptr_false): New.
19026 * hooks.h (hook_bool_gsiptr_false): New.
19027 * target.def (fold_stmt): New.
19028 * doc/tm.texi: Regenerate.
19029
19030 2013-04-22 Vladimir Makarov <vmakarov@redhat.com>
19031
19032 PR target/57018
19033 * lra-eliminations.c (mark_not_eliminable): Prevent elimination of
19034 a set sp if no stack realignment.
19035
19036 2013-04-22 Nick Clifton <nickc@redhat.com>
19037
19038 * config.gcc (tilegx-linux): Extend extra_objs rather than
19039 overwriting it.
19040 (tilepro-linux): Likewise.
19041
19042 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19043
19044 * config/aarch64/aarch64-builtins.c
19045 (CF): Remove.
19046 (CF0, CF1, CF2, CF3, CF4, CF10): New.
19047 (VAR<1-12>): Add MAP parameter.
19048 (BUILTIN_*): Likewise.
19049 * config/aarch64/aarch64-simd-builtins.def: Set MAP parameter.
19050 * config/aarch64/aarch64-simd.md (aarch64_sshl_n<mode>): Remove.
19051 (aarch64_ushl_n<mode>): Likewise.
19052 (aarch64_sshr_n<mode>): Likewise.
19053 (aarch64_ushr_n<mode>): Likewise.
19054 (aarch64_<maxmin><mode>): Likewise.
19055 (aarch64_sqrt<mode>): Likewise.
19056 * config/aarch64/arm_neon.h (vshl<q>_n_*): Use new builtin names.
19057 (vshr<q>_n_*): Likewise.
19058
19059 2013-04-22 James Greenhalgh <james.greenhalgh@arm.com>
19060
19061 * config/aarch64/aarch64-builtins.c
19062 (aarch64_simd_builtin_type_mode): Handle SF types.
19063 (sf_UP): Define.
19064 (BUILTIN_GPF): Define.
19065 (aarch64_init_simd_builtins): Handle SF types.
19066 * config/aarch64/aarch64-simd-builtins.def (frecpe): Add support.
19067 (frecps): Likewise.
19068 (frecpx): Likewise.
19069 * config/aarch64/aarch64-simd.md
19070 (simd_types): Update simd_frcp<esx> to simd_frecp<esx>.
19071 (aarch64_frecpe<mode>): New.
19072 (aarch64_frecps<mode>): Likewise.
19073 * config/aarch64/aarch64.md (unspec): Add UNSPEC_FRECP<ESX>.
19074 (v8type): Add frecp<esx>.
19075 (aarch64_frecp<FRECP:frecp_suffix><mode>): New.
19076 (aarch64_frecps<mode>): Likewise.
19077 * config/aarch64/iterators.md (FRECP): New.
19078 (frecp_suffix): Likewise.
19079 * config/aarch64/arm_neon.h
19080 (vrecp<esx><qsd>_<fd><32, 64>): Convert to using builtins.
19081
19082 2013-04-22 Christian Bruel <christian.bruel@st.com>
19083
19084 PR target/56995
19085 * config/sh/sh.h (enum reg_class): Remove DF_HI_REGS.
19086 (REG_CLASS_NAMES): Idem.
19087 (REG_CLASS_CONTENTS): Idem.
19088 (REGCLASS_HAS_FP_REG): Idem.
19089 * config/sh/sh.c (sh_cannot_change_mode_class): Idem.
19090 (sh_conditional_register_usage): Idem.
19091
19092 2013-04-21 Jeff Law <law@redhat.com>
19093
19094 * tree-ssa-forwprop.c (simplify_conversion_from_bitmask): New function.
19095 (ssa_forward_propagate_and_combine): Use it.
19096
19097 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19098
19099 * lra.c: Update the flow chart diagram.
19100
19101 2013-04-19 Vladimir Makarov <vmakarov@redhat.com>
19102
19103 PR rtl-optimization/56847
19104 * lra-constraints.c (process_alt_operands): Discourage alternative
19105 with non-matche doffsettable memory constraint fro memory with
19106 known offset.
19107
19108 2013-04-19 Richard Biener <rguenther@suse.de>
19109
19110 PR tree-optimization/56982
19111 * builtins.def (BUILT_IN_LONGJMP): longjmp is not a leaf
19112 function.
19113 * gimplify.c (gimplify_call_expr): Notice special calls.
19114 (gimplify_modify_expr): Likewise.
19115 * tree-cfg.c (make_abnormal_goto_edges): Handle setjmp-like
19116 abnormal control flow receivers.
19117 (call_can_make_abnormal_goto): Handle cfun->calls_setjmp
19118 in the same way as cfun->has_nonlocal_labels.
19119 (gimple_purge_dead_abnormal_call_edges): Likewise.
19120 (stmt_starts_bb_p): Make setjmp-like abnormal control flow
19121 receivers start a basic-block.
19122
19123 2013-04-19 Richard Biener <rguenther@suse.de>
19124
19125 * tree-vectorizer.h (struct _slp_instance): Move load_permutation
19126 member ...
19127 (struct _slp_tree): ... here. Make it a vector of unsigned ints.
19128 (SLP_INSTANCE_LOAD_PERMUTATION): Remove.
19129 (SLP_TREE_LOAD_PERMUTATION): Add.
19130 (vect_transform_slp_perm_load): Adjust prototype.
19131 * tree-vect-slp.c (vect_free_slp_tree): Adjust.
19132 (vect_free_slp_instance): Likewise.
19133 (vect_create_new_slp_node): Likewise.
19134 (vect_supported_slp_permutation_p): Remove.
19135 (vect_slp_rearrange_stmts): Adjust.
19136 (vect_supported_load_permutation_p): Likewise. Inline
19137 vect_supported_slp_permutation_p here.
19138 (vect_analyze_slp_instance): Compute load permutations per
19139 slp node instead of per instance.
19140 (vect_get_slp_defs): Adjust.
19141 (vect_transform_slp_perm_load): Likewise.
19142 (vect_schedule_slp_instance): Remove redundant code.
19143 (vect_schedule_slp): Remove hack for PR56270, add it ...
19144 * tree-vect-stmts.c (vectorizable_load): ... here, do not
19145 CSE loads for SLP. Adjust.
19146
19147 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19148
19149 * config/arm/arm.c (load_multiple_sequence, ldm_stm_operation_p): Fix
19150 spelling in two comments.
19151
19152 2013-04-19 Greta Yorsh <Greta.Yorsh@arm.com>
19153
19154 PR target/56797
19155 * config/arm/arm.c (load_multiple_sequence): Require SP
19156 as base register for loads if SP is in the register list.
19157
19158 2013-04-19 Martin Jambor <mjambor@suse.cz>
19159
19160 PR tree-optimization/56718
19161 * ipa-cp.c (ipa_value_from_known_type_jfunc): Moved...
19162 * ipa-prop.c (ipa_binfo_from_known_type_jfunc): ...here, renamed
19163 and made public. Adjusted all callers.
19164 (ipa_intraprocedural_devirtualization): New function.
19165 * ipa-prop.h (ipa_binfo_from_known_type_jfunc): Declare.
19166 (ipa_intraprocedural_devirtualization): Likewise.
19167 * Makefile.in (tree-ssa-pre.o): Add ipa-prop.h to dependencies.
19168
19169 2013-04-19 Richard Biener <rguenther@suse.de>
19170
19171 PR tree-optimization/57000
19172 * tree-ssa-reassoc.c (pass_reassoc): Add TODO_update_ssa_only_virtuals.
19173
19174 2013-04-19 Terry Guo <terry.guo@arm.com>
19175
19176 * config/arm/cortex-m4-fpu.md (cortex_m4_v): Delete cpu unit.
19177 Replace with ...
19178 (cortex_m4_v_a, cortex_m4_v_b): ... new cpu units.
19179 (cortex_m4_v, cortex_m4_exa_va, cortex_m4_exb_vb): New reservations.
19180 (cortex_m4_fmacs): Use new reservations.
19181 (cortex_m4_f_load, cortex_m4_f_store): Likewise.
19182
19183 2013-04-18 Vladimir Makarov <vmakarov@redhat.com>
19184
19185 PR rtl-optimization/56999
19186 * lra-coalesce.c (coalescable_pseudo_p): Remove 2nd parameter and
19187 related code.
19188 (lra_coalesce): Remove split_origin_bitmap and related code.
19189 * lra.c (lra): Coalesce after undoing inheritance. Recreate live
19190 ranges if necessary.
19191
19192 2013-04-18 Uros Bizjak <ubizjak@gmail.com>
19193
19194 * config/i386/i386.c (x86_64_ms_sysv_extra_clobbered_registers):
19195 New array.
19196 (ix86_expand_call): Remove clobbered_registers array and use
19197 x86_64_ms_sysv_extra_clobbered_registers instead.
19198 * config/i386/i386.h (x86_64_ms_sysv_extra_clobbered_registers):
19199 Declare here.
19200 * config/i386/predicates.md (call_rex64_ms_sysv_operation): New
19201 predicate.
19202 * config/i386/i386.md (*call_rex64_ms_sysv): Use
19203 call_rex64_ms_sysv_operation predicate. Remove explicit clobbers.
19204 (*call_value_rex64_ms_sysv): Ditto.
19205
19206 2013-04-18 Cary Coutant <ccoutant@google.com>
19207
19208 * dwarf2out.c (output_pubnames): Check die_perennial_p of
19209 parent instead of die_mark.
19210
19211 2013-04-18 Diego Novillo <dnovillo@google.com>
19212
19213 * gimple.c (create_gimple_tmp): New.
19214 (get_expr_type): New.
19215 (build_assign): New.
19216 (build_type_cast): New.
19217 * gimple.h (enum ssa_mode): Define.
19218 (gimple_seq_set_location): New.
19219 * asan.c (build_check_stmt): Change some gimple_build_* calls
19220 to use build_assign and build_type_cast.
19221
19222 2013-04-18 Richard Biener <rguenther@suse.de>
19223
19224 * tree-vect-data-refs.c (vect_analyze_group_access): Properly
19225 handle negative step. Remove redundant checks.
19226 (vect_create_data_ref_ptr): Avoid ICEs with non-constant steps.
19227 * tree-vect-stmts.c (vectorizable_load): Instead of asserting
19228 for negative step and grouped loads fail to vectorize.
19229
19230 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19231
19232 * emit-rtl.c (reset_insn_used_flags): New function.
19233 (reset_all_used_flags): Use it.
19234 (verify_insn_sharing): New function.
19235 (verify_rtl_sharing): Fix verification for SEQUENCEs.
19236
19237 2013-04-18 Jakub Jelinek <jakub@redhat.com>
19238
19239 PR tree-optimization/56984
19240 * tree-vrp.c (register_edge_assert_for_2): For (x >> M) < N
19241 and (x >> M) >= N don't register any assertion if N << M is the
19242 minimum value.
19243
19244 2013-04-18 Steven Bosscher <steven@gcc.gnu.org>
19245
19246 * lower-subreg.c (resolve_simple_move): If called self-recursive,
19247 do not delete_insn insns that have not yet been emitted, only
19248 unlink them with remove_insn.
19249 * df-scan.c (df_insn_delete): Revert r197492.
19250
19251 2013-04-17 Steven Bosscher <steven@gcc.gnu.org>
19252
19253 * emit-rtl.c (link_insn_into_chain): Handle chaining of SEQUENCEs.
19254 * reorg.c (emit_delay_sequence): Simplify with emit-rtl API.
19255
19256 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19257
19258 * config/arm/arm.md (movsicc_insn): Convert define_insn into
19259 define_insn_and_split.
19260 (and_scc,ior_scc,negscc): Likewise.
19261 (cmpsi2_addneg, subsi3_compare): Convert to named patterns.
19262
19263 2013-04-17 Greta Yorsh <Greta.Yorsh@arm.com>
19264
19265 * config/arm/arm.c (use_return_insn): Return 0 for targets that
19266 can benefit from using a sequence of LDRD instructions in epilogue
19267 instead of a single LDM instruction.
19268
19269 2013-04-17 Manuel López-Ibáñez <manu@gcc.gnu.org>
19270
19271 PR 45688
19272 * doc/extend.texi: Fix typo.
19273
19274 2013-04-17 Richard Biener <rguenther@suse.de>
19275
19276 * tree-vect-slp.c (vect_build_slp_tree_1): Split out from ...
19277 (vect_build_slp_tree): ... here.
19278 (vect_build_slp_tree_1): Compute which stmts of the SLP group
19279 match. Remove special-casing of mismatched complex loads.
19280 (vect_build_slp_tree): Based on the result from vect_build_slp_tree_1
19281 re-try the match with swapped commutative operands.
19282 (vect_supported_load_permutation_p): Remove special-casing of
19283 mismatched complex loads.
19284 (vect_analyze_slp_instance): Adjust.
19285
19286 2013-04-17 Richard Biener <rguenther@suse.de>
19287
19288 PR rtl-optimization/56921
19289 * cfgloop.h (struct loop): Add simple_loop_desc member.
19290 (struct niter_desc): Mark with GTY(()).
19291 (simple_loop_desc): Do not use aux field but simple_loop_desc.
19292 * loop-iv.c (get_simple_loop_desc): Likewise.
19293 (free_simple_loop_desc): Likewise.
19294
19295 Revert
19296 2013-04-16 Richard Biener <rguenther@suse.de>
19297
19298 PR rtl-optimization/56921
19299 * loop-init.c (pass_rtl_move_loop_invariants): Add
19300 TODO_do_not_ggc_collect to todo_flags_finish.
19301 (pass_rtl_unswitch): Same.
19302 (pass_rtl_unroll_and_peel_loops): Same.
19303 (pass_rtl_doloop): Same.
19304
19305 2013-04-17 Eric Botcazou <ebotcazou@adacore.com>
19306
19307 * tree-ssa-alias.c (nonoverlapping_component_refs_of_decl_p): New.
19308 (decl_refs_may_alias_p): Add REF1 and REF2 parameters.
19309 Use nonoverlapping_component_refs_of_decl_p to disambiguate component
19310 references.
19311 (refs_may_alias_p_1): Adjust call to decl_refs_may_alias_p.
19312 * tree-streamer.c (record_common_node): Adjust reference in comment.
19313
19314 2013-04-17 Terry Guo <terry.guo@arm.com>
19315
19316 * config/arm/cortex-m4.md: Add a new bypass.
19317
19318 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19319
19320 * config/aarch64/aarch64.md (*adds_<optab><mode>_multp2):
19321 New pattern.
19322 (*subs_<optab><mode>_multp2): New pattern.
19323 (*adds_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19324 (*subs_<optab><ALLX:mode>_<GPI:mode>): New pattern.
19325
19326 2013-04-16 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19327
19328 * config/aarch64/aarch64.md (*adds_mul_imm_<mode>): New pattern.
19329 (*subs_mul_imm_<mode>): New pattern.
19330
19331 2013-04-16 David Edelsohn <dje.gcc@gmail.com>
19332
19333 PR target/56948
19334 * config/rs6000/vsx.md (vsx_mov<mode>): Add j->r alternative.
19335 (vsx_movti_64bit): Change j->wa to O->wa. Add n->r alternative.
19336 (vsx_movti_32bit): Change j->wa to O->wa.
19337
19338 2013-04-16 Richard Biener <rguenther@suse.de>
19339
19340 PR rtl-optimization/56921
19341 * loop-init.c (pass_rtl_move_loop_invariants): Add
19342 TODO_do_not_ggc_collect to todo_flags_finish.
19343 (pass_rtl_unswitch): Same.
19344 (pass_rtl_unroll_and_peel_loops): Same.
19345 (pass_rtl_doloop): Same.
19346
19347 2013-04-16 Greta Yorsh <Greta.Yorsh@arm.com>
19348
19349 * config/arm/arm.c (emit_multi_reg_push): New declaration
19350 for an existing function.
19351 (arm_emit_strd_push): New function.
19352 (arm_expand_prologue): Used here.
19353 (arm_emit_ldrd_pop): New function.
19354 (arm_expand_epilogue): Used here.
19355 (arm_get_frame_offsets): Update condition.
19356 (arm_emit_multi_reg_pop): Add a special case for load of a single
19357 register with writeback.
19358
19359 2013-04-16 Uros Bizjak <ubizjak@gmail.com>
19360
19361 * doc/invoke.texi (i386 Option): Reword -mstack-protector-guard
19362 description.
19363
19364 2013-04-16 Richard Biener <rguenther@suse.de>
19365
19366 PR tree-optimization/56756
19367 * tree-ssa-loop-im.c (struct first_mem_ref_loc_1): New functor.
19368 (first_mem_ref_loc): New.
19369 (execute_sm): Place the load temporarily before a previous
19370 access instead of in the latch edge to ensure its SSA dependencies
19371 are defined at points dominating the load.
19372
19373 2013-04-16 Steven Bosscher <steven@gcc.gnu.org>
19374
19375 * cfgrtl.c (cfg_layout_merge_blocks): Revert r184005, implement
19376 correct fix by moving header and footer insn to the footer of
19377 the merged basic block. Clear BB_END of the merged-away block.
19378
19379 PR middle-end/43631
19380 * emit-rtl.c (make_note_raw): New function.
19381 (link_insn_into_chain): New static inline function.
19382 (add_insn): Use it.
19383 (add_insn_before, add_insn_after): Factor insn chain linking code...
19384 (add_insn_before_nobb, add_insn_after_nobb): ...here, new functions
19385 using link_insn_into_chain.
19386 (note_outside_basic_block_p): New helper function for emit_note_after
19387 and emit_note_before.
19388 (emit_note_after): Use nobb variant of add_insn_after if the note
19389 should not be contained in a basic block.
19390 (emit_note_before): Use nobb variant of add_insn_before if the note
19391 should not be contained in a basic block.
19392 (emit_note_copy): Use make_note_raw.
19393 (emit_note): Likewise.
19394 * bb-reorder.c (insert_section_boundary_note): Remove hack to set
19395 BLOCK_FOR_INSN to NULL manually for NOTE_INSN_SWITCH_TEXT_SECTIONS.
19396 * jump.c (cleanup_barriers): Use reorder_insns_nobb to avoid making
19397 the moved barrier the tail of the basic block it follows.
19398 * var-tracking.c (pass_variable_tracking): Add TODO_verify_flow.
19399
19400 2013-04-15 Jakub Jelinek <jakub@redhat.com>
19401
19402 PR tree-optimization/56962
19403 * gimple-ssa-strength-reduction.c (record_increment): Only set
19404 initializer if gimple_assign_rhs_code is {,POINTER_}PLUS_EXPR and
19405 either rhs1 or rhs2 is equal to c->base_expr.
19406
19407 2013-04-15 Richard Biener <rguenther@suse.de>
19408
19409 PR tree-optimization/56933
19410 * tree-vectorizer.h (struct _stmt_vec_info): Remove read_write_dep
19411 member.
19412 (GROUP_READ_WRITE_DEPENDENCE): Remove.
19413 (STMT_VINFO_GROUP_READ_WRITE_DEPENDENCE): Likewise.
19414 * tree-vect-data-refs.c (vect_analyze_group_access): Move
19415 dependence check ...
19416 vect_analyze_data_ref_dependence (vect_analyze_data_ref_dependence):
19417 ... here.
19418 * tree-vect-stmts.c (new_stmt_vec_info): Do not initialize
19419 GROUP_READ_WRITE_DEPENDENCE.
19420
19421 2013-04-15 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19422
19423 * emit-rtl.c (reset_all_used_flags): New function.
19424 (verify_rtl_sharing): Call reset_all_used_flags before and after
19425 performing the checks.
19426
19427 2013-04-15 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19428
19429 * config/arm/arm.c (const_ok_for_dimode_op): Handle AND case.
19430 * config/arm/arm.md (*anddi3_insn): Change to insn_and_split.
19431 * config/arm/constraints.md (De): New constraint.
19432 * config/arm/neon.md (anddi3_neon): Delete.
19433 (neon_vand<mode>): Expand to standard anddi3 pattern.
19434 * config/arm/predicates.md (imm_for_neon_inv_logic_operand):
19435 Move earlier in the file.
19436 (neon_inv_logic_op2): Likewise.
19437 (arm_anddi_operand_neon): New predicate.
19438
19439 2013-04-15 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
19440
19441 * configure.ac (gcc_cv_ld_as_needed): Set
19442 gcc_cv_ld_as_needed_option, gcc_cv_no_as_needed_option.
19443 Use -z ignore, -z record on *-*-solaris2*.
19444 (HAVE_LD_AS_NEEDED): Update comment.
19445 (LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION): Define.
19446 * configure: Regenerate.
19447 * config.in: Regenerate.
19448 * gcc.c (init_gcc_specs) [USE_LD_AS_NEEDED]: Use
19449 LD_AS_NEEDED_OPTION, LD_NO_AS_NEEDED_OPTION.
19450 * config/sol2.h [HAVE_LD_AS_NEEDED] (USE_LD_AS_NEEDED): Define.
19451 * doc/tm.texi.in (USE_LD_AS_NEEDED): Allow for --as-needed
19452 equivalents. Fix markup.
19453 * doc/tm.texi: Regenerate.
19454
19455 2013-04-15 Andrew Hsieh <andrewhsieh.google.com>
19456
19457 * config/i386/i386.opt: New option mstack-protector-guard=.
19458 * config/i386/i386-opts.h: Add enum stack_protector_guard.
19459 * config/i386/i386.h: Define TARGET_SSP_GLOBAL_GUARD and
19460 TARGET_SSP_TLS_GUARD.
19461 * config/i386/i386.c (ix86_option_override_internal): Set
19462 ix86_stack_protector_guard.
19463 * config/i386/i386.md (stack_protect_set): Enable for
19464 TARGET_SSP_TLS_GUARD only.
19465 (stack_protect_set_<mode>): Ditto.
19466 (stack_protect_test): Ditto.
19467 (stack_protect_test_<mode>): Ditto.
19468 * doc/invoke.texi (i386 Option): Document.
19469
19470 2013-04-15 Eric Botcazou <ebotcazou@adacore.com>
19471
19472 PR target/56890
19473 * config/sparc/sparc.c (enum sparc_mode_class): Add H_MODE value.
19474 (S_MODES): Set H_MODE bit.
19475 (SF_MODES): Set only S_MODE and SF_MODE bits.
19476 (DF_MODES): Set SF_MODES and only D_MODE and DF_MODE bits.
19477 (sparc_init_modes) <MODE_INT>: Set H_MODE bit for sub-word modes.
19478 <MODE_VECTOR_INT>: Do not set SF_MODE for sub-word modes.
19479 <MODE_FLOAT>: Likewise.
19480
19481 2013-04-15 Joey Ye <joey.ye@arm.com>
19482
19483 * config/arm/arm.c (thumb_far_jump_used_p): Fix typo in comments.
19484
19485 2013-04-15 Joey Ye <joey.ye@arm.com>
19486
19487 * config/arm/arm.c (thumb1_final_prescan_insn): Assert lr save
19488 for real far jump.
19489 (thumb_far_jump_used_p): Count instruction size and set
19490 far_jump_used.
19491
19492 2013-04-14 Eric Botcazou <ebotcazou@adacore.com>
19493
19494 * reorg.c (fill_simple_delay_slots): Reindent block of code.
19495 * resource.c (mark_target_live_regs): Reformat conditional block.
19496
19497 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19498
19499 * sched-deps.c (deps_analyze_insn): Do not check for EH_REGION insn
19500 notes, they are emitted only just before final.
19501 * sched-int.h: Include insn-attr.h before checking INSN_SCHEDULING.
19502
19503 2013-04-13 Steven Bosscher <steven@gcc.gnu.org>
19504
19505 * emit-rtl.c (remove_insn): Do not call df_insn_delete here.
19506 * cfgrtl.c (delete_insn): Call it here instead.
19507 * lra-spills.c (lra_final_code_change): Use delete_insn.
19508 * haifa-sched.c (sched_remove_insn): Likewise.
19509 * sel-sched-ir.c (return_nop_to_pool): Clear INSN_DELETED_P for nops
19510 returning to the nop pool.
19511 (sel_remove_insn): Simplify the only_disconnect case via remove_insn,
19512 use delete_insn for definitive removal. Clear BLOCK_FOR_INSN.
19513
19514 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19515
19516 * doc/tm.texi.in (LOOP_ALIGN): Remove loop note references.
19517 * doc/tm.texi: Regenerated.
19518
19519 2013-04-12 Uros Bizjak <ubizjak@gmail.com>
19520
19521 * config/i386/i386.c (ix86_hard_regno_mode_ok): Use ANY_QI_REGNO_P in
19522 QImode checks.
19523
19524 2013-04-12 Steven Bosscher <steven@gcc.gnu.org>
19525
19526 * df-core.c (df_find_def): Compare register numbers.
19527 (df_find_use): Likewise.
19528
19529 2013-04-12 Vladimir Makarov <vmakarov@redhat.com>
19530
19531 PR target/56903
19532 * config/i386/i386.c (ix86_hard_regno_mode_ok): Add
19533 lra_in_progress for return.
19534
19535 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19536
19537 * config/arm/arm.md (mov_scc,mov_negscc,mov_notscc): Convert
19538 define_insn into define_insn_and_split and emit movsicc patterns.
19539
19540 2013-04-12 Greta Yorsh <Greta.Yorsh@arm.com>
19541
19542 * config/arm/arm.c (gen_operands_ldrd_strd): Initialize "base".
19543
19544 2013-04-12 Richard Biener <rguenther@suse.de>
19545
19546 * tree-pass.h (TODO_do_not_ggc_collect): New.
19547 * passes.c (execute_one_ipa_transform_pass): Honor
19548 TODO_do_not_ggc_collect.
19549 (execute_one_pass): Likewise.
19550
19551 Revert
19552 2013-04-10 Richard Biener <rguenther@suse.de>
19553
19554 * passes.c (init_optimization_passes): Remove reload pass.
19555 * ira.c (do_reload): Merge into ...
19556 (ira): ... this.
19557 (rest_of_handle_reload): Remove.
19558 (pass_reload): Likewise.
19559 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19560 of reload for vzeroupper pass placement.
19561
19562 2013-04-12 Jakub Jelinek <jakub@redhat.com>
19563
19564 PR tree-optimization/56918
19565 PR tree-optimization/56920
19566 * fold-const.c (int_const_binop_1): Use op1.mul_with_sign (op2, ...)
19567 instead of op1 - op2. Pass 2 * TYPE_PRECISION (type) as second
19568 argument to rshift method. For 2 * HOST_BITS_PER_WIDE_INT precision
19569 use wide_mul_with_sign method.
19570
19571 2013-04-12 Richard Biener <rguenther@suse.de>
19572
19573 * gimple.c (is_gimple_constant): Vector CONSTRUCTORs should
19574 not be considered a gimple constant.
19575
19576 2013-04-12 Marc Glisse <marc.glisse@inria.fr>
19577
19578 * fold-const.c (const_binop): Handle vector shifts by a scalar.
19579 (fold_binary_loc): Call const_binop also for mixed vector-scalar
19580 operations.
19581
19582 2013-04-12 Manuel López-Ibáñez <manu@gcc.gnu.org>
19583 Jakub Jelinek <jakub@redhat.com>
19584
19585 * opts.c: Include diagnostic-color.h.
19586 (common_handle_option): Handle OPT_fdiagnostics_color_.
19587 * Makefile.in (OBJS-libcommon): Add diagnostic-color.o.
19588 (diagnostic.o, opts.o, pretty-print.o): Depend on diagnostic-color.h.
19589 (diagnostic-color.o): New.
19590 * common.opt (fdiagnostics-color, fdiagnostics-color=): New options.
19591 (diagnostic_color_rule): New enum.
19592 * dwarf2out.c (gen_producer_string): Don't print -fdiagnostics-color*.
19593 * langhooks.c (lhd_print_error_function): Add %r "locus" and %R around
19594 the location string.
19595 * diagnostic.def: Add 3rd argument to DEFINE_DIAGNOSTIC_KIND macros,
19596 either NULL, or color kind.
19597 * diagnostic-color.c: New file.
19598 * diagnostic-color.h: New file.
19599 * diagnostic-core.h (DEFINE_DIAGNOSTIC_KIND): Adjust macro for 3
19600 arguments.
19601 * doc/invoke.texi (-fdiagnostics-color): Document.
19602 * pretty-print.h (pp_show_color): Define.
19603 (struct pretty_print_info): Add show_color field.
19604 * diagnostic.c: Include diagnostic-color.h.
19605 (diagnostic_build_prefix): Adjust for 3 argument DEFINE_DIAGNOSTIC_KIND
19606 macros. Colorize error:, warning: etc. strings and also the location
19607 string.
19608 (diagnostic_show_locus): Colorize the caret line.
19609 * pretty-print.c: Include diagnostic-color.h.
19610 (pp_base_format): Handle %r and %R format specifiers. Colorize strings
19611 inside of %< %> quotes or quoted through q format modifier.
19612
19613 2013-04-12 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
19614
19615 * ifcvt.c (end_ifcvt_sequence): Mark a and b for unsharing as well.
19616
19617 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19618
19619 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Allow NEG
19620 code in CC_NZ mode.
19621 * config/aarch64/aarch64.md (*neg_<shift><mode>3_compare0): New
19622 pattern.
19623
19624 2013-04-11 Marek Polacek <polacek@redhat.com>
19625
19626 PR tree-optimization/48184
19627 * params.def (PARAM_ALIGN_THRESHOLD): Increase the minimum value to 1.
19628
19629 2013-04-11 Eric Botcazou <ebotcazou@adacore.com>
19630
19631 * stor-layout.c (skip_simple_constant_arithmetic): Move to...
19632 * tree.c (skip_simple_constant_arithmetic): ...here and make public.
19633 (skip_simple_arithmetic): Tidy up.
19634 * tree.h (skip_simple_constant_arithmetic): Declare.
19635
19636 2013-04-11 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
19637
19638 * config/aarch64/aarch64.h (REVERSIBLE_CC_MODE): Define.
19639
19640 2013-04-11 Richard Biener <rguenther@suse.de>
19641
19642 * tree-vect-loop.c (get_initial_def_for_induction): Properly
19643 generate vector constants.
19644
19645 2013-04-11 Richard Biener <rguenther@suse.de>
19646
19647 PR tree-optimization/56878
19648 * tree-flow.h (outermost_invariant_loop_for_expr): Declare.
19649 * tree-ssa-loop-ivopts.c (outermost_invariant_loop_for_expr):
19650 New function.
19651 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment):
19652 Prefer to align the DR with the most invariant base address.
19653
19654 2013-04-11 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19655
19656 * opts.c (common_handle_option): Fix formatting and add FALLTHRU
19657 comment.
19658
19659 2013-04-11 James Greenhalgh <james.greenhalgh@arm.com>
19660
19661 * config/aarch64/aarch64-simd.md (aarch64_vcond_internal): Fix
19662 floating-point vector comparisons against 0.
19663
19664 2013-04-11 Jakub Jelinek <jakub@redhat.com>
19665
19666 PR tree-optimization/56899
19667 * fold-const.c (extract_muldiv_1): Apply distributive law
19668 only if TYPE_OVERFLOW_WRAPS (ctype).
19669
19670 2013-04-11 Bin Cheng <bin.cheng@arm.com>
19671
19672 PR target/56124
19673 * ira-costs.c (scan_one_insn): Check whether the source rtx of
19674 loading has side effect.
19675
19676 2013-04-10 Steven Bosscher <steven@gcc.gnu.org>
19677
19678 * config/sparc/sparc.c: Include tree-pass.h.
19679 (TARGET_MACHINE_DEPENDENT_REORG): Do not redefine.
19680 (sparc_reorg): Rename to sparc_do_work_around_errata. Move to
19681 head of file. Change return type. Split off gate function.
19682 (sparc_gate_work_around_errata): New function.
19683 (pass_work_around_errata): New pass definition.
19684 (insert_pass_work_around_errata) New pass insert definition to
19685 insert pass_work_around_errata just after delayed-branch scheduling.
19686 (sparc_option_override): Insert the pass.
19687 * config/sparc/t-sparc (sparc.o): Add TREE_PASS_H dependence.
19688
19689 2013-04-10 David S. Miller <davem@davemloft.net>
19690
19691 * config/sparc/sparc.h (ASM_CPU_SPEC): Pass -Av8 if -mcpu=supersparc
19692 or -mcpu=hypersparc.
19693
19694 * target.def (cstore_mode): New hook.
19695 * target.h: Include insn-codes.h
19696 * targhooks.c: Likewise.
19697 (default_cstore_mode): New function.
19698 * targhooks.h: Declare it.
19699 * doc/tm.texi.in: New hook slot for TARGET_CSTORE_MODE.
19700 * doc/tm.texi: Rebuild.
19701 * expmed.c (emit_cstore): Obtain cstore boolean result mode using
19702 target hook, rather than inspecting the insn_data.
19703 * config/sparc/sparc.c (sparc_cstore_mode): New function.
19704 (TARGET_CSTORE_MODE): Redefine.
19705 (emit_scc_insn): When TARGET_ARCH64, emit new 64-bit boolean
19706 result patterns.
19707 * config/sparc/predicates.md (cstore_result_operand): New special
19708 predicate.
19709 * config/sparc/sparc.md (cstoresi4, cstoredi4, cstore<F:mode>4):
19710 Use it for operand 0.
19711 (*seqsi_special): Rewrite using 'P' mode iterator on operand 0.
19712 (*snesi_special): Likewise.
19713 (*snesi_zero): Likewise.
19714 (*seqsi_zero): Likewise.
19715 (*sltu_insn): Likewise.
19716 (*sgeu_insn): Likewise.
19717 (*seqdi_special): Make operand 0 and comparison operation be of
19718 DImode.
19719 (*snedi_special): Likewise.
19720 (*snedi_special_vis3): Likewise.
19721 (*neg_snesi_zero): Rename to *neg_snesisi_zero.
19722 (*neg_snesi_sign_extend): Rename to *neg_snesidi_zero.
19723 (*snesi_zero_extend): Delete, covered by 'P' mode iterator.
19724 (*neg_seqsi_zero): Rename to *neg_seqsisi_zero.
19725 (*neg_seqsi_sign_extend): Rename to *neg_seqsidi_zero.
19726 (*seqsi_zero_extend): Delete, covered by 'P' mode iterator.
19727 (*sltu_extend_sp64): Likewise.
19728 (*neg_sltu_insn): Rename to *neg_sltusi_insn.
19729 (*neg_sltu_extend_sp64): Rename to *neg_sltudi_insn.
19730 (*sgeu_extend_sp64): Delete, covered by 'P' mode iterator.
19731 (*neg_sgeu_insn): Rename to *neg_sgeusi_insn.
19732 (*neg_sgeu_extend_sp64): Rename to *neg_sgeudi_insn.
19733
19734 2013-04-10 Yufeng Zhang <yufeng.zhang@arm.com>
19735
19736 * config/aarch64/aarch64.c (aarch64_print_extension): New function.
19737 (aarch64_start_file): Use the new function.
19738
19739 2013-04-10 Senthil Kumar Selvaraj <senthil_kumar.selvaraj@atmel.com>
19740 Jason Merrill <jason@redhat.com>
19741
19742 * common.opt: Add -gdwarf.
19743 * opts.c (common_handle_option): Handle it.
19744 * gcc.c (ASM_DEBUG_SPEC): Don't expect "-2" for DWARF.
19745
19746 2013-04-10 Richard Biener <rguenther@suse.de>
19747
19748 * passes.c (execute_todo): Do not call ggc_collect conditional here.
19749 (execute_one_ipa_transform_pass): But unconditionally here.
19750 (execute_one_pass): And here.
19751 (init_optimization_passes): Remove reload pass.
19752 * tree-pass.h (TODO_ggc_collect): Remove.
19753 (pass_reload): Likewise.
19754 * ira.c (do_reload): Merge into ...
19755 (ira): ... this.
19756 (rest_of_handle_reload): Remove.
19757 (pass_reload): Likewise.
19758 * config/i386/i386.c (ix86_option_override): Refer to ira instead
19759 of reload for vzeroupper pass placement.
19760 * <everywhere>: Remove TODO_ggc_collect from todo_flags_start
19761 and todo_flags_finish of all passes.
19762
19763 2013-04-10 Richard Biener <rguenther@suse.de>
19764
19765 * tree-vectorizer.h (struct _slp_oprnd_info): Remove
19766 first_const_oprnd field, rename first_def_type to first_op_type.
19767 * tree-vect-slp.c (vect_create_oprnd_info): Adjust.
19768 (vect_get_and_check_slp_defs): Always use the type of the
19769 operand. Allow mixed vect_external_def, vect_constant_def types.
19770 (vect_get_constant_vectors): Handle mixed vect_external_def,
19771 vect_constant_def types.
19772
19773 2013-04-10 Joern Rennecke <joern.rennecke@embecosm.com>
19774
19775 PR tree-optimization/55524
19776 * tree-ssa-math-opts.c
19777 (convert_mult_to_fma): Don't use an fms construct
19778 when we don't have an fms operation, but fnma, and it looks
19779 likely that we'll be able to use the latter.
19780
19781 2013-04-10 Zhouyi Zhou <yizhouzhou@ict.ac.cn>
19782
19783 * cif-code.def (OVERWRITABLE): Correct the comment for overwritable
19784 function.
19785 * ipa-inline.c (can_inline_edge_p): Let dump mechanism report the
19786 inline fail caused by overwritable functions.
19787
19788 2013-04-10 Chung-Ju Wu <jasonwucj@gmail.com>
19789
19790 * combine.c (simplify_compare_const): Use GET_MODE_MASK to filter out
19791 unnecessary bits in the constant power of two case.
19792
19793 2013-04-10 Richard Biener <rguenther@suse.de>
19794
19795 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove
19796 broken code swapping operands.
19797 (vect_build_slp_tree): Do not compute load permutations here.
19798 (vect_analyze_slp_instance): Compute load permutations here,
19799 after building the SLP tree.
19800
19801 2013-04-09 Christian Bruel <christian.bruel@st.com>
19802
19803 * config/sh/sh.md (barrier_align): Use next/prev_active_insn instead
19804 of next/prev_real_insn.
19805
19806 2013-04-09 Jan Hubicka <jh@suse.cz>
19807
19808 * ipa.c (cgraph_externally_visible_p, varpool_externally_visible_p):
19809 Drop aliased parameter.
19810 (function_and_variable_visibility): Do not handle alias pairs.
19811 * cgraph.c (varpool_externally_visible_p): Update prototype.
19812 * varpool.c (varpool_add_new_variable): Update.
19813
19814 2013-04-09 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
19815
19816 * config/arm/arm.md (minmax_arithsi_non_canon): New pattern.
19817
19818 2013-04-09 Steven Bosscher <steven@gcc.gnu.org>
19819
19820 * sched-vis.c (print_pattern): Print SEQUENCE of insns as insns.
19821
19822 * config/sparc/sparc.md: Use define_c_enum for "unspec" and "unspecv".
19823
19824 2013-04-09 Marek Polacek <polacek@redhat.com>
19825
19826 PR tree-optimization/48762
19827 * params.def (PARAM_MAX_CSE_INSNS): Increase the minimum value to 1.
19828
19829 2013-04-09 Richard Biener <rguenther@suse.de>
19830
19831 * tree-vect-slp.c (vect_get_and_check_slp_defs): Remove code
19832 dealing with cost.
19833 (vect_build_slp_tree): Likewise.
19834 (vect_analyze_slp_cost_1, vect_analyze_slp_cost): New functions
19835 calculating the cost of a SLP instance.
19836 (vect_analyze_slp_instance): Use it from here, after building
19837 the SLP tree.
19838
19839 2013-04-09 Jakub Jelinek <jakub@redhat.com>
19840
19841 PR middle-end/56883
19842 * omp-low.c (expand_omp_for_generic, expand_omp_for_static_nochunk,
19843 expand_omp_for_static_chunk): Use simple_p = true in
19844 force_gimple_operand_gsi calls when assigning to addressable decls.
19845
19846 2013-04-09 Jeff Law <law@redhat.com>
19847
19848 * tree-vrp.c (simplify_cond_using_ranges): Simplify test of boolean
19849 when the boolean was created by converting a wider object which
19850 had a boolean range.
19851
19852 2013-04-09 Richard Biener <rguenther@suse.de>
19853
19854 * tree-vectorizer.h (slp_void_p): Remove.
19855 (slp_tree): Typedef before _slp_tree declaration.
19856 (struct _slp_tree): Use a vector of slp_tree as children.
19857 (vect_get_place_in_interleaving_chain): Remove.
19858 * tree-vect-data-refs.c (vect_get_place_in_interleaving_chain):
19859 Move ...
19860 * tree-vect-slp.c (vect_get_place_in_interleaving_chain): ... here
19861 and make static.
19862 (vect_free_slp_tree, vect_print_slp_tree, vect_mark_slp_stmts,
19863 vect_mark_slp_stmts_relevant, vect_slp_rearrange_stmts,
19864 vect_detect_hybrid_slp_stmts, vect_slp_analyze_node_operations,
19865 vect_schedule_slp_instance, vect_remove_slp_scalar_calls):
19866 Use slp_node instead of slp_void_p and adjust.
19867
19868 2013-04-09 Richard Biener <rguenther@suse.de>
19869
19870 * tree-ssa-loop-manip.c (rewrite_into_loop_closed_ssa): Avoid
19871 work that is not necessary.
19872
19873 2013-04-09 Jakub Jelinek <jakub@redhat.com>
19874
19875 PR tree-optimization/56854
19876 * tree-ssa-forwprop.c (forward_propagate_addr_expr_1): Don't
19877 forward into clobber stmts if it would change MEM_REF lhs into
19878 non-MEM_REF.
19879
19880 2013-04-09 Maxim Kuvyrkov <maxim@kugelworks.com>
19881
19882 * tree.c (type_hash_lookup, type_hash_add): Make static.
19883 * tree.h (type_hash_lookup, type_hash_add): Remove global declarations.
19884
19885 2013-04-09 Richard Biener <rguenther@suse.de>
19886
19887 * tree.h (unsave_expr_now): Remove.
19888 * tree-inline.c (mark_local_for_remap_r): Remove.
19889 (unsave_expr_1): Likewise.
19890 (unsave_r): Likewise.
19891 (unsave_expr_now): Likewise.
19892 * tree-ssa-copy.c (replace_exp_1): Use unshare_expr.
19893 (propagate_tree_value): Likewise.
19894
19895 2013-04-08 Steven Bosscher <steven@gcc.gnu.org>
19896
19897 * doc/rtl.texi (sequence): Rewrite documentation to match the
19898 current use of SEQUENCE rtl objects.
19899 * rtl.def (SEQUENCE): Likewise.
19900
19901 * doc/rtl.texi (NOTE_INSN_EH_REGION_BEG, NOTE_INSN_EH_REGION_END):
19902 Update documentation.
19903 (NOTE_INSN_LOOP_BEG, NOTE_INSN_LOOP_END, NOTE_INSN_LOOP_CONT,
19904 NOTE_INSN_LOOP_VTOP): Remove documentation for non-existing notes.
19905
19906 * reg-notes.def (REG_EH_CONTEXT): Remove unused note.
19907
19908 2013-04-08 Teresa Johnson <tejohnson@google.com>
19909
19910 * basic-block.h (GCOV_COMPUTE_SCALE): Define.
19911 * ipa-inline-analysis.c (param_change_prob): Use helper rounding divide
19912 methods.
19913 (estimate_edge_size_and_time): Add comment to suggest using rounding
19914 methods.
19915 (estimate_node_size_and_time): Ditto.
19916 (remap_edge_change_prob): Use helper rounding divide methods.
19917 * value-prof.c (gimple_divmod_fixed_value_transform): Ditto.
19918 (gimple_mod_pow2_value_transform): Ditto.
19919 (gimple_mod_subtract_transform): Ditto.
19920 (gimple_ic_transform): Ditto.
19921 (gimple_stringops_transform): Ditto.
19922 * stmt.c (conditional_probability): Ditto.
19923 (emit_case_dispatch_table): Ditto.
19924 * lto-cgraph.c (merge_profile_summaries): Ditto.
19925 * tree-optimize.c (execute_fixup_cfg): Ditto.
19926 * cfgcleanup.c (try_forward_edges): Ditto.
19927 * cfgloopmanip.c (scale_loop_profile): Ditto.
19928 (loopify): Ditto.
19929 (duplicate_loop_to_header_edge): Ditto.
19930 (lv_adjust_loop_entry_edge): Ditto.
19931 * tree-vect-loop.c (vect_transform_loop): Ditto.
19932 * profile.c (compute_branch_probabilities): Ditto.
19933 * cfgbuild.c (compute_outgoing_frequencies): Ditto.
19934 * lto-streamer-in.c (input_cfg): Ditto.
19935 * gimple-streamer-in.c (input_bb): Ditto.
19936 * ipa-cp.c (update_profiling_info): Ditto.
19937 (update_specialized_profile): Ditto.
19938 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge): Ditto.
19939 * cfg.c (update_bb_profile_for_threading): Add comment to suggest using
19940 rounding methods.
19941 * sched-rgn.c (compute_dom_prob_ps): Ditto.
19942 (compute_trg_info): Ditto.
19943 * cfgrtl.c (force_nonfallthru_and_redirect): Ditto.
19944 (purge_dead_edges): Ditto.
19945 * loop-unswitch.c (unswitch_loop): Ditto.
19946 * cgraphclones.c (cgraph_clone_edge): Ditto.
19947 (cgraph_clone_node): Ditto.
19948 * tree-inline.c (copy_bb): Ditto.
19949 (copy_edges_for_bb): Ditto.
19950 (initialize_cfun): Ditto.
19951 (copy_cfg_body): Ditto.
19952 (expand_call_inline): Ditto.
19953
19954 2013-04-08 Kai Tietz <ktietz@redhat.com>
19955
19956 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Replaced
19957 TARGET_CYGWIN64 by TARGET_64BIT.
19958
19959 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
19960
19961 * config/epiphany/epiphany.md (GPR_1): New constant.
19962 (define_expand "mov<mode>cc): FAIL if gen_compare_reg returned 0.
19963 * config/epiphany/epiphany.c (gen_compare_reg):
19964 For flag_finite_math_only, avoid swapping operands when r0 and/or r1
19965 is already in place.
19966 Use GPR_0 / GPR_1 instead of 0/1 for r0/r1 register numbers.
19967 Don't require being called during rtl expansion; If y operlaps r0,
19968 return 0.
19969 (epiphany_compute_frame_size, epiphany_expand_prologue): Use GPR_1.
19970 (epiphany_expand_epilogue): Likewise.
19971
19972 * config/epiphany/epiphany.c (epiphany_select_cc_mode):
19973 Don't use CC_FPmode for ORDERED / UNORDERED.
19974 * config/epiphany/epiphany.md (cmpsf_ord): Make pattern unconditional.
19975
19976 * config/epiphany/constraints.md (CnL): New constraint.
19977 * config/epiphany/epiphany.md (addsi3_i): Add r/r/CnL alternative.
19978 * config/epiphany/predicates.md (add_operand): Allow 1024.
19979
19980 * config/epiphany/epiphany.md (logical_op): New code iterator.
19981 (op_mnc): New code attribute.
19982 (<op_mnc>_f, mov_f, cstoresi4): New patterns.
19983 (mov_f+1, mov_f+2): New peephole2 patterns.
19984
19985 * config/epiphany/epiphany.md (mov_f+2): New peephole2 pattern.
19986 (cstoresi4): Also allow re-use of zero result when doing a NE
19987 comparison to a non-zero operand.
19988 Use (clobber (scratch)) for first insn if the gpr output is not needed.
19989
19990 * config/epiphany/epiphany.md (<insn_opname>v2si3):
19991 Use gen_addsi3_i / gen_subsi3_i.
19992
19993 2013-04-08 Jakub Jelinek <jakub@redhat.com>
19994
19995 PR c++/34949
19996 PR c++/50243
19997 * tree-eh.c (optimize_clobbers): Only remove clobbers if bb doesn't
19998 contain anything but clobbers, at most one __builtin_stack_restore,
19999 optionally debug stmts and final resx, and if it has at least one
20000 incoming EH edge. Don't check for SSA_NAME on LHS of a clobber.
20001 (sink_clobbers): Don't check for SSA_NAME on LHS of a clobber.
20002 Instead of moving clobbers with MEM_REF LHS with SSA_NAME address
20003 which isn't defaut definition, remove them.
20004 (unsplit_eh, cleanup_empty_eh): Use single_{pred,succ}_{p,edge}
20005 instead of EDGE_COUNT comparisons or EDGE_{PRED,SUCC}.
20006 * tree-ssa-ccp.c (execute_fold_all_builtins): Remove clobbers
20007 with MEM_REF LHS with SSA_NAME address.
20008
20009 2013-04-08 Jeff Law <law@redhat.com>
20010
20011 * gimple.c (canonicalize_cond_expr_cond): Rewrite x ^ y into x != y.
20012
20013 2013-04-08 Richard Biener <rguenther@suse.de>
20014
20015 * gimple-pretty-print.c (debug_gimple_stmt): Do not print
20016 extra newline.
20017 * tree-vect-loop.c (vect_determine_vectorization_factor): Dump
20018 determined vector type.
20019 (vect_analyze_data_refs): Likewise.
20020 (vect_get_new_vect_var): Adjust.
20021 (vect_create_destination_var): Preserve SSA name versions.
20022 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): Do
20023 not dump anything here.
20024
20025 2013-04-08 Joern Rennecke <joern.rennecke@embecosm.com>
20026
20027 * config/epiphany/epiphany.h (struct GTY (()) machine_function):
20028 Add member lr_slot_known.
20029 * config/epiphany/epiphany.md (reload_insi_ra): Compute lr_slot_offs
20030 if necessary.
20031 * config/epiphany/epiphany.c (epiphany_compute_frame_size):
20032 Remove code that sets lr_slot_offset according to what a previous
20033 version of epiphany_emit_save_restore used to do.
20034 (epiphany_emit_save_restore): When doing an lr save or restore,
20035 set/verify lr_slot_known and lr_slot_offset.
20036
20037 2013-04-08 Xinyu Qi <xyqi@marvell.com>
20038
20039 PR target/54338
20040 * config/arm/arm.h (REG_CLASS_CONTENTS): Include IWMMXT_GR_REGS
20041 in ALL_REGS.
20042
20043 2013-04-08 Richard Biener <rguenther@suse.de>
20044
20045 * alias.c (find_base_term): Fix thinko in previous change.
20046
20047 2013-04-08 Jakub Jelinek <jakub@redhat.com>
20048
20049 * tree-loop-distribution.c (const_with_all_bytes_same): New function.
20050 (generate_memset_builtin): Only handle integer_all_onesp as -1 val if
20051 TYPE_PRECISION is equal to mode bitsize. Use const_with_all_bytes_same
20052 if possible to compute val.
20053 (classify_partition): Verify CONSTRUCTOR doesn't have any elts.
20054 For QImode integers don't require anything about precision. Use
20055 const_with_all_bytes_same to find out if the constant doesn't have
20056 repeated bytes in it.
20057
20058 2013-04-08 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
20059
20060 * config/s390/s390.c (s390_expand_insv): Only accept insertions
20061 within mode size.
20062
20063 2013-04-08 Marek Polacek <polacek@redhat.com>
20064
20065 PR rtl-optimization/48182
20066 * params.def (PARAM_MIN_CROSSJUMP_INSNS): Increase the minimum
20067 value to 1.
20068
20069 2013-04-06 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
20070
20071 PR target/55487
20072 * config/pa/pa.c (legitimize_pic_address): Before incrementing label
20073 nuses, make sure we have a label.
20074
20075 2013-04-05 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
20076
20077 PR target/56843
20078 * config/rs6000/rs6000.c (rs6000_emit_swdiv_high_precision): Remove.
20079 (rs6000_emit_swdiv_low_precision): Remove.
20080 (rs6000_emit_swdiv): Rewrite to handle between one and four
20081 iterations of Newton-Raphson generally; modify required number of
20082 iterations for some cases.
20083 * config/rs6000/rs6000.h (RS6000_RECIP_HIGH_PRECISION_P): Remove.
20084
20085 2013-04-05 Steven Bosscher <steven@gcc.gnu.org>
20086
20087 * bb-reorder.c (fix_crossing_unconditional_branches): Remove a
20088 set-but-unused variable.
20089
20090 * cgraph.c (cgraph_release_function_body): Clear cfun->cfg to make
20091 basic blocks of released function bodies garbage-collectable.
20092
20093 * ree.c (find_and_remove_re): Do not call df_finish_pass here.
20094 (struct rtl_opt_pass): Add TODO_df_finish.
20095
20096 * rtl.def (DEFINE_SUBST, DEFINE_SUBST_ATTR): Add documentation.
20097
20098 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20099
20100 * config/arm/constraints.md (q): New constraint.
20101 * config/arm/ldrdstrd.md: New file.
20102 * config/arm/arm.md (ldrdstrd.md) New include.
20103 (arm_movdi): Use "q" instead of "r" constraint
20104 for double-word memory access.
20105 (movdf_soft_insn): Likewise.
20106 * config/arm/vfp.md (movdi_vfp): Likewise.
20107 * config/arm/t-arm (MD_INCLUDES): Add ldrdstrd.md.
20108 * config/arm/arm-protos.h (gen_operands_ldrd_strd): New declaration.
20109 * config/arm/arm.c (gen_operands_ldrd_strd): New function.
20110 (mem_ok_for_ldrd_strd): Likewise.
20111 (output_move_double): Update assertion.
20112
20113 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20114
20115 * config/arm/arm.md: Comment on splitting Thumb1 patterns.
20116
20117 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20118
20119 * config/arm/arm.md (arm_smax_insn): Convert define_insn into
20120 define_insn_and_split.
20121 (arm_smin_insn,arm_umaxsi3,arm_uminsi3): Likewise.
20122
20123 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20124
20125 * config/arm/arm.md (arm_ashldi3_1bit): Convert define_insn into
20126 define_insn_and_split.
20127 (arm_ashrdi3_1bit,arm_lshrdi3_1bit): Likewise.
20128 (shiftsi3_compare): New pattern.
20129 (rrx): New pattern.
20130 * config/arm/unspecs.md (UNSPEC_RRX): New.
20131
20132 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20133
20134 * config/arm/arm.md (negdi_extendsidi): New pattern.
20135 (negdi_zero_extendsidi): Likewise.
20136
20137 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20138
20139 * config/arm/arm.md (andsi_iorsi3_notsi): Convert define_insn into
20140 define_insn_and_split.
20141 (arm_negdi2,arm_abssi2,arm_neg_abssi2): Likewise.
20142 (arm_cmpdi_insn,arm_cmpdi_unsigned): Likewise.
20143
20144 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20145
20146 * config/arm/arm.md (arm_subdi3): Convert define_insn into
20147 define_insn_and_split.
20148 (subdi_di_zesidi,subdi_di_sesidi): Likewise.
20149 (subdi_zesidi_di,subdi_sesidi_di,subdi_zesidi_zesidi): Likewise.
20150
20151 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20152
20153 * config/arm/arm.md (subsi3_carryin): New pattern.
20154 (subsi3_carryin_const): Likewise.
20155 (subsi3_carryin_compare,subsi3_carryin_compare_const): Likewise.
20156 (subsi3_carryin_shift,rsbsi3_carryin_shift): Likewise.
20157
20158 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20159
20160 * config/arm/arm.md (incscc,arm_incscc,decscc,arm_decscc): Delete.
20161
20162 2013-04-05 Greta Yorsh <Greta.Yorsh@arm.com>
20163
20164 * config/arm/arm.md (addsi3_carryin_<optab>): Set attribute predicable.
20165 (addsi3_carryin_alt2_<optab>,addsi3_carryin_shift_<optab>): Likewise.
20166
20167 2013-04-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20168
20169 * config/arm/arm.c (arm_expand_builtin): Change fcode
20170 type to unsigned int.
20171
20172 2013-04-05 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20173
20174 * doc/invoke.texi (ARM Options): Document cortex-a53 support.
20175
20176 2013-04-04 Ian Lance Taylor <iant@google.com>
20177
20178 * doc/standards.texi (Standards): The Go frontend supports the Go 1
20179 language standard.
20180
20181 2013-04-04 Steven Bosscher <steven@gcc.gnu.org>
20182
20183 PR middle-end/56729
20184 * df-scan.c (df_insn_delete): Disable failing assert.
20185
20186 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20187
20188 * config/arm/arm-protos.h (arm_builtin_vectorized_function):
20189 New function prototype.
20190 * config/arm/arm.c (TARGET_VECTORIZE_BUILTINS): Define.
20191 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Likewise.
20192 (arm_builtin_vectorized_function): New function.
20193
20194 2013-04-04 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20195
20196 * config/arm/arm_neon_builtins.def: New file.
20197 * config/arm/arm.c (neon_builtin_data): Move contents to
20198 arm_neon_builtins.def.
20199 (enum arm_builtins): Include neon builtin definitions.
20200 (ARM_BUILTIN_NEON_BASE): Move from enum to macro.
20201 * config/arm/t-arm (arm.o): Add dependency on arm_neon_builtins.def.
20202
20203 2013-04-04 Marek Polacek <polacek@redhat.com>
20204
20205 PR tree-optimization/48186
20206 * predict.c (maybe_hot_frequency_p): Return false if
20207 HOT_BB_FREQUENCY_FRACTION is 0.
20208 (cgraph_maybe_hot_edge_p): Likewise.
20209
20210 2013-04-04 Richard Biener <rguenther@suse.de>
20211
20212 PR tree-optimization/56826
20213 * tree-vect-slp.c (vect_build_slp_tree): Compute ncopies
20214 more accurately.
20215
20216 2013-04-04 Richard Biener <rguenther@suse.de>
20217
20218 PR tree-optimization/56213
20219 * tree-vect-data-refs.c (vect_check_strided_load): Remove.
20220 (vect_analyze_data_refs): Allow all non-nested loads as strided loads.
20221
20222 2013-04-04 Richard Biener <rguenther@suse.de>
20223
20224 PR tree-optimization/56837
20225 * tree-loop-distribution.c (classify_partition): For non-zero
20226 values require that the value has the same precision as its
20227 mode to be useful as memset value.
20228
20229 2013-04-03 Nick Clifton <nickc@redhat.com>
20230
20231 * config/v850/v850e3v5.md (fmasf4): Use fmaf.s on E3V5 architectures.
20232 (fmssf4): Use fmsf.s on E3V5 architectures.
20233 (fnmasf4): Use fnmaf.s on E3V5 architectures.
20234 (fnmssf4): Use fnmsf.s on E3V5 architectures.
20235
20236 2013-04-03 Jeff Law <law@redhat.com>
20237
20238 * Makefile.in (lra-constraints.o): Depend on $(OPTABS_H).
20239 (lra-eliminations.o): Likewise.
20240
20241 2013-04-03 Teresa Johnson <tejohnson@google.com>
20242
20243 * gcov-io.c (compute_working_sets): Moved most of body of old
20244 compute_working_sets here from profile.c.
20245 * gcov-io.h (NUM_GCOV_WORKING_SETS): Moved here from profile.c.
20246 (gcov_working_set_t): Moved typedef here from basic-block.h
20247 (compute_working_set): Declare.
20248 * profile.c (NUM_GCOV_WORKING_SETS): Moved to gcov-io.h.
20249 (get_working_sets): Renamed from compute_working_set,
20250 replace most of body with call to new compute_working_sets.
20251 (get_exec_counts): Replace call to compute_working_sets
20252 to get_working_sets.
20253 * profile.h (get_working_sets): Renamed from compute_working_set.
20254 * lto-cgraph.c (input_symtab): Replace call to compute_working_sets
20255 to get_working_sets.
20256 * basic-block.h (gcov_working_set_t): Moved to gcov-io.h.
20257 * gcov-dump.c (dump_working_sets): New function.
20258
20259 2013-04-03 Kenneth Zadeck <zadeck@naturalbridge.com>
20260
20261 * hwint.c (sext_hwi, zext_hwi): New functions.
20262 * hwint.h (HOST_BITS_PER_HALF_WIDE_INT, HOST_HALF_WIDE_INT,
20263 HOST_HALF_WIDE_INT_PRINT, HOST_HALF_WIDE_INT_PRINT_C,
20264 HOST_HALF_WIDE_INT_PRINT_DEC, HOST_HALF_WIDE_INT_PRINT_DEC_C,
20265 HOST_HALF_WIDE_INT_PRINT_UNSIGNED, HOST_HALF_WIDE_INT_PRINT_HEX,
20266 HOST_HALF_WIDE_INT_PRINT_HEX_PURE): New symbols.
20267 (sext_hwi, zext_hwi): New functions.
20268
20269 2013-04-03 Jeff Law <law@redhat.com>
20270
20271 PR tree-optimization/56799
20272 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Bring
20273 back test for widening conversion erroneously dropped in prior change.
20274
20275 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20276
20277 PR target/56809
20278 * config/aarch64/aarch64.c (is_jump_table): Use next_active_insn
20279 instead of next_real_insn.
20280
20281 2013-04-03 Marek Polacek <polacek@redhat.com>
20282
20283 PR sanitizer/55702
20284 * tsan.c (instrument_func_exit): Allow BUILT_IN_RETURN functions.
20285
20286 2013-04-03 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
20287
20288 PR target/56809
20289 * config/arm/arm.c (is_jump_table): Use next_active_insn instead of
20290 next_real_insn.
20291 (thumb1_output_casesi): Likewise.
20292 (thumb2_output_casesi): Likewise.
20293
20294 2013-04-03 Richard Biener <rguenther@suse.de>
20295
20296 PR tree-optimization/56817
20297 * tree-ssa-loop-ivcanon.c (tree_unroll_loops_completely):
20298 Split out ...
20299 (tree_unroll_loops_completely_1): ... new function to manually
20300 walk the loop tree, properly defering outer loops of unrolled
20301 loops to later iterations.
20302
20303 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20304
20305 * tree-vect-stmts.c (vectorizable_store): Accept BIT_FIELD_REF.
20306 (vectorizable_load): Likewise.
20307 * tree-vect-slp.c (vect_build_slp_tree): Likewise.
20308 * tree-vect-data-refs.c (vect_create_data_ref_ptr): Handle VECTOR_TYPE.
20309
20310 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20311
20312 * tree-flow-inline.h (get_addr_base_and_unit_offset_1): Handle
20313 BIT_FIELD_REF.
20314
20315 2013-04-03 Ulrich Weigand <Ulrich.Weigand@de.ibm.com>
20316
20317 * config/spu/spu.c (emit_nop_for_insn): Handle JUMP_TABLE_DATA.
20318
20319 2013-04-03 Bin Cheng <bin.cheng@arm.com>
20320
20321 * rtl.h (AUTO_INC_DEC): Fix typo of HAVE_POST_MODIFY_DISP.
20322
20323 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20324
20325 PR tree-optimization/56790
20326 * fold-const.c (fold_ternary_loc) <VEC_COND_EXPR>: Add constant
20327 folding.
20328
20329 2013-04-03 Marc Glisse <marc.glisse@inria.fr>
20330
20331 * simplify-rtx.c (simplify_binary_operation_1) <VEC_SELECT>:
20332 Handle VEC_MERGE.
20333 (simplify_ternary_operation) <VEC_MERGE>: Use unsigned HOST_WIDE_INT
20334 for masks. Test for side effects. Handle nested VEC_MERGE. Handle
20335 equal arguments.
20336
20337 2013-04-03 Jakub Jelinek <jakub@redhat.com>
20338
20339 PR c/19449
20340 * tree.h (force_folding_builtin_constant_p): New decl.
20341 * builtins.c (force_folding_builtin_constant_p): New variable.
20342 (fold_builtin_constant_p): Fold immediately also if
20343 force_folding_builtin_constant_p.
20344
20345 2013-04-03 Richard Biener <rguenther@suse.de>
20346
20347 PR tree-optimization/56812
20348 * tree-vect-data-refs.c (vect_slp_analyze_data_ref_dependence):
20349 DRs of the same interleaving chain are independent.
20350
20351 2013-04-02 Jason Merrill <jason@redhat.com>
20352
20353 * gdbinit.in (pbb): Use debug fn.
20354
20355 2013-04-02 Lawrence Crowl <crowl@google.com>
20356
20357 * sese.h (struct ivtype_map_elt_s): Remove unused.
20358 (extern debug_ivtype_map): Remove unused.
20359 (extern eq_ivtype_map_elts): Remove unused.
20360 * sese.c (debug_ivtype_map): Removed unused.
20361 (debug_ivtype_map_1): Removed unused.
20362 (debug_ivtype_elt): Remove unused.
20363 (eq_ivtype_map_elts): Remove unused.
20364
20365
20366 2013-04-02 Kai Tietz <ktietz@redhat.com>
20367
20368 PR target/52790
20369 * config/i386/cygming.h (SUB_TARGET_RECORD_STUB): New sub-target macro.
20370 * config/i386/i386-protos.h (i386_pe_record_stub): Add new prototype.
20371 * config/i386/i386.c (legitimize_pe_coff_extern_decl): New static
20372 function.
20373 (legitimize_pe_coff_symbol): Likewise.
20374 (is_imported_p): New helper-function.
20375 (ix86_option_override_internal): Make MEDIUM_PIC the default code-model
20376 for Windows x64 targets.
20377 (ix86_expand_prologue): Optimize for pe-coff targets.
20378 (ix86_expand_split_stack_prologue): Adjust for pe-coff targets.
20379 (legitimate_pic_address_disp_p): Adjust for x64 pe-coff to support
20380 medium/large code-model.
20381 (legitimize_pic_address): Likewise.
20382 (legitimize_tls_address): Likewise.
20383 (ix86_expand_call): Likewise.
20384 (x86_output_mi_thunk): Likewise.
20385 (get_dllimport_decl): Add new beimport argument.
20386 (construct_plt_address): Don't assert for x64 pe-coff targets.
20387 * config/i386/i386.h (PIC_OFFSET_TABLE_REGNUM): Adjust for x64 pe-coff
20388 targets.
20389 (SYMBOL_FLAG_STUBVAR): New macro.
20390 (SYMBOL_REF_STUBVAR_P): Likewise.
20391 * config/i386/winnt.c (stub_list): New structure.
20392 (stub_head): New local variable.
20393 (i386_pe_record_stub): New function.
20394 (i386_pe_file_end): Emit refptr-stubs.
20395
20396 2013-04-02 Jakub Jelinek <jakub@redhat.com>
20397
20398 PR rtl-optimization/56745
20399 * ifcvt.c (cond_exec_find_if_block): Don't try to optimize
20400 if then_bb has no successors and else_bb is EXIT_BLOCK_PTR.
20401
20402 PR c++/34949
20403 * tree-ssa-alias.c (stmt_kills_ref_p_1): If base != ref->base
20404 and both of them are MEM_REFs, just compare first argument for
20405 equality and attempt to deal even with differing offsets.
20406
20407 PR c++/34949
20408 * tree-cfg.c (verify_gimple_assign_single): Allow lhs
20409 of gimple_clobber_p to be MEM_REF.
20410 * gimplify.c (gimplify_modify_expr): Gimplify *to_p of
20411 an assignment from TREE_CLOBBER_P. Allow it to be MEM_REF
20412 after gimplification.
20413 * asan.c (get_mem_ref_of_assignment): Don't instrument
20414 gimple_clobber_p stmts.
20415 * tree-ssa-dse.c (dse_optimize_stmt): Allow DSE of
20416 gimple_clobber_p stmt if they have MEM_REF lhs and
20417 are dead because of another gimple_clobber_p stmt.
20418 * tree-ssa-live.c (clear_unused_block_pointer): Treat
20419 gimple_clobber_p stmts like debug stmts.
20420 (remove_unused_locals): Remove clobbers with MEM_REF lhs
20421 that refer to unused VAR_DECLs or uninitialized values.
20422 * tree-sra.c (sra_ipa_reset_debug_stmts): Also remove
20423 gimple_clobber_p stmts if they refer to removed parameters.
20424 (get_repl_default_def_ssa_name, sra_ipa_modify_expr): Fix up
20425 formatting.
20426
20427 2013-04-02 Uros Bizjak <ubizjak@gmail.com>
20428
20429 * config/i386/i386.md (*testqi_ext_3): Merge with *testqi_ext_3_rex64
20430 using SWI48 mode attribute.
20431
20432 2013-04-02 Wei Mi <wmi@google.com>
20433
20434 * config/i386/i386.c (ix86_rtx_costs): Set proper rtx cost for
20435 ashl<mode>3_mask, *<shift_insn><mode>3_mask and
20436 *<rotate_insn><mode>3_mask in i386.md.
20437
20438 2013-04-02 Alexander Ivchenko <alexander.ivchenko@intel.com>
20439
20440 * config.gcc (arm*-*-linux-*): Remove duplicate t-linux-android.
20441
20442 2013-04-02 Richard Biener <rguenther@suse.de>
20443
20444 PR tree-optimization/56778
20445 * tree-vect-data-refs.c (vect_analyze_data_ref_dependence):
20446 Runtime alias tests are not supported for gather loads.
20447 * tree-vect-loop-manip.c (vect_loop_versioning): Insert
20448 stmts referenced from SSA operands before updating SSA form.
20449
20450 2013-04-02 Ian Caulfield <ian.caulfield@arm.com>
20451 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
20452
20453 * config/arm/arm-arches.def (armv8-a): Default to cortex-a53.
20454 * config/arm/t-arm (MD_INCLUDES): Depend on cortex-a53.md.
20455 * config/arm/cortex-a53.md: New file.
20456 * config/arm/bpabi.h (BE8_LINK_SPEC): Handle cortex-a53.
20457 * config/arm/arm.md (generic_sched, generic_vfp): Handle cortex-a53.
20458 * config/arm/arm.c (arm_issue_rate): Likewise.
20459 * config/arm/arm-tune.md: Regenerate
20460 * config/arm/arm-tables.opt: Regenerate.
20461 * config/arm/arm-cores.def: Add cortex-a53.
20462
20463 2013-04-02 Zhenqiang Chen <zhenqiang.chen@arm.com>
20464
20465 * config/arm/uclinux-elf.h: Add %L to LINK_GCC_C_SEQUENCE_SPEC for
20466 non-static link.
20467
20468 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20469
20470 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add variants for
20471 scalar load/store operations using B/H registers.
20472 (*zero_extend<SHORT:mode><GPI:mode>2_aarch64): Likewise.
20473
20474 2013-04-02 Sofiane Naci <sofiane.naci@arm.com>
20475
20476 * config/aarch64/aarch64.md (*mov<mode>_aarch64): Add alternatives for
20477 scalar move.
20478 * config/aarch64/aarch64.c
20479 (aarch64_simd_scalar_immediate_valid_for_move): New.
20480 * config/aarch64/aarch64-protos.h
20481 (aarch64_simd_scalar_immediate_valid_for_move): New.
20482 * config/aarch64/constraints.md (Dh, Dq): New.
20483 * config/aarch64/iterators.md (hq): New.
20484
20485 2013-04-02 Eric Botcazou <ebotcazou@adacore.com>
20486
20487 * reorg.c (get_branch_condition): Deal with conditional returns.
20488 (fill_simple_delay_slots): Remove dead code dealing with jumps.
20489
20490 2013-04-01 Wei Mi <wmi@google.com>
20491
20492 * config/i386/i386.md (*ashl<mode>3_mask): Rewrite as define_insn.
20493 Truncate operand 2 using %b asm operand modifier.
20494 (*<shift_insn><mode>3_mask): Ditto.
20495 (*<rotate_insn><mode>3_mask): Ditto.
20496
20497 2013-04-01 Steven Bosscher <steven@gcc.gnu.org>
20498
20499 PR middle-end/56798
20500 * cfgbuild.c (inside_basic_block_p): Restore check broken at r197234.
20501
20502 2013-03-31 Kaz Kojima <kkojima@gcc.gnu.org>
20503
20504 * config/sh/sh.md (casesi_worker_1): Use next_active_insn instead
20505 of next_real_insn.
20506 (casesi_worker_2, casesi_shift_media, casesi_load_media): Likewise.
20507
20508 2013-03-30 Lawrence Crowl <crowl@google.com>
20509
20510 * dse.c (clear_alias_sets): Remove never set.
20511 (disqualified_clear_alias_sets): Remove never set.
20512 (clear_alias_mode_pool): Remove never set.
20513 (dse_step0): Remove condition that is never true.
20514 (canon_address): Remove condition that is never true.
20515 (dse_step7): Remove condition that is never true.
20516 (rest_of_handle_dse): Remove condition that is never true.
20517 (rest_of_handle_dse::did_global): Remove never read from above.
20518 (dse_step2_spill): Remove never called from above.
20519 (dse_step5_spill): Remove never called from above.
20520
20521 2013-03-30 Steven Bosscher <steven@gcc.gnu.org>
20522
20523 * doc/md.texi (Standard Names) <casesi>: Update documentation for
20524 JUMP_TABLE_DATA changes.
20525 * doc/tm.texi.in (Dispatch Tables) <ASM_OUTPUT_CASE_LABEL>: Likewise.
20526 * doc/rtl.texi (Flags) <INSN_DELETED_P, SCHED_GROUP_P>: Likewise.
20527 (Insns) <jump_table_data>: New entry.
20528 * doc/tm.texi: Regenerate.
20529
20530 * cfgrtl.c (fixup_reorder_chain): Do not emit barriers to BB_FOOTER.
20531
20532 * postreload-gcse.c (bb_has_well_behaved_predecessors): Correct test
20533 for table jump at the end of a basic block using tablejump_p.
20534 * targhooks.c (default_invalid_within_doloop): Likewise.
20535 * config/rs6000/rs6000.c (TARGET_INVALID_WITHIN_DOLOOP): Remove
20536 target hook implementation that is identical to the default hook.
20537 (rs6000_invalid_within_doloop): Remove.
20538
20539 * bb-reorder.c (fix_crossing_unconditional_branches): Remove set but
20540 unused variable from tablejump_p call.
20541
20542 * rtl.def (JUMP_TABLE_DATA): New RTX_INSN object.
20543 * rtl.h (RTX_PREV, RTX_NEXT): Adjust for new JUMP_TABLE_DATA.
20544 (INSN_DELETED_P): Likewise.
20545 (emit_jump_table_data): New prototype.
20546 * gengtype.c (adjust_field_rtx_def): Handle JUMP_TABLE_DATA fields
20547 after 4th as unused.
20548 * print-rtl.c (print_rtl): Handle JUMP_TABLE_DATA.
20549 * sched-vis.c (print_insn): Likewise.
20550 * emit-rtl.c (active_insn_p): Consider JUMP_TABLE_DATA an active
20551 insn for compatibility with back ends that use next_active_insn to
20552 identify jump table data.
20553 (set_insn_deleted): Remove no longer useful JUMP_TABLE_DATA_P check.
20554 (remove_insn): Likewise.
20555 (emit_insn): Do not accept JUMP_TABLE_DATA objects in insn chains
20556 to be emitted.
20557 (emit_debug_insn, emit_jump_insn, emit_call_insn, emit_label): Idem.
20558 (emit_jump_table_data): New function.
20559
20560 * cfgbuild.c (inside_basic_block_p): A JUMP_INSN is always inside a
20561 basic block, a JUMP_TABLE_DATA never is.
20562 (control_flow_insn_p): JUMP_TABLE_DATA is not a control flow insn.
20563 * cfgrtl.c (duplicate_insn_chain): Split handling of JUMP_TABLE_DATA
20564 off from code handling real insns.
20565 * final.c (get_attr_length_1): Simplify for JUMP_INSNs.
20566 * function.c (instantiate_virtual_regs): Remove JUMP_TABLE_DATA_P
20567 test, now redundant because JUMP_TABLE_DATA is not an INSN_P insn.
20568 * gcse.c (insert_insn_end_basic_block): Likewise, JUMP_TABLE_DATA_P
20569 is not a NONDEBUG_INSN_P.
20570 * ira-costs.c (scan_one_insn): Likewise.
20571 * jump.c (mark_all_labels): Likewise.
20572 (mark_jump_label_1): Likewise.
20573 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20574 * lra.c (get_insn_freq): Expect all insns reaching here to be in
20575 a basic block.
20576 (check_rtl): Remove JUMP_TABLE_DATA_P test, not a NONDEBUG_INSN_P insn.
20577 * predict.c (expensive_function_p): Use FOR_BB_INSNS.
20578 * reload1.c (calculate_needs_all_insns): Call set_label_offsets for
20579 JUMP_TABLE_DATA_P insns.
20580 (calculate_elim_costs_all_insns): Likewise.
20581 (set_label_offsets): Recurse on the PATTERN of JUMP_TABLE_DATA insns.
20582 (elimination_costs_in_insn): Remove redundant JUMP_TABLE_DATA_P test.
20583 (delete_output_reload): Code style fixups.
20584 * reorg.c (dbr_schedule): Move JUMP_TABLE_DATA_P up to avoid setting
20585 insn flags on this non-insn.
20586 * sched-rgn.c (add_branch_dependences): Treat JUMP_TABLE_DATA insns
20587 as scheduling barriers, for pre-change compatibility.
20588 * stmt.c (emit_case_dispatch_table): Emit jump table data not as
20589 JUMP_INSN objects but instead as JUMP_TABLE_DATA objects.
20590
20591 * config/alpha/alpha.c (alpha_does_function_need_gp): Remove
20592 redundant JUMP_TABLE_DATA_P test.
20593 * config/arm/arm.c (thumb_far_jump_used_p): Likewise.
20594 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20595 (frv_for_each_packet): Likewise.
20596 * config/i386/i386.c (min_insn_size): Likewise.
20597 (ix86_avoid_jump_mispredicts): Likewise.
20598 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20599 * config/mep/mep.c (mep_reorg_erepeat): Likewise.
20600 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20601 (mips16_insn_length): Robustify.
20602 (mips_has_long_branch_p): Remove redundant JUMP_TABLE_DATA_P test.
20603 (mips16_split_long_branches): Likewise.
20604 * config/pa/pa.c (pa_combine_instructions): Likewise.
20605 * config/rs6000/rs6000.c (get_next_active_insn): Treat
20606 JUMP_TABLE_DATA objects as active insns, like in active_insn_p.
20607 * config/s390/s390.c (s390_chunkify_start): Treat JUMP_TABLE_DATA
20608 as contributing to pool range lengths.
20609 * config/sh/sh.c (find_barrier): Restore check for ADDR_DIFF_VEC.
20610 Remove redundant JUMP_TABLE_DATA_P test.
20611 (sh_loop_align): Likewise.
20612 (split_branches): Likewise.
20613 (sh_insn_length_adjustment): Likewise.
20614 * config/spu/spu.c (get_branch_target): Likewise.
20615
20616 2013-03-29 Jan Hubicka <jh@suse.cz>
20617
20618 * lto-cgraph.c (output_profile_summary, input_profile_summary): Use
20619 gcov streaming; stream hot bb threshold to ltrans.
20620 * predict.c (get_hot_bb_threshold): Break out from ....
20621 (maybe_hot_count_p): ... here.
20622 (set_hot_bb_threshold): New function.
20623 * lto-section-in.c (lto_section_name): Add profile.
20624 * profile.h (get_hot_bb_threshold, set_hot_bb_threshold): Declare.
20625 * ipa.c: Include hash-table.h, tree-inline.h, profile.h, lto-streamer.h
20626 and data-streamer.h
20627 (histogram_entry): New structure.
20628 (histogram, histogram_pool): New global vars.
20629 (histogram_hash): New structure.
20630 (histogram_hash::hash): New method.
20631 (histogram_hash::equal): Likewise.
20632 (account_time_size): New function.
20633 (cmp_counts): New function.
20634 (dump_histogram): New function.
20635 (ipa_profile_generate_summary): New function.
20636 (ipa_profile_write_summary): New function.
20637 (ipa_profile_read_summary): New function.
20638 (ipa_profile): Decide on threshold.
20639 (pass_ipa_profile): Add ipa_profile_write_summary and
20640 ipa_profile_read_summary.
20641 * Makefile.in (ipa.o): Update dependencies.
20642 * lto-streamer.h (LTO_section_ipa_profile): New section.
20643
20644 2013-03-29 Gabriel Dos Reis <gdr@integrable-solutions.net>
20645
20646 * tree.h (VAR_P): New.
20647
20648 2013-03-29 Paolo Carlini <paolo.carlini@oracle.com>
20649
20650 PR lto/56777
20651 * doc/invoke.texi ([-fwhole-program]): Fix typo.
20652
20653 2013-03-29 Steven Bosscher <steven@gcc.gnu.org>
20654
20655 * cfgbuild.c (inside_basic_block_p): Use JUMP_TABLE_DATA_P in lieu
20656 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20657 (control_flow_insn_p): Likewise.
20658 * cfgrtl.c (duplicate_insn_chain): Likewise.
20659 * final.c (get_attr_length_1): Likewise.
20660 (shorten_branches): Likewise.
20661 (final_scan_insn): Likewise.
20662 * function.c (instantiate_virtual_regs): Likewise.
20663 * gcse.c (insert_insn_end_basic_block): Likewise.
20664 * ira-costs.c (scan_one_insn): Likewise.
20665 * lra-eliminations.c (eliminate_regs_in_insn): Likewise.
20666 * lra.c (check_rtl): Likewise.
20667 * reload1.c (elimination_costs_in_insn): Likewise.
20668 * reorg.c (follow_jumps): Likewise.
20669
20670 * config/arm/arm.c (is_jump_table): Use JUMP_TABLE_DATA_P in lieu
20671 of tests for JUMP_P and a ADDR_DIFF_VEC or ADDR_VEC pattern.
20672 (thumb_far_jump_used_p): Likewise.
20673 * config/bfin/bfin.c (workaround_rts_anomaly): Likewise.
20674 (workaround_speculation): Likewise.
20675 (add_sched_insns_for_speculation): Likewise.
20676 * config/c6x/c6x.c (reorg_emit_nops): Likewise.
20677 * config/frv/frv.c (frv_function_contains_far_jump): Likewise.
20678 (frv_for_each_packet): Likewise.
20679 * config/i386/i386.c (ix86_avoid_jump_mispredicts): Likewise.
20680 * config/ia64/ia64.c (emit_all_insn_group_barriers): Likewise.
20681 (final_emit_insn_group_barriers): Likewise.
20682 * config/m32r/m32r.c (m32r_is_insn): Likewise.
20683 * config/mips/mips.c (USEFUL_INSN_P): Likewise.
20684 (mips16_insn_length): Likewise.
20685 * config/pa/pa.c (pa_reorg): Likewise.
20686 (pa_combine_instructions): Likewise.
20687 * config/rs6000/rs6000.c (rs6000_invalid_within_doloop): Likewise.
20688 * config/sh/sh.c (fixup_addr_diff_vecs): Likewise.
20689 (sh_reorg): Likewise.
20690 (split_branches): Likewise.
20691 * config/spu/spu.c (get_branch_target): Likewise.
20692
20693 * config/s390/s390.c (s390_chunkify_start): Simplify logic using
20694 JUMP_TABLE_DATA_P.
20695
20696 2013-03-29 Kirill Yukhin <kirill.yukhin@intel.com>
20697
20698 * config/i386/avx2intrin.h (_mm256_broadcastsi128_si256):
20699 Fix declaration name.
20700
20701 2013-03-28 Lawrence Crowl <crowl@google.com>
20702
20703 * graphds.h (struct graph.indicies): Remove unused.
20704 * graphite-poly.h (struct graph.original_pddrs): Remove unused.
20705 (SCOP_ORIGINAL_PDDRS): Remove unused.
20706 * sese.h (extern insert_loop_close_phis): Removed unused.
20707 (extern insert_guard_phis): Removed unused.
20708 (extern ivtype_map_elt_info): Removed unused.
20709 (new_ivtype_map_elt): Removed unused.
20710 * sese.c (ivtype_map_elt_info): Removed unused.
20711
20712 2013-03-28 Lawrence Crowl <crowl@google.com>
20713
20714 * Makefile.in: Add several missing include dependences.
20715 (DUMPFILE_H): New.
20716 (test-dump.o): New. This object is not added to any executable,
20717 but is present for ad-hoc testing.
20718 * bitmap.c
20719 (debug (const bitmap_head_def &)): New.
20720 (debug (const bitmap_head_def *)): New.
20721 * bitmap.h
20722 (extern debug (const bitmap_head_def &)): New.
20723 (extern debug (const bitmap_head_def *)): New.
20724 * cfg.c
20725 (debug (edge_def &)): New.
20726 (debug (edge_def *)): New.
20727 * cfghooks.c
20728 (debug (basic_block_def &)): New.
20729 (debug (basic_block_def *)): New.
20730 * dumpfile.h
20731 (dump_node (const_tree, int, FILE *)): Correct source file.
20732 * dwarf2out.c
20733 (debug (die_struct &)): New.
20734 (debug (die_struct *)): New.
20735 * dwarf2out.h
20736 (extern debug (die_struct &)): New.
20737 (extern debug (die_struct *)): New.
20738 * gimple-pretty-print.c
20739 (debug (gimple_statement_d &)): New.
20740 (debug (gimple_statement_d *)): New.
20741 * gimple-pretty-print.h
20742 (extern debug (gimple_statement_d &)): New.
20743 (extern debug (gimple_statement_d *)): New.
20744 * ira-build.c
20745 (debug (ira_allocno_copy &)): New.
20746 (debug (ira_allocno_copy *)): New.
20747 (debug (ira_allocno &)): New.
20748 (debug (ira_allocno *)): New.
20749 * ira-int.h
20750 (extern debug (ira_allocno_copy &)): New.
20751 (extern debug (ira_allocno_copy *)): New.
20752 (extern debug (ira_allocno &)): New.
20753 (extern debug (ira_allocno *)): New.
20754 * ira-lives.c
20755 (debug (live_range &)): New.
20756 (debug (live_range *)): New.
20757 * lra-int.h
20758 (debug (lra_live_range &)): New.
20759 (debug (lra_live_range *)): New.
20760 * lra-lives.c
20761 (debug (lra_live_range &)): New.
20762 (debug (lra_live_range *)): New.
20763 * omega.c
20764 (debug (omega_pb_d &)): New.
20765 (debug (omega_pb_d *)): New.
20766 * omega.h
20767 (extern debug (omega_pb_d &)): New.
20768 (extern debug (omega_pb_d *)): New.
20769 * print-rtl.c
20770 (debug (const rtx_def &)): New.
20771 (debug (const rtx_def *)): New.
20772 * print-tree.c
20773 (debug_tree (tree): Move within file.
20774 (debug_raw (const tree_node &)): New.
20775 (debug_raw (const tree_node *)): New.
20776 (dump_tree_via_hooks (const tree_node *, int)): New.
20777 (debug (const tree_node &)): New.
20778 (debug (const tree_node *)): New.
20779 (debug_verbose (const tree_node &)): New.
20780 (debug_verbose (const tree_node *)): New.
20781 (debug_head (const tree_node &)): New.
20782 (debug_head (const tree_node *)): New.
20783 (debug_body (const tree_node &)): New.
20784 (debug_body (const tree_node *)): New.
20785 (debug_vec_tree (tree): Move and reimplement in terms of dump.
20786 (debug (vec<tree, va_gc> &)): New.
20787 (debug (vec<tree, va_gc> *)): New.
20788 * rtl.h
20789 (extern debug (const rtx_def &)): New.
20790 (extern debug (const rtx_def *)): New.
20791 * sbitmap.c
20792 (debug_raw (simple_bitmap_def &)): New.
20793 (debug_raw (simple_bitmap_def *)): New.
20794 (debug (simple_bitmap_def &)): New.
20795 (debug (simple_bitmap_def *)): New.
20796 * sbitmap.h
20797 (extern debug (simple_bitmap_def &)): New.
20798 (extern debug (simple_bitmap_def *)): New.
20799 (extern debug_raw (simple_bitmap_def &)): New.
20800 (extern debug_raw (simple_bitmap_def *)): New.
20801 * sel-sched-dump.c
20802 (debug (vinsn_def &)): New.
20803 (debug (vinsn_def *)): New.
20804 (debug_verbose (vinsn_def &)): New.
20805 (debug_verbose (vinsn_def *)): New.
20806 (debug (expr_def &)): New.
20807 (debug (expr_def *)): New.
20808 (debug_verbose (expr_def &)): New.
20809 (debug_verbose (expr_def *)): New.
20810 (debug (vec<rtx> &)): New.
20811 (debug (vec<rtx> *)): New.
20812 * sel-sched-dump.h
20813 (extern debug (vinsn_def &)): New.
20814 (extern debug (vinsn_def *)): New.
20815 (extern debug_verbose (vinsn_def &)): New.
20816 (extern debug_verbose (vinsn_def *)): New.
20817 (extern debug (expr_def &)): New.
20818 (extern debug (expr_def *)): New.
20819 (extern debug_verbose (expr_def &)): New.
20820 (extern debug_verbose (expr_def *)): New.
20821 (extern debug (vec<rtx> &)): New.
20822 (extern debug (vec<rtx> *)): New.
20823 * sel-sched-ir.h
20824 (_list_iter_cond_expr): Make inline instead of static.
20825 * sreal.c
20826 (debug (sreal &)): New.
20827 (debug (sreal *)): New.
20828 * sreal.h
20829 (extern debug (sreal &)): New.
20830 (extern debug (sreal *)): New.
20831 * tree.h
20832 (extern debug_raw (const tree_node &)): New.
20833 (extern debug_raw (const tree_node *)): New.
20834 (extern debug (const tree_node &)): New.
20835 (extern debug (const tree_node *)): New.
20836 (extern debug_verbose (const tree_node &)): New.
20837 (extern debug_verbose (const tree_node *)): New.
20838 (extern debug_head (const tree_node &)): New.
20839 (extern debug_head (const tree_node *)): New.
20840 (extern debug_body (const tree_node &)): New.
20841 (extern debug_body (const tree_node *)): New.
20842 (extern debug (vec<tree, va_gc> &)): New.
20843 (extern debug (vec<tree, va_gc> *)): New.
20844 * tree-cfg.c
20845 (debug (struct loop &)): New.
20846 (debug (struct loop *)): New.
20847 (debug_verbose (struct loop &)): New.
20848 (debug_verbose (struct loop *)): New.
20849 * tree-dump.c: Add header dependence.
20850 * tree-flow.h
20851 (extern debug (struct loop &)): New.
20852 (extern debug (struct loop *)): New.
20853 (extern debug_verbose (struct loop &)): New.
20854 (extern debug_verbose (struct loop *)): New.
20855 * tree-data-ref.c
20856 (debug (data_reference &)): New.
20857 (debug (data_reference *)): New.
20858 (debug (vec<data_reference_p> &)): New.
20859 (debug (vec<data_reference_p> *)): New.
20860 (debug (vec<ddr_p> &)): New.
20861 (debug (vec<ddr_p> *)): New.
20862 * tree-data-ref.h
20863 (extern debug (data_reference &)): New.
20864 (extern debug (data_reference *)): New.
20865 (extern debug (vec<data_reference_p> &)): New.
20866 (extern debug (vec<data_reference_p> *)): New.
20867 (extern debug (vec<ddr_p> &)): New.
20868 (extern debug (vec<ddr_p> *)): New.
20869 * tree-ssa-alias.c
20870 (debug (pt_solution &)): New.
20871 (debug (pt_solution *)): New.
20872 * tree-ssa-alias.h
20873 (extern debug (pt_solution &)): New.
20874 (extern debug (pt_solution *)): New.
20875 * tree-ssa-alias.c
20876 (debug (_var_map &)): New.
20877 (debug (_var_map *)): New.
20878 (debug (tree_live_info_d &)): New.
20879 (debug (tree_live_info_d *)): New.
20880 * tree-ssa-alias.h
20881 (extern debug (_var_map &)): New.
20882 (extern debug (_var_map *)): New.
20883 (extern debug (tree_live_info_d &)): New.
20884 (extern debug (tree_live_info_d *)): New.
20885
20886 2013-03-28 Jan Hubicka <jh@suse.cz>
20887
20888 * lto-cgraph.c (merge_profile_summaries): Fix overflows.
20889
20890 2013-03-28 Ian Bolton <ian.bolton@arm.com>
20891
20892 * config/aarch64/aarch64.md (aarch64_can_eliminate): Keep frame
20893 record only when desired or required.
20894
20895 2013-03-28 Uros Bizjak <ubizjak@gmail.com>
20896
20897 * config/i386/i386.md (*vec_extract2vdi_1): Merge with
20898 *vec_extractv2di_1_rex64. Use x64 isa attribute.
20899
20900 2013-03-28 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
20901
20902 * config/aarch64/aarch64.md (*and<mode>3_compare0): New pattern.
20903 (*andsi3_compare0_uxtw): New pattern.
20904 (*and_<SHIFT:optab><mode>3_compare0): New pattern.
20905 (*and_<SHIFT:optab>si3_compare0_uxtw): New pattern.
20906
20907 2013-03-28 Jan Hubicka <jh@suse.cz>
20908
20909 * data-streamer-in.c (streamer_read_gcov_count): New function.
20910 * gimple-streamer-out.c: Include value-prof.h.
20911 (output_gimple_stmt): Output histogram.
20912 (output_bb): Use streamer_write_gcov_count.
20913 * value-prof.c: Include data-streamer.h
20914 (dump_histogram_value): Add HIST_TYPE_MAX.
20915 (stream_out_histogram_value): New function.
20916 (stream_in_histogram_value): New function.
20917 * value-prof.h (enum hist_type): Add HIST_TYPE_MAX.
20918 (stream_out_histogram_value, stream_in_histogram_value): Declare.
20919 * data-streamer-out.c (streamer_write_gcov_count): New function.
20920 (streamer_write_gcov_count_stream): New function.
20921 * lto-cgraph.c (lto_output_edge): Update counter streaming.
20922 (lto_output_node): Likewise.
20923 (input_node, input_edge): Likewise.
20924 * lto-streamer-out.c (output_cfg): Update streaming.
20925 * lto-streamer-in.c (input_cfg): Likewise.
20926 * data-streamer.h (streamer_write_gcov_count,
20927 streamer_write_gcov_count_stream, streamer_read_gcov_count): Declare.
20928 * gimple-streamer-in.c: Include value-prof.h
20929 (input_gimple_stmt): Input histograms.
20930 (input_bb): Update profile streaming.
20931
20932 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20933
20934 * genmodes.c (emit_max_int): New function.
20935 (emit_insn_modes_h): Added call to emit_max_function.
20936 * doc/rtl.texi (MAX_BITSIZE_MODE_ANY_INT, MAX_BITSIZE_MODE_ANY_MODE):
20937 Added doc.
20938 * machmode.def: Fixed comment.
20939
20940 2013-03-28 Kenneth Zadeck <zadeck@naturalbridge.com>
20941
20942 * combine.c (try_combine): Removed useless assert.
20943 * cselib.c (rtx_equal_for_cselib_1): Removed unnecessary parens.
20944
20945 2013-03-28 Marek Polacek <polacek@redhat.com>
20946 Richard Biener <rguenther@suse.de>
20947
20948 PR tree-optimization/56695
20949 * tree-vect-stmts.c (vectorizable_condition): Unconditionally
20950 build signed result of a vector comparison.
20951 * tree-cfg.c (verify_gimple_comparison): Check that a result
20952 of a vector comparison has signed type.
20953
20954 2013-03-28 Richard Biener <rguenther@suse.de>
20955
20956 PR tree-optimization/37021
20957 * tree-vect-slp.c (vect_build_slp_tree): When not unrolling
20958 do not restrict gaps between groups.
20959 * tree-vect-stmts.c (vectorizable_load): Properly account for
20960 a gap between groups.
20961
20962 2013-03-28 Eric Botcazou <ebotcazou@adacore.com>
20963
20964 * toplev.c (process_options): Do not disable -fomit-frame-pointer on a
20965 general basis if unwind info is requested and ACCUMULATE_OUTGOING_ARGS
20966 is not enabled.
20967
20968 2013-03-27 Gerald Pfeifer <gerald@pfeifer.com>
20969
20970 * doc/invoke.texi (AVR Options): Tweak link for AVR-LibC user manual.
20971 * doc/extend.texi (Named Address Spaces): Ditto.
20972 (Variable Attributes): Ditto.
20973
20974 2013-03-27 Kai Tietz <ktietz@redhat.com>
20975
20976 * config.build: Add support for cygwin x64 target.
20977 * config.gcc: Likewise.
20978 * config.host: Likewise.
20979 * configure.ac: Likewise
20980 * configure: Regenerated.
20981
20982 2013-03-27 Kai Tietz <ktietz@redhat.com>
20983
20984 * config/i386/cygwin-stdint.h: Add support for cygwin x64 target.
20985 * config/i386/t-cygwin-w64: New file.
20986 * config/i386/cygwin-w64.h: New file.
20987 * config/i386/cygwin.h (EXTRA_OS_CPP_BUILTINS): Extend
20988 and add support for x64-cygwin target.
20989 (CPP_SPEC): Likewise.
20990 (CXX_WRAP_SPEC_LIST): Undefine before define.
20991 (LIBGCJ_SONAME): Use 15 as version.
20992
20993 2013-03-27 Richard Biener <rguenther@suse.de>
20994
20995 PR tree-optimization/56716
20996 * tree-ssa-structalias.c (perform_var_substitution): Adjust
20997 dumping for ref nodes.
20998
20999 2013-03-27 Martin Jambor <mjambor@suse.cz>
21000
21001 PR tree-optimization/55334
21002 * ipa-cp.c (initialize_node_lattices): Allow IPA-CP through and to
21003 restricted pointers to arrays.
21004
21005 2013-03-27 Gabriel Dos Reis <gdr@integrable-solutions.net>
21006
21007 * Makefile.in (.SUFFIXES): Add .cc.
21008 (.c.o): Apply same recipe for implicit rule .cc.o.
21009
21010 2013-03-27 Richard Biener <rguenther@suse.de>
21011
21012 PR tree-optimization/37021
21013 * tree-vect-data-refs.c (vect_check_strided_load): Allow
21014 REALPART/IMAGPART_EXPRs around the supported refs.
21015 * tree-ssa-structalias.c (find_func_aliases): Assume that
21016 floating-point values are not used to transfer pointers.
21017
21018 2013-03-27 Alexander Ivchenko <alexander.ivchenko@intel.com>
21019
21020 * target.def (TARGET_HAS_IFUNC_P): New target hook.
21021 * doc/tm.texi.in (TARGET_HAS_IFUNC_P): New.
21022 * doc/tm.texi: Regenerate.
21023 * targhooks.h (default_has_ifunc_p): New.
21024 * targhooks.c (default_has_ifunc_p): Ditto.
21025 * config/linux-protos.h: New file.
21026 * config/linux-android.h (TARGET_HAS_IFUNC_P): Using version of this
21027 hook for linux which disables support of indirect functions in android.
21028 * config/linux-android.c: New file.
21029 * config/t-linux-android.c: Ditto.
21030 * config.gcc: Added new object file linux-android.o.
21031 * config/i386/i386.c (ix86_get_function_versions_dispatcher):
21032 Using TARGET_HAS_IFUNC hook instead of HAVE_GNU_INDIRECT_FUNCTION.
21033 * varasm.c (do_assemble_alias): Likewise.
21034 * configure.ac: Define HAVE_GNU_INDIRECT_FUNCTION as zero if the target
21035 doesn't support indirect functions.
21036 * configure: Regenerate.
21037
21038 2013-03-27 Bin Cheng <bin.cheng@arm.com>
21039
21040 PR target/56102
21041 * config/arm/arm.c (thumb1_rtx_costs, thumb1_size_rtx_costs): Fix
21042 rtx costs for SET/ASHIFT/ASHIFTRT/LSHIFTRT/ROTATERT patterns with
21043 mult-word mode.
21044
21045 2013-03-27 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
21046
21047 * config/s390/s390.h (TARGET_FLT_EVAL_METHOD): Define.
21048
21049 2013-03-27 Terry Guo <terry.guo@arm.com>
21050
21051 * config/arm/arm-cores.def: Added core cortex-r7.
21052 * config/arm/arm-tune.md: Regenerated.
21053 * config/arm/arm-tables.opt: Regenerated.
21054 * doc/invoke.texi: Added entry for core cortex-r7.
21055
21056 2013-03-27 Walter Lee <walt@tilera.com>
21057
21058 * config/tilegx/tilegx.c (tilegx_expand_prologue): Avoid
21059 double-decrement of next_scratch_regno.
21060
21061 2013-03-27 Walter Lee <walt@tilera.com>
21062
21063 * config/tilegx/tilegx.md (insn_v1mulu): Fix predicates on
21064 input operands.
21065 (insn_v1mulus): Ditto.
21066 (insn_v2muls): Ditto.
21067
21068 2013-03-27 Walter Lee <walt@tilera.com>
21069
21070 * config/tilegx/tilegx.h (ASM_OUTPUT_ADDR_VEC_ELT): Delete extra tab.
21071 (ASM_OUTPUT_ADDR_DIFF_ELT): Ditto.
21072
21073 2013-03-27 Walter Lee <walt@tilera.com>
21074
21075 * config/tilegx/tilegx.md (*sibcall_insn): Fix type atribute for jr.
21076 (*sibcall_value): Ditto.
21077
21078 2013-03-27 Walter Lee <walt@tilera.com>
21079
21080 * config/tilegx/tilegx.md (insn_mnz_<mode>): Replaced by ...
21081 (insn_mnz_v8qi): ... this ...
21082 (insn_mnz_v4hi): ... and this. Replace (const_int 0) with the
21083 vector equivalent.
21084 (insn_v<n>mnz): Replaced by ...
21085 (insn_v1mnz): ... this ...
21086 (insn_v2mnz): ... and this. Replace (const_int 0) with the vector
21087 equivalent.
21088 (insn_mz_<mode>): Replaced by ...
21089 (insn_mz_v8qi): ... this ...
21090 (insn_mz_v4hi): ... and this. Replace (const_int 0) with the
21091 vector equivalent.
21092 (insn_v<n>mz): Replaced by ...
21093 (insn_v1mz): ... this ...
21094 (insn_v2mz): ... and this. Replace (const_int 0) with the vector
21095 equivalent.
21096
21097 2013-03-26 Eric Botcazou <ebotcazou@adacore.com>
21098
21099 * doc/invoke.texi (SPARC options): Remove -mlittle-endian.
21100
21101 2013-03-26 Roland McGrath <mcgrathr@google.com>
21102
21103 * config/arm/arm.c (arm_print_operand: case 'w'): Use fputs rather
21104 than fprintf with a non-constant, non-format string.
21105
21106 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21107
21108 * config/i386/i386.md (*cmpqi_ext_1): Merge with *cmpqi_ext_1_rex64
21109 using nox64 isa attribute. Use nonimmediate_x86nomem_operand as
21110 operand 0 predicate.
21111 (*cmpqi_ext_3): Merge with *cmpqi_ext_3_rex64 using nox64 isa
21112 attribute. Use general_x64nomem_operand as operand 1 predicate.
21113 (*movqi_extv_1): Merge with *movqi_extv_1_rex64 using nox64 isa
21114 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21115 (*movqi_extzv_2): Merge with *movqi_extzv_2_rex64 using nox64 isa
21116 attribute. Use nonimmediate_x64nomem_operand as operand 0 predicate.
21117 (mov<mode>_insv_1): Remove expander. Merge insn with
21118 movsi_insv_1 using SWI48 mode iterator and nox64 isa attribute.
21119 Use general_x64nomem_operand as operand 1 predicate.
21120 (addqi_ext_1): Merge with *addqi_ext_1_rex64 using nox64 isa attribute.
21121 (*testqi_ext_1): Merge with *testqi_ext_1_rex64 using nox64 isa
21122 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21123 (*andqi_ext_1): Merge with *andqi_ext_1_rex64 using nox64 isa
21124 attribute. Use nonimmediate_x64nomem_operand as operand 2 predicate.
21125 (*<code>qi_ext_1): Merge with *<code>qi_ext_1_rex64 using nox64 isa
21126 attribute. Use nonimmediate_x64nomem_operand as operand 1 predicate.
21127 (*xorqi_cc_ext_1): Merge with *xorqi_cc_ext_1_rex64 using nox64
21128 isa attribute. Use general_x64nomem_operand as operand 2 predicate.
21129 * config/i386/predicates.md (nonimmediate_x64nomem_operand): New.
21130 (general_x64nomem_operand): Ditto.
21131
21132 2013-03-26 Sebastian Huber <sebastian.huber@embedded-brains.de>
21133
21134 * config/rtems.opt: Add -pthread option.
21135
21136 2013-03-26 Richard Biener <rguenther@suse.de>
21137
21138 * alias.c (find_base_term): Avoid redundant and not used recursion.
21139 (base_alias_check): Get the initial base term from the caller.
21140 (true_dependence_1): Compute and pass base terms to base_alias_check.
21141 (write_dependence_p): Likewise.
21142 (may_alias_p): Likewise.
21143
21144 2013-03-26 Sofiane Naci <sofiane.naci@arm.com>
21145
21146 * config/aarch64/aarch64.c (aarch64_classify_address): Support
21147 PC-relative load in SI modes and above only.
21148
21149 2013-03-26 Xinyu Qi <xyqi@marvell.com>
21150
21151 * config/arm/arm.h (FIRST_IWMMXT_GR_REGNUM): Add comment.
21152 * config/arm/iwmmxt.md (WCGR0): Update.
21153 (WCGR1, WCGR2, WCGR3): Likewise.
21154
21155 2013-03-26 Uros Bizjak <ubizjak@gmail.com>
21156
21157 * config/i386/i386.md (*movdfcc_1): Merge with *movdfcc_1_rex64.
21158 Use x64 and nox64 isa attributes.
21159
21160 2013-03-26 Richard Biener <rguenther@suse.de>
21161
21162 * emit-rtl.c (set_mem_attributes_minus_bitpos): Remove
21163 alignment computations and rely on get_object_alignment_1
21164 for the !TYPE_P case.
21165 Commonize DECL/COMPONENT_REF handling in the ARRAY_REF path.
21166
21167 2013-03-26 Walter Lee <walt@tilera.com>
21168
21169 * config/tilegx/tilegx.h (PROFILE_BEFORE_PROLOGUE): Define.
21170 * config/tilegx/tilepro.h (PROFILE_BEFORE_PROLOGUE): Define.
21171
21172 2013-03-25 Jeff Law <law@redhat.com>
21173
21174 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Add missing
21175 check for INTEGRAL_TYPE_P that was missing due to checking in
21176 wrong version of prior patch.
21177
21178 2013-03-25 Walter Lee <walt@tilera.com>
21179
21180 * config/tilegx/tilegx-builtins.h (enum tilegx_builtin): Add
21181 TILEGX_INSN_SHUFFLEBYTES1.
21182 * config/tilegx/tilegx.c (tilegx_builtin_info): Add entry for
21183 shufflebytes1.
21184 (tilegx_builtins): Ditto.
21185 * config/tilegx/tilegx.md (insn_shufflebytes1): New pattern.
21186
21187 2013-03-25 Walter Lee <walt@tilera.com>
21188
21189 * config/tilegx/tilegx.md (floatsisf2): New pattern.
21190 (floatunssisf2): New pattern.
21191 (floatsidf2): New pattern.
21192 (floatunssidf2): New pattern.
21193
21194 2013-03-25 Walter Lee <walt@tilera.com>
21195
21196 * config/tilegx/tilegx.c (expand_set_cint64_one_inst): Inline
21197 tests for constraint J, K, N, P.
21198
21199 2013-03-25 Walter Lee <walt@tilera.com>
21200
21201 * config/tilegx/tilegx.c (tilegx_asm_preferred_eh_data_format):
21202 Use indirect/pcrel encoding.
21203 * config/tilepro/tilepro.c (tilepro_asm_preferred_eh_data_format):
21204 Ditto.
21205
21206 2013-03-25 Steve Ellcey <sellcey@mips.com>
21207
21208 * config/mips/mmips-cpus.def (74kc, 74kf2_1, 74kf, 74kf, 74kf1_1,
21209 74kfx, 74kx, 74kf3_2): Add PTF_AVOID_IMADD.
21210 * config/mips/mips.c (mips_option_override): Set IMADD default.
21211 * config/mips/mips.h (PTF_AVOID_IMADD): New.
21212 (ISA_HAS_MADD_MSUB): Remove MIPS16 check.
21213 (GENERATE_MADD_MSUB): Remove TUNE_74K check, add MIPS16 check.
21214 * config/mips/mips.md (mimadd): New flag for integer madd/msub.
21215 * doc/invoke.texi (-mimadd/-mno-imadd): New.
21216
21217 2013-03-25 Jeff Law <law@redhat.com>
21218
21219 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Rework
21220 slightly to avoid creating and folding useless trees. Simplify
21221 slightly by restricting to INTEGER_CSTs and using int_fits_type_p.
21222
21223 2013-03-25 Uros Bizjak <ubizjak@gmail.com>
21224
21225 * config/i386/i386.md (*zero_extendsidi2): Merge with
21226 *zero_extendsidi2_rex64. Use x64 and nox64 isa attributes.
21227 * config/i386/predicates.md (x86_64_zext_operand): Rename from
21228 x86_64_zext_general_operand. Use nonimmediate_operand on 32bit
21229 targets. Clarify comment.
21230
21231 2013-03-25 Martin Jambor <mjambor@suse.cz>
21232
21233 * ipa-prop.c (ipa_write_jump_function): Stream simple and aritmetic
21234 pass-through jump functions differently.
21235 (ipa_read_jump_function): Likewise. Also use setter functions to set
21236 up jump functions.
21237
21238 2013-03-25 Martin Jambor <mjambor@suse.cz>
21239
21240 * ipa-cp.c (ipa_get_indirect_edge_target): Renamed to
21241 ipa_get_indirect_edge_target_1, added parameter agg_reps and ability to
21242 process it.
21243 (ipa_get_indirect_edge_target): New function.
21244 (devirtualization_time_bonus): New parameter known_aggs, pass it to
21245 ipa_get_indirect_edge_target. Update all callers.
21246 (ipcp_discover_new_direct_edges): New parameter aggvals. Pass it to
21247 ipa_get_indirect_edge_target_1 instead of calling
21248 ipa_get_indirect_edge_target.
21249 (create_specialized_node): Pass aggvlas to
21250 ipcp_discover_new_direct_edges.
21251
21252 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21253
21254 * config/arm/arm.md (f_sels, f_seld): New types.
21255 (*cmov<mode>): New pattern.
21256 * config/arm/predicates.md (arm_vsel_comparison_operator): New
21257 predicate.
21258
21259 2013-03-25 Kai Tietz <ktietz@redhat.com>
21260
21261 * config/i386/xm-mingw32.h (__USE_MINGW_ANSI_STDIO): Enable
21262 POSIX-printf for mingw-hosted builds.
21263
21264 2013-03-25 Richard Biener <rguenther@suse.de>
21265
21266 PR middle-end/56694
21267 * tree-eh.c (lower_eh_must_not_throw): Strip BLOCKs from the
21268 must-not-throw stmt location.
21269
21270 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21271
21272 * config/arm/arm.c (arm_emit_load_exclusive): Add acq parameter.
21273 Emit load-acquire versions when acq is true.
21274 (arm_emit_store_exclusive): Add rel parameter.
21275 Emit store-release versions when rel is true.
21276 (arm_split_compare_and_swap): Use acquire-release instructions
21277 instead.
21278 of barriers when appropriate.
21279 (arm_split_atomic_op): Likewise.
21280 * config/arm/arm.h (TARGET_HAVE_LDACQ): New macro.
21281 * config/arm/unspecs.md (VUNSPEC_LAX): New unspec.
21282 (VUNSPEC_SLX): Likewise.
21283 (VUNSPEC_LDA): Likewise.
21284 (VUNSPEC_STL): Likewise.
21285 * config/arm/sync.md (atomic_load<mode>): New pattern.
21286 (atomic_store<mode>): Likewise.
21287 (arm_load_acquire_exclusive<mode>): Likewise.
21288 (arm_load_acquire_exclusivesi): Likewise.
21289 (arm_load_acquire_exclusivedi): Likewise.
21290 (arm_store_release_exclusive<mode>): Likewise.
21291
21292 2013-03-25 Catherine Moore <clm@codesourcery.com>
21293
21294 * config/mips/constraints.md (u, Udb7 Uead, Uean, Uesp, Uib3,
21295 Uuw6, Usb4, ZS, ZT, ZU, ZV, ZW): New constraints.
21296 * config/mip/predicates.md (lwsp_swsp_operand,
21297 lw16_sw16_operand, lhu16_sh16_operand, lbu16_operand,
21298 sb16_operand, db4_operand, db7_operand, ib3_operand,
21299 sb4_operand, ub4_operand, uh4_operand, uw4_operand,
21300 uw5_operand, uw6_operand, addiur2_operand, addiusp_operand,
21301 andi16_operand): New predicates.
21302 * config/mips/mips.md (compression): New attribute.
21303 (enabled): New attribute.
21304 (length): Consider compression in computing length.
21305 (shift_compression): New code attribute.
21306 (*add<mode>3): New operands. Record compression.
21307 (sub<mode>3): Likewise.
21308 (one_cmpl<mode>2): Likewise.
21309 (*and<mode>3): Likewise.
21310 (*ior<mode>3): Likewise.
21311 (unnamed pattern for xor): Likewise.
21312 (*zero_extend<SHORT:mode><GPR:mode>2): Likewise.
21313 (*<optab><mode>3): Likewise.
21314 (*mov<mode>_internal: Likewise.
21315 * config/mips/mips-protos.h (mips_signed_immediate_p): New.
21316 (mips_unsigned_immediate_p): New.
21317 (umips_lwsp_swsp_address_p): New.
21318 (m16_based_address_p): New.
21319 * config/mips/mips-protos.h (mips_signed_immediate_p): New prototype.
21320 (mips_unsigned_immediate_p): New prototype.
21321 (lwsp_swsp_address_p): New prototype.
21322 (m16_based_address_p): New prototype.
21323 * config/mips/mips.c (mips_unsigned_immediate_p): New function.
21324 (mips_signed_immediate_p): New function.
21325 (m16_based_address_p): New function.
21326 (lwsp_swsp_address_p): New function.
21327 (mips_print_operand_punctuation): Recognize short delay slot insns
21328 for microMIPS.add<mode>3"
21329
21330 2013-03-25 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
21331
21332 PR target/56720
21333 * config/arm/iterators.md (v_cmp_result): New mode attribute.
21334 * config/arm/neon.md (vcond<mode><mode>): Handle unordered cases.
21335
21336 2013-03-25 Richard Biener <rguenther@suse.de>
21337
21338 PR tree-optimization/56689
21339 * tree-vrp.c (execute_vrp): Mark loops for fixup if we removed
21340 any edge.
21341
21342 2013-03-25 Richard Biener <rguenther@suse.de>
21343
21344 * tree-ssa-loop-im.c (struct mem_ref): Use bitmap_head instead
21345 of bitmap.
21346 (memory_references): Likewise.
21347 (outermost_indep_loop, mem_ref_alloc, mark_ref_stored,
21348 gather_mem_refs_stmt, record_dep_loop, ref_indep_loop_p_1,
21349 ref_indep_loop_p_2, find_refs_for_sm): Adjust.
21350 (gather_mem_refs_in_loops): Fold into ...
21351 (analyze_memory_references): ... this. Move initialization
21352 to tree_ssa_lim_initialize.
21353 (fill_always_executed_in): Rename to ...
21354 (fill_always_executed_in_1): ... this.
21355 (fill_always_executed_in): Move contains_call computation to
21356 this new function from ...
21357 (tree_ssa_lim_initialize): ... here.
21358 (tree_ssa_lim): Call fill_always_executed_in.
21359
21360 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21361
21362 * postreload.c (reload_combine): Fix code detecting returns.
21363
21364 2013-03-25 Eric Botcazou <ebotcazou@adacore.com>
21365
21366 * function.c (emit_use_return_register_into_block): On cc0 targets,
21367 do not emit the sequence between cc0 setter and user.
21368
21369 2013-03-25 Kai Tietz <ktietz@redhat.com>
21370
21371 * config/i386/predicates.md (local_symbolic_operand): Interpret
21372 dll-imported symbols as none-local.
21373
21374 2013-03-25 Richard Biener <rguenther@suse.de>
21375
21376 * tree-ssa-loop-im.c (struct depend): Remove.
21377 (struct lim_aux_data): Make depends a vec of gimples.
21378 (free_lim_aux_data): Adjust.
21379 (add_dependency): Likewise.
21380 (set_level): Likewise.
21381
21382 2013-03-25 Richard Biener <rguenther@suse.de>
21383
21384 PR middle-end/56434
21385 * calls.c (expand_call): Use MALLOC_ABI_ALIGNMENT to annotate
21386 the pointer returned by calls with ECF_MALLOC set.
21387
21388 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21389
21390 * config/i386/mmx.md (mov<mode>): Add ?!Ym,r and r,?!Ym alternatives.
21391
21392 2013-03-24 Uros Bizjak <ubizjak@gmail.com>
21393
21394 * config/i386/mmx.md (mov<mode>): Merge with movv2sf expander
21395 using MMXMODE mode iterator.
21396 (*move<mode>_internal): Merge with *movv2sf_internal and
21397 *movv2sf_internal_rex64 using MMXMODE mode iterator.
21398
21399 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21400
21401 * gcse.c (oprs_unchanged_p): Respect flag_gcse_lm.
21402 (record_last_mem_set_info): Likewise.
21403
21404 * df-core.c (rest_of_handle_df_initialize): Use XCNEWVEC instead
21405 of XNEWVEC followed by memset.
21406 (df_worklist_dataflow): Use XNEWVEC instead of xmalloc with a cast.
21407
21408 2013-03-23 Steven Bosscher <steven@gcc.gnu.org>
21409
21410 * config/avr/avr.c, config/bfin/bfin.c, config/c6x/c6x.c,
21411 config/epiphany/epiphany.c, config/frv/frv.c, config/ia64/ia64.c,
21412 config/iq2000/iq2000.c, config/mcore/mcore.c, config/mep/mep.c,
21413 config/mmix/mmix.c, config/pa/pa.c, config/rs6000/rs6000.c,
21414 config/s390/s390.c, config/sparc/sparc.c, config/spu/spu.c,
21415 config/stormy16/stormy16.c, config/v850/v850.c, config/xtensa/xtensa.c,
21416 dwarf2out.c, hw-doloop.c, resource.c, rtl.h : Where applicable, use
21417 the predicates NOTE_P, NONJUMP_INSN_P, JUMP_P, CALL_P, LABEL_P, and
21418 BARRIER_P instead of GET_CODE.
21419
21420 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21421
21422 * config/sparc/sparc.c (sparc_emit_probe_stack_range): Fix small
21423 inaccuracy in the probing code.
21424
21425 * config/sparc/sparc.md (ctrapsi4): Add predicate for operand #3.
21426 (ctrapdi4): Likewise.
21427
21428 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21429
21430 * calls.c (expand_call): Add missing guard to code handling return
21431 of non-BLKmode structures in MSB.
21432 * function.c (expand_function_end): Likewise.
21433
21434 2013-03-23 Eric Botcazou <ebotcazou@adacore.com>
21435
21436 * combine.c (try_combine): Adjust comment. Do not add the set of
21437 insn #0 if the destination indirectly is set or dies in insn #2.
21438 Tidy up code to distribute a new note.
21439
21440 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21441
21442 * config/i386/i386.md (*movdi_internal): Set prefix_rex attribute
21443 also for alternatives 16 and 17.
21444
21445 2013-03-22 Uros Bizjak <ubizjak@gmail.com>
21446
21447 * config/i386/sse.md (*mov<mode>_internal): Merge with
21448 *mov<mode>_internal_rex64. Use x64 and nox64 isa attributes.
21449 Emit insn template depending on type attribute. Use
21450 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21451 movd instead of movq mnemonic for interunit moves. Rewrite mode
21452 attribute calculation. Remove unit attribute calculation.
21453 Set prefix attribute to maybe_vex for sselog1 and ssemov types.
21454 Set prefix_data16 attribute for DImode ssemov types.
21455 Use Ym instead of y for SSE-MMX conversion alternatives.
21456 Reorder operand constraints.
21457
21458 2013-03-22 Steven Bosscher <steven@gcc.gnu.org>
21459
21460 * df.h (df_insn_delete): Adjust prototype.
21461 * emit-rtl.c (remove_insn): Pass a basic block to df_insn_delete
21462 and let it decide whether mark the basic block dirty.
21463 (set_insn_deleted): Only pass INSN_P insns to df_insn_delete.
21464 * df-scan.c (df_insn_info_delete): New helper function, split
21465 off from df_insn_delete.
21466 (df_scan_free_bb_info): Use it.
21467 (df_insn_rescan, df_insn_rescan_all, df_process_deferred_rescans):
21468 Likewise.
21469 (df_insn_delete): Likewise. Take insn rtx as argument. Verify
21470 that the insn is actually an insn and it has a non-NULL basic block.
21471 Do not mark basic block dirty if only deleting a DEBUG_INSN.
21472
21473 2013-03-22 Richard Biener <rguenther@suse.de>
21474
21475 * tree-ssa-loop-im.c (struct mem_ref): Remove indep_ref and
21476 dep_ref members.
21477 (mem_ref_alloc): Do not allocate them.
21478 (refs_independent_p): Do not query or maintain a cache.
21479
21480 2013-03-22 Richard Biener <rguenther@suse.de>
21481
21482 * tree-ssa-loop-im.c (memory_references): Drop all_refs_in_loop.
21483 (gather_mem_refs_in_loops): Do not compute it.
21484 (analyze_memory_references): Do not allocate it.
21485 (tree_ssa_lim_finalize): Do not free it.
21486 (for_all_locs_in_loop): Do not query all_refs_in_loop.
21487
21488 2013-03-22 Richard Biener <rguenther@suse.de>
21489
21490 * is-a.h (as_a): Use gcc_checking_assert.
21491
21492 2013-03-22 Ian Bolton <ian.bolton@arm.com>
21493
21494 * config/aarch64/aarch64.c (aarch64_print_operand): New
21495 format specifier for printing a constant in hex.
21496 * config/aarch64/aarch64.md (insv_imm<mode>): Use the X
21497 format specifier for printing second operand.
21498
21499 2013-03-22 Richard Biener <rguenther@suse.de>
21500
21501 * tree-ssa-loop-im.c (memory_references): Add refs_stored_in_loop
21502 bitmaps.
21503 (gather_mem_refs_in_loops): Perform store accumulation here.
21504 (create_vop_ref_mapping_loop): Remove.
21505 (create_vop_ref_mapping): Likewise.
21506 (analyze_memory_references): Initialize refs_stored_in_loop.
21507 (LOOP_DEP_BIT): New define to map to bits in (in)dep_loop bitmaps.
21508 (record_indep_loop): Remove.
21509 (record_dep_loop): New function.
21510 (ref_indep_loop_p_1): Adjust to only walk over references
21511 in the loop, not its subloops.
21512 (ref_indep_loop_p): Rename to ...
21513 (ref_indep_loop_p_2): ... this and recurse over the loop tree,
21514 maintaining a more fine-grained cache.
21515 (ref_indep_loop_p): Wrap ref_indep_loop_p_2.
21516 (tree_ssa_lim_finalize): Free refs_stored_in_loop.
21517
21518 2013-03-22 Richard Biener <rguenther@suse.de>
21519
21520 * tree-ssa-loop-im.c (struct mem_ref_locs): Remove.
21521 (struct mem_ref): Make accesses_in_loop a vec of a vec of
21522 aggregate mem_ref_loc.
21523 (free_mem_ref_locs): Inline into ...
21524 (memref_free): ... this and adjust.
21525 (mem_ref_alloc): Adjust.
21526 (mem_ref_locs_alloc): Remove.
21527 (record_mem_ref_loc): Adjust.
21528 (get_all_locs_in_loop): Rewrite into ...
21529 (for_all_locs_in_loop): ... this iterator.
21530 (rewrite_mem_ref_loc): New functor.
21531 (rewrite_mem_refs): Use for_all_locs_in_loop.
21532 (sm_set_flag_if_changed): New functor.
21533 (execute_sm_if_changed_flag_set): Use for_all_locs_in_loop.
21534 (ref_always_accessed): New functor.
21535 (ref_always_accessed_p): Use for_all_locs_in_loop.
21536
21537 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21538
21539 * tree-pass.h (PROP_gimple_lvec): New.
21540 * passes.c (dump_properties): Handle PROP_gimple_lvec.
21541 (init_optimization_passes): Move pass_lower_vector.
21542 * tree-vect-generic.c (gate_expand_vector_operations_ssa): Test
21543 PROP_gimple_lvec.
21544 (pass_lower_vector): Provide PROP_gimple_lvec.
21545 (pass_lower_vector_ssa): Likewise.
21546 * cfgexpand.c (pass_expand): Require PROP_gimple_lvec.
21547
21548 2013-03-21 Mark Wielaard <mjw@redhat.com>
21549
21550 * dwarf2out.c (size_of_aranges): Skip DECL_IGNORED_P functions.
21551
21552 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21553
21554 * config/i386/i386.md (*movdi_internal): Disparage slightly
21555 all MMX moves to/from memory. Use Yi instead of x for SSE-MMX
21556 conversion alternatives.
21557
21558 2013-03-21 Jakub Jelinek <jakub@redhat.com>
21559
21560 PR middle-end/48087
21561 * diagnostic.def (DK_WERROR): New kind.
21562 * diagnostic.h (werrorcount): Define.
21563 * diagnostic.c (diagnostic_report_diagnostic): For DK_WARNING
21564 promoted to DK_ERROR, increment DK_WERROR counter instead of
21565 DK_ERROR counter.
21566 * toplev.c (toplev_main): Call print_ignored_options even if
21567 just werrorcount is non-zero. Exit with FATAL_EXIT_CODE
21568 even if just werrorcount is non-zero.
21569
21570 PR debug/55608
21571 * dwarf2out.c (tree_add_const_value_attribute): Call ggc_free (array)
21572 on failure.
21573 (resolve_one_addr): Fail if referenced STRING_CST hasn't been written.
21574 (string_cst_pool_decl): New function.
21575 (optimize_one_addr_into_implicit_ptr): New function.
21576 (resolve_addr_in_expr): Optimize DWARF location expression
21577 DW_OP_addr DW_OP_stack_value where DW_OP_addr refers to some variable
21578 which doesn't live in memory, but has DW_AT_location or
21579 DW_AT_const_value, or refers to a string literal, into
21580 DW_OP_GNU_implicit_pointer.
21581 (optimize_location_into_implicit_ptr): New function.
21582 (resolve_addr): If removing DW_AT_location of a variable because
21583 it was DW_OP_addr of address of the variable, but the variable doesn't
21584 live in memory, try to emit const value attribute for the initializer.
21585
21586 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21587
21588 * tree.h (VECTOR_TYPE_P): New macro.
21589 (VECTOR_INTEGER_TYPE_P, VECTOR_FLOAT_TYPE_P, FLOAT_TYPE_P,
21590 TYPE_MODE): Use it.
21591 * fold-const.c (fold_cond_expr_with_comparison): Use build_zero_cst.
21592 VEC_COND_EXPR cannot be lvalues.
21593 (fold_ternary_loc) <VEC_COND_EXPR>: Merge with the COND_EXPR case.
21594
21595 2013-03-21 Marc Glisse <marc.glisse@inria.fr>
21596
21597 * simplify-rtx.c (simplify_binary_operation_1) <VEC_CONCAT>:
21598 Restrict the transformation to equal modes.
21599
21600 2013-03-21 Richard Biener <rguenther@suse.de>
21601
21602 PR tree-optimization/39326
21603 * tree-ssa-loop-im.c (UNANALYZABLE_MEM_ID): New define.
21604 (MEM_ANALYZABLE): Adjust.
21605 (record_mem_ref_loc): Move bitmap ops ...
21606 (gather_mem_refs_stmt): ... here. Use the shared mem-ref for
21607 unanalyzable refs, do not record locations for it.
21608 (analyze_memory_references): Allocate ref zero as shared
21609 unanalyzable ref.
21610 (refs_independent_p): Do not test for unanalyzed mems here.
21611 (ref_indep_loop_p_1): Special-case disambiguation against
21612 the unanalyzed ref.
21613 (ref_indep_loop_p): Assert we are not queried for the unanalyzed mem.
21614
21615 2013-03-21 Christophe Lyon <christophe.lyon@linaro.org>
21616
21617 * config/arm/arm-protos.h (tune_params): Add
21618 prefer_neon_for_64bits field.
21619 * config/arm/arm.c (prefer_neon_for_64bits): New variable.
21620 (arm_slowmul_tune): Default prefer_neon_for_64bits to false.
21621 (arm_fastmul_tune, arm_strongarm_tune, arm_xscale_tune): Ditto.
21622 (arm_9e_tune, arm_v6t2_tune, arm_cortex_tune): Ditto.
21623 (arm_cortex_a15_tune, arm_cortex_a5_tune): Ditto.
21624 (arm_cortex_a9_tune, arm_v6m_tune, arm_fa726te_tune): Ditto.
21625 (arm_option_override): Handle -mneon-for-64bits new option.
21626 * config/arm/arm.h (TARGET_PREFER_NEON_64BITS): New macro.
21627 (prefer_neon_for_64bits): Declare new variable.
21628 * config/arm/arm.md (arch): Rename neon_onlya8 and neon_nota8 to
21629 avoid_neon_for_64bits and neon_for_64bits. Remove onlya8 and nota8.
21630 (arch_enabled): Handle new arch types. Remove support for onlya8
21631 and nota8.
21632 (one_cmpldi2): Use new arch names.
21633 (zero_extend<mode>di2, extend<mode>di2): Ditto.
21634 * config/arm/arm.opt (mneon-for-64bits): Add option.
21635 * config/arm/neon.md (adddi3_neon, subdi3_neon, iordi3_neon)
21636 (anddi3_neon, xordi3_neon, ashldi3_neon, <shift>di3_neon): Use
21637 neon_for_64bits instead of nota8 and avoid_neon_for_64bits instead
21638 of onlya8.
21639 * doc/invoke.texi (-mneon-for-64bits): Document.
21640
21641 2013-03-21 Richard Biener <rguenther@suse.de>
21642
21643 PR tree-optimization/39326
21644 * tree-ssa-loop-im.c (bb_loop_postorder): New global static.
21645 (sort_bbs_in_loop_postorder_cmp): New function.
21646 (gather_mem_refs_in_loops): Assign mem-ref IDs in loop postorder.
21647
21648 2013-03-21 Richard Biener <rguenther@suse.de>
21649
21650 * tree-vect-data-refs.c (vect_update_interleaving_chain): Remove.
21651 (vect_insert_into_interleaving_chain): Likewise.
21652 (vect_drs_dependent_in_basic_block): Inline ...
21653 (vect_slp_analyze_data_ref_dependence): ... here. New function,
21654 split out from ...
21655 (vect_analyze_data_ref_dependence): ... here. Simplify.
21656 (vect_check_interleaving): Simplify.
21657 (vect_analyze_data_ref_dependences): Likewise. Split out ...
21658 (vect_slp_analyze_data_ref_dependences): ... this new function.
21659 (dr_group_sort_cmp): New function.
21660 (vect_analyze_data_ref_accesses): Compute data-reference groups
21661 here instead of in vect_analyze_data_ref_dependence. Use
21662 a more efficient algorithm.
21663 * tree-vect-slp.c (vect_slp_analyze_bb_1): Use
21664 vect_slp_analyze_data_ref_dependences. Call
21665 vect_analyze_data_ref_accesses earlier.
21666 * tree-vect-loop.c (vect_analyze_loop_2): Likewise.
21667 * tree-vectorizer.h (vect_analyze_data_ref_dependences): Adjust.
21668 (vect_slp_analyze_data_ref_dependences): New prototype.
21669
21670 2013-03-21 Richard Biener <rguenther@suse.de>
21671
21672 * tree-ssa-loop-im.c (can_sm_ref_p): Do not test whether
21673 ref is stored in the loop.
21674 (find_refs_for_sm): Walk only over all stores.
21675 (store_motion_loop): Allocate from lim_bitmap_obstack.
21676 (store_motion): Likewise.
21677
21678 2013-03-21 Richard Biener <rguenther@suse.de>
21679
21680 * tree-vect-loop-manip.c (slpeel_tree_peel_loop_to_edge):
21681 Update virtual SSA form.
21682
21683 2013-03-21 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
21684
21685 * configure.ac (gcc_cv_ld_eh_frame_ciev3): New test.
21686 * configure: Regenerate.
21687 * config.in: Regenerate.
21688 * config/sol2.c (solaris_override_options): Only enforce DWARF 2
21689 if !HAVE_LD_EH_FRAME_CIEV3.
21690
21691 2013-03-21 Richard Biener <rguenther@suse.de>
21692
21693 * tree-cfg.c (verify_expr_no_block): New function.
21694 (verify_expr_location_1): Verify that neither DECL_DEBUG_EXPR
21695 nor DECL_VALUE_EXPR have locations with associated blocks.
21696 * tree-ssa-live.c (clear_unused_block_pointer_1): Remove.
21697 (clear_unused_block_pointer): Remove code dealing with
21698 blocks in DECL_DEBUG_EXPR locations.
21699
21700 2013-03-21 Richard Biener <rguenther@suse.de>
21701
21702 * tree.h (DECL_DEBUG_EXPR_IS_FROM): Rename to ...
21703 (DECL_HAS_DEBUG_EXPR_P): ... this. Guard properly.
21704 * tree.c (copy_node_stat): Do not copy DECL_HAS_DEBUG_EXPR_P.
21705 * dwarf2out.c (add_var_loc_to_decl): Use DECL_HAS_DEBUG_EXPR_P
21706 instead of DECL_DEBUG_EXPR_IS_FROM.
21707 * gimplify.c (gimplify_modify_expr): Likewise.
21708 * tree-cfg.c (verify_expr_location_1): Likewise.
21709 * tree-complex.c (create_one_component_var): Likewise.
21710 * tree-sra.c (create_access_replacement): Likewise.
21711 * tree-ssa-live.c (clear_unused_block_pointer_1): Likewise.
21712 (clear_unused_block_pointer): Likewise.
21713 * tree-streamer-in.c (unpack_ts_decl_common_value_fields): Likewise.
21714 * tree-streamer-out.c (pack_ts_decl_common_value_fields): Likewise.
21715 * var-tracking.c (var_debug_decl): Likewise.
21716 (track_expr_p): Likewise.
21717 * tree-inline.c (add_local_variables): Likewise. Set
21718 DECL_HAS_DEBUG_EXPR_P after copying it.
21719 * tree-diagnostic.c (default_tree_printer): Use DECL_HAS_DEBUG_EXPR_P
21720 instead of DECL_DEBUG_EXPR_IS_FROM. Guard properly.
21721
21722 2013-03-21 Uros Bizjak <ubizjak@gmail.com>
21723
21724 PR bootstrap/56656
21725 * configure.ac (HAVE_AS_IX86_INTERUNIT_MOVQ): New test.
21726 * configure: Regenerate.
21727 * config.in: Regenerate.
21728 * config/i386/i386.md (*movdf_internal): Use
21729 HAVE_AS_IX86_INTERUNIT_MOVQ to handle broken assemblers that require
21730 movd instead of movq mnemonic for interunit moves.
21731 (*movdi_internal): Ditto.
21732
21733 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21734
21735 * config/aarch64/aarch64-simd.md (simd_fabd): New Attribute.
21736 (abd<mode>_3): New pattern.
21737 (aba<mode>_3): New pattern.
21738 (fabd<mode>_3): New pattern.
21739
21740 2013-03-21 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
21741
21742 * config/aarch64/aarch64-elf.h (REGISTER_PREFIX): Remove.
21743 * config/aarch64/aarch64.c (aarch64_print_operand): Remove all
21744 occurrence of REGISTER_PREFIX as its empty string.
21745
21746 2013-03-20 Jeff Law <law@redhat.com>
21747
21748 * tree-ssa-dom.c (record_equivalences_from_incoming_edge): Record
21749 addititional equivalences for equality comparisons between an SSA_NAME
21750 and a constant where the SSA_NAME was set from a widening conversion.
21751
21752 2013-03-20 Walter Lee <walt@tilera.com>
21753
21754 * config/tilegx/sync.md (atomic_test_and_set): New pattern.
21755
21756 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21757
21758 * config/i386/i386.md (*movoi_internal_avx): Emit insn template
21759 depending on type attribute.
21760 (*movti_internal): Ditto.
21761 (*movtf_internal): Ditto.
21762 (*movxf_internal): Ditto.
21763 (*movdf_internal): Ditto.
21764 (*movsf_internal): Ditto.
21765
21766 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
21767
21768 * config/i386/i386.md (*movti_internal): Set prefix attribute to
21769 maybe_vex for sselog1 and ssemov types.
21770 (*movdi_internal): Reorder operand constraints.
21771 (*movsi_internal): Ditto. Set prefix attribute to
21772 maybe_vex for sselog1 and ssemov types.
21773 (*movtf_internal): Set prefix attribute to maybe_vex
21774 for sselog1 and ssemov types.
21775 (*movdf_internal): Ditto. Set prefix_data16 attribute for
21776 DImode ssemov types. Reorder operand constraints.
21777 (*movsf_internal): Set type of alternatives 3,4 to imov. Set prefix
21778 attribute to maybe_vex for sselog1 and ssemov types. Set prefix_data16
21779 attribute for SImode ssemov types. Reorder operand constraints.
21780
21781 2013-03-20 Martin Jambor <mjambor@suse.cz>
21782
21783 * params.def (PARAM_IPA_CP_ARRAY_INDEX_HINT_BONUS): New parameter.
21784 * ipa-cp.c (hint_time_bonus): Add abonus for known array indices.
21785
21786 2013-03-20 Pat Haugen <pthaugen@us.ibm.com>
21787
21788 * config/rs6000/predicates.md (indexed_address, update_address_mem
21789 update_indexed_address_mem): New predicates.
21790 * config/rs6000/vsx.md (vsx_extract_<mode>_zero): Set correct "type"
21791 attribute for load/store instructions.
21792 * config/rs6000/dfp.md (movsd_store): Likewise.
21793 (movsd_load): Likewise.
21794 * config/rs6000/rs6000.md (zero_extend<mode>di2_internal1): Likewise.
21795 (unnamed HI->DI extend define_insn): Likewise.
21796 (unnamed SI->DI extend define_insn): Likewise.
21797 (unnamed QI->SI extend define_insn): Likewise.
21798 (unnamed QI->HI extend define_insn): Likewise.
21799 (unnamed HI->SI extend define_insn): Likewise.
21800 (unnamed HI->SI extend define_insn): Likewise.
21801 (extendsfdf2_fpr): Likewise.
21802 (movsi_internal1): Likewise.
21803 (movsi_internal1_single): Likewise.
21804 (movhi_internal): Likewise.
21805 (movqi_internal): Likewise.
21806 (movcc_internal1): Correct mnemonic for stw insn. Set correct "type"
21807 attribute for load/store instructions.
21808 (mov<mode>_hardfloat): Set correct "type" attribute for load/store
21809 instructions.
21810 (mov<mode>_softfloat): Likewise.
21811 (mov<mode>_hardfloat32): Likewise.
21812 (mov<mode>_hardfloat64): Likewise.
21813 (mov<mode>_softfloat64): Likewise.
21814 (movdi_internal32): Likewise.
21815 (movdi_internal64): Likewise.
21816 (probe_stack_<mode>): Likewise.
21817
21818 2013-03-20 Michael Meissner <meissner@linux.vnet.ibm.com>
21819
21820 * config/rs6000/vector.md (VEC_R): Add 32-bit integer, binary
21821 floating point, and decimal floating point to reload iterator.
21822
21823 * config/rs6000/constraints.md (wl constraint): New constraints to
21824 return FLOAT_REGS if certain options are used to reduce the number
21825 of separate patterns that exist in the file.
21826 (wx constraint): Likewise.
21827 (wz constraint): Likewise.
21828
21829 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21830 -mdebug=reg, print wg, wl, wx, and wz constraints.
21831 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21832 Initialize the reload functions for 64-bit binary/decimal floating
21833 point types.
21834 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21835 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21836 create the buffer on the stack to overcome not having a 32-bit
21837 load and store.
21838 (rs6000_emit_move): Likewise.
21839 (rs6000_secondary_memory_needed_rtx): Likewise.
21840 (rs6000_alloc_sdmode_stack_slot): Likewise.
21841 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21842 via xxlxor, just like DFmode 0.0.
21843
21844 * config/rs6000/rs6000.h (TARGET_NO_SDMODE_STACK): New macro,
21845 define as 1 if we are running on a power7 or newer.
21846 (enum r6000_reg_class_enum): Add new constraints.
21847
21848 * config/rs6000/dfp.md (movsd): Delete, combine with binary
21849 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21850 with other moves by using conditional constraits (wg). Use LFIWZX
21851 and STFIWX for loading SDmode on power7. Use xxlxor to create 0.0f.
21852 (movsd splitter): Likewise.
21853 (movsd_hardfloat): Likewise.
21854 (movsd_softfloat): Likewise.
21855
21856 * config/rs6000/rs6000.md (FMOVE32): New iterators to combine
21857 binary and decimal floating point moves.
21858 (fmove_ok): New attributes to combine binary and decimal floating
21859 point moves, and to combine power6x (mfpgpr) moves along normal
21860 floating moves.
21861 (real_value_to_target): Likewise.
21862 (f32_lr): Likewise.
21863 (f32_lm): Likewise.
21864 (f32_li): Likewise.
21865 (f32_sr): Likewise.
21866 (f32_sm): Likewise.
21867 (f32_si): Likewise.
21868 (movsf): Combine binary and decimal floating point moves. Combine
21869 power6x (mfpgpr) moves with other moves by using conditional
21870 constraits (wg). Use LFIWZX and STFIWX for loading SDmode on power7.
21871 (mov<mode> for SFmode/SDmode); Likewise.
21872 (SFmode/SDmode splitters): Likewise.
21873 (movsf_hardfloat): Likewise.
21874 (mov<mode>_hardfloat for SFmode/SDmode): Likewise.
21875 (movsf_softfloat): Likewise.
21876 (mov<mode>_softfloat for SFmode/SDmode): Likewise.
21877
21878 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wl,
21879 wx and wz constraints.
21880
21881 * config/rs6000/constraints.md (wg constraint): New constraint to
21882 return FLOAT_REGS if -mmfpgpr (power6x) was used.
21883
21884 * config/rs6000/rs6000.h (enum r6000_reg_class_enum): Add wg
21885 constraint.
21886
21887 * config/rs6000/rs6000.c (rs6000_debug_reg_global): If
21888 -mdebug=reg, print wg, wl, wx, and wz constraints.
21889 (rs6000_init_hard_regno_mode_ok): Initialize new constraints.
21890 Initialize the reload functions for 64-bit binary/decimal floating
21891 point types.
21892 (reg_offset_addressing_ok_p): If we are on a power7 or later, use
21893 LFIWZX and STFIWX to load/store 32-bit decimal types, and don't
21894 create the buffer on the stack to overcome not having a 32-bit
21895 load and store.
21896 (rs6000_emit_move): Likewise.
21897 (rs6000_secondary_memory_needed_rtx): Likewise.
21898 (rs6000_alloc_sdmode_stack_slot): Likewise.
21899 (rs6000_preferred_reload_class): On VSX, we can create SFmode 0.0f
21900 via xxlxor, just like DFmode 0.0.
21901
21902 * config/rs6000/dfp.md (movdd): Delete, combine with binary
21903 floating point moves in rs6000.md. Combine power6x (mfpgpr) moves
21904 with other moves by using conditional constraits (wg). Use LFIWZX
21905 and STFIWX for loading SDmode on power7.
21906 (movdd splitters): Likewise.
21907 (movdd_hardfloat32): Likewise.
21908 (movdd_softfloat32): Likewise.
21909 (movdd_hardfloat64_mfpgpr): Likewise.
21910 (movdd_hardfloat64): Likewise.
21911 (movdd_softfloat64): Likewise.
21912
21913 * config/rs6000/rs6000.md (FMOVE64): New iterators to combine
21914 64-bit binary and decimal floating point moves.
21915 (FMOVE64X): Likewise.
21916 (movdf): Combine 64-bit binary and decimal floating point moves.
21917 Combine power6x (mfpgpr) moves with other moves by using
21918 conditional constraits (wg).
21919 (mov<mode> for DFmode/DDmode): Likewise.
21920 (DFmode/DDmode splitters): Likewise.
21921 (movdf_hardfloat32): Likewise.
21922 (mov<mode>_hardfloat32 for DFmode/DDmode): Likewise.
21923 (movdf_softfloat32): Likewise.
21924 (movdf_hardfloat64_mfpgpr): Likewise.
21925 (movdf_hardfloat64): Likewise.
21926 (mov<mode>_hardfloat64 for DFmode/DDmode): Likewise.
21927 (movdf_softfloat64): Likewise.
21928 (mov<mode>_softfloat64 for DFmode/DDmode): Likewise.
21929 (reload_<mode>_load): Move to later in the file so they aren't in
21930 the middle of the floating point move insns.
21931 (reload_<mode>_store): Likewise.
21932
21933 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wg
21934 constraint.
21935
21936 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print out wg
21937 constraint if -mdebug=reg.
21938 (rs6000_initi_hard_regno_mode_ok): Enable wg constraint if -mfpgpr.
21939 Enable using dd reload support if needed.
21940
21941 * config/rs6000/dfp.md (movtd): Delete, combine with 128-bit
21942 binary and decimal floating point moves in rs6000.md.
21943 (movtd_internal): Likewise.
21944
21945 * config/rs6000/rs6000.md (FMOVE128): Combine 128-bit binary and
21946 decimal floating point moves.
21947 (movtf): Likewise.
21948 (movtf_internal): Likewise.
21949 (mov<mode>_internal, TDmode/TFmode): Likewise.
21950 (movtf_softfloat): Likewise.
21951 (mov<mode>_softfloat, TDmode/TFmode): Likewise.
21952
21953 * config/rs6000/rs6000.md (movdi_mfpgpr): Delete, combine with
21954 movdi_internal64, using wg constraint for move direct operations.
21955 (movdi_internal64): Likewise.
21956
21957 * config/rs6000/rs6000.c (rs6000_debug_reg_global): Print
21958 MODES_TIEABLE_P for selected modes. Print the numerical value of
21959 the various virtual registers. Use GPR/FPR first/last values,
21960 instead of hard coding the register numbers. Print which modes
21961 have reload functions registered.
21962 (rs6000_option_override_internal): If -mdebug=reg, trace the options
21963 settings before/after setting cpu, target and subtarget settings.
21964 (rs6000_secondary_reload_trace): Improve the RTL dump for -mdebug=addr
21965 and for secondary reload failures in rs6000_secondary_reload_inner.
21966 (rs6000_secondary_reload_fail): Likewise.
21967 (rs6000_secondary_reload_inner): Likewise.
21968
21969 * config/rs6000/rs6000.md (FIRST_GPR_REGNO): Add convenience
21970 macros for first/last GPR and FPR registers.
21971 (LAST_GPR_REGNO): Likewise.
21972 (FIRST_FPR_REGNO): Likewise.
21973 (LAST_FPR_REGNO): Likewise.
21974
21975 * config/rs6000/vector.md (mul<mode>3): Use the combined macro
21976 VECTOR_UNIT_ALTIVEC_OR_VSX_P instead of separate calls to
21977 VECTOR_UNIT_ALTIVEC_P and VECTOR_UNIT_VSX_P.
21978 (vcond<mode><mode>): Likewise.
21979 (vcondu<mode><mode>): Likewise.
21980 (vector_gtu<mode>): Likewise.
21981 (vector_gte<mode>): Likewise.
21982 (xor<mode>3): Don't allow logical operations on TImode in 32-bit
21983 to prevent the compiler from converting DImode operations to TImode.
21984 (ior<mode>3): Likewise.
21985 (and<mode>3): Likewise.
21986 (one_cmpl<mode>2): Likewise.
21987 (nor<mode>3): Likewise.
21988 (andc<mode>3): Likewise.
21989
21990 * config/rs6000/constraints.md (wt constraint): New constraint
21991 that returns VSX_REGS if TImode is allowed in VSX registers.
21992
21993 * config/rs6000/predicates.md (easy_fp_constant): 0.0f is an easy
21994 constant under VSX.
21995
21996 * config/rs6000/rs6000-modes.def (PTImode): Define, PTImode is
21997 similar to TImode, but it is restricted to being in the GPRs.
21998
21999 * config/rs6000/rs6000.opt (-mvsx-timode): New switch to allow
22000 TImode to occupy a single VSX register.
22001
22002 * config/rs6000/rs6000-cpus.def (ISA_2_6_MASKS_SERVER): Default to
22003 -mvsx-timode for power7/power8.
22004 (power7 cpu): Likewise.
22005 (power8 cpu): Likewise.
22006
22007 * config/rs6000/rs6000.c (rs6000_hard_regno_nregs_internal): Make
22008 sure that TFmode/TDmode take up two registers if they are ever
22009 allowed in the upper VSX registers.
22010 (rs6000_hard_regno_mode_ok): If -mvsx-timode, allow TImode in VSX
22011 registers.
22012 (rs6000_init_hard_regno_mode_ok): Likewise.
22013 (rs6000_debug_reg_global): Add debugging for PTImode and wt
22014 constraint. Print if LRA is turned on.
22015 (rs6000_option_override_internal): Give an error if -mvsx-timode
22016 and VSX is not enabled.
22017 (invalid_e500_subreg): Handle PTImode, restricting it to GPRs. If
22018 -mvsx-timode, restrict TImode to reg+reg addressing, and PTImode
22019 to reg+offset addressing. Use PTImode when checking offset
22020 addresses for validity.
22021 (reg_offset_addressing_ok_p): Likewise.
22022 (rs6000_legitimate_offset_address_p): Likewise.
22023 (rs6000_legitimize_address): Likewise.
22024 (rs6000_legitimize_reload_address): Likewise.
22025 (rs6000_legitimate_address_p): Likewise.
22026 (rs6000_eliminate_indexed_memrefs): Likewise.
22027 (rs6000_emit_move): Likewise.
22028 (rs6000_secondary_reload): Likewise.
22029 (rs6000_secondary_reload_inner): Handle PTImode. Allow 64-bit
22030 reloads to fpr registers to continue to use reg+offset addressing,
22031 but 64-bit reloads to altivec registers need reg+reg addressing.
22032 Drop test for PRE_MODIFY, since VSX loads/stores no longer support
22033 it. Treat LO_SUM like a PLUS operation.
22034 (rs6000_secondary_reload_class): If type is 64-bit, prefer to use
22035 FLOAT_REGS instead of VSX_RGS to allow use of reg+offset addressing.
22036 (rs6000_cannot_change_mode_class): Do not allow TImode in VSX
22037 registers to share a register with a smaller sized type, since VSX
22038 puts scalars in the upper 64-bits.
22039 (print_operand): Add support for PTImode.
22040 (rs6000_register_move_cost): Use VECTOR_MEM_VSX_P instead of
22041 VECTOR_UNIT_VSX_P to catch types that can be loaded in VSX
22042 registers, but don't have arithmetic support.
22043 (rs6000_memory_move_cost): Add test for VSX.
22044 (rs6000_opt_masks): Add -mvsx-timode.
22045
22046 * config/rs6000/vsx.md (VSm): Change to use 64-bit aligned moves
22047 for TImode.
22048 (VSs): Likewise.
22049 (VSr): Use wt constraint for TImode.
22050 (VSv): Drop TImode support.
22051 (vsx_movti): Delete, replace with versions for 32-bit and 64-bit.
22052 (vsx_movti_64bit): Likewise.
22053 (vsx_movti_32bit): Likewise.
22054 (vec_store_<mode>): Use VSX iterator instead of vector iterator.
22055 (vsx_and<mode>3): Delete use of '?' constraint on inputs, just put
22056 one '?' on the appropriate output constraint. Do not allow TImode
22057 logical operations on 32-bit systems.
22058 (vsx_ior<mode>3): Likewise.
22059 (vsx_xor<mode>3): Likewise.
22060 (vsx_one_cmpl<mode>2): Likewise.
22061 (vsx_nor<mode>3): Likewise.
22062 (vsx_andc<mode>3): Likewise.
22063 (vsx_concat_<mode>): Likewise.
22064 (vsx_xxpermdi_<mode>): Fix thinko for non V2DF/V2DI modes.
22065
22066 * config/rs6000/rs6000.h (MASK_VSX_TIMODE): Map from
22067 OPTION_MASK_VSX_TIMODE.
22068 (enum rs6000_reg_class_enum): Add RS6000_CONSTRAINT_wt.
22069 (STACK_SAVEAREA_MODE): Use PTImode instead of TImode.
22070
22071 * config/rs6000/rs6000.md (INT mode attribute): Add PTImode.
22072 (TI2 iterator): New iterator for TImode, PTImode.
22073 (wd mode attribute): Add values for vector types.
22074 (movti_string): Replace TI move operations with operations for TImode
22075 and PTImode. Add support for TImode being allowed in VSX registers.
22076 (mov<mode>_string, TImode/PTImode): Likewise.
22077 (movti_ppc64): Likewise.
22078 (mov<mode>_ppc64, TImode/PTImode): Likewise.
22079 (TI mode splitters): Likewise.
22080
22081 * doc/md.texi (PowerPC and IBM RS6000 constraints): Document wt
22082 constraint.
22083
22084 2013-03-20 Marc Glisse <marc.glisse@inria.fr>
22085
22086 PR tree-optimization/56355
22087 * fold-const.c (tree_binary_nonnegative_warnv_p) <MULT_EXPR>:
22088 Also handle integers with undefined overflow.
22089
22090 2013-03-20 Catherine Moore <clm@codesourcery.com>
22091 Maciej W. Rozycki <macro@codesourcery.com>
22092 Tom de Vries <tom@codesourcery.com>
22093 Nathan Sidwell <nathan@codesourcery.com>
22094 Iain Sandoe <iain@codesourcery.com>
22095 Nathan Froyd <froydnj@codesourcery.com>
22096 Chao-ying Fu <fu@mips.com>
22097
22098 * doc/extend.texi: (micromips, nomicromips, nocompression):
22099 Document new function attributes.
22100 * doc/invoke.texi (minterlink-compressed, mmicromips,
22101 m14k, m14ke, m14kec): Document new options.
22102 (minterlink-mips16): Update documentation.
22103 * doc/md.texi (ZC, ZD): Document new constraints.
22104 * configure.ac (gcc_cv_as_micromips): Check if linker
22105 supports the .set micromips directive.
22106 * configure: Regenerate.
22107 * config.in: Regenerate.
22108 * config/mips/mips-tables.opt: Regenerate.
22109 * config/mips/micromips.md: New file.
22110 * constraints.md (ZC, ZD): New constraints.
22111 * config/mips/predicates.md (movep_src_register): New predicate.
22112 (movep_src_operand): New predicate.
22113 (non_volatile_mem_operand): New predicate.
22114 * config/mips/mips.md (multimem): New type.
22115 (length): Differentiate between 17-bit and 18-bit branch offsets.
22116 (MOVEP1, MOVEP2): New mode iterator.
22117 (mov_<load>l): Use ZC constraint.
22118 (mov_<load>r): Likewise.
22119 (mov_<store>l): Likewise.
22120 (mov_<store>r): Likewise.
22121 (*branch_equality<mode>_inverted): Add microMIPS support.
22122 (*branch_equality<mode>): Likewise.
22123 (*jump_absolute): Likewise.
22124 (indirect_jump_<mode>): Likewise.
22125 (tablejump_<mode>): Likewise.
22126 (<optab>_internal): Likewise.
22127 (sibcall_internal): Likewise.
22128 (sibcall_value_internal): Likewise.
22129 (prefetch): Use constraint ZD.
22130 * config/mips/mips.opt (minterlink-compressed): New option.
22131 (minterlink-mips16): Now an alias for minterlink-compressed.
22132 (mmicromips): New option.
22133 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
22134 (compare_and_swap_12): Likewise.
22135 (sync_add<mode>): Likewise.
22136 (sync_<optab>_12): Likewise.
22137 (sync_old_<optab>_12): Likewise.
22138 (sync_new_<optab>_12): Likewise.
22139 (sync_nand_12): Likewise.
22140 (sync_old_nand_12): Likewise.
22141 (sync_new_nand_12): Likewise.
22142 (sync_sub<mode>): Likewise.
22143 (sync_old_add<mode>): Likewise.
22144 (sync_old_sub<mode>): Likewise.
22145 (sync_new_add<mode>): Likewise.
22146 (sync_new_sub<mode>): Likewise.
22147 (sync_<optab><mode>): Likewise.
22148 (sync_old_<optab><mode>): Likewise.
22149 (sync_new_<optab><mode>): Likewise.
22150 (sync_nand<mode>): Likewise.
22151 (sync_old_nand<mode>): Likewise.
22152 (sync_new_nand<mode>): Likewise.
22153 (sync_lock_test_and_set<mode>): Likewise.
22154 (test_and_set_12): Likewise.
22155 (atomic_compare_and_swap<mode>): Likewise.
22156 (atomic_exchange<mode>_llsc): Likewise.
22157 (atomic_fetch_add<mode>_llsc): Likewise.
22158 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
22159 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
22160 (umips_save_restore_pattern_p): Likewise.
22161 (umips_load_store_pair_p): Likewise.
22162 (umips_output_load_store_pair): Likewise.
22163 (umips_movep_target_p): Likewise.
22164 (umips_12bit_offset_address_p): Likewise.
22165 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
22166 (mips_base_mips16): Rename this...
22167 (mips_base_compression_flags): ...to this. Update all uses.
22168 (mips_attribute_table): Add micromips, nomicromips and nocompression.
22169 (mips_mips16_decl_p): Delete.
22170 (mips_nomips16_decl_p): Delete.
22171 (mips_get_compress_on_flags): New function.
22172 (mips_get_compress_off_flags): New function.
22173 (mips_get_compress_mode): New function.
22174 (mips_get_compress_on_name): New function.
22175 (mips_get_compress_off_name): New function.
22176 (mips_insert_attributes): Support multiple compression types.
22177 (mips_merge_decl_attributes): Likewise.
22178 (umips_12bit_offset_address_p): New function.
22179 (mips_start_function_definition): Emit .set micromips directive.
22180 (mips_call_may_need_jalx_p): New function.
22181 (mips_function_ok_for_sibcall): Add microMIPS support.
22182 (mips_print_operand_punctuation): Support short delay slots and
22183 compact jumps.
22184 (umips_swm_mask, umips_swm_encoding): New.
22185 (umips_build_save_restore): New function.
22186 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
22187 (was_mips16_p): Remove.
22188 (old_compression_mode): New.
22189 (mips_set_compression_mode): New function.
22190 (mips_set_current_function): Add microMIPS support.
22191 (mips_option_override): Likewise.
22192 (umips_save_restore_pattern_p): New function.
22193 (umips_output_save_restore): New function.
22194 (umips_load_store_pair_p_1): New function.
22195 (umips_load_store_pair_p): New function.
22196 (umips_output_load_store_pair_1): New function.
22197 (umips_output_load_store_pair): New function.
22198 (umips_movep_target_p) New function.
22199 (mips_prepare_pch_save): Add microMIPS support.
22200 * config/mips/mips.h (TARGET_COMPRESSION): New.
22201 (TARGET_CPU_CPP_BUILTINS): Update macro
22202 to use new compression flags and to support microMIPS.
22203 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
22204 (MIPS_ARCH_FLOAT_SPEC): Likewise.
22205 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
22206 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
22207 (ASM_SPEC): Support mmicromips and mno-micromips.
22208 (M16STORE_REG_P): New macro.
22209 (MIPS_CALL): Support TARGET_MICROMIPS.
22210 (MICROMIPS_J): New macro.
22211 (mips_base_mips16): Rename this...
22212 (mips_base_compression_flags): ...to this.
22213 (UMIPS_12BIT_OFFSET_P): New macro.
22214 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
22215 (MULTILIB_DIRNAMES): Likewise.
22216 2013-03-20 Richard Biener <rguenther@suse.de>
22217
22218 PR tree-optimization/56661
22219 * tree-ssa-sccvn.c (visit_use): Only value-number calls if
22220 the result does not have to be distinct.
22221
22222 2013-03-20 Richard Biener <rguenther@suse.de>
22223
22224 * tree-inline.c (copy_tree_body_r): Sync MEM_REF code with
22225 remap_gimple_op_r.
22226
22227 2013-03-20 Bill Schmidt <wschmidt@linux.vnet.ibm.com>
22228 Steven Bosscher <steven@gcc.gnu.org>
22229
22230 PR rtl-optimization/56605
22231 * loop-iv.c (implies_p): Handle equal RTXs and subregs.
22232
22233 2013-03-20 Uros Bizjak <ubizjak@gmail.com>
22234
22235 PR bootstrap/56656
22236 * config/i386/i386.md (*movdi_internal): Handle broken assemblers
22237 that require movd instead of movq.
22238
22239 2013-03-20 Richard Biener <rguenther@suse.de>
22240
22241 * tree-ssa-structalias.c (struct variable_info): Add pointer
22242 to the first field of an aggregate with sub-vars. Make
22243 this and the pointer to the next subfield its ID.
22244 (vi_next): New function.
22245 (nothing_id, anything_id, readonly_id, escaped_id, nonlocal_id,
22246 storedanything_id, integer_id): Increment by one.
22247 (new_var_info, get_call_vi, lookup_call_clobber_vi,
22248 get_call_clobber_vi): Adjust.
22249 (solution_set_expand): Simplify and speedup.
22250 (solution_set_add): Inline into ...
22251 (set_union_with_increment): ... this. Adjust accordingly.
22252 (do_sd_constraint): Likewise.
22253 (do_ds_constraint): Likewise.
22254 (do_complex_constraint): Simplify.
22255 (build_pred_graph): Adjust.
22256 (solve_graph): Likewise. Simplify and speedup.
22257 (get_constraint_for_ssa_var, get_constraint_for_ptr_offset,
22258 get_constraint_for_component_ref, get_constraint_for_1,
22259 first_vi_for_offset, first_or_preceding_vi_for_offset,
22260 create_function_info_for, create_variable_info_for_1,
22261 create_variable_info_for, intra_create_variable_infos): Adjust.
22262 (init_base_vars): Push NULL for ID zero.
22263 (compute_points_to_sets): Adjust.
22264
22265 2013-03-20 Richard Biener <rguenther@suse.de>
22266
22267 * cfgloop.c (verify_loop_structure): Streamline and avoid
22268 ICEing on corrupt loop tree.
22269 * graph.c (draw_cfg_nodes_for_loop): Avoid ICEing on corrupt
22270 loop tree.
22271
22272 2013-03-20 Richard Biener <rguenther@suse.de>
22273
22274 * tree-vect-loop-manip.c (slpeel_can_duplicate_loop_p): Do not
22275 check whether an SSA update is needed.
22276
22277 2013-03-20 Richard Sandiford <rdsandiford@googlemail.com>
22278
22279 * config/mips/constraints.md (T): Rename to...
22280 (Yf): ...this.
22281 (U): Rename to...
22282 (Yd): ...this.
22283 * config/mips/mips.md (*movdi_64bit, *movdi_64bit_mips16)
22284 (*mov<mode>_internal, *mov<mode>_mips16): Update accordingly.
22285
22286 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22287
22288 * config/aarch64/aarch64.md (*sub<mode>3_carryin): New pattern.
22289 (*subsi3_carryin_uxtw): Likewise.
22290
22291 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22292
22293 * config/aarch64/aarch64.md (*ror<mode>3_insn): New pattern.
22294 (*rorsi3_insn_uxtw): Likewise.
22295
22296 2013-03-19 Ian Bolton <ian.bolton@arm.com>
22297
22298 * config/aarch64/aarch64.md (*extr<mode>5_insn): New pattern.
22299 (*extrsi5_insn_uxtw): Likewise.
22300
22301 2013-03-19 Richard Biener <rguenther@suse.de>
22302
22303 PR tree-optimization/56273
22304 * passes.c (init_optimization_passes): Move second VRP after DOM.
22305
22306 2013-03-19 Uros Bizjak <ubizjak@gmail.com>
22307
22308 * config/i386/i386.md (*movti_internal): Merge from
22309 *movti_internal_rex64 and *movti_internal_sse. Use x64 isa attribute.
22310 (*movdi_internal): Merge with *movdi_internal_rex64. Use x64 and
22311 nox64 isa attributes.
22312
22313 2013-03-18 Richard Biener <rguenther@suse.de>
22314
22315 * tree-ssa-structalias.c (find): Use gcc_checking_assert.
22316 (unite): Likewise.
22317 (merge_node_constraints): Likewise.
22318 (build_succ_graph): Likewise.
22319 (valid_graph_edge): Inline into single caller.
22320 (unify_nodes): Likewise. Use bitmap_set_bit return value
22321 and cache varinfo.
22322 (scc_visit): Fix formatting and variable use.
22323 (do_sd_constraint): Use gcc_checking_assert.
22324 (do_ds_constraint): Likewise.
22325 (do_complex_constraint): Likewise.
22326 (condense_visit): Likewise. Cleanup.
22327 (dump_pred_graph): New function.
22328 (perform_var_substitution): Dump the pred-graph before
22329 variable substitution.
22330 (find_equivalent_node): Use gcc_checking_assert.
22331 (rewrite_constraints): Guard checking loop with ENABLE_CHECKING.
22332
22333 2013-03-18 Richard Biener <rguenther@suse.de>
22334
22335 * tree-vect-loop-manip.c (vect_create_cond_for_alias_checks):
22336 Remove cond_expr_stmt_list argument and do not gimplify the
22337 built expression.
22338 (vect_loop_versioning): Adjust.
22339 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref):
22340 Cleanup to use less temporaries.
22341 (vect_create_data_ref_ptr): Cleanup.
22342
22343 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22344
22345 PR tree-optimization/56635
22346 * fold-const.c (operand_equal_p): For MEM_REF and TARGET_MEM_REF,
22347 require types_compatible_p types.
22348
22349 2013-03-18 Nick Clifton <nickc@redhat.com>
22350
22351 * config/stormy16/stormy16.c (xstormy16_expand_prologue): Remove
22352 spurious backslash.
22353
22354 * config/mn10300/mn10300.c (mn10300_get_live_callee_saved_regs):
22355 Add missing line to comment describing function.
22356
22357 2013-03-18 Richard Biener <rguenther@suse.de>
22358
22359 PR tree-optimization/56210
22360 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
22361 Handle string / character search functions.
22362 * tree-ssa-alias.c (ref_maybe_used_by_call_p_1): Likewise.
22363
22364 2013-03-18 Richard Biener <rguenther@suse.de>
22365
22366 PR middle-end/56483
22367 * cfgexpand.c (expand_gimple_cond): Inline gimple_cond_single_var_p
22368 and implement properly.
22369 * gimple.h (gimple_cond_single_var_p): Remove.
22370
22371 2013-03-18 Richard Biener <rguenther@suse.de>
22372
22373 * tree-data-ref.h (find_data_references_in_loop): Declare.
22374 * tree-data-ref.c (get_references_in_stmt): Use a stack
22375 vector pre-allocated in the callers.
22376 (find_data_references_in_stmt): Adjust.
22377 (graphite_find_data_references_in_stmt): Likewise.
22378 (create_rdg_vertices): Likewise.
22379 (find_data_references_in_loop): Export.
22380 * tree-vect-data-refs.c (vect_analyze_data_ref_dependences):
22381 Compute dependences here...
22382 (vect_analyze_data_refs): ...not here. When we encounter
22383 a non-vectorizable data reference in basic-block vectorization
22384 truncate the data reference vector. Do not bother to
22385 fixup data-dependence information for gather loads.
22386 * tree-vect-slp.c (vect_slp_analyze_bb_1): Check the number
22387 of data references, as reported.
22388
22389 2013-03-18 Richard Biener <rguenther@suse.de>
22390
22391 PR tree-optimization/3713
22392 * tree-ssa-sccvn.c (visit_copy): Simplify. Always propagate
22393 has_constants and expr.
22394 (stmt_has_constants): Properly valueize SSA names when deciding
22395 whether the stmt has constants.
22396
22397 2013-03-18 Richard Biener <rguenther@suse.de>
22398
22399 * tree-ssa-loop-manip.c (find_uses_to_rename): Do not scan the
22400 whole function when there is nothing to do.
22401 * tree-ssa-loop.c (pass_vectorize): Remove TODO_update_ssa.
22402 * tree-vectorizer.c (vectorize_loops): Update virtual and
22403 loop-closed SSA once.
22404 * tree-vect-loop.c (vect_transform_loop): Do not update SSA here.
22405
22406 2013-03-18 Richard Biener <rguenther@suse.de>
22407
22408 PR middle-end/56113
22409 * domwalk.c (bb_postorder): New global static.
22410 (cmp_bb_postorder): New function.
22411 (walk_dominator_tree): Replace scheme imposing an order for
22412 visiting dominator sons by one sorting them at the time they
22413 are pushed on the stack.
22414
22415 2013-03-18 Richard Biener <rguenther@suse.de>
22416
22417 PR tree-optimization/39326
22418 * tree-ssa-loop-im.c (refs_independent_p): Exploit symmetry.
22419 (struct mem_ref): Replace mem member with ao_ref typed member.
22420 (MEM_ANALYZABLE): Adjust.
22421 (memref_eq): Likewise.
22422 (mem_ref_alloc): Likewise.
22423 (gather_mem_refs_stmt): Likewise.
22424 (mem_refs_may_alias_p): Use the ao_ref to query the alias oracle.
22425 (execute_sm_if_changed_flag_set): Adjust.
22426 (execute_sm): Likewise.
22427 (ref_always_accessed_p): Likewise.
22428 (refs_independent_p): Likewise.
22429 (can_sm_ref_p): Likewise.
22430
22431 2013-03-18 Jakub Jelinek <jakub@redhat.com>
22432
22433 PR c/56566
22434 * tree.c (tree_int_cst_min_precision): For integer_zerop (value)
22435 return 1 even for !unsignedp.
22436
22437 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22438
22439 * config/i386/i386.md (isa): Add x64 and nox64.
22440 (enabled): Define x64 for TARGET_64BIT and nox64 for !TARGET_64BIT.
22441 (*pushtf): Enable *roF alternative for x64 isa only.
22442 (*pushxf): Merge with *pushxf_nointeger. Use Yx*r constraint. Set
22443 mode attribute of integer alternatives to DImode for TARGET_64BIT.
22444 (*pushdf): Merge with *pushdf_rex64. Use x64 and nox64 isa attributes.
22445 (*movtf_internal): Merge from *movtf_internal_rex64 and
22446 *movtf_internal_sse. Use x64 and nox64 isa attributes.
22447 (*movxf_internal): Merge with *movxf_internal_rex64. Use x64 and
22448 nox64 isa attributes.
22449 (*movdf_internal): Merge with *movdf_internal_rex64. Use x64 and
22450 nox64 isa attributes.
22451 * config/i386/constraints.md (Yd): Do not set for TARGET_64BIT.
22452
22453 2013-03-17 Uros Bizjak <ubizjak@gmail.com>
22454
22455 * config/alpha/alpha.c (TARGET_LRA_P): New define.
22456
22457 2013-03-17 Jakub Jelinek <jakub@redhat.com>
22458
22459 PR target/56640
22460 * config/arm/arm.h (REG_CLASS_NAMES): Add "SFP_REG" and "AFP_REG"
22461 class names. Remove trailing comma after "ALL_REGS".
22462
22463 2013-03-16 Jan Hubicka <jh@suse.cz>
22464
22465 * cgraph.h (cgraph_get_create_real_symbol_node): Declare.
22466 * cgraph.c (cgraph_get_create_real_symbol_node): New function.
22467 * cgrpahbuild.c: Use cgraph_get_create_real_symbol_node instead
22468 of cgraph_get_create_node.
22469 * ipa-prop.c (ipa_make_edge_direct_to_target): Likewise.
22470
22471 2013-03-16 Jason Merrill <jason@redhat.com>
22472
22473 PR debug/49090
22474 * dwarf2out.c (gen_generic_params_dies): Indicate default arguments
22475 with DW_AT_default_value.
22476
22477 2013-03-16 Jakub Jelinek <jakub@redhat.com>
22478
22479 * BASE-VER: Set to 4.9.0.
22480
22481 2013-03-14 Andi Kleen <ak@linux.intel.com>
22482
22483 PR target/56619
22484 * doc/extend.texi: Document __ATOMIC_HLE_ACQUIRE,
22485 __ATOMIC_HLE_RELEASE. Document __builtin_ia32 TSX intrincs.
22486 Document _x* TSX intrinsics.
22487
22488 2013-03-14 Edgar E. Iglesias <edgar.iglesias@xilinx.com>
22489 David Holsgrove <david.holsgrove@xilinx.com>
22490
22491 * configure.ac: Add MicroBlaze TLS support detection.
22492 * configure: Regenerate.
22493 * config/microblaze/microblaze-protos.h
22494 (microblaze_cannot_force_const_mem, microblaze_tls_referenced_p,
22495 symbol_mentioned_p, label_mentioned_p): Add prototypes.
22496 * config/microblaze/microblaze.c (microblaze_address_type): Add
22497 ADDRESS_TLS and tls_reloc address types.
22498 (microblaze_address_info): Add tls_reloc.
22499 (TARGET_HAVE_TLS): Define.
22500 (get_tls_get_addr, microblaze_tls_symbol_p, microblaze_tls_operand_p_1,
22501 microblaze_tls_referenced_p, microblaze_cannot_force_const_mem,
22502 symbol_mentioned_p, label_mentioned_p, tls_mentioned_p,
22503 load_tls_operand, microblaze_call_tls_get_addr,
22504 microblaze_legitimize_tls_address): New functions.
22505 (microblaze_classify_unspec): Handle UNSPEC_TLS.
22506 (get_base_reg): Use microblaze_tls_symbol_p.
22507 (microblaze_classify_address): Handle TLS.
22508 (microblaze_legitimate_pic_operand): Use symbol_mentioned_p,
22509 label_mentioned_p and microblaze_tls_referenced_p.
22510 (microblaze_legitimize_address): Handle TLS.
22511 (microblaze_address_insns): Handle ADDRESS_TLS.
22512 (pic_address_needs_scratch): Handle TLS.
22513 (print_operand_address): Handle TLS.
22514 (microblaze_expand_prologue): Check TLS_NEEDS_GOT.
22515 (microblaze_expand_move): Handle TLS.
22516 (microblaze_legitimate_constant_p): Check
22517 microblaze_cannot_force_const_mem and microblaze_tls_symbol_p.
22518 (TARGET_CANNOT_FORCE_CONST_MEM): Define.
22519 * config/microblaze/microblaze.h (TLS_NEEDS_GOT): Define
22520 (PIC_OFFSET_TABLE_REGNUM): Set.
22521 * config/microblaze/linux.h (TLS_NEEDS_GOT): Define.
22522 * config/microblaze/microblaze.md (UNSPEC_TLS): Define.
22523 (addsi3, movsi_internal2, movdf_internal): Update constraints
22524 * config/microblaze/predicates.md (arith_plus_operand): Define
22525 (move_operand): Redefine as move_src_operand,
22526 check microblaze_tls_referenced_p.
22527
22528 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22529
22530 * config/aarch64/aarch64.md: (*and<mode>3nr_compare0): Use CC_NZ.
22531 (*and_<SHIFT:optab><mode>3nr_compare0): Likewise.
22532
22533 2013-03-14 Ian Bolton <ian.bolton@arm.com>
22534
22535 * config/aarch64/aarch64.c (aarch64_select_cc_mode): Return correct
22536 CC mode for AND.
22537
22538 2013-03-14 Jakub Jelinek <jakub@redhat.com>
22539
22540 PR tree-optimization/53265
22541 * common.opt (Waggressive-loop-optimizations): New option.
22542 * tree-ssa-loop-niter.c: Include tree-pass.h.
22543 (do_warn_aggressive_loop_optimizations): New function.
22544 (record_estimate): Call it. Don't add !is_exit bounds to loop->bounds
22545 if number_of_latch_executions returned constant.
22546 (estimate_numbers_of_iterations_loop): Call number_of_latch_executions
22547 early. If number_of_latch_executions returned constant, set
22548 nb_iterations_upper_bound back to it.
22549 * cfgloop.h (struct loop): Add warned_aggressive_loop_optimizations
22550 field.
22551 * Makefile.in (tree-ssa-loop-niter.o): Depend on $(TREE_PASS_H).
22552 * doc/invoke.texi (-Wno-aggressive-loop-optimizations): Document.
22553
22554 * config/aarch64/t-aarch64-linux (MULTARCH_DIRNAME): Remove.
22555 (MULTILIB_OSDIRNAMES): Set.
22556 * genmultilib: If defaultosdirname doesn't start with :: , set
22557 defaultosdirname2 instead, clear it and emit two . multilib_raw
22558 entries instead of just one.
22559
22560 2013-03-14 Kaz Kojima <kkojima@gcc.gnu.org>
22561
22562 * config/sh/linux.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22563 (SUBTARGET_OVERRIDE_OPTIONS): Set TARGET_USERMODE as default.
22564 * config/sh/netbsd-elf.h (TARGET_DEFAULT): Remove MASK_USERMODE.
22565 (SUBTARGET_OVERRIDE_OPTIONS): New.
22566
22567 2013-03-13 Oleg Endo <olegendo@gcc.gnu.org>
22568
22569 PR target/49880
22570 * config/sh/sh.opt (FPU_SINGLE_ONLY): New mask.
22571 (musermode): Convert to Var(TARGET_USERMODE).
22572 * config/sh/sh.h (SELECT_SH2A_SINGLE_ONLY, SELECT_SH4_SINGLE_ONLY,
22573 MASK_ARCH): Add MASK_FPU_SINGLE_ONLY.
22574 * config/sh/sh.c (sh_option_override): Use
22575 TARGET_FPU_DOUBLE || TARGET_FPU_SINGLE_ONLY for call-fp case.
22576 * config/sh/sh.md (udivsi3_i1, divsi3_i1): Remove ! TARGET_SH4
22577 condition.
22578 (udivsi3_i4, divsi3_i4): Use TARGET_FPU_DOUBLE condition instead of
22579 TARGET_SH4.
22580 (udivsi3_i4_single, divsi3_i4_single): Use
22581 TARGET_FPU_SINGLE_ONLY || TARGET_FPU_DOUBLE instead of TARGET_HARD_SH4.
22582
22583 2013-03-13 Dave Korn <dave.korn.cygwin@....>
22584
22585 * config/i386/cygwin.h (SHARED_LIBGCC_SPEC): Make shared libgcc the
22586 default setting.
22587
22588 2013-03-13 Richard Biener <rguenther@suse.de>
22589
22590 PR tree-optimization/56608
22591 * tree-vect-slp.c (vect_schedule_slp): Do not remove scalar
22592 calls when vectorizing basic-blocks.
22593
22594 2013-03-13 Jakub Jelinek <jakub@redhat.com>
22595
22596 PR plugins/45078
22597 * config.gcc: On arm, mips, sh and sparc add vxworks-dummy.h to
22598 tm_file.
22599
22600 2013-03-12 Jakub Jelinek <jakub@redhat.com>
22601
22602 * doc/invoke.texi (-Waddr-space-convert): Move into the table earlier.
22603
22604 2013-03-11 Jan Hubicka <jh@suse.cz>
22605
22606 PR lto/56557
22607 * lto-streamer-out.c (output_symbol_p): Skip references from
22608 constructors of external variables.
22609
22610 2013-03-11 Jan Hubicka <jh@suse.cz>
22611
22612 PR middle-end/56571
22613 * valtrack.c (cleanup_auto_inc_dec): Unshare clobbers originating
22614 from pseudos.
22615 * emit-rtl.c (verify_rtx_sharing): Likewise.
22616 (copy_insn_1): Likewise.
22617 * rtl.c (copy_rtx): Likewise.
22618
22619 2013-03-11 Georg-Johann Lay <avr@gjlay.de>
22620
22621 PR target/56591
22622 * config/avr/avr.c (avr_print_operand): Add space after '%c' in
22623 output_operand_lossage message.
22624
22625 2013-03-11 Richard Earnshaw <rearnsha@arm.com>
22626
22627 PR target/56470
22628 * arm.c (shift_op): Validate RTL pattern on the fly.
22629 (arm_print_operand, case 'S'): Don't use shift_operator to validate
22630 the RTL.
22631
22632 2013-03-10 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
22633
22634 PR target/56347
22635 * config/pa/pa.md (call_value): Check for calls to powf and direct to
22636 new call patterns that clobber %fr12.
22637 (call_val_powf, call_val_powf_pic, call_val_powf_64bit): New insn,
22638 split and postreload patterns.
22639 * config/pa/pa.c (pa_conditional_register_usage): Revert marking
22640 registers %fr12 and %fr12R as call used.
22641
22642 2013-03-09 Steven Bosscher <steven@gcc.gnu.org>
22643
22644 * dse.c (delete_dead_store_insn): Respect TDF_DETAILS.
22645 (canon_address, record_store, replace_read, check_mem_read_rtx,
22646 scan_insn, dse_step1, dse_step2_init, dse_step2_spill,
22647 dse_step4, dse_step5_nospill, dse_step5_spill, dse_step6,
22648 rest_of_handle_dse): Likewise.
22649
22650 2013-03-09 Richard Sandiford <rdsandiford@googlemail.com>
22651
22652 PR middle-end/56524
22653 * tree.h (tree_optimization_option): Rename target_optabs to optabs.
22654 Add base_optabs.
22655 (TREE_OPTIMIZATION_OPTABS): Update after previous field change.
22656 (TREE_OPTIMIZATION_BASE_OPTABS): New macro.
22657 (save_optabs_if_changed): Replace with...
22658 (init_tree_optimization_optabs): ...this.
22659 * optabs.c (save_optabs_if_changed): Rename to...
22660 (init_tree_optimization_optabs): ...this. Take the optimization node
22661 as argument. Do nothing if the base optabs are already correct.
22662 Reuse the existing TREE_OPTIMIZATION_OPTABS memory if we need
22663 to recompute optabs.
22664 * function.h (function): Remove optabs field.
22665 * function.c (invoke_set_current_function_hook): Call
22666 init_tree_optimization_optabs. Use the result to initialize
22667 this_fn_optabs.
22668
22669 2013-02-27 Aldy Hernandez <aldyh@redhat.com>
22670
22671 * trans-mem.c (expand_transaction): Do not set PR_INSTRUMENTEDCODE
22672 if GTMA_HAS_NO_INSTRUMENTATION.
22673 (generate_tm_state): Keep GTMA_HAS_NO_INSTRUMENTATION bit.
22674 (ipa_tm_transform_transaction): Set GTMA_HAS_NO_INSTRUMENTATION.
22675 * gimple.h (GTMA_HAS_NO_INSTRUMENTATION): Define.
22676 * gimple-pretty-print.c (dump_gimple_transaction): Handle
22677 GTMA_HAS_NO_INSTRUMENTATION.
22678
22679 2013-03-08 Jakub Jelinek <jakub@redhat.com>
22680
22681 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Don't link against
22682 libasan_preinit.o.
22683
22684 2013-03-08 Marek Polacek <polacek@redhat.com>
22685 Jakub Jelinek <jakub@redhat.com>
22686
22687 PR tree-optimization/56478
22688 * predict.c (is_comparison_with_loop_invariant_p): Change the
22689 type of loop_step to tree.
22690 (predict_loops): Adjust.
22691 (predict_iv_comparison): Perform the computations on double_ints.
22692
22693 2013-03-08 Richard Biener <rguenther@suse.de>
22694
22695 PR tree-optimization/56570
22696 * tree-cfg.c (verify_expr_location_1): Verify locations for
22697 DECL_DEBUG_EXPR.
22698 * tree-sra.c (create_access_replacement): Strip locations
22699 from DECL_DEBUG_EXPRs.
22700
22701 2013-03-08 Richard Biener <rguenther@suse.de>
22702
22703 * tree-inline.c (expand_call_inline): Do not associate
22704 a BLOCK with the location in BLOCK_SOURCE_LOCATION.
22705 * tree-cfg.c (verify_location): Verify BLOCK_SOURCE_LOCATION.
22706
22707 2013-03-08 Richard Biener <rguenther@suse.de>
22708
22709 * tree-ssa-ter.c (is_replaceable_p): Do not TER across location
22710 or block changes with -Og. Fix for location / block encoding
22711 changes and PHI arguments with locations.
22712
22713 2013-03-07 Steven Bosscher <steven@gcc.gnu.org>
22714
22715 * bitmap.c (struct bitmap_descriptor_d): Use unsigned HOST_WIDEST_INT
22716 for all counters.
22717 (struct output_info): Likewise.
22718 (register_overhead): Remove bad gcc_assert.
22719 (bitmap_find_bit): If there is only a single bitmap element, do not
22720 count a miss as a search.
22721 (print_statistics): Update for counter type changes.
22722 (dump_bitmap_statistics): Likewise. Print headers such that they
22723 are properly lined up with the printed counters.
22724
22725 2013-03-07 Jakub Jelinek <jakub@redhat.com>
22726
22727 PR tree-optimization/56559
22728 * tree-ssa-reassoc.c (zero_one_operation): When looking at rhs2,
22729 check that it has only a single use.
22730
22731 2013-03-07 Richard Biener <rguenther@suse.de>
22732
22733 * doc/invoke.texi (fwhole-program): Discourage use in combination
22734 with -flto.
22735
22736 2013-03-06 Jakub Jelinek <jakub@redhat.com>
22737
22738 * config/arm/t-arm (TM_H, OPTIONS_H_EXTRA): Add arm-cores.def.
22739
22740 PR tree-optimization/56539
22741 * tree-tailcall.c (adjust_return_value_with_ops): Use GSI_SAME_STMT
22742 instead of GSI_CONTINUE_LINKING as last argument to
22743 force_gimple_operand_gsi. Adjust function comment.
22744
22745 * config/aarch64/t-aarch64 (TM_H, OPTIONS_H_EXTRA): Add
22746 aarch64-cores.def.
22747
22748 PR middle-end/56548
22749 * expr.c (expand_cond_expr_using_cmove): When expanding cmove in
22750 promoted mode, convert the result back to the original mode.
22751
22752 2013-03-06 Richard Biener <rguenther@suse.de>
22753
22754 PR middle-end/56294
22755 * tree-into-ssa.c (insert_phi_nodes_for): Add dumping.
22756 (insert_updated_phi_nodes_compare_uids): New function.
22757 (update_ssa): Sort symbols_to_rename after UID before
22758 traversing it to insert PHI nodes.
22759
22760 2013-03-06 Richard Biener <rguenther@suse.de>
22761
22762 PR middle-end/50494
22763 * tree-vect-data-refs.c (vect_can_force_dr_alignment_p):
22764 Do not adjust alignment of DECL_IN_CONSTANT_POOL decls.
22765
22766 Revert
22767 2013-02-13 Richard Biener <rguenther@suse.de>
22768
22769 PR lto/50494
22770 * varasm.c (output_constant_def_1): Get the decl representing
22771 the constant as argument.
22772 (output_constant_def): Wrap output_constant_def_1.
22773 (make_decl_rtl): Use output_constant_def_1 with the decl
22774 representing the constant.
22775 (build_constant_desc): Optionally re-use a decl already
22776 representing the constant.
22777 (tree_output_constant_def): Adjust.
22778
22779 2013-03-06 Joey Ye <joey.ye@arm.com>
22780
22781 PR lto/50293
22782 * gcc.c (convert_white_space): New function.
22783 (main): Handles white space in function name.
22784
22785 2013-03-06 Oleg Endo <olegendo@gcc.gnu.org>
22786
22787 PR target/56529
22788 * config/sh/sh.c (sh_option_override): Check for TARGET_DYNSHIFT
22789 instead of TARGET_SH2 for call-table case. Do not set sh_div_strategy
22790 to SH_DIV_CALL_TABLE for TARGET_SH2.
22791 * config.gcc (sh_multilibs): Add m2 and m2a to sh*-*-linux* multilib
22792 list.
22793 * doc/invoke.texi (SH options): Document mdiv= call-div1, call-fp,
22794 call-table options.
22795
22796 2013-03-05 Sterling Augustine <saugustine@google.com>
22797 Cary Coutant <ccoutant@google.com>
22798
22799 PR debug/55364
22800 * dwarf2out.c (resolve_addr): Don't call
22801 remove_loc_list_addr_table_entries a second time for the same
22802 expression.
22803
22804 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22805
22806 PR debug/56510
22807 * cfgexpand.c (expand_debug_parm_decl): Call copy_rtx on incoming.
22808 (avoid_complex_debug_insns): New function.
22809 (expand_debug_locations): Call it.
22810
22811 PR rtl-optimization/56484
22812 * ifcvt.c (noce_process_if_block): If else_bb is NULL, avoid extending
22813 lifetimes of hard registers on small register class machines.
22814
22815 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
22816
22817 * config/microblaze/microblaze-protos.h: Rename
22818 microblaze_is_interrupt_handler to microblaze_is_interrupt_variant.
22819 * config/microblaze/microblaze.c (microblaze_attribute_table): Add
22820 fast_interrupt.
22821 (microblaze_fast_interrupt_function_p): New function.
22822 (microblaze_is_interrupt_handler): Rename to
22823 microblaze_is_interrupt_variant and add fast_interrupt check.
22824 (microblaze_must_save_register): Use microblaze_is_interrupt_variant.
22825 (save_restore_insns): Likewise.
22826 (compute_frame_size): Likewise.
22827 (microblaze_function_prologue): Add FAST_INTERRUPT_NAME.
22828 (microblaze_globalize_label): Likewise.
22829 * config/microblaze/microblaze.h: Define FAST_INTERRUPT_NAME.
22830 * config/microblaze/microblaze.md: Use wrapper
22831 microblaze_is_interrupt_variant.
22832
22833 2013-03-05 Kai Tietz <ktietz@redhat.com>
22834
22835 * sdbout.c (sdbout_one_type): Switch to current function's section
22836 supporting cold/hot.
22837
22838 2013-03-05 David Holsgrove <david.holsgrove@xilinx.com>
22839
22840 * doc/invoke.texi (MicroBlaze): Add -mbig-endian, -mlittle-endian,
22841 -mxl-reorder.
22842
22843 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22844
22845 PR middle-end/56461
22846 * ggc-common.c (gt_pch_save): For ENABLE_VALGRIND_CHECKING,
22847 if VALGRIND_GET_VBITS is defined, temporarily make object
22848 memory all defined, and restore previous valgrind addressability
22849 and definability afterwards. Free this_object at the end.
22850
22851 PR middle-end/56461
22852 * lra.c (lra): Call lra_clear_live_ranges if live_p,
22853 right before calling lra_create_live_ranges, also call it
22854 when clearing live_p. Only call lra_clear_live_ranges
22855 at the end if live_p.
22856
22857 PR middle-end/56461
22858 * sched-deps.c (delete_dep_node): Free DEP_REPLACE.
22859
22860 2013-03-05 Richard Biener <rguenther@suse.de>
22861
22862 PR tree-optimization/56521
22863 * tree-ssa-sccvn.c (set_value_id_for_result): Always initialize
22864 value-id.
22865
22866 2013-03-05 Steven Bosscher <steven@gcc.gnu.org>
22867
22868 PR c++/55135
22869 * except.h (remove_unreachable_eh_regions): New prototype.
22870 * except.c (remove_eh_handler_splicer): New function, split out
22871 of remove_eh_handler.
22872 (remove_eh_handler): Use remove_eh_handler_splicer. Add comment
22873 warning about running it on many EH regions one at a time.
22874 (remove_unreachable_eh_regions_worker): New function, walk the
22875 EH tree in depth-first order and remove non-marked regions.
22876 (remove_unreachable_eh_regions): New function.
22877 * tree-eh.c (mark_reachable_handlers): New function, split out
22878 from remove_unreachable_handlers.
22879 (remove_unreachable_handlers): Use mark_reachable_handlers and
22880 remove_unreachable_eh_regions.
22881 (remove_unreachable_handlers_no_lp): Use mark_reachable_handlers
22882 and remove_unreachable_eh_regions.
22883
22884 2013-03-05 Richard Biener <rguenther@suse.de>
22885
22886 PR middle-end/56525
22887 * loop-init.c (fix_loop_structure): Remove loops in two stages,
22888 not freeing them until the end.
22889
22890 2013-03-05 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
22891
22892 * config/s390/s390.h: Define DWARF2_ASM_LINE_DEBUG_INFO.
22893
22894 2013-03-05 Richard Biener <rguenther@suse.de>
22895
22896 PR tree-optimization/56270
22897 * tree-vect-slp.c (vect_schedule_slp): Clear vectorized stmts
22898 of loads after scheduling an SLP instance.
22899
22900 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22901
22902 * Makefile.in (dg_target_exps): Add aarch64.exp, epiphany.exp and
22903 tic6x.exp.
22904 (check_gcc_parallelize): Run guality.exp as a separate job from
22905 vect.exp with unsorted.exp and $(dg_target_exps) separately from
22906 struct-layout-1.exp with stackalign.exp.
22907
22908 * alias.c (init_alias_analysis): Clear reg_known_equiv_p bitmap.
22909
22910 PR middle-end/56461
22911 * tree-vect-slp.c (vect_supported_load_permutation_p): Free
22912 load_index sbitmap even if some bit in it isn't set.
22913
22914 PR middle-end/56461
22915 * tree-ssa-loop-niter.c (bb_queue): Remove typedef.
22916 (discover_iteration_bound_by_body_walk): Change queues to
22917 vec<vec<basic_block> > and queue to vec<basic_block>. Fix up
22918 spelling in comment. Call safe_push on queues[bound_index] directly.
22919 Release queues[queue_index] in every iteration unconditionally.
22920 Release bounds vector.
22921
22922 PR middle-end/56461
22923 * tree-vect-stmts.c (free_stmt_vec_info_vec): Call
22924 free_stmt_vec_info on any left-over stmt_vec_info in the vector.
22925 * tree-vect-loop.c (vect_create_epilog_for_reduction): Release
22926 inner_phis vector.
22927
22928 2013-03-05 Richard Biener <rguenther@suse.de>
22929
22930 PR lto/56515
22931 * tree-inline.c (remap_blocks_to_null): New function.
22932 (expand_call_inline): When expanding a call stmt without
22933 an associated block inline remap all callee blocks to NULL.
22934
22935 2013-03-05 Jakub Jelinek <jakub@redhat.com>
22936
22937 PR rtl-optimization/56494
22938 * simplify-rtx.c (simplify_truncation): If C is narrower than A,
22939 optimize (truncate:A (subreg:B (truncate:C X) 0)) into
22940 (subreg:A (truncate:C X) 0) instead of (truncate:A X).
22941
22942 PR middle-end/56461
22943 * sel-sched-ir.c (free_sched_pools): Release
22944 succs_info_pool.stack[succs_info_pool.max_top] vectors too
22945 if succs_info_pool.max_top isn't -1.
22946
22947 PR bootstrap/56509
22948 * opts.c (opts_obstack, opts_concat): Moved to...
22949 * opts-common.c (opts_obstack, opts_concat): ... here.
22950
22951 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22952
22953 PR middle-end/56461
22954 * diagnostic.c (diagnostic_append_note): Save and restore old prefix.
22955
22956 2013-03-04 Martin Jambor <mjambor@suse.cz>
22957
22958 * tree-dfa.c (get_or_create_ssa_default_def): Use parameter fn in
22959 all appropriate places.
22960
22961 2013-01-04 Eric Botcazou <ebotcazou@adacore.com>
22962
22963 PR tree-optimization/56424
22964 * ipa-split.c (split_function): Do not set the RSO flag if result is
22965 not by reference and its type is a register type.
22966
22967 2013-03-04 David Holsgrove <david.holsgrove@xilinx.com>
22968
22969 * config/microblaze/microblaze.c (microblaze_valid_pic_const): New
22970 (microblaze_legitimate_pic_operand): Likewise
22971 * config/microblaze/microblaze.h (LEGITIMATE_PIC_OPERAND_P): calls
22972 new function microblaze_legitimate_pic_operand
22973 * config/microblaze/microblaze-protos.h
22974 (microblaze_legitimate_pic_operand): Declare.
22975
22976 2013-03-04 Edgar E. Iglesias <edgar.iglesias@gmail.com>
22977
22978 * config/microblaze/predicates.md (call_insn_simple_operand):
22979 New predicate for supported rtx code types.
22980 * config/microblaze/microblaze.md (call_internal1): Use
22981 call_insn_simple_operand predicate.
22982
22983 2013-03-04 Jakub Jelinek <jakub@redhat.com>
22984
22985 PR middle-end/56461
22986 * tree-loop-distribution.c (ldist_gen): Call partition_free after each
22987 partitions.ordered_remove.
22988
22989 PR middle-end/56461
22990 * tree-vect-stmts.c (vectorizable_conversion): Don't call
22991 vec_oprnds0.create (1) for modifier == NONE.
22992
22993 PR middle-end/56461
22994 * tree-vect-stmts.c (vectorizable_shift): Don't call create methods
22995 on vec_oprnds0 or vec_oprnds1 before loop, only call it on
22996 vec_oprnds1 right before pushing anything to it for
22997 scalar_shift_arg.
22998
22999 PR middle-end/56461
23000 * tree-vect-loop.c (destroy_loop_vec_info): For !clean_stmts, just
23001 set nbbs to 0 instead of having separate code path.
23002 (vect_analyze_loop_form): Call destroy_loop_vec_info with true
23003 instead of false as last argument if returning NULL.
23004
23005 2013-03-03 Sandra Loosemore <sandra@codesourcery.com>
23006
23007 * target.def (TARGET_OPTION_VALID_ATTRIBUTE_P): Update comments;
23008 the attribute is now called "target" instead of "option".
23009 (TARGET_OPTION_PRAGMA_PARSE): Likewise, for the pragma.
23010 * doc/tm.texi.in (Target Attributes): Likewise document the correct
23011 attribute/pragma name for TARGET_OPTION_VALID_P and
23012 TARGET_OPTION_PRAGMA_PARSE. Also copy-edit and correct markup.
23013 * doc/tm.texi: Regenerated.
23014
23015 2013-03-02 David Holsgrove <david.holsgrove@xilinx.com>
23016
23017 * config/microblaze/microblaze.c:
23018 Check mcpu, pcmp requirement and set TARGET_REORDER to 0 if not met.
23019 * config/microblaze/microblaze.h: Add -mxl-reorder to
23020 DRIVER_SELF_SPECS.
23021 * config/microblaze/microblaze.md: New bswapsi2 and bswaphi2.
23022 instructions emitted if TARGET_REORDER.
23023 * config/microblaze/microblaze.opt: New option -mxl-reorder set to 1
23024 or 0 for -m/-mno case, but initialises as 2 to detect default use case
23025 separately.
23026
23027 2013-03-01 Xinliang David Li <davidxl@google.com>
23028
23029 * tree-ssa-uninit.c (compute_control_dep_chain): Limit post-dom
23030 walk length.
23031
23032 2013-03-01 Jakub Jelinek <jakub@redhat.com>
23033
23034 PR middle-end/56461
23035 * tree-ssa-loop-ivcanon.c (tree_estimate_loop_size): Release path
23036 vector even when returning true. Fix up function comment formatting.
23037
23038 PR middle-end/56461
23039 * ira-build.c (ira_loop_nodes_count): New variable.
23040 (create_loop_tree_nodes): Initialize it.
23041 (finish_loop_tree_nodes): Use it instead of looking at current_loops.
23042
23043 PR middle-end/56461
23044 * tree-vect-data-refs.c (vect_permute_store_chain): Avoid using copy
23045 method on dr_chain and result_chain.
23046 * tree-vect-stmts.c (vectorizable_store): Only call
23047 result_chain.create if j == 0.
23048
23049 PR middle-end/56461
23050 * tree-vect-stmts.c (vect_create_vectorized_promotion_stmts): Call
23051 vec_oprnds0->release (); rather than vec_oprnds0->truncate (0)
23052 before overwriting it.
23053
23054 2013-03-01 Tobias Burnus <burnus@net-b.de>
23055
23056 * doc/extended.texi (C Extensions): Change order in @menu
23057 to match @node.
23058 (Other MIPS Built-in Functions): Move last MIPS entry before
23059 "picoChip Built-in Functions".
23060 (SH Built-in Functions): Move after RX Built-in Functions.
23061 * doc/gcc.texi (Introduction): Change order in @menu
23062 to match @node.
23063 * doc/md.texi (Constraints): Ditto.
23064 * gty.texi (Type Information): Ditto.
23065 (User-provided marking routines for template types): Make
23066 subsection.
23067 * doc/invoke.texi (AArch64 Options): Move before
23068 "Adapteva Epiphany Options".
23069
23070 2013-02-28 Konstantin Serebryany <konstantin.s.serebryany@gmail.com>
23071 Jakub Jelinek <jakub@redhat.com>
23072
23073 PR sanitizer/56454
23074 * asan.c (gate_asan): Lookup no_sanitize_address instead of
23075 no_address_safety_analysis attribute.
23076 * doc/extend.texi (no_address_safety_attribute): Rename to
23077 no_sanitize_address attribute, mention no_address_safety_analysis
23078 attribute as deprecated alias.
23079
23080 2013-02-28 Jakub Jelinek <jakub@redhat.com>
23081
23082 PR middle-end/56461
23083 * tree-vectorizer.h (vect_get_slp_defs): Change 3rd argument
23084 type to vec<vec<tree> > *.
23085 * tree-vect-slp.c (vect_get_slp_defs): Likewise. Change vec_defs
23086 to be vec<tree> instead of vec<tree> *, set vec_defs
23087 to vNULL and call vec_defs.create (number_of_vects), adjust other
23088 uses of vec_defs.
23089 * tree-vect-stmts.c (vect_get_vec_defs, vectorizable_call,
23090 vectorizable_condition): Adjust vect_get_slp_defs callers.
23091
23092 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23093
23094 * config/aarch64/aarch64.c
23095 (aarch64_float_const_representable): Remove unused variable.
23096
23097 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23098
23099 * config/aarch64/aarch64.c (aarch64_mangle_type): Make static.
23100
23101 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23102
23103 * config/aarch64/aarch64-builtins.c
23104 (aarch64_init_simd_builtins): Make static.
23105
23106 2013-02-28 James Greenhalgh <james.greenhalgh@arm.com>
23107
23108 * config/aarch64/aarch64.c
23109 (aarch64_simd_make_constant): Make static.
23110
23111 2013-02-28 Martin Jambor <mjambor@suse.cz>
23112
23113 * tree-sra.c (load_assign_lhs_subreplacements): Do not put replacements
23114 with no initialization to the RHS of debug statements.
23115
23116 2013-02-28 Martin Jambor <mjambor@suse.cz>
23117
23118 PR tree-optimization/56294
23119 * tree-sra.c (analyze_access_subtree): Create replacement declarations.
23120 Adjust dumping.
23121 (get_access_replacement): Do not call create_access_replacement.
23122 Assert a replacement exists.
23123 (get_repl_default_def_ssa_name): Create the replacement declaration
23124 itself.
23125
23126 2013-02-28 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
23127
23128 * config/arm/arm.c (arm_output_mi_thunk): Call final_start_function and
23129 final_end_function.
23130
23131 2013-02-28 Marek Polacek <polacek@redhat.com>
23132
23133 PR rtl-optimization/56466
23134 * loop-unroll.c (unroll_and_peel_loops): Call fix_loop_structure
23135 if we're changing a loop.
23136 (peel_loops_completely): Likewise.
23137
23138 2013-02-28 Paolo Carlini <paolo.carlini@oracle.com>
23139
23140 PR c++/55813
23141 * doc/invoke.texi ([-Wctor-dtor-privacy]): Complete.
23142
23143 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23144
23145 PR target/56445
23146 * config/avr/avr.c (avr_init_builtins): Use 'n' instead of empty
23147 macro parameters with: FX_FTYPE_FX, FX_FTYPE_FX_INT, INT_FTYPE_FX,
23148 INTX_FTYPE_FX, FX_FTYPE_INTX.
23149 * config/avr/builtins.def: Adjust respective DEF_BUILTIN.
23150
23151 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23152
23153 * avr/avr-mcus.def (ata5272, ata5505, attiny1634, ata6285)
23154 (ata6286, atmega8a, atmega48pa, ata5790, ata5790n, ata5795)
23155 (atmega164pa, atmega165pa, atmega168pa, atmega16hva, atmega16hvb)
23156 (atmega16hvbrevb, atmega16m1, atmega16u4, atmega26hvg, atmega32a)
23157 (atmega32a, atmega3250pa, atmega3290pa, atmega32c1, atmega32m1)
23158 (atmega32u4, atmega32u6, atmega64a, atmega6490a, atmega6490p)
23159 (atmega64c1, atmega64m1, atmega64rfa2, atmega64rfr2, atmega32hvb)
23160 (atmega32hvbrevb, atmega16hva2, atmega48hvf, at90pwm161)
23161 (atmega128a, atmega1284, atmxt112sl, atmxt224, atmxt224e)
23162 (atmxt336s, atxmega16a4u, atxmega16c4, atxmega32a4u, atxmega32c4)
23163 (atxmega32e5, atxmega64a3u, atxmega64a4u, atxmega64b1, atxmega64b3)
23164 (atxmega64c3, atxmega64d4, atxmega128a3u, atxmega128b1)
23165 (atxmega128b3, atxmega128c3, atxmega128d4, atmxt540s, atmxt540sreva)
23166 (atxmega192a3u, atxmega192c3, atxmega256a3u, atxmega256c3)
23167 (atxmega384c3, atxmega384d3, atxmega128a4u): New AVR_MCU.
23168 (avrxmega6): Increase max flash segments from 5 to 6.
23169 * config/avr/t-multilib: Regenerate.
23170 * config/avr/avr-tables.opt: Regenerate.
23171 * doc/avr-mmcu.texi: Regenerate.
23172
23173 2013-02-28 Georg-Johann Lay <avr@gjlay.de>
23174
23175 * config/avr/avr.h (device_to_arch): Rename to device_to_ld.
23176 (avr_device_to_arch): Rename to avr_device_to_ld.
23177 (avr_device_to_as): New prototype.
23178 (EXTRA_SPEC_FUNCTIONS): Add device_to_as.
23179 (ASM_SPEC): Use device_to_as to get -mmcu= and -mno-skip-bug=.
23180 * config/avr/driver-avr.c (avr_device_to_as): New.
23181 (avr_device_to_arch): Rename to avr_device_to_ld.
23182
23183 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23184
23185 PR middle-end/56461
23186 * tree-vect-data-refs.c (vect_permute_load_chain): Avoid using copy
23187 method on dr_chain and result_chain.
23188
23189 PR middle-end/56461
23190 * tree-ssa-loop-niter.c (maybe_lower_iteration_bound): Call
23191 pointer_set_destroy on not_executed_last_iteration.
23192
23193 PR middle-end/56461
23194 * tree-vect-loop.c (vectorizable_reduction): Release vect_defs vector.
23195
23196 PR middle-end/56461
23197 * ipa-pure-const.c (propagate): Use FOR_EACH_FUNCTION instead of
23198 FOR_EACH_DEFINED_FUNCTION when freeing state.
23199
23200 PR middle-end/56461
23201 * df-scan.c (df_insn_delete): Use df_scan_free_mws_vec before
23202 pool_free.
23203 (df_insn_rescan_debug_internal): Use df_scan_free_mws_vec before
23204 overwriting it.
23205
23206 PR middle-end/56461
23207 * ipa-cp.c (decide_whether_version_node): Call vec_free on
23208 known_aggs[i].items and release known_aggs vector.
23209
23210 PR middle-end/56461
23211 * ipa-reference.c (propagate): Free node_info even for alias nodes.
23212
23213 2013-02-27 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23214
23215 * config/microblaze/microblaze.c (microblaze_emit_compare):
23216 Use xor for EQ/NE comparisions.
23217 * config/microblaze/microblaze.md (cstoresf4): Add constraints
23218 (cbranchsf4): Adjust operator to comparison_operator.
23219
23220 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23221
23222 PR middle-end/56461
23223 * tree-flow.h (edge_var_map_vector): Change into va_heap, vl_embed
23224 vector.
23225 * tree-ssa.c (redirect_edge_var_map_add): Use vec_safe_reserve and
23226 vec_safe_push, always update *slot.
23227 (redirect_edge_var_map_clear): Use vec_free.
23228 (redirect_edge_var_map_dup): Use vec_safe_copy and vec_safe_reserve.
23229 (free_var_map_entry): Use vec_free.
23230 * tree-cfgcleanup.c (remove_forwarder_block_with_phi): Use
23231 FOR_EACH_VEC_SAFE_ELT instead of FOR_EACH_VEC_ELT.
23232
23233 2013-02-27 Andrey Belevantsev <abel@ispras.ru>
23234
23235 PR middle-end/45472
23236 * sel-sched-ir.c (merge_expr): Also change vinsn of merged expr
23237 when the may_trap_p bit of the exprs being merged differs.
23238 Reorder tests for speculativeness in the logical and operator.
23239
23240 2013-02-27 Jakub Jelinek <jakub@redhat.com>
23241
23242 * incpath.c (add_standard_paths): Use reconcat instead of concat
23243 where appropriate and avoid leaking memory.
23244
23245 * opts.h: Include obstack.h.
23246 (opts_concat): New prototype.
23247 (opts_obstack): New declaration.
23248 * opts.c (opts_concat): New function.
23249 (opts_obstack): New variable.
23250 (init_options_struct): Call gcc_init_obstack on opts_obstack.
23251 (finish_options): Use opts_concat instead of concat
23252 and XOBNEWVEC instead of XNEWVEC.
23253 * opts-common.c (generate_canonical_option, decode_cmdline_option,
23254 generate_option): Likewise.
23255 * Makefile.in (OPTS_H): Depend on $(OBSTACK_H).
23256 * lto-wrapper.c (main): Call gcc_init_obstack on opts_obstack.
23257
23258 PR target/56455
23259 * stmt.c (expand_switch_as_decision_tree_p): If flag_pic
23260 and ASM_OUTPUT_ADDR_DIFF_ELT isn't defined, return true.
23261
23262 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23263
23264 PR middle-end/56461
23265 * lra-spills.c (lra_spill): Free spill_hard_reg at the end.
23266
23267 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23268
23269 * config/arm/arm.c (const_ok_for_dimode_op): Back out last change.
23270 (arm_block_move_unaligned_straight): Likewise.
23271 (arm_adjust_block_mem): Likewise.
23272
23273 2013-02-26 Joern Rennecke <joern.rennecke@embecosm.com>
23274
23275 PR target/48901
23276 * config/lm32/lm32.c (gen_int_relational): Remove unused variables
23277 temp, cond and label.
23278 * config/lm32/lm32.md (ashlsi3): Remove unused variable one.
23279
23280 PR target/52500
23281 * config/c6x/c6x.c (dbx_register_map): Change to unsigned.
23282 * config/c6x/c6x.h (dbx_register_map): Update declaration.
23283
23284 PR target/52501
23285 * config/cr16/cr16-protos.h: Move end of RTX_CODE guard below end
23286 of prologue/epilogue functions.
23287
23288 PR target/52550
23289 * config/tilegx/tilegx.c (tilegx_expand_prologue):
23290 Remove unused variable cfa_offset.
23291 * config/tilepro/tilepro.c (tilepro_expand_prologue): Likewise.
23292
23293 PR target/54639
23294 * config/mn10300/mn10300.c (mn10300_expand_epilogue): Avoid offset
23295 type promotion to unsigned.
23296
23297 PR target/54640
23298 * config/arm/arm.c (const_ok_for_dimode_op): Make code consistent
23299 for HOST_WIDE_INT of 32 bit / same size as int.
23300 (arm_block_move_unaligned_straight): Likewise.
23301 (arm_adjust_block_mem): Likewise.
23302
23303 PR target/54662
23304 * config/mep/t-mep (mep-pragma.o): Use ALL_COMPILERFLAGS instead of
23305 ALL_CFLAGS.
23306
23307 2013-02-26 Marek Polacek <polacek@redhat.com>
23308
23309 PR tree-optimization/56426
23310 * tree-ssa-loop.c (tree_ssa_loop_init): Always call scev_initialize.
23311
23312 2013-02-26 Richard Biener <rguenther@suse.de>
23313
23314 PR target/56444
23315 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc): Remove
23316 unused variable loops.
23317
23318 2013-02-26 Jakub Jelinek <jakub@redhat.com>
23319
23320 PR tree-optimization/56448
23321 * fold-const.c (operand_equal_p) <case tcc_reference>: Don't look at
23322 TREE_SIDE_EFFECTS if flags contain OEP_CONSTANT_ADDRESS_OF.
23323 Clear OEP_CONSTANT_ADDRESS_OF from flags before recursing on second or
23324 later operands of the references, or even first operand for
23325 INDIRECT_REF, TARGET_MEM_REF or MEM_REF.
23326
23327 PR tree-optimization/56443
23328 * tree-vect-stmts.c (get_vectype_for_scalar_type_and_size): For
23329 overaligned types, pass TYPE_UNSIGNED (scalar_type) as second argument
23330 to type_for_mode langhook.
23331
23332 2013-02-25 Matt Turner <mattst88@gmail.com>
23333
23334 * doc/invoke.texi: Document r4700.
23335
23336 2013-02-25 Richard Biener <rguenther@suse.de>
23337
23338 PR tree-optimization/56175
23339 * tree-ssa-forwprop.c (hoist_conversion_for_bitop_p): New predicate,
23340 split out from ...
23341 (simplify_bitwise_binary): ... here. Also guard the conversion
23342 of (type) X op CST to (type) (X op ((type-x) CST)) with it.
23343
23344 2013-02-25 Catherine Moore <clm@codesourcery.com>
23345
23346 Revert:
23347 2013-02-24 Catherine Moore <clm@codesourcery.com>
23348 Maciej W. Rozycki <macro@codesourcery.com>
23349 Tom de Vries <tom@codesourcery.com>
23350 Nathan Sidwell <nathan@codesourcery.com>
23351 Iain Sandoe <iain@codesourcery.com>
23352 Nathan Froyd <froydnj@codesourcery.com>
23353 Chao-ying Fu <fu@mips.com>
23354
23355 * doc/extend.texi: (micromips, nomicromips, nocompression):
23356 Document new function attributes.
23357 * doc/invoke.texi (minterlink-compressed, mmicromips,
23358 m14k, m14ke, m14kec): Document new options.
23359 (minterlink-mips16): Update documentation.
23360 * doc/md.texi (ZC, ZD): Document new constraints.
23361 * configure.ac (gcc_cv_as_micromips): Check if linker
23362 supports the .set micromips directive.
23363 * configure: Regenerate.
23364 * config.in: Regenerate.
23365 * config/mips/mips-tables.opt: Regenerate.
23366 * config/mips/micromips.md: New file.
23367 * constraints.md (ZC, AD): New constraints.
23368 * config/mips/predicates.md (movep_src_register): New predicate.
23369 (movep_src_operand): New predicate.
23370 (non_volatile_mem_operand): New predicate.
23371 * config/mips/mips.md (multimem): New type.
23372 (length): Differentiate between 17-bit and 18-bit branch offsets.
23373 (MOVEP1, MOVEP2): New mode iterator.
23374 (mov_<load>l): Use ZC constraint.
23375 (mov_<load>r): Likewise.
23376 (mov_<store>l): Likewise.
23377 (mov_<store>r): Likewise.
23378 (*branch_equality<mode>_inverted): Add microMIPS support.
23379 (*branch_equality<mode>): Likewise.
23380 (*jump_absolute): Likewise.
23381 (indirect_jump_<mode>): Likewise.
23382 (tablejump_<mode>): Likewise.
23383 (<optab>_internal): Likewise.
23384 (sibcall_internal): Likewise.
23385 (sibcall_value_internal): Likewise.
23386 (prefetch): Use constraint ZD.
23387 * config/mips/mips.opt (minterlink-compressed): New option.
23388 (minterlink-mips16): Now an alias for minterlink-compressed.
23389 (mmicromips): New option.
23390 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23391 (compare_and_swap_12): Likewise.
23392 (sync_add<mode>): Likewise.
23393 (sync_<optab>_12): Likewise.
23394 (sync_old_<optab>_12): Likewise.
23395 (sync_new_<optab>_12): Likewise.
23396 (sync_nand_12): Likewise.
23397 (sync_old_nand_12): Likewise.
23398 (sync_new_nand_12): Likewise.
23399 (sync_sub<mode>): Likewise.
23400 (sync_old_add<mode>): Likewise.
23401 (sync_old_sub<mode>): Likewise.
23402 (sync_new_add<mode>): Likewise.
23403 (sync_new_sub<mode>): Likewise.
23404 (sync_<optab><mode>): Likewise.
23405 (sync_old_<optab><mode>): Likewise.
23406 (sync_new_<optab><mode>): Likewise.
23407 (sync_nand<mode>): Likewise.
23408 (sync_old_nand<mode>): Likewise.
23409 (sync_new_nand<mode>): Likewise.
23410 (sync_lock_test_and_set<mode>): Likewise.
23411 (test_and_set_12): Likewise.
23412 (atomic_compare_and_swap<mode>): Likewise.
23413 (atomic_exchange<mode>_llsc): Likewise.
23414 (atomic_fetch_add<mode>_llsc): Likewise.
23415 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23416 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23417 (umips_save_restore_pattern_p): Likewise.
23418 (umips_load_store_pair_p): Likewise.
23419 (umips_output_load_store_pair): Likewise.
23420 (umips_movep_target_p): Likewise.
23421 (umips_12bit_offset_address_p): Likewise.
23422 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23423 (mips_base_mips16): Rename this...
23424 (mips_base_compression_flags): ...to this. Update all uses.
23425 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23426 (mips_mips16_decl_p): Delete.
23427 (mips_nomips16_decl_p): Delete.
23428 (mips_get_compress_on_flags): New function.
23429 (mips_get_compress_off_flags): New function.
23430 (mips_get_compress_mode): New function.
23431 (mips_get_compress_on_name): New function.
23432 (mips_get_compress_off_name): New function.
23433 (mips_insert_attributes): Support multiple compression types.
23434 (mips_merge_decl_attributes): Likewise.
23435 (umips_12bit_offset_address_p): New function.
23436 (mips_start_function_definition): Emit .set micromips directive.
23437 (mips_call_may_need_jalx_p): New function.
23438 (mips_function_ok_for_sibcall): Add microMIPS support.
23439 (mips_print_operand_punctuation): Support short delay slots and
23440 compact jumps.
23441 (umips_swm_mask, umips_swm_encoding): New.
23442 (umips_build_save_restore): New function.
23443 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23444 (was_mips16_p): Remove.
23445 (old_compression_mode): New.
23446 (mips_set_compression_mode): New function.
23447 (mips_set_current_function): Add microMIPS support.
23448 (mips_option_override): Likewise.
23449 (umips_save_restore_pattern_p): New function.
23450 (umips_output_save_restore): New function.
23451 (umips_load_store_pair_p_1): New function.
23452 (umips_load_store_pair_p): New function.
23453 (umips_output_load_store_pair_1): New function.
23454 (umips_output_load_store_pair): New function.
23455 (umips_movep_target_p) New function.
23456 (mips_prepare_pch_save): Add microMIPS support.
23457 * config/mips/mips.h (TARGET_COMPRESSION): New.
23458 (TARGET_CPU_CPP_BUILTINS): Update macro
23459 to use new compression flags and to support microMIPS.
23460 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23461 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23462 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23463 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23464 (ASM_SPEC): Support mmicromips and mno-micromips.
23465 (M16STORE_REG_P): New macro.
23466 (MIPS_CALL): Support TARGET_MICROMIPS.
23467 (MICROMIPS_J): New macro.
23468 (mips_base_mips16): Rename this...
23469 (mips_base_compression_flags): ...to this.
23470 (UMIPS_12BIT_OFFSET_P): New macro.
23471 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23472 (MULTILIB_DIRNAMES): Likewise.
23473
23474 2013-02-25 Tom de Vries <tom@codesourcery.com>
23475
23476 PR rtl-optimization/56131
23477 * insn-notes.def (INSN_NOTE_BASIC_BLOCK): Update comment.
23478 * cfgrtl.c (delete_insn): Don't reorder NOTE_INSN_DELETED_LABEL and
23479 NOTE_INSN_BASIC_BLOCK if BLOCK_FOR_INSN == NULL.
23480
23481 2013-02-25 Tobias Burnus <burnus@net-b.de>
23482
23483 * doc/invoke.texi (-fsanitize=): Move from optimization
23484 to debugging options.
23485
23486 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23487
23488 * sched-deps.c (sched_analyze_insn): Fix typo in comment.
23489
23490 2013-02-25 Andrey Belevantsev <abel@ispras.ru>
23491 Alexander Monakov <amonakov@ispras.ru>
23492
23493 PR middle-end/56077
23494 * sched-deps.c (sched_analyze_insn): When reg_pending_barrier,
23495 flush pending lists also on non-jumps. Adjust comment.
23496
23497 2013-02-24 Catherine Moore <clm@codesourcery.com>
23498 Maciej W. Rozycki <macro@codesourcery.com>
23499 Tom de Vries <tom@codesourcery.com>
23500 Nathan Sidwell <nathan@codesourcery.com>
23501 Iain Sandoe <iain@codesourcery.com>
23502 Nathan Froyd <froydnj@codesourcery.com>
23503 Chao-ying Fu <fu@mips.com>
23504
23505 * doc/extend.texi: (micromips, nomicromips, nocompression):
23506 Document new function attributes.
23507 * doc/invoke.texi (minterlink-compressed, mmicromips,
23508 m14k, m14ke, m14kec): Document new options.
23509 (minterlink-mips16): Update documentation.
23510 * doc/md.texi (ZC, ZD): Document new constraints.
23511 * configure.ac (gcc_cv_as_micromips): Check if linker
23512 supports the .set micromips directive.
23513 * configure: Regenerate.
23514 * config.in: Regenerate.
23515 * config/mips/mips-tables.opt: Regenerate.
23516 * config/mips/micromips.md: New file.
23517 * constraints.md (ZC, AD): New constraints.
23518 * config/mips/predicates.md (movep_src_register): New predicate.
23519 (movep_src_operand): New predicate.
23520 (non_volatile_mem_operand): New predicate.
23521 * config/mips/mips.md (multimem): New type.
23522 (length): Differentiate between 17-bit and 18-bit branch offsets.
23523 (MOVEP1, MOVEP2): New mode iterator.
23524 (mov_<load>l): Use ZC constraint.
23525 (mov_<load>r): Likewise.
23526 (mov_<store>l): Likewise.
23527 (mov_<store>r): Likewise.
23528 (*branch_equality<mode>_inverted): Add microMIPS support.
23529 (*branch_equality<mode>): Likewise.
23530 (*jump_absolute): Likewise.
23531 (indirect_jump_<mode>): Likewise.
23532 (tablejump_<mode>): Likewise.
23533 (<optab>_internal): Likewise.
23534 (sibcall_internal): Likewise.
23535 (sibcall_value_internal): Likewise.
23536 (prefetch): Use constraint ZD.
23537 * config/mips/mips.opt (minterlink-compressed): New option.
23538 (minterlink-mips16): Now an alias for minterlink-compressed.
23539 (mmicromips): New option.
23540 * config/mips/sync.md (sync_compare_and_swap<mode>): Use ZR constraint.
23541 (compare_and_swap_12): Likewise.
23542 (sync_add<mode>): Likewise.
23543 (sync_<optab>_12): Likewise.
23544 (sync_old_<optab>_12): Likewise.
23545 (sync_new_<optab>_12): Likewise.
23546 (sync_nand_12): Likewise.
23547 (sync_old_nand_12): Likewise.
23548 (sync_new_nand_12): Likewise.
23549 (sync_sub<mode>): Likewise.
23550 (sync_old_add<mode>): Likewise.
23551 (sync_old_sub<mode>): Likewise.
23552 (sync_new_add<mode>): Likewise.
23553 (sync_new_sub<mode>): Likewise.
23554 (sync_<optab><mode>): Likewise.
23555 (sync_old_<optab><mode>): Likewise.
23556 (sync_new_<optab><mode>): Likewise.
23557 (sync_nand<mode>): Likewise.
23558 (sync_old_nand<mode>): Likewise.
23559 (sync_new_nand<mode>): Likewise.
23560 (sync_lock_test_and_set<mode>): Likewise.
23561 (test_and_set_12): Likewise.
23562 (atomic_compare_and_swap<mode>): Likewise.
23563 (atomic_exchange<mode>_llsc): Likewise.
23564 (atomic_fetch_add<mode>_llsc): Likewise.
23565 * config/mips/mips-cpus.def (m14kc, m14k): New processors.
23566 * config/mips/mips-protos.h (umips_output_save_restore): New prototype.
23567 (umips_save_restore_pattern_p): Likewise.
23568 (umips_load_store_pair_p): Likewise.
23569 (umips_output_load_store_pair): Likewise.
23570 (umips_movep_target_p): Likewise.
23571 (umips_12bit_offset_address_p): Likewise.
23572 * config/mips/mips.c (MIPS_MAX_FIRST_STEP): Update for microMIPS.
23573 (mips_base_mips16): Rename this...
23574 (mips_base_compression_flags): ...to this. Update all uses.
23575 (mips_attribute_table): Add micromips, nomicromips and nocompression.
23576 (mips_mips16_decl_p): Delete.
23577 (mips_nomips16_decl_p): Delete.
23578 (mips_get_compress_on_flags): New function.
23579 (mips_get_compress_off_flags): New function.
23580 (mips_get_compress_mode): New function.
23581 (mips_get_compress_on_name): New function.
23582 (mips_get_compress_off_name): New function.
23583 (mips_insert_attributes): Support multiple compression types.
23584 (mips_merge_decl_attributes): Likewise.
23585 (umips_12bit_offset_address_p): New function.
23586 (mips_start_function_definition): Emit .set micromips directive.
23587 (mips_call_may_need_jalx_p): New function.
23588 (mips_function_ok_for_sibcall): Add microMIPS support.
23589 (mips_print_operand_punctuation): Support short delay slots and
23590 compact jumps.
23591 (umips_swm_mask, umips_swm_encoding): New.
23592 (umips_build_save_restore): New function.
23593 (mips_for_each_saved_gpr_and_fpr): Add microMIPS support.
23594 (was_mips16_p): Remove.
23595 (old_compression_mode): New.
23596 (mips_set_compression_mode): New function.
23597 (mips_set_current_function): Add microMIPS support.
23598 (mips_option_override): Likewise.
23599 (umips_save_restore_pattern_p): New function.
23600 (umips_output_save_restore): New function.
23601 (umips_load_store_pair_p_1): New function.
23602 (umips_load_store_pair_p): New function.
23603 (umips_output_load_store_pair_1): New function.
23604 (umips_output_load_store_pair): New function.
23605 (umips_movep_target_p) New function.
23606 (mips_prepare_pch_save): Add microMIPS support.
23607 * config/mips/mips.h (TARGET_COMPRESSION): New.
23608 (TARGET_CPU_CPP_BUILTINS): Update macro
23609 to use new compression flags and to support microMIPS.
23610 (MIPS_ISA_LEVEL_SPEC): Add m14k processors.
23611 (MIPS_ARCH_FLOAT_SPEC): Likewise.
23612 (ISA_HAS_LWXS): Include TARGET_MICROMIPS.
23613 (ISA_HAS_LOAD_DELAY): Exclude TARGET_MICROMIPS.
23614 (ASM_SPEC): Support mmicromips and mno-micromips.
23615 (M16STORE_REG_P): New macro.
23616 (MIPS_CALL): Support TARGET_MICROMIPS.
23617 (MICROMIPS_J): New macro.
23618 (mips_base_mips16): Rename this...
23619 (mips_base_compression_flags): ...to this.
23620 (UMIPS_12BIT_OFFSET_P): New macro.
23621 * config/mips/t-sde: (MULTILIB_OPTIONS): Add microMIPS.
23622 (MULTILIB_DIRNAMES): Likewise.
23623
23624 2013-02-24 Jakub Jelinek <jakub@redhat.com>
23625
23626 PR target/52555
23627 * target-globals.c (save_target_globals): For init_reg_sets and
23628 target_reinit remporarily set this_fn_optabs to this_target_optabs.
23629
23630 2013-02-22 James Grennahlgh <james.greenhalgh@arm.com>
23631
23632 * config/aarch64/aarch64-simd-builtins.def: Add copyright header.
23633 * config/aarch64/t-aarch64
23634 (aarch64-builtins.o): Depend on aarch64-simd-builtins.def.
23635
23636 2013-02-22 Vladimir Makarov <vmakarov@redhat.com>
23637
23638 PR inline-asm/56148
23639 * lra-constraints.c (process_alt_operands): Reload operand
23640 conflicting with earlier clobber only if no more other conflicting
23641 operands.
23642
23643 2013-02-22 Jakub Jelinek <jakub@redhat.com>
23644
23645 PR sanitizer/56393
23646 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Link in libasan_preinit.o
23647 if not linking a shared library.
23648
23649 2013-02-22 Seth LaForge <sethml@google.com>
23650
23651 * config.gcc (arm*-*-eabi*): Treat arm*eb as big-endian.
23652
23653 2013-02-22 Greta Yorsh <Greta.Yorsh@arm.com>
23654
23655 * config/arm/arm.md (split for extendsidi): Update condition.
23656 (zero_extend<mode>di2,extend<mode>di2): Add an alternative.
23657 * config/arm/iterators.md (qhs_extenddi_cstr): Likewise.
23658 (qhs_zextenddi_cstr): Likewise.
23659
23660 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23661
23662 PR middle-end/56420
23663 * expmed.c (EXACT_POWER_OF_2_OR_ZERO_P): Do subtraction in uhwi, to
23664 avoid signed wrapping.
23665 (expand_mult): Handle properly multiplication by
23666 ((dword_type) -1) << (BITS_PER_WORD - 1). Improve multiplication by
23667 ((dword_type) 1) << (BITS_PER_WORD - 1). Avoid undefined behavior
23668 in the compiler if coeff is HOST_WIDE_INT_MIN.
23669 (expand_divmod): Don't make ext_op1 static, change it's type to
23670 uhwi. Avoid undefined behavior in -INTVAL (op1).
23671
23672 PR rtl-optimization/50339
23673 * lower-subreg.h (struct lower_subreg_choices): Add splitting_ashiftrt
23674 field.
23675 * lower-subreg.c (compute_splitting_shift): Handle ASHIFTRT.
23676 (compute_costs): Call compute_splitting_shift also for ASHIFTRT
23677 into splitting_ashiftrt field.
23678 (find_decomposable_shift_zext, resolve_shift_zext): Handle also
23679 ASHIFTRT.
23680 (dump_choices): Fix up printing LSHIFTRT choices, print ASHIFTRT
23681 choices.
23682
23683 2013-02-20 Aldy Hernandez <aldyh@redhat.com>
23684
23685 PR middle-end/56108
23686 * trans-mem.c (execute_tm_mark): Do not expand transactions that
23687 are sure to go irrevocable.
23688
23689 2013-02-21 Hans-Peter Nilsson <hp@axis.com>
23690
23691 * doc/rtl.texi (vec_concat, vec_duplicate): Mention that
23692 scalars are valid operands.
23693
23694 2013-02-21 Martin Jambor <mjambor@suse.cz>
23695
23696 PR tree-optimization/56310
23697 * ipa-cp.c (agg_replacements_to_vector): New parameter index, copy
23698 only matching indices and non-negative final offsets.
23699 (intersect_aggregates_with_edge): Pass src_idx to
23700 agg_replacements_to_vector. Pass src_idx insstead of index to
23701 intersect_with_agg_replacements.
23702
23703 2013-02-21 Martin Jambor <mjambor@suse.cz>
23704
23705 * ipa-cp.c (good_cloning_opportunity_p): Dump the real threshold
23706 instead of hard-wired defaults.
23707
23708 2013-02-21 Maciej W. Rozycki <macro@codesourcery.com>
23709
23710 * doc/invoke.texi (MIPS Options): Update documentation of the
23711 floating-point multiply-accumulate instruction restrictions.
23712
23713 2013-02-21 Kostya Serebryany <kcc@google.com>
23714
23715 * config/i386/i386.c (ix86_asan_shadow_offset): Use 0x7fff8000 as
23716 asan_shadow_offset on x86_64 linux.
23717
23718 2013-02-21 Richard Biener <rguenther@suse.de>
23719
23720 PR tree-optimization/56415
23721 Revert
23722 2013-02-11 Richard Biener <rguenther@suse.de>
23723
23724 PR tree-optimization/56273
23725 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
23726 first VRP run.
23727
23728 2013-02-21 Jakub Jelinek <jakub@redhat.com>
23729
23730 PR bootstrap/56258
23731 * doc/invoke.texi (-fdump-rtl-pro_and_epilogue): Use @item
23732 instead of @itemx.
23733
23734 PR inline-asm/56405
23735 * expr.c (expand_expr_real_1) <case TARGET_MEM_REF, MEM_REF>: Don't
23736 use movmisalign or extract_bit_field for EXPAND_MEMORY modifier.
23737
23738 2013-02-20 Jan Hubicka <jh@suse.cz>
23739
23740 PR tree-optimization/56265
23741 * ipa-prop.c (ipa_make_edge_direct_to_target): Fixup callgraph
23742 when target is referenced for first time.
23743
23744 2013-02-20 Richard Biener <rguenther@suse.de>
23745
23746 * tree-call-cdce.c (tree_call_cdce): Do not remove unused locals.
23747 * tree-ssa-forwprop.c (ssa_forward_propagate_and_combine): Likewise.
23748 * tree-ssa-dce.c (perform_tree_ssa_dce): Likewise.
23749 * tree-ssa-copyrename.c (copy_rename_partition_coalesce): Do
23750 not return anything.
23751 (rename_ssa_copies): Do not remove unused locals.
23752 * tree-ssa-ccp.c (do_ssa_ccp): Likewise.
23753 * tree-ssanames.c (pass_release_ssa_names): Remove unused locals first.
23754 * passes.c (execute_function_todo): Do not schedule unused locals
23755 removal if cleanup_tree_cfg did something.
23756 * tree-ssa-live.c (remove_unused_locals): Dump statistics
23757 about the number of removed locals.
23758
23759 2013-02-20 Richard Biener <rguenther@suse.de>
23760
23761 PR tree-optimization/56398
23762 * tree-vect-loop-manip.c (adjust_debug_stmts): Skip SSA default defs.
23763
23764 2013-02-20 Martin Jambor <mjambor@suse.cz>
23765
23766 PR tree-optimization/55334
23767 * ipa-cp.c (initialize_node_lattices): Disable IPA-CP through and to
23768 restricted pointers to arrays.
23769
23770 2013-02-20 Richard Biener <rguenther@suse.de>
23771 Jakub Jelinek <jakub@redhat.com>
23772
23773 PR tree-optimization/56396
23774 * tree-ssa-ccp.c (n_const_val): New static variable.
23775 (get_value): Return NULL for SSA names we don't have a lattice
23776 entry for.
23777 (ccp_initialize): Initialize n_const_val.
23778 * tree-ssa-copy.c (n_copy_of): New static variable.
23779 (init_copy_prop): Initialize n_copy_of.
23780 (get_value): Return NULL_TREE for SSA names we don't have a
23781 lattice entry for.
23782
23783 2013-02-20 Martin Jambor <mjambor@suse.cz>
23784
23785 * ipa-cp.c (initialize_node_lattices): Fix dumping condition.
23786
23787 2013-02-20 Richard Biener <rguenther@suse.de>
23788
23789 * genpreds.c (write_lookup_constraint): Do not compare first
23790 letter of the constraint again.
23791
23792 2013-02-20 Richard Biener <rguenther@suse.de>
23793
23794 * tree-ssa-loop-ivopts.c (alloc_use_cost_map): Use bitmap_count_bits
23795 and ceil_log2.
23796 (get_use_iv_cost): Terminate hashtable walk when coming across
23797 an empty entry.
23798
23799 2013-02-20 Igor Zamyatin <igor.zamyatin@intel.com>
23800
23801 * config/i386/i386.c (initial_ix86_tune_features): Turn on fp
23802 reassociation for avx2 targets.
23803
23804 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23805
23806 * config/microblaze/microblaze.c: microblaze_has_clz = 0
23807 Add version check for v8.10.a to enable microblaze_has_clz
23808 * config/microblaze/microblaze.h: Add TARGET_HAS_CLZ as combined
23809 version and TARGET_PATTERN_COMPARE check
23810 * config/microblaze/microblaze.md: New clzsi2 instruction
23811
23812 2012-02-19 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23813
23814 * config/microblaze/microblaze.md (call_value_intern): Check symbol is
23815 function before branching.
23816
23817 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
23818
23819 * sel-sched-dump.c (dump_insn_rtx_flags): Explicitly set
23820 DUMP_INSN_RTX_UID.
23821 (dump_insn_rtx_1): Pass PATTERN (insn) to str_pattern_slim.
23822
23823 2012-02-19 Andrey Belevantsev <abel@ispras.ru>
23824
23825 PR middle-end/55889
23826 * sel-sched.c: Include ira.h.
23827 (implicit_clobber_conflict_p): New function.
23828 (moveup_expr): Use it.
23829 * Makefile.in (sel-sched.o): Depend on ira.h.
23830
23831 2013-02-19 Richard Biener <rguenther@suse.de>
23832
23833 PR tree-optimization/56384
23834 * tree-ssa-sccvn.h (struct vn_phi_s): Add type member.
23835 (vn_hash_type): Split out from ...
23836 (vn_hash_constant_with_type): ... here.
23837 * tree-ssa-sccvn.c (vn_phi_compute_hash): Use vn_hash_type.
23838 (vn_phi_eq): Compare types from vn_phi_s structure.
23839 (vn_phi_lookup): Populate vn_phi_s type.
23840 (vn_phi_insert): Likewise.
23841
23842 2013-02-19 Jakub Jelinek <jakub@redhat.com>
23843
23844 PR tree-optimization/56350
23845 * tree-vect-loop.c (vectorizable_reduction): If orig_stmt, return false
23846 if haven't found reduction or nested cycle operand, rather than
23847 asserting we must find it.
23848
23849 PR tree-optimization/56381
23850 * tree-ssa-pre.c (create_expression_by_pieces): Fix up last argument
23851 to fold_build3.
23852
23853 2013-02-18 Aldy Hernandez <aldyh@redhat.com>
23854 Jakub Jelinek <jakub@redhat.com>
23855
23856 PR target/52555
23857 * genopinit.c (raw_optab_handler): Use this_fn_optabs.
23858 (swap_optab_enable): Same.
23859 (init_all_optabs): Use argument instead of global.
23860 * tree.h (struct tree_optimization_option): New field target_optabs.
23861 * expr.h (init_all_optabs): Add argument to prototype.
23862 (TREE_OPTIMIZATION_OPTABS): New.
23863 (save_optabs_if_changed): Protoize.
23864 * optabs.h: Declare this_fn_optabs.
23865 * optabs.c (save_optabs_if_changed): New.
23866 Declare this_fn_optabs.
23867 (init_optabs): Add argument to init_all_optabs() call.
23868 * function.c (invoke_set_current_function_hook): Handle per
23869 function optabs.
23870 * function.h (struct function): New field optabs.
23871 * config/mips/mips.c (mips_set_mips16_mode): Handle when
23872 optimization_current_node has changed.
23873 * target-globals.h (save_target_globals_default_opts): Protoize.
23874 * target-globals.c (save_target_globals_default_opts): New.
23875
23876 2013-02-18 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
23877
23878 PR target/56347
23879 * config/pa/pa.c (pa_conditional_register_usage): On HP-UX, mark
23880 registers %fr12 and %fr12R as call used.
23881
23882 PR target/56214
23883 * config/pa/predicates.md (base14_operand): Except for BLKmode, QImode
23884 and HImode, require all displacements to be an integer multiple of
23885 their mode size.
23886 * config/pa/pa.c (pa_legitimate_address_p): For REG+BASE addresses,
23887 only allow QImode and HImode when reload is in progress and strict is
23888 true. Likewise for symbolic addresses. Use base14_operand to check
23889 displacements in REG+BASE addresses.
23890
23891 2013-02-18 Richard Biener <rguenther@suse.de>
23892
23893 PR tree-optimization/56366
23894 * tree-vect-loop.c (get_initial_def_for_induction): Properly
23895 handle sign-conversion of outer-loop initial induction value.
23896
23897 2013-02-18 Richard Biener <rguenther@suse.de>
23898
23899 PR middle-end/56349
23900 * cfghooks.c (merge_blocks): If we merge a latch into another
23901 block adjust references to it.
23902 * cfgloop.c (flow_loops_find): Reset latch before recomputing it.
23903 (verify_loop_structure): Verify that a recorded latch is in fact
23904 a latch.
23905
23906 2013-02-18 Richard Biener <rguenther@suse.de>
23907
23908 PR tree-optimization/56321
23909 * tree-ssa-reassoc.c (propagate_op_to_single_use): Properly
23910 order SSA name release and virtual operand unlinking.
23911
23912 2013-02-17 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23913
23914 * config/microblaze/microblaze.md (save_stack_block): Define.
23915 (restore_stack_block): Likewise.
23916
23917 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23918
23919 * config/microblaze/linux.h (TARGET_SUPPORTS_PIC): Define as 1.
23920 * config/microblaze/microblaze.h (TARGET_SUPPORTS_PIC): Define as 1.
23921 * config/microblaze/microblaze.c (microblaze_option_override):
23922 Bail out early for PIC modes when target does not support PIC.
23923
23924 2013-02-16 Edgar E. Iglesias <edgar.iglesias@gmail.com>
23925
23926 * config/microblaze/microblaze.c (microblaze_asm_trampoline_template):
23927 Replace with a microblaze version.
23928 (microblaze_trampoline_init): Adapt for microblaze.
23929 * config/microblaze/microblaze.h (TRAMPOLINE_SIZE): Adapt for
23930 microblaze.
23931
23932 2013-02-16 Jakub Jelinek <jakub@redhat.com>
23933 Dodji Seketeli <dodji@redhat.com>
23934
23935 PR asan/56330
23936 * asan.c (get_mem_refs_of_builtin_call): White space and style cleanup.
23937 (instrument_mem_region_access): Do not forget to always put
23938 instrumentation of the of 'base' and 'base + len' in a "if (len !=
23939 0) statement, even for cases where either 'base' or 'base + len'
23940 are not instrumented -- because they have been previously
23941 instrumented. Simplify the logic by putting all the statements
23942 instrument 'base + len' inside a sequence, and then insert that
23943 sequence right before the current insertion point. Then, to
23944 instrument 'base + len', just get an iterator on that statement.
23945 And do not forget to update the pointer to iterator the function
23946 received as argument.
23947
23948 2013-02-15 Vladimir Makarov <vmakarov@redhat.com>
23949
23950 PR rtl-optimization/56348
23951 * lra-assigns.c (reload_pseudo_compare_func): Prefer bigger pseudos.
23952
23953 2013-02-15 Steven Bosscher <steven@gcc.gnu.org>
23954
23955 * graph.c (start_graph_dump): Print dumpfile base as digraph label.
23956 (clean_graph_dump_file): Pass base to start_graph_dump.
23957
23958 2013-02-14 Richard Henderson <rth@redhat.com>
23959
23960 PR target/55941
23961 * lower-subreg.c (simple_move): Check dest mode instead of src mode.
23962
23963 2013-02-14 Steven Bosscher <steven@gcc.gnu.org>
23964
23965 * collect2-aix.h: Define F_LOADONLY.
23966
23967 2013-02-14 Richard Biener <rguenther@suse.de>
23968
23969 PR lto/50494
23970 * varasm.c (output_constant_def_1): Get the decl representing
23971 the constant as argument.
23972 (output_constant_def): Wrap output_constant_def_1.
23973 (make_decl_rtl): Use output_constant_def_1 with the decl
23974 representing the constant.
23975 (build_constant_desc): Optionally re-use a decl already
23976 representing the constant.
23977 (tree_output_constant_def): Adjust.
23978
23979 2013-02-14 Dodji Seketeli <dodji@redhat.com>
23980
23981 Fix an asan crash
23982 * asan.c (instrument_builtin_call): Really put the length of the
23983 second source argument into src1_len.
23984
23985 2013-02-13 Jakub Jelinek <jakub@redhat.com>
23986
23987 * asan.c (create_cond_insert_point): Add create_then_fallthru_edge
23988 argument. If it is false, don't create edge from then_bb to
23989 fallthru_bb.
23990 (insert_if_then_before_iter): Pass true to it.
23991 (build_check_stmt): Pass false to it.
23992 (transform_statements): Flush hash table only on extended basic
23993 block boundaries, rather than at the beginning of every bb.
23994 Don't flush hash table on nonfreeing_call_p calls.
23995 * tree-flow.h (nonfreeing_call_p): New prototype.
23996 * tree-ssa-phiopt.c (nonfreeing_call_p): No longer static.
23997
23998 2013-02-13 David S. Miller <davem@davemloft.net>
23999
24000 * expmed.c (expand_shift_1): Only strip scalar integer subregs.
24001
24002 2013-02-13 Vladimir Makarov <vmakarov@redhat.com>
24003
24004 PR target/56184
24005 * ira.c (max_regno_before_ira): Move from ...
24006 (ira): ... here.
24007 (fix_reg_equiv_init): Use max_regno_before_ira instead of
24008 vec_safe_length.
24009
24010 2013-02-13 Jakub Jelinek <jakub@redhat.com>
24011
24012 * config/i386/i386.c (ix86_asan_shadow_offset): Revert last change.
24013
24014 2013-02-13 Richard Biener <rguenther@suse.de>
24015
24016 PR lto/56295
24017 * gimple-streamer-out.c (output_gimple_stmt): Undo wrapping
24018 globals in MEM_REFs.
24019
24020 2013-02-13 Richard Biener <rguenther@suse.de>
24021
24022 * loop-init.c (loop_optimizer_init): Clear loop state when
24023 re-initializing preserved loops.
24024 * loop-unswitch.c (unswitch_single_loop): Return whether
24025 we unswitched the loop. Do not verify loop state here.
24026 (unswitch_loops): When we unswitched a loop discover new loops.
24027
24028 2013-02-13 Kostya Serebryany <kcc@google.com>
24029
24030 * config/i386/i386.c: Use 0x7fff8000 as asan_shadow_offset
24031 on x86_64 linux.
24032 * sanitizer.def: Rename __asan_init to __asan_init_v1.
24033
24034 2013-02-12 Dodji Seketeli <dodji@redhat.com>
24035
24036 Avoid instrumenting duplicated memory access in the same basic block
24037 * Makefile.in (asan.o): Add new dependency on hash-table.h
24038 * asan.c (struct asan_mem_ref, struct mem_ref_hasher): New types.
24039 (asan_mem_ref_init, asan_mem_ref_get_end, get_mem_ref_hash_table)
24040 (has_stmt_been_instrumented_p, empty_mem_ref_hash_table)
24041 (free_mem_ref_resources, has_mem_ref_been_instrumented)
24042 (has_stmt_been_instrumented_p, update_mem_ref_hash_table)
24043 (get_mem_ref_of_assignment): New functions.
24044 (get_mem_refs_of_builtin_call): Extract from
24045 instrument_builtin_call and tweak a little bit to make it fit with
24046 the new signature.
24047 (instrument_builtin_call): Use the new
24048 get_mem_refs_of_builtin_call. Use gimple_call_builtin_p instead
24049 of is_gimple_builtin_call.
24050 (instrument_derefs, instrument_mem_region_access): Insert the
24051 instrumented memory reference into the hash table.
24052 (maybe_instrument_assignment): Renamed instrument_assignment into
24053 this, and change it to advance the iterator when instrumentation
24054 actually happened and return true in that case. This makes it
24055 homogeneous with maybe_instrument_assignment, and thus give a
24056 chance to callers to be more 'regular'.
24057 (transform_statements): Clear the memory reference hash table
24058 whenever we enter a new BB, when we cross a function call, or when
24059 we are done transforming statements. Use
24060 maybe_instrument_assignment instead of instrumentation. No more
24061 need to special case maybe_instrument_assignment and advance the
24062 iterator after calling it; it's now handled just like
24063 maybe_instrument_call. Update comment.
24064
24065 2013-02-13 Richard Biener <rguenther@suse.de>
24066
24067 * config/mn10300/mn10300.c (mn10300_scan_for_setlb_lcc):
24068 Fix loop discovery code.
24069
24070 2013-02-12 Vladimir Makarov <vmakarov@redhat.com>
24071
24072 PR inline-asm/56148
24073 * lra-constraints.c (process_alt_operands): Match early clobber
24074 operand with itself. Check conflicts with earlyclobber only if
24075 the operand is not reloaded. Prefer to reload conflicting operand
24076 if earlyclobber and matching operands are the same.
24077
24078 2013-02-12 Richard Biener <rguenther@suse.de>
24079
24080 PR lto/56297
24081 * lto-streamer-out.c (write_symbol): Do not output symbols
24082 for hard register variables.
24083
24084 2013-02-12 Georg-Johann Lay <avr@gjlay.de>
24085
24086 PR target/54222
24087 * config/avr/avr-dimode.md (umulsidi3, mulsidi3): New expanders.
24088 (umulsidi3_insn, mulsidi3_insn): New insns.
24089
24090 2013-02-12 Christophe Lyon <christophe.lyon@linaro.org>
24091
24092 * config/arm/arm-protos.h (struct cpu_vec_costs): New struct type.
24093 (struct tune_params): Add vec_costs field.
24094 * config/arm/arm.c (arm_builtin_vectorization_cost)
24095 (arm_add_stmt_cost): New functions.
24096 (TARGET_VECTORIZE_BUILTIN_VECTORIZATION_COST)
24097 (TARGET_VECTORIZE_ADD_STMT_COST): Define.
24098 (arm_default_vec_cost): New struct of type cpu_vec_costs.
24099 (arm_slowmul_tune, arm_fastmul_tune, arm_strongarm_tune)
24100 (arm_xscale_tune, arm_9e_tune, arm_v6t2_tune, arm_cortex_tune)
24101 (arm_cortex_a15_tune, arm_cortex_a5_tune, arm_cortex_a9_tune)
24102 (arm_v6m_tune, arm_fa726te_tune): Define new vec_costs field.
24103
24104 2013-02-12 Richard Biener <rguenther@suse.de>
24105
24106 PR lto/56295
24107 * gimple-streamer-in.c (input_gimple_stmt): Strip MEM_REFs off
24108 decls again if possible.
24109
24110 2013-02-12 Richard Biener <rguenther@suse.de>
24111
24112 PR middle-end/56288
24113 * tree-ssa.c (verify_ssa_name): Fix check, move
24114 SSA_NAME_IN_FREE_LIST check up.
24115
24116 2013-02-12 Jakub Jelinek <jakub@redhat.com>
24117 Steven Bosscher <steven@gcc.gnu.org>
24118
24119 PR rtl-optimization/56151
24120 * optabs.c (add_equal_note): Don't return 0 if target is a MEM,
24121 equal to op0 or op1, and last_insn pattern is CODE operation
24122 with MEM dest and one of the operands matches that MEM.
24123
24124 2013-02-11 Sriraman Tallam <tmsriramgoogle.com>
24125
24126 * doc/extend.texi: Document Function Multiversioning and "default"
24127 parameter string to target attribute.
24128 * config/i386/i386.c (get_builtin_code_for_version): Return 0 if
24129 target attribute parameter is "default".
24130 (ix86_compare_version_priority): Remove checks for target attribute.
24131 (ix86_mangle_function_version_assembler_name): Change error to sorry.
24132 Remove check for target attribute equal to NULL. Add assert.
24133 (ix86_generate_version_dispatcher_body): Change error to sorry.
24134
24135 2013-02-11 Iain Sandoe <iain@codesourcery.com>
24136 Jack Howarth <howarth@bromo.med.uc.edu>
24137 Patrick Marlier <patrick.marlier@gmail.com>
24138
24139 PR libitm/55693
24140 * config/darwin.h: Replace ENDFILE_SPEC with TM_DESTRUCTOR and
24141 define ENDFILE_SPEC as TM_DESTRUCTOR.
24142 * config/i386/darwin.h (ENDFILE_SPEC): Use TM_DESTRUCTOR.
24143
24144 2013-02-11 Alexander Potapenko <glider@google.com>
24145 Jack Howarth <howarth@bromo.med.uc.edu>
24146 Jakub Jelinek <jakub@redhat.com>
24147
24148 PR sanitizer/55617
24149 * config/darwin.c (cdtor_record): Rename ctor_record.
24150 (sort_cdtor_records): Rename sort_ctor_records.
24151 (finalize_dtors): New routine to sort destructors by
24152 priority before use in assemble_integer.
24153 (machopic_asm_out_destructor): Use finalize_dtors if needed.
24154
24155 2013-02-11 Uros Bizjak <ubizjak@gmail.com>
24156
24157 PR rtl-optimization/56275
24158 * simplify-rtx.c (avoid_constant_pool_reference): Check that
24159 offset is non-negative and less than cmode size before
24160 calling simplify_subreg.
24161
24162 2013-02-11 Richard Biener <rguenther@suse.de>
24163
24164 PR tree-optimization/56264
24165 * cfgloop.h (fix_loop_structure): Adjust prototype.
24166 * loop-init.c (fix_loop_structure): Return the number of
24167 newly discovered loops.
24168 * tree-cfgcleanup.c (repair_loop_structures): When new loops
24169 are discovered, do a full loop-closed SSA rewrite.
24170
24171 2013-02-11 Richard Biener <rguenther@suse.de>
24172
24173 PR tree-optimization/56273
24174 * tree-vrp.c (simplify_cond_using_ranges): Disable for the
24175 first VRP run.
24176 (check_array_ref): Fix missing newline in dumps.
24177 (search_for_addr_array): Likewise.
24178
24179 2013-02-09 David Edelsohn <dje.gcc@gmail.com>
24180
24181 * config/rs6000/aix61.h (OS_MISSING_ALTIVEC): Undefine.
24182
24183 2013-02-09 Jakub Jelinek <jakub@redhat.com>
24184
24185 PR target/56256
24186 * config/rs6000/rs6000.h (ASSEMBLER_DIALECT): Define.
24187
24188 2013-02-08 Vladimir Makarov <vmakarov@redhat.com>
24189
24190 PR rtl-optimization/56246
24191 * lra-constraints.c (simplify_operand_subreg): Try to reuse
24192 reload pseudo.
24193 * lra.c (lra): Clear lra_optional_reload_pseudos only when all
24194 constraints are satisfied.
24195
24196 2013-02-08 Jeff Law <law@redhat.com>
24197
24198 PR debug/53948
24199 * emit-rtl.c (reg_is_parm_p): New function.
24200 * regs.h (reg_is_parm_p): New prototype.
24201 * ira-conflicts.c (ira_build_conflicts): Allow parameters in
24202 callee-clobbered registers.
24203
24204 2013-02-08 Michael Meissner <meissner@linux.vnet.ibm.com>
24205
24206 PR target/56043
24207 * config/rs6000/rs6000.c (rs6000_builtin_vectorized_libmass):
24208 If there is no implicit builtin declaration, just return NULL.
24209
24210 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24211
24212 * config/i386/sse.md (FMAMODEM): New mode iterator.
24213 (fma<mode>4, fms<mode>4, fnma<mode>4, fnms<mode>4): Use FMAMODEM
24214 mode iterator. Do not use TARGET_SSE_MATH in insn constraint.
24215
24216 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24217
24218 * config/i386/gnu-user.h (TARGET_CAN_SPLIT_STACK): Define only
24219 when HAVE_GAS_CFI_PERSONALITY_DIRECTIVE is set.
24220 * config/i386/gnu-user64.h (TARGET_CAN_SPLIT_STACK): Ditto.
24221
24222 2013-02-08 Edgar E. Iglesias <edgar.iglesias@gmail.com>
24223
24224 * config.gcc (microblaze*-linux*): Add TARGET_BIG_ENDIAN_DEFAULT.
24225 (microblaze*-*-elf): Likewise.
24226 * config/microblaze/linux.h: Add -mbig-endian / -mlittle-endian to
24227 LINK_SPEC.
24228 * config/microblaze/microblaze-c.c: Add builtin defines for
24229 _LITTLE_ENDIAN and _BIG_ENDIAN.
24230 * config/microblaze/microblaze.h: Add TARGET_ENDIAN_DEFAULT and
24231 add to TARGET_DEFAULT flags.
24232 Expand ASM_SPEC and LINK_SPEC.
24233 Update BYTES_BIG_ENDIAN and WORDS_BIG_ENDIAN.
24234 * config/microblaze/microblaze.md: Update extendsidi2 and
24235 movdi_internal instructions to use low-order / high-order reg
24236 print_operands.
24237 * config/microblaze/microblaze.opt: Add mbig-endian and mlittle-endian
24238 options and inversemask / mask of LITTLE_ENDIAN.
24239 * config/microblaze/t-microblaze: Expand multilib options to
24240 include mlittle-endian (le) and update exceptions patterns.
24241
24242 2013-02-08 Jakub Jelinek <jakub@redhat.com>
24243
24244 PR rtl-optimization/56195
24245 * lra-constraints.c (get_reload_reg): Don't reuse regs
24246 if they have smaller mode than requested, if they have
24247 wider mode than requested, try to return a SUBREG.
24248
24249 PR tree-optimization/56250
24250 * fold-const.c (extract_muldiv_1) <case NEGATE_EXPR>: Don't optimize
24251 if type is unsigned and code isn't MULT_EXPR.
24252
24253 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24254
24255 PR tree-optimization/56064
24256 * fixed-value.c (fixed_from_double_int): Sign/zero extend payload
24257 bits according to mode.
24258 * fixed-value.h (fixed_from_double_int)
24259 (const_fixed_from_double_int): Adjust comments.
24260
24261 2013-02-08 Richard Biener <rguenther@suse.de>
24262
24263 PR lto/56231
24264 * lto-streamer.h (struct data_in): Remove current_file, current_line
24265 and current_col members.
24266 * lto-streamer-out.c (lto_output_location): Stream changed bits
24267 en-block for efficiency.
24268 * lto-streamer-in.c (clear_line_info): Remove.
24269 (lto_input_location): Cache current file, line and column
24270 globally via local statics. Read changed bits en-block.
24271 (input_function): Do not call clear_line_info.
24272 (lto_read_body): Likewise.
24273 (lto_input_toplevel_asms): Likewise.
24274
24275 2013-02-08 Michael Matz <matz@suse.de>
24276
24277 PR tree-optimization/52448
24278 * tree-ssa-phiopt.c (struct name_to_bb): Add phase member.
24279 (nt_call_phase): New static.
24280 (add_or_mark_expr): Only mark accesses with newer phase than any
24281 call seen.
24282 (nonfreeing_call_p): New.
24283 (nt_init_block): Update nt_call_phase, mark blocks as visited.
24284 (nt_fini_block): Keep blocks marked as visited.
24285 (get_non_trapping): Initialize nt_call_phase, and reset aux pointer.
24286
24287 2013-02-08 Richard Biener <rguenther@suse.de>
24288
24289 * ira.c (ira): Free broken dominator information.
24290
24291 2013-02-08 Uros Bizjak <ubizjak@gmail.com>
24292
24293 * config/i386/i386.c (ix86_spill_class): Use INTEGER_CLASS_P macro.
24294
24295 2013-02-08 Marek Polacek <polacek@redhat.com>
24296
24297 * cfgloop.c (verify_loop_structure): Add more checking of headers.
24298
24299 2013-02-08 Richard Biener <rguenther@suse.de>
24300
24301 PR middle-end/56181
24302 * cfgloop.h (flow_loops_find): Adjust.
24303 (bb_loop_header_p): Declare.
24304 * cfgloop.c (bb_loop_header_p): New function split out from ...
24305 (flow_loops_find): ... here. Adjust function signature,
24306 support incremental loop structure update.
24307 (verify_loop_structure): Cleanup. Verify a loop is a loop.
24308 * cfgloopmanip.c (fix_loop_structure): Move ...
24309 * loop-init.c (fix_loop_structure): ... here.
24310 (apply_loop_flags): Split out from ...
24311 (loop_optimizer_init): ... here.
24312 (fix_loop_structure): Use apply_loop_flags. Use flow_loops_find
24313 in incremental mode, only remove dead loops here.
24314
24315 2013-02-08 Georg-Johann Lay <avr@gjlay.de>
24316
24317 PR target/54222
24318 * config/avr/avr.md (unspec) <UNSPEC_ROUND>: Add.
24319 * config/avr/avr-fixed.md (ALL4QA, ALL124QA): New mode iterators.
24320 (round<mode>3, round<mode>3_const): New expanders for fixed-mode.
24321 (*round<mode>3.libgcc): New insns for fixed-modes.
24322 * config/avr/builtins.def (ABSxx): Use a non-NULL LIBNAME.
24323 (ROUNDxx, COUNTLSxx, BITSxx, xxBITS): New DEF_BUILTINs.
24324 (ROUNDFX, COUNTLSFX, ABSFX): New DEF_BUILTINs.
24325 * config/avr/stdfix.h (absFX, bitsFX, FXbits): Remove inline
24326 implementations. Define to __builtin_avr_absFX,
24327 __builtin_avr_bitsFX, __builtin_avr_FXbits, respectively.
24328 (roundFX, countlsFX): Define to __builtin_avr_roundFX,
24329 __builtin_avr_countlsFX, respectively.
24330 * config/avr/avr-c.c (target.h): Include it.
24331 (enum avr_builtin_id): New enum.
24332 (avr_resolve_overloaded_builtin): New static function.
24333 (avr_register_target_pragmas): Use it to set
24334 targetm.resolve_overloaded_builtin.
24335 * config/avr/avr.c (avr_init_builtins): Supply myriads of local
24336 tree nodes used by DEF_BUILTIN.
24337 (avr_expand_builtin) <AVR_BUILTIN_ROUNDxx>: Sanity-check them.
24338 (avr_fold_builtin) <AVR_BUILTIN_BITSxx>: Fold to VIEW_COVERT_EXPR.
24339 <AVR_BUILTIN_xxBITS>: Same.
24340
24341 2013-02-08 Richard Biener <rguenther@suse.de>
24342
24343 * cfgloop.c (verify_loop_structure): Properly handle
24344 a loop exiting to another loop header.
24345 * ira-int.h (ira_loops): Remove.
24346 * ira.c (ira_loops): Remove.
24347 (ira): Use loop_optimizer_init and loop_optimizer_finalize.
24348 (do_reload): Use loop_optimizer_finalize.
24349 * ira-build.c (create_loop_tree_nodes): Use get_loops and
24350 number_of_loops to access the loop tree.
24351 (more_one_region_p): Likewise.
24352 (finish_loop_tree_nodes): Likewise.
24353 (rebuild_regno_allocno_maps): Likewise.
24354 (mark_loops_for_removal): Likewise.
24355 (mark_all_loops_for_removal): Likewise.
24356 (remove_unnecessary_regions): Likewise.
24357 (ira_build): Likewise.
24358 * ira-emit.c (setup_entered_from_non_parent_p): Likewise.
24359
24360 2013-02-08 Richard Biener <rguenther@suse.de>
24361
24362 * Makefile.in (tree-tailcall.o): Add $(CFGLOOP_H) dependency.
24363 * ipa-pure-const.c (analyze_function): Avoid calling
24364 mark_irreducible_loops twice.
24365 * tree-tailcall.c (tree_optimize_tail_calls_1): Mark loops for fixup.
24366
24367 2013-02-07 David S. Miller <davem@davemloft.net>
24368
24369 * dwarf2out.c (based_loc_descr): Perform leaf register remapping
24370 on 'reg'.
24371 * var-tracking.c (vt_add_function_parameter): Test the presence of
24372 HAVE_window_save properly and do not remap argument registers when
24373 we have a leaf function.
24374
24375 2013-02-07 Uros Bizjak <ubizjak@gmail.com>
24376
24377 PR bootstrap/56227
24378 * ggc-page.c (ggc_print_statistics): Use HOST_LONG_LONG_FORMAT
24379 instead of "ll".
24380 * config/i386/i386.c (ix86_print_operand): Ditto.
24381
24382 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24383
24384 * lra-constraints.c (process_alt_operands): Fix recently added comment.
24385
24386 2013-02-07 Vladimir Makarov <vmakarov@redhat.com>
24387
24388 PR rtl-optimization/56225
24389 * lra-constraints.c (process_alt_operands): Check that reload hard
24390 reg can hold value for strict_low_part.
24391
24392 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24393
24394 PR debug/56154
24395 * dwarf2out.c (dwarf2_debug_hooks): Set end_function hook to
24396 dwarf2out_end_function.
24397 (in_first_function_p, maybe_at_text_label_p,
24398 first_loclabel_num_not_at_text_label): New variables.
24399 (dwarf2out_var_location): In the first function find out
24400 lowest loclabel_num N where .LVLN is known not to be equal to .Ltext0.
24401 (find_empty_loc_ranges_at_text_label, dwarf2out_end_function): New
24402 functions.
24403
24404 2013-02-07 Eric Botcazou <ebotcazou@adacore.com>
24405
24406 PR rtl-optimization/56178
24407 * cse.c (cse_insn): Do not create a REG_EQUAL note if the source is a
24408 SUBREG of a register. Tidy up related block of code.
24409 * fwprop.c (forward_propagate_and_simplify): Do not create a REG_EQUAL
24410 note if the source is a register or a SUBREG of a register.
24411
24412 2013-02-07 Jakub Jelinek <jakub@redhat.com>
24413
24414 PR target/56228
24415 * config/rs6000/rs6000.md (ptrm): New mode attr.
24416 (call_indirect_aix<ptrsize>, call_indirect_aix<ptrsize>_nor11,
24417 call_value_indirect_aix<pttrsize>,
24418 call_value_indirect_aix<pttrsize>_nor11): Use <ptrm> instead of
24419 m in constraints.
24420
24421 2013-02-07 Michael Haubenwallner <michael.haubenwallner@salomon.at>
24422
24423 * collect2.c (main): Set aix64_flag for -G and -bsvr4 too, disable
24424 if -bnortl. Convert to strcmp and strncmp.
24425
24426 2013-02-07 Alan Modra <amodra@gmail.com>
24427
24428 PR target/54009
24429 * config/rs6000/rs6000.c (mem_operand_gpr): Check that LO_SUM
24430 addresses won't wrap when offsetting.
24431 (rs6000_secondary_reload): Provide secondary reloads needed for
24432 wrapping LO_SUM addresses.
24433
24434 2013-02-06 Thomas Schwinge <thomas@codesourcery.com>
24435
24436 * config/gnu.h (GNU_USER_TARGET_OS_CPP_BUILTINS): Never define
24437 MACH, just __MACH__.
24438
24439 2013-02-06 Richard Biener <rguenther@suse.de>
24440
24441 * tracer.c (tracer): Mark loops with LOOPS_NEED_FIXUP
24442 instead of calling fix_loop_structure.
24443
24444 2013-02-06 Jakub Jelinek <jakub@redhat.com>
24445
24446 PR middle-end/56217
24447 * omp-low.c (use_pointer_for_field): Return false if
24448 lower_send_shared_vars doesn't generate any copy-out code.
24449
24450 2013-02-06 Tom de Vries <tom@codesourcery.com>
24451
24452 PR rtl-optimization/56131
24453 * cfgrtl.c (delete_insn): Use NOTE_BASIC_BLOCK instead of BLOCK_FOR_INSN
24454 to get the bb of a NOTE_INSN_BASIC_BLOCK. Handle the case that the bb
24455 of the label is NULL. Add comment.
24456
24457 2013-02-05 Jakub Jelinek <jakub@redhat.com>
24458
24459 * tree.h (struct tree_decl_with_vis): Remove thread_local field.
24460
24461 PR sanitizer/55374
24462 * config/gnu-user.h (LIBTSAN_EARLY_SPEC): Define.
24463 (STATIC_LIBTSAN_LIBS): Likewise.
24464 * gcc.c (ADD_STATIC_LIBTSAN_LIBS, LIBTSAN_EARLY_SPEC): Define.
24465 (LIBTSAN_SPEC): Add ADD_STATIC_LIBTSAN_LIBS, if LIBTSAN_EARLY_SPEC
24466 is defined, don't add anything else beyond that.
24467 (SANITIZER_EARLY_SPEC, SANITIZER_SPEC): Define.
24468 (LINK_COMMAND_SPEC): Use them.
24469
24470 PR tree-optimization/56205
24471 * tree-stdarg.c (check_all_va_list_escapes): Return true if
24472 there are any PHI nodes that set non-va_list_escape_vars SSA_NAME
24473 and some va_list_escape_vars SSA_NAME appears in some PHI argument.
24474
24475 2013-02-05 Richard Biener <rguenther@suse.de>
24476
24477 PR tree-optimization/53342
24478 PR tree-optimization/53185
24479 * tree-vectorizer.h (vect_check_strided_load): Remove.
24480 * tree-vect-data-refs.c (vect_enhance_data_refs_alignment): Do
24481 not disallow peeling for vectorized strided loads.
24482 (vect_check_strided_load): Make static and simplify.
24483 (vect_analyze_data_refs): Adjust.
24484 * tree-vect-stmts.c (vectorizable_load): Handle peeled loops
24485 correctly when vectorizing strided loads.
24486
24487 2013-02-05 Richard Biener <rguenther@suse.de>
24488
24489 * doc/install.texi: Refer to ISL, not PPL.
24490
24491 2013-02-05 Jan Hubicka <jh@suse.cz>
24492
24493 PR tree-optimization/55789
24494 * params.def (PARAM_EARLY_INLINER_MAX_ITERATIONS): Drop to 1.
24495
24496 2013-02-05 Jan Hubicka <jh@suse.cz>
24497
24498 PR tree-optimization/55789
24499 * cgraphclones.c (cgraph_remove_node_and_inline_clones): Remove
24500 the dead call anyway.
24501
24502 2013-02-05 Eric Botcazou <ebotcazou@adacore.com>
24503
24504 PR sanitizer/55374
24505 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Add missing guard.
24506
24507 2013-02-04 Alexander Potapenko <glider@google.com>
24508 Jack Howarth <howarth@bromo.med.uc.edu>
24509 Jakub Jelinek <jakub@redhat.com>
24510
24511 PR sanitizer/55617
24512 * config/darwin.c (sort_ctor_records): Stabilized qsort
24513 on constructor priority by using original position.
24514 (finalize_ctors): New routine to sort constructors by
24515 priority before use in assemble_integer.
24516 (machopic_asm_out_constructor): Use finalize_ctors if needed.
24517
24518 2013-02-04 Jakub Jelinek <jakub@redhat.com>
24519
24520 PR libstdc++/54314
24521 * config/i386/winnt.c (i386_pe_assemble_visibility): Don't warn
24522 about visibility on artificial decls.
24523 * config/sol2.c (solaris_assemble_visibility): Likewise.
24524
24525 2013-02-04 Kai Tietz <ktietz@redhat.com>
24526
24527 PR target/56186
24528 * config/i386/i386.c (function_value_ms_64): Add additional valtype
24529 argument and improve checking of return-argument types for 16-byte
24530 modes.
24531 (ix86_function_value_1): Add additional valtype argument on call
24532 of function_value_64.
24533 (return_in_memory_ms_64): Sync 16-byte sized mode handling with
24534 handling infunction_value_64 function.
24535
24536 2013-02-04 Matthew Gretton-Dann <matthew.gretton-dann@linaro.org>
24537
24538 * reload.c (subst_reloads): Fix DEBUG_RELOAD build issue.
24539
24540 2013-02-04 Richard Biener <rguenther@suse.de>
24541
24542 PR tree-optimization/56188
24543 * tree-ssa-structalias.c (label_visit): Consider case with
24544 initially non-empty points-to set.
24545 (perform_var_substitution): Dump node mapping and clean up.
24546
24547 2013-02-04 Richard Guenther <rguenther@suse.de>
24548
24549 PR lto/56168
24550 * lto-symtab.c (lto_symtab_merge_decls_1): Make non-builtin
24551 node prevail as last resort.
24552 (lto_symtab_merge_decls): Remove guard on LTRANS here.
24553 (lto_symtab_prevailing_decl): Builtins are their own prevailing decl.
24554
24555 2013-02-04 Richard Biener <rguenther@suse.de>
24556
24557 PR tree-optimization/56113
24558 * tree-ssa-structalias.c (equiv_class_lookup, equiv_class_add):
24559 Merge into ...
24560 (equiv_class_lookup_or_add): ... this.
24561 (label_visit): Adjust and fix error in previous patch.
24562 (perform_var_substitution): Adjust.
24563
24564 2013-02-03 Oleg Endo <olegendo@gcc.gnu.org>
24565
24566 * config/sh/divtab.c: Fix formatting and comments throughout the file.
24567 * config/sh/sh4-300.md: Likewise.
24568 * config/sh/sh4a.md: Likewise.
24569 * config/sh/constraints.md: Likewise.
24570 * config/sh/sh.md: Likewise.
24571 * config/sh/netbsd-elf.h: Likewise.
24572 * config/sh/predicates.md: Likewise.
24573 * config/sh/sh-protos.h: Likewise.
24574 * config/sh/ushmedia.h: Likewise.
24575 * config/sh/linux.h: Likewise.
24576 * config/sh/sh.c: Likewise.
24577 * config/sh/superh.h: Likewise.
24578 * config/sh/elf.h: Likewise.
24579 * config/sh/sh4.md: Likewise.
24580 * config/sh/sh.h: Likewise.
24581
24582 2013-02-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
24583
24584 * config/pa/constraints.md: Adjust unused letters. Change "T"
24585 constraint to match_test floating_point_store_memory_operand().
24586 * config/pa/predicates.md (reg_plus_base_memory_operand): New.
24587 (base14_operand): New.
24588 (floating_point_store_memory_operand): New.
24589 (integer_store_memory_operand): Revise to use base14_operand and
24590 reg_plus_base_memory_operand.
24591 (move_dest_operand): Allow symbolic_memory_operands.
24592 (symbolic_memory_operand): Check for LO_SOM.
24593 (symbolic_operand): Change default case to break.
24594 * config/pa/pa.md: Remove unamed DFmode and SFmode patterns to force
24595 CONST_DOUBLE values to be reloaded by putting them into memory when
24596 the destination is a floating point register.
24597 (movdf): Remove code to handle CONST_DOUBLE.
24598 (movsf): Likewise.
24599 (reload_indf_r1): New.
24600 (reload_insf_r1): New.
24601 Consistently use "Q" and "T" constraints with integer and floating
24602 point move instructions, respectively.
24603 (movdi): Remove FAIL.
24604 Change predicate for source operand unamed DImode move from
24605 general_operand to move_src_operand.
24606 (umulsidi3): Change predicate for destination operand to
24607 register_operand.
24608 Likewise for similar unamed patterns.
24609 * config/pa/pa-protos.h (pa_legitimize_reload_address): Declare.
24610 * config/pa/pa.c (pa_symbolic_expression_p): Remove extra parenthesis.
24611 (hppa_legitimize_address): Simplify mask calculation.
24612 (pa_emit_move_sequence): Revised handling of secondary reloads from
24613 REG+D addresses for floating point loads and stores. Directly handle
24614 loading CONST0_RTX (mode) to a floating point register.
24615 (pa_secondary_reload): Handle reloading DF and SFmode constant values
24616 to floating point registers. Don't restrict secondary reloads to
24617 floating point registers to integer modes. Revise some comments and
24618 cleanup some code.
24619 (TARGET_LEGITIMATE_ADDRESS_P): Define.
24620 (pa_legitimate_address_p): New.
24621 (pa_legitimize_reload_address): New.
24622 * config/pa/pa.h (STRICT_REG_OK_FOR_INDEX_P): New.
24623 (STRICT_REG_OK_FOR_BASE_P): New.
24624 (GO_IF_LEGITIMATE_ADDRESS): Delete. Update some related comments.
24625 (LEGITIMIZE_RELOAD_ADDRESS): Revise to use pa_legitimize_reload_address.
24626
24627 2013-02-03 David Edelsohn <dje.gcc@gmail.com>
24628 Andrew Dixie <andrewd@gentrack.com>
24629
24630 * collect2.c (GCC_CHECK_HDR): Do not scan objects with F_LOADONLY
24631 flag set.
24632
24633 2013-02-03 Richard Sandiford <rdsandiford@googlemail.com>
24634
24635 * expmed.c (extract_bit_field_1): Pass the full width of the
24636 structure to get_best_reg_extraction_insn.
24637
24638 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24639
24640 PR target/54601
24641 * configure.ac (use_cxa_atexit): Add AIX.
24642 * configure: Regenerate.
24643
24644 * config/rs6000/aix61.h (STARTFILE_SPEC): Add crtcxa.o.
24645
24646 2013-02-01 Jakub Jelinek <jakub@redhat.com>
24647
24648 PR debug/54793
24649 * final.c (need_profile_function): New variable.
24650 (final_start_function): Drop ATTRIBUTE_UNUSED from first argument.
24651 If first of NOTE_INSN_BASIC_BLOCK or NOTE_INSN_FUNCTION_BEG
24652 is only preceeded by NOTE_INSN_VAR_LOCATION or NOTE_INSN_DELETED
24653 notes, targetm.asm_out.function_prologue doesn't emit anything,
24654 HAVE_prologue and profiler should be emitted before prologue,
24655 set need_profile_function instead of emitting it.
24656 (final_scan_insn): If need_profile_function, emit
24657 profile_function on the first NOTE_INSN_BASIC_BLOCK or
24658 NOTE_INSN_FUNCTION_BEG note.
24659
24660 2013-02-01 Richard Henderson <rth@redhat.com>
24661
24662 * config/rs6000/rs6000.md (smulditi3): New.
24663 (umulditi3): New.
24664
24665 * config/alpha/alpha.md (umulditi3): New.
24666
24667 2013-02-01 David Edelsohn <dje.gcc@gmail.com>
24668
24669 * config/rs6000/xcoff.h (ASM_OUTPUT_ALIGNED_COMMON): Use floor_log2.
24670 (ASM_OUTPUT_ALIGNED_LOCAL): New.
24671
24672 2013-02-01 Richard Biener <rguenther@suse.de>
24673
24674 PR tree-optimization/56113
24675 * tree-ssa-structalias.c (label_visit): Reduce work for
24676 single-predecessor nodes.
24677
24678 2013-02-01 Eric Botcazou <ebotcazou@adacore.com>
24679
24680 * fold-const.c (make_range_step) <TRUTH_NOT_EXPR>: Bail out if the
24681 range isn't testing for zero.
24682
24683 2013-01-31 Steven Bosscher <steven@gcc.gnu.org>
24684
24685 PR middle-end/56113
24686 * fwprop.c (fwprop_init): Set up loops without CFG modifications.
24687
24688 2013-01-31 Hiroyuki Ono <hiroyuki.ono.jc@renesas.com>
24689 Nick Clifton <nickc@redhat.com>
24690
24691 * config/v850/constraints.md (Q): Define as a memory constraint.
24692 * config/v850/predicates.md (label_ref_operand): New predicate.
24693 (e3v5_shift_operand): New predicate.
24694 (ior_operator): New predicate.
24695 * config/v850/t-v850: Add e3v5 multilib.
24696 * config/v850/v850-protos.h (v850_adjust_insn_length): Prototype.
24697 (v850_gen_movdi): Prototype.
24698 * config/v850/v850.c: Add support for e3v5 architecture.
24699 Rename all uses of TARGET_V850E || TARGET_V850E2_ALL to
24700 TARGET_V850E_UP.
24701 (construct_save_jarl): Add e3v5 long JARL support.
24702 (v850_adjust_insn_length): New function. Adjust length of call
24703 insns when using e3v5 instructions.
24704 (v850_gen_movdi): New function: Generate instructions to move a
24705 DImode value.
24706 * config/v850/v850.h (TARGET_CPU_v850e3v5): Define.
24707 (CPP_SPEC): Define __v850e3v5__ as appropriate.
24708 (TARGET_USE_FPU): Enable for e3v5.
24709 (CONST_OK_FOR_W): New macro.
24710 (ADJUST_INSN_LENGTH): Define.
24711 * config/v850/v850.md (UNSPEC_LOOP): Define.
24712 (attr cpu): Add v850e3v5.
24713 Rename all uses of TARGET_V850E2 to TARGET_V850E2V3_UP.
24714 (movdi): New pattern.
24715 (movdi_internal): New pattern.
24716 (cbranchsf4): Conditionalize on TARGET_USE_FPU.
24717 (cbranchdf4): Conditionalize on TARGET_USE_FPU.
24718 (cstoresf4): Likewise.
24719 (cstoredf4): Likewise.
24720 (insv): New pattern.
24721 (rotlso3_a): New pattern.
24722 (rotlsi3_b): New pattern
24723 (rotlsi3_v850e3v5): New pattern.
24724 (doloop_begin): New pattern.
24725 (fix_loop_counter): New pattern.
24726 (doloop_end): New pattern.
24727 (branch_normal): Add e3v5 long branch support.
24728 (branch_invert): Likewise.
24729 (branch_z_normal): Likewise.
24730 (branch_z_invert): Likewise.
24731 (branch_nz_normal): Likewise.
24732 (branch_nz_invert): Likewise.
24733 (call_internal_short): Add e3v5 register-indirect JARL support.
24734 (call_internal_long): Likewise.
24735 (call_value_internal_short): Likewise.
24736 (call_value_internal_long): Likewise.
24737 * config/v850/v850.opt (mv850e3v5, mv850e2v4): New options.
24738 (mloop): New option.
24739 * config.gcc: Add support for configuring v840e3v5 target.
24740 * doc/invoke.texi: Document new v850 specific command line options.
24741
24742 2013-01-31 Paul Koning <ni1d@arrl.net>
24743
24744 PR debug/55059
24745 PR debug/54508
24746 * dwarf2out.c (prune_unused_types_mark): Mark all of parent's
24747 children if parent is a class.
24748 (prune_unused_types_prune): Don't add DW_AT_declaration.
24749
24750 2013-01-31 Richard Biener <rguenther@suse.de>
24751
24752 PR tree-optimization/56157
24753 * tree-vect-slp.c (vect_get_slp_defs): More thoroughly try to
24754 match up operand with SLP child.
24755
24756 2013-01-31 Jason Merrill <jason@redhat.com>
24757
24758 PR debug/54410
24759 * dwarf2out.c (gen_struct_or_union_type_die): Always schedule template
24760 parameters the first time.
24761 (gen_scheduled_generic_parms_dies): Check completeness here.
24762
24763 2013-01-31 Richard Biener <rguenther@suse.de>
24764
24765 PR middle-end/53073
24766 * common.opt (faggressive-loop-optimizations): New flag,
24767 enabled by default.
24768 * doc/invoke.texi (faggressive-loop-optimizations): Document.
24769 * tree-ssa-loop-niter.c (estimate_numbers_of_iterations_loop): Guard
24770 infer_loop_bounds_from_undefined by it.
24771
24772 2013-01-31 Richard Biener <rguenther@suse.de>
24773
24774 PR tree-optimization/56150
24775 * tree-ssa-loop-manip.c (find_uses_to_rename_stmt): Do not
24776 visit virtual operands.
24777 (find_uses_to_rename_bb): Likewise.
24778
24779 2013-01-31 Richard Biener <rguenther@suse.de>
24780
24781 PR tree-optimization/56150
24782 * tree-ssa-tail-merge.c (gimple_equal_p): Properly handle
24783 mixed store non-store stmts.
24784
24785 2013-01-30 Jakub Jelinek <jakub@redhat.com>
24786
24787 PR sanitizer/55374
24788 * gcc.c (LIBASAN_SPEC): Define just to ADD_STATIC_LIBASAN_LIBS if
24789 LIBASAN_EARLY_SPEC is defined.
24790 (LIBASAN_EARLY_SPEC): Define to empty string if not already defined.
24791 (LINK_COMMAND_SPEC): Add LIBASAN_EARLY_SPEC for -fsanitize=address,
24792 before %o.
24793 * config/gnu-user.h (LIBASAN_EARLY_SPEC): Define.
24794
24795 PR c++/55742
24796 * config/i386/i386.c (ix86_valid_target_attribute_inner_p): Diagnose
24797 invalid args instead of ICEing on it.
24798 (ix86_valid_target_attribute_tree): Return error_mark_node if
24799 ix86_valid_target_attribute_inner_p failed.
24800 (ix86_valid_target_attribute_p): Return false only if
24801 ix86_valid_target_attribute_tree returned error_mark_node. Allow
24802 target("default") attribute.
24803 (sorted_attr_string): Change argument from const char * to tree,
24804 merge in all target attribute arguments rather than just one.
24805 Formatting fix. Use XNEWVEC instead of xmalloc and XDELETEVEC
24806 instead of free. Avoid using strcat.
24807 (ix86_mangle_function_version_assembler_name): Mangle
24808 target("default") as if no target attribute is present. Adjust
24809 sorted_attr_string caller. Avoid leaking memory. Use XNEWVEC
24810 instead of xmalloc and XDELETEVEC instead of free.
24811 (ix86_function_versions): Don't return true if one of the decls
24812 doesn't have target attribute. If they don't and one of the decls
24813 is DECL_FUNCTION_VERSIONED, report an error. Adjust
24814 sorted_attr_string caller. Use XDELETEVEC instead of free.
24815 (ix86_supports_function_versions): Remove.
24816 (make_name): Fix up formatting.
24817 (make_dispatcher_decl): Remove resolver_name and its initialization.
24818 Avoid leaking memory.
24819 (is_function_default_version): Return true if there is
24820 target("default") attribute rather than no target attribute at all.
24821 (make_resolver_func): Avoid leaking memory.
24822 (ix86_generate_version_dispatcher_body): Likewise.
24823 (TARGET_OPTION_SUPPORTS_FUNCTION_VERSIONS): Remove.
24824 * target.def (supports_function_versions): Remove.
24825 * doc/tm.texi.in (SUPPORTS_FUNCTION_VERSIONS): Remove.
24826 * doc/tm.texi: Regenerated.
24827
24828 2013-01-30 Vladimir Makarov <vmakarov@redhat.com>
24829
24830 PR rtl-optimization/56144
24831 * lra-constraints.c (get_reload_reg): Don't reuse reload pseudo
24832 for values with side effects.
24833
24834 2013-01-30 Richard Biener <rguenther@suse.de>
24835
24836 * sparseset.h (sparseset_bit_p): Use gcc_checking_assert.
24837 (sparseset_pop): Likewise.
24838 * cfganal.c (compute_idf): Likewise. Increase work-stack size
24839 to be able to use quick_push in the worker loop.
24840
24841 2013-01-30 Marek Polacek <polacek@redhat.com>
24842
24843 * cfgcleanup.c (cleanup_cfg): Don't mark affected BBs.
24844
24845 2013-01-30 Richard Biener <rguenther@suse.de>
24846
24847 PR lto/56147
24848 * lto-symtab.c (lto_symtab_merge_decls_1): Guard DECL_BUILT_IN check.
24849
24850 2013-01-30 Georg-Johann Lay <avr@gjlay.de>
24851
24852 PR tree-optimization/56064
24853 * fixed-value.c (fixed_from_double_int): New function.
24854 * fixed-value.h (fixed_from_double_int): New prototype.
24855 (const_fixed_from_double_int): New static inline function.
24856 * fold-const.c (native_interpret_fixed): New static function.
24857 (native_interpret_expr) <FIXED_POINT_TYPE>: Use it.
24858 (can_native_interpret_type_p) <FIXED_POINT_TYPE>: Return true.
24859 (native_encode_fixed): New static function.
24860 (native_encode_expr) <FIXED_CST>: Use it.
24861 (native_interpret_int): Move double_int worker code to...
24862 * double-int.c (double_int::from_buffer): ...this new static method.
24863 * double-int.h (double_int::from_buffer): Prototype it.
24864
24865 2013-01-30 Richard Biener <rguenther@suse.de>
24866
24867 * tree-ssa-structalias.c (final_solutions, final_solutions_obstack):
24868 New pointer-map and obstack.
24869 (init_alias_vars): Allocate pointer-map and obstack.
24870 (delete_points_to_sets): Free them.
24871 (find_what_var_points_to): Cache result.
24872 (find_what_p_points_to): Adjust for changed interface of
24873 find_what_var_points_to.
24874 (compute_points_to_sets): Likewise.
24875 (ipa_pta_execute): Likewise.
24876
24877 2013-01-30 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
24878
24879 * configure.ac (HAVE_AS_SPARC_NOBITS): New test.
24880 * configure: Regenerate.
24881 * config.in: Regenerate.
24882 * config/sparc/sparc.c (sparc_solaris_elf_asm_named_section): Emit
24883 #nobits/#progbits if supported.
24884
24885 2013-01-29 Oleg Endo <olegendo@gcc.gnu.org>
24886
24887 PR target/56121
24888 * config/sh/sh.md (bclr_m2a, bset_m2a, bst_m2a, bld_m2a, bldsign_m2a,
24889 bld_reg, *bld_regqi, band_m2a, bandreg_m2a, bor_m2a, borreg_m2a,
24890 bxor_m2a, bxorreg_m2a): Add satisfies_constraint_K03 condition.
24891
24892 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24893
24894 * config/arm/cortex-a7.md (cortex_a7_neon, cortex_a7_all): Remove.
24895 (cortex_a7_idiv): Use cortex_a7_both instead of cortex_a7_all.
24896
24897 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24898
24899 * config/arm/arm.c (cortexa7_younger): Return true for TYPE_CALL.
24900 * config/arm/cortex-a7.md (cortex_a7_call): Update required units.
24901
24902 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24903
24904 * config/arm/arm-protos.h (arm_mac_accumulator_is_result): New
24905 declaration.
24906 * config/arm/arm.c (arm_mac_accumulator_is_result): New function.
24907 * config/arm/cortex-a7.md: New bypasses using
24908 arm_mac_accumulator_is_result.
24909
24910 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24911
24912 * config/arm/cortex-a7.md (cortex_a7_neon_mul): New reservation.
24913 (cortex_a7_neon_mla): Likewise.
24914 (cortex_a7_fpfmad): New reservation.
24915 (cortex_a7_fpmacs): Use ffmas and update required units.
24916 (cortex_a7_fpmuld): Update required units and latency.
24917 (cortex_a7_fpmacd): Likewise.
24918 (cortex_a7_fdivs, cortex_a7_fdivd): Likewise.
24919 (cortex_a7_neon). Likewise.
24920 (bypass) Update participating units.
24921
24922 2013-01-29 Greta Yorsh <Greta.Yorsh@arm.com>
24923
24924 * config/arm/arm.md (type): Add ffmas and ffmad to "type" attribute.
24925 * config/arm/vfp.md (fma,fmsub,fnmsub,fnmadd): Change type
24926 from fmac to ffma.
24927 * config/arm/vfp11.md (vfp_farith): Use ffmas.
24928 (vfp_fmul): Use ffmad.
24929 * config/arm/cortex-r4f.md (cortex_r4_fmacs): Use ffmas.
24930 (cortex_r4_fmacd): Use ffmad.
24931 * config/arm/cortex-m4-fpu.md (cortex_m4_fmacs): Use ffmas.
24932 * config/arm/cortex-a9.md (cortex_a9_fmacs): Use ffmas.
24933 (cortex_a9_fmacd): Use ffmad.
24934 * config/arm/cortex-a8-neon.md (cortex_a8_vfp_macs): Use ffmas.
24935 (cortex_a8_vfp_macd): Use ffmad.
24936 * config/arm/cortex-a5.md (cortex_a5_fpmacs): Use ffmas.
24937 (cortex_a5_fpmacd): Use ffmad.
24938 * config/arm/cortex-a15-neon.md (cortex_a15_vfp_macs) Use ffmas.
24939 (cortex_a15_vfp_macd): Use ffmad.
24940 * config/arm/arm1020e.md (v10_fmul): Use ffmas and ffmad.
24941
24942 2013-01-29 Jason Merrill <jason@redhat.com>
24943
24944 PR libstdc++/54314
24945 * varasm.c (default_assemble_visibility): Don't warn about
24946 visibility on artificial decls.
24947
24948 2013-01-29 Richard Biener <rguenther@suse.de>
24949
24950 PR tree-optimization/56113
24951 * tree-ssa-structalias.c (equiv_class_lookup): Also return
24952 the bitmap leader.
24953 (label_visit): Free duplicate bitmaps and record the leader instead.
24954 (perform_var_substitution): Adjust.
24955
24956 2013-01-29 Richard Biener <rguenther@suse.de>
24957
24958 PR tree-optimization/55270
24959 * tree-ssa-dom.c (eliminate_degenerate_phis): If we changed
24960 the CFG, schedule loops for fixup.
24961
24962 2013-01-29 Nick Clifton <nickc@redhat.com>
24963
24964 * config/rl78/rl78.c (rl78_regno_mode_code_ok_for_base_p): Allow
24965 SP_REG.
24966
24967 2013-01-28 Leif Ekblad <leif@rdos.net>
24968
24969 * config.gcc (i[34567]86-*-rdos*, x86_64-*-rdos*): New targets.
24970 * config/i386/i386.h (TARGET_RDOS): New macro.
24971 (DEFAULT_LARGE_SECTION_THRESHOLD): New macro.
24972 * config/i386/i386.c (ix86_option_override_internal): For 64bit
24973 TARGET_RDOS, set ix86_cmodel to CM_MEDIUM_PIC and flag_pic to 1.
24974 * config/i386/i386.opt (mlarge-data-threshold): Initialize to
24975 DEFAULT_LARGE_SECTION_THRESHOLD.
24976 * config/i386/i386.md (R14_REG, R15_REG): New constants.
24977 * config/i386/rdos.h: New file.
24978 * config/i386/rdos64.h: New file.
24979
24980 2013-01-28 Bernd Schmidt <bernds@codesourcery.com>
24981
24982 PR other/54814
24983 * reload.c (find_valid_class_1): Use in_hard_reg_set_p instead of
24984 TEST_HARD_REG_BIT.
24985
24986 2013-01-28 Jakub Jelinek <jakub@redhat.com>
24987
24988 PR rtl-optimization/56117
24989 * sched-deps.c (sched_analyze_2) <case PREFETCH>: For use_cselib
24990 call cselib_lookup_from_insn on the MEM before calling
24991 add_insn_mem_dependence.
24992
24993 2013-01-28 Richard Biener <rguenther@suse.de>
24994
24995 * tree-inline.c (remap_gimple_stmt): Do not assing a BLOCK
24996 to a stmt that didn't have one.
24997 (copy_phis_for_bb): Likewise for PHI arguments.
24998 (copy_debug_stmt): Likewise for debug stmts.
24999
25000 2013-01-28 Richard Biener <rguenther@suse.de>
25001
25002 PR tree-optimization/56034
25003 * tree-loop-distribution.c (enum partition_kind): Add PKIND_REDUCTION.
25004 (partition_builtin_p): Adjust.
25005 (generate_code_for_partition): Handle PKIND_REDUCTION. Assert
25006 it is the last partition.
25007 (rdg_flag_uses): Check SSA_NAME_IS_DEFAULT_DEF before looking
25008 up the vertex for the definition.
25009 (classify_partition): Classify whether a partition is a
25010 PKIND_REDUCTION, thus has uses outside of the loop.
25011 (ldist_gen): Inherit PKIND_REDUCTION when merging partitions.
25012 Merge all PKIND_REDUCTION partitions into the last partition.
25013 (tree_loop_distribution): Seed partitions from reductions as well.
25014
25015 2013-01-28 Jakub Jelinek <jakub@redhat.com>
25016
25017 PR tree-optimization/56125
25018 * tree-ssa-math-opts.c (gimple_expand_builtin_pow): Don't optimize
25019 pow(x,c) into sqrt(x) * powi(x, n/2) or
25020 1.0 / (sqrt(x) * powi(x, abs(n/2))) if c is an integer or when
25021 optimizing for size.
25022 Don't optimize pow(x,c) into powi(x, n/3) * powi(cbrt(x), n%3) or
25023 1.0 / (powi(x, abs(n)/3) * powi(cbrt(x), abs(n)%3)) if 2c is an
25024 integer.
25025
25026 PR tree-optimization/56094
25027 * gimplify.c (force_gimple_operand_1): Temporarily set input_location
25028 to UNKNOWN_LOCATION while gimplifying expr.
25029
25030 2013-01-27 Uros Bizjak <ubizjak@gmail.com>
25031
25032 PR target/56114
25033 * config/i386/i386.md (*movabs<mode>_1): Add square brackets around
25034 operand 0 in movabs insn template for -masm=intel asm alternative.
25035 (*movabs<mode>_2): Ditto for operand 1.
25036
25037 2013-01-26 David Holsgrove <david.holsgrove@xilinx.com>
25038
25039 PR target/54663
25040 * config.gcc (microblaze*-linux*): Add tmake_file to allow building
25041 of microblaze-c.o
25042
25043 2013-01-26 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25044
25045 * config.gcc (microblaze*-*-*): Rename microblaze*-*-elf, update
25046 tm_file.
25047
25048 2013-01-25 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25049
25050 * config/aarch64/aarch64.c (TARGET_FIXED_CONDITION_CODE_REGS):
25051 Undef to avoid warning.
25052
25053 2013-01-25 Michael Haubenwallner <michael.haubenwallner@salomon.at>
25054
25055 * configure.ac (gcc_cv_ld_static_dynamic): Define for AIX native ld.
25056 * configure: Regenerate.
25057
25058 2013-01-25 Jakub Jelinek <jakub@redhat.com>
25059
25060 PR tree-optimization/56098
25061 * tree-ssa-phiopt.c (nt_init_block): Don't call add_or_mark_expr
25062 for stmts with volatile ops.
25063 (cond_store_replacement): Don't optimize if assign has volatile ops.
25064 (cond_if_else_store_replacement_1): Don't optimize if either
25065 then_assign or else_assign have volatile ops.
25066 (hoist_adjacent_loads): Don't optimize if either def1 or def2 have
25067 volatile ops.
25068
25069 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25070
25071 * doc/invoke.texi (AVR Built-in Macros): Document __XMEGA__.
25072
25073 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25074
25075 * doc/extend.texi (Example of asm with clobbered asm reg): Fix
25076 missing ':' in asm example.
25077
25078 2013-01-25 Tejas Belagod <tejas.belagod@arm.com>
25079
25080 * config/aarch64/aarch64-simd-builtins.def: Separate sq<r>dmulh_lane
25081 entries into lane and laneq entries.
25082 * config/aarch64/aarch64-simd.md (aarch64_sq<r>dmulh_lane<mode>):
25083 Remove AdvSIMD scalar modes.
25084 (aarch64_sq<r>dmulh_laneq<mode>): New.
25085 (aarch64_sq<r>dmulh_lane<mode>): New RTL pattern for Scalar AdvSIMD
25086 modes.
25087 * config/aarch64/arm_neon.h: Fix all the vq<r>dmulh_lane* intrinsics'
25088 builtin implementations to relfect changes in RTL in aarch64-simd.md.
25089 * config/aarch64/iterators.md (VCOND): New.
25090 (VCONQ): New.
25091
25092 2013-01-25 Georg-Johann Lay <avr@gjlay.de>
25093
25094 PR target/54222
25095 * config/avr/builtins.def (DEF_BUILTIN): Add LIBNAME argument.
25096 Add NULL LIBNAME argument to existing definitions.
25097 (ABSHR, ABSR, ABSLR, ABSLLR, ABSHK, ABSK, ABSLK, ABSLLK): New.
25098 * config/avr/avr-c.c (DEF_BUILTIN): Add LIBNAME argument.
25099 * config/avr/avr.c (DEF_BUILTIN): Same.
25100 (avr_init_builtins): Pass down LIBNAME to add_builtin_function.
25101 (avr_expand_builtin): Expand to a vanilla call if a libgcc
25102 implementation is available (DECL_ASSEMBLER_NAME is set).
25103 (avr_fold_absfx): New static function.
25104 (avr_fold_builtin): Use it to handle: AVR_BUILTIN_ABSHR,
25105 AVR_BUILTIN_ABSR, AVR_BUILTIN_ABSLR, AVR_BUILTIN_ABSLLR,
25106 AVR_BUILTIN_ABSHK, AVR_BUILTIN_ABSK, AVR_BUILTIN_ABSLK,
25107 AVR_BUILTIN_ABSLLK.
25108 * config/avr/stdfix.h (abshr, absr, abslr, absllr)
25109 (abshk, absk, abslk, absllk): Provide as static inline functions.
25110
25111 2013-01-25 Marek Polacek <polacek@redhat.com>
25112
25113 PR tree-optimization/56035
25114 * cfgloopmanip.c (fix_loop_structure): Remove redundant condition.
25115
25116 2012-01-24 Uros Bizjak <ubizjak@gmail.com>
25117
25118 * config/i386/i386.md (*movti_internal_rex64): Add (o,e) alternative.
25119 (*movtf_internal_rex64): Add (!o,C) alternative
25120 (*movxf_internal_rex64): Ditto.
25121 (*movdf_internal_rex64): Add (?r,C) and (?m,C) alternatives.
25122
25123 2013-01-24 Shenghou Ma <minux.ma@gmail.com>
25124
25125 * doc/invoke.texi: fix typo.
25126 * doc/objc.texi: fix typo.
25127
25128 2013-01-24 Richard Sandiford <rdsandiford@googlemail.com>
25129
25130 * config/mips/mips.md (*and<mode>3_mips16): Use the "W" constraint
25131 for the first two alternatives.
25132
25133 2013-01-24 Diego Novillo <dnovillo@google.com>
25134
25135 * Makefile.in (GGC): Remove. Replace all instances with ggc-page.o.
25136 (ggc-zone.o): Remove.
25137 * configure.ac: Remove option --with-gc.
25138 * configure: Re-generate.
25139 * doc/install.texi: Remove documentation for --with-gc.
25140 * gengtype.c (write_enum_defn): Remove. Update all users.
25141 (write_Types_process_field): Remove generation of gt_e_* argument.
25142 (output_type_enum): Remove. Update all users.
25143 (write_enum_defn): Remove. Update all users.
25144 (enum alloc_zone): Remove. Update all users.
25145 (write_splay_tree_allocator_def): Remove generation of gt_e_* argument.
25146 * ggc-common.c (ggc_splay_alloc): Remove first argument.
25147 Update all callers.
25148 (struct ptr_data): Remove field TYPE. Update all users.
25149 (gt_pch_note_object): Remove argument TYPE. Update all users.
25150 * ggc-internal.h (ggc_pch_alloc_object): Remove last argument.
25151 Update all users.
25152 * ggc-none.c (ggc_alloc_typed_stat): Remove.
25153 (struct alloc_zone): Remove.
25154 (ggc_internal_alloc_zone_stat): Remove.
25155 (ggc_internal_cleared_alloc_zone_stat): Remove.
25156 * ggc-page.c (ggc_alloc_typed_stat): Remove.
25157 (ggc_pch_count_object): Remove last argument. Update all users.
25158 (ggc_pch_alloc_object): Remove last argument. Update all users.
25159 (struct alloc_zone): Remove.
25160 * ggc-zone.c: Remove.
25161 * ggc.h (gt_pch_note_object): Remove last argument. Update all users.
25162 (struct alloc_zone): Remove.
25163 (ggc_alloc_typed_stat): Remove.
25164 (ggc_alloc_typed): Remove.
25165 (ggc_splay_alloc): Remove first argument.
25166 (rtl_zone): Remove. Update all users.
25167 (tree_zone): Remove. Update all users.
25168 (tree_id_zone): Remove. Update all users.
25169 (ggc_internal_zone_alloc_stat): Remove. Update all users.
25170 (ggc_internal_zone_cleared_alloc_stat): Remove. Update all users.
25171 (ggc_internal_zone_vec_alloc_stat): Remove. Update all users.
25172 * tree-ssanames.c: Remove references to zone allocator in comments.
25173
25174 2013-01-24 Georg-Johann Lay <avr@gjlay.de>
25175
25176 * config/avr/avr.c (avr_out_fract): Make register numbers that
25177 might be outside of source operand signed.
25178
25179 2013-01-24 Uros Bizjak <ubizjak@gmail.com>
25180
25181 * config/i386/constraints.md (Yf): New constraint.
25182 * config/i386/i386.md (*movdf_internal_rex64): Use Yf*f instead
25183 of f constraint to conditionaly disable x87 register preferences.
25184 (*movdf_internal): Ditto.
25185 (*movsf_internal): Ditto.
25186
25187 2013-01-24 Steven Bosscher <steven@gcc.gnu.org>
25188
25189 PR inline-asm/55934
25190 * lra-assigns.c (assign_by_spills): Throw away the pattern of asms
25191 that have operands with impossible constraints.
25192 Add a FIXME for a speed-up opportunity.
25193 * lra-constraints.c (process_alt_operands): Verify that a class
25194 selected from constraints on asms is valid for the operand mode.
25195 (curr_insn_transform): Remove incorrect comment.
25196
25197 2013-01-23 David Edelsohn <dje.gcc@gmail.com>
25198
25199 * config/rs6000/rs6000.c (rs6000_delegitimize_address): Check that
25200 TOC operand is a valid symbol ref in the constant pool.
25201
25202 2013-01-23 Edgar E. Iglesias <edgar.iglesias@gmail.com>
25203
25204 * config/microblaze/linux.h: Add TARGET_OS_CPP_BUILTINS
25205
25206 2013-01-23 Georg-Johann Lay <avr@gjlay.de>
25207
25208 PR target/54222
25209 * config/avr/stdfix.h: New file.
25210 * t-avr (stdfix-gcc.h): New rule to build it.
25211 (EXTRA_HEADERS): Set it to install stdfix.h, stdfix-gcc.h.
25212
25213 2013-01-23 Kostya Serebryany <kcc@google.com>
25214
25215 * config/darwin.h: remove dependency on
25216 CoreFoundation (asan on Mac OS).
25217
25218 2013-01-23 Jakub Jelinek <jakub@redhat.com>
25219
25220 PR target/49069
25221 * config/arm/arm.md (cbranchdi4, cstoredi4): Use s_register_operand
25222 instead of cmpdi_operand for first comparison operand.
25223 Don't assert that comparison operands aren't both constants.
25224
25225 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25226
25227 * doc/install.texi (Downloading the Source): Update references to
25228 downloading separate components.
25229
25230 2013-01-22 Jonathan Wakely <jwakely.gcc@gmail.com>
25231
25232 * doc/extend.texi (__int128): Improve grammar.
25233
25234 2013-01-22 Uros Bizjak <ubizjak@gmail.com>
25235
25236 PR target/56028
25237 * config/i386/i386.md (*movti_internal_rex64): Change (o,riF)
25238 alternative to (o,r).
25239 (*movdi_internal_rex64): Remove (!o,n) alternative.
25240 (DImode immediate->memory splitter): Remove.
25241 (DImode immediate->memory peephole2): Remove.
25242 (movtf): Enable for TARGET_64BIT || TARGET_SSE.
25243 (*movtf_internal_rex64): Rename from *movtf_internal. Change (!o,F*r)
25244 alternative to (!o,*r).
25245 (*movtf_internal_sse): New pattern.
25246 (*movxf_internal_rex64): New pattern.
25247 (*movxf_internal): Disable for TARGET_64BIT.
25248 (*movdf_internal_rex64): Remove (!o,F) alternative.
25249
25250 2013-01-22 Jakub Jelinek <jakub@redhat.com>
25251
25252 PR middle-end/56074
25253 * dumpfile.c (dump_loc): Only print loc if LOCATION_LOCUS (loc)
25254 isn't UNKNOWN_LOCATION nor BUILTINS_LOCATION.
25255 * tree-vect-loop-manip.c (find_loop_location): Also ignore
25256 stmt locations where LOCATION_LOCUS of the stmt location is
25257 UNKNOWN_LOCATION or BUILTINS_LOCATION.
25258
25259 PR target/55686
25260 * config/i386/i386.md (UNSPEC_STOS): New.
25261 (strset_singleop, *strsetdi_rex_1, *strsetsi_1, *strsethi_1,
25262 *strsetqi_1): Add UNSPEC_STOS.
25263
25264 2013-01-22 Paolo Carlini <paolo.carlini@oracle.com>
25265
25266 PR c++/56067
25267 * doc/invoke.texi: Remove left over -Wsynth example.
25268
25269 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25270
25271 PR tree-optimization/56051
25272 * fold-const.c (fold_binary_loc): Don't fold
25273 X < (cast) (1 << Y) into (X >> Y) != 0 if cast is either
25274 a narrowing conversion, or widening conversion from signed
25275 to unsigned.
25276
25277 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25278
25279 PR rtl-optimization/56023
25280 * haifa-sched.c (fix_inter_tick): Do not update ticks of instructions,
25281 dependent on debug instruction.
25282
25283 2013-01-21 Martin Jambor <mjambor@suse.cz>
25284
25285 PR middle-end/56022
25286 * function.c (allocate_struct_function): Call
25287 invoke_set_current_function_hook earlier.
25288
25289 2013-01-21 Jakub Jelinek <jakub@redhat.com>
25290
25291 * reload1.c (init_reload): Only initialize reload_obstack
25292 during the first call.
25293
25294 2013-01-21 Marek Polacek <polacek@redhat.com>
25295
25296 * cfgloop.c (verify_loop_structure): Fix up grammar.
25297
25298 2013-01-21 Yi-Hsiu Hsu <ahsu@marvell.com>
25299
25300 * config/arm/marvell-pj4.md (pj4_shift_conds, pj4_alu_shift,
25301 pj4_alu_shift_conds, pj4_shift): Handle simple_alu_shift.
25302
25303 2013-01-21 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25304
25305 PR target/56058
25306 * config/arm/marvell-pj4.md: Update copyright year.
25307 Fix up use of alu to alu_reg and simple_alu_imm.
25308
25309 2013-01-21 Uros Bizjak <ubizjak@gmail.com>
25310
25311 * config/i386/i386.md (enabled): Do not disable fma4 for TARGET_FMA.
25312
25313 2013-01-20 Vladimir Makarov <vmakarov@redhat.com>
25314
25315 PR target/55433
25316 * lra-constraints.c (curr_insn_transform): Don't reuse original
25317 insn for secondary memory move when memory mode should be different.
25318
25319 2013-01-20 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25320
25321 * config/pa/pa.md (atomic_loaddi, atomic_loaddi_1, atomic_storedi,
25322 atomic_storedi_1): New patterns.
25323
25324 2013-01-20 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25325
25326 btver2 pipeline descriptions.
25327 * config/i386/i386.c: Enable CPU_BTVER2 to use btver2 pipeline
25328 descriptions.
25329 * config/i386/i386.md (btver2_decode): New type attributes.
25330 * config/i386/sse.md (btver2_decode, btver2_sse_attr): New
25331 type attributes.
25332 * config/i386/btver2.md: New file describing btver2 pipelines.
25333
25334 2013-01-19 Andrew Pinski <apinski@cavium.com>
25335
25336 PR tree-optimization/52631
25337 * tree-ssa-sccvn (visit_use): Before looking up the original
25338 statement, try looking up the simplified expression.
25339
25340 2013-01-19 Anthony Green <green@moxielogic.com>
25341
25342 * config/moxie/moxie.c (moxie_expand_prologue): Set
25343 current_function_static_stack_size.
25344
25345 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25346
25347 PR tree-optimization/56029
25348 * tree-phinodes.c (reserve_phi_args_for_new_edge): Set
25349 gimple_phi_arg_location for the new arg to UNKNOWN_LOCATION.
25350
25351 2013-01-18 Sharad Singhai <singhai@google.com>
25352
25353 PR tree-optimization/55995
25354 * dumpfile.c (dump_loc): Print location only if available.
25355 * tree-vectorizer.c (increase_alignment): Intialize vect_location.
25356
25357 2013-01-18 Vladimir Makarov <vmakarov@redhat.com>
25358
25359 PR target/55433
25360 * lra-constraints.c (curr_insn_transform): Reuse original insn for
25361 secondary memory move.
25362 (inherit_reload_reg): Use rclass instead of cl for
25363 check_secondary_memory_needed_p.
25364
25365 2013-01-18 Jakub Jelinek <jakub@redhat.com>
25366
25367 PR middle-end/56015
25368 * expr.c (expand_expr_real_2) <case COMPLEX_EXPR>: Handle
25369 the case where writing real complex part of target modifies op1.
25370
25371 2013-01-18 James Greenhalgh <james.greenhalgh@arm.com>
25372
25373 * config/aarch64/aarch64-simd.md
25374 (aarch64_vcond_internal<mode>): Handle unordered cases.
25375 * config/aarch64/iterators.md (v_cmp_result): New.
25376
25377 2013-01-18 Yi-Hsiu Hsu <ahsu@marvell.com>
25378 Ramana Radhakrishnan <ramana.radhakrishnan@arm.com>
25379
25380 * config/arm/marvell-pj4.md: New file.
25381 * config/arm/arm.c (arm_issue_rate): Add marvell_pj4.
25382 * config/arm/arm.md (generic_sched): Add marvell_pj4.
25383 (generic_vfp): Likewise.
25384 * config/arm/arm-cores.def: Add marvell-pj4.
25385 * config/arm/arm-tune.md: Regenerate.
25386 * config/arm/arm-tables.opt: Regenerate.
25387 * config/arm/bpabi.h (BE8_LINK_SPEC): Add marvell_pj4.
25388 * doc/invoke.texi: Document marvell-pj4.
25389
25390 2013-01-18 Tejas Belagod <tejas.belagod@arm.com>
25391
25392 * config/aarch64/arm_neon.h: Map scalar types to standard types.
25393
25394 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25395
25396 PR debug/54114
25397 PR debug/54402
25398 PR debug/49888
25399 * var-tracking.c (negative_power_of_two_p): New.
25400 (global_get_addr_cache, local_get_addr_cache): New.
25401 (get_addr_from_global_cache, get_addr_from_local_cache): New.
25402 (vt_canonicalize_addr): Rewrite using the above. Adjust the
25403 heading comment.
25404 (vt_stack_offset_p): Remove.
25405 (vt_canon_true_dep): Always canonicalize loc's address.
25406 (clobber_overlapping_mems): Make sure we have a MEM.
25407 (local_get_addr_clear_given_value): New.
25408 (val_reset): Clear local cached entries.
25409 (compute_bb_dataflow): Create and release the local cache.
25410 Disable duplicate MEMs clobbering.
25411 (emit_notes_in_bb): Clobber MEMs likewise.
25412 (vt_emit_notes): Create and release the local cache.
25413 (vt_initialize, vt_finalize): Create and release the global
25414 cache, respectively.
25415 * alias.c (rtx_equal_for_memref_p): Compare operands of ENTRY_VALUEs.
25416
25417 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25418
25419 PR libmudflap/53359
25420 * tree-mudflap.c (mudflap_finish_file): Skip deferred decls
25421 not found in the symtab.
25422
25423 2013-01-18 Alexandre Oliva <aoliva@redhat.com>
25424
25425 PR debug/56006
25426 PR rtl-optimization/55547
25427 PR rtl-optimization/53827
25428 PR debug/53671
25429 PR debug/49888
25430 * alias.c (offset_overlap_p): New, factored out of...
25431 (memrefs_conflict_p): ... this. Use absolute sizes. Retain
25432 the conservative special case for symbolic constants. Don't
25433 adjust zero sizes on alignment.
25434
25435 2013-01-18 Bernd Schmidt <bernds@codesourcery.com>
25436
25437 PR rtl-optimization/52573
25438 * regrename.c (build_def_use): Ignore REG_DEAD notes if there is a
25439 REG_UNUSED for the same register.
25440
25441 2013-01-17 Richard Biener <rguenther@suse.de>
25442 Marek Polacek <polacek@redhat.com>
25443
25444 PR rtl-optimization/55833
25445 * loop-unswitch.c (unswitch_loops): Move loop verification...
25446 (unswitch_single_loop): ...here. Call mark_irreducible_loops.
25447 * cfgloopmanip.c (fix_loop_placement): Add IRRED_INVALIDATED parameter.
25448 Set it to true when we're removing a loop from hierarchy tree in
25449 an irreducible region.
25450 (fix_bb_placements): Adjust caller.
25451 (fix_loop_placements): Likewise.
25452
25453 2013-01-17 Georg-Johann Lay <avr@gjlay.de>
25454
25455 * config/avr/builtins.def (DEF_BUILTIN): Factor out
25456 "__builtin_avr_" from NAME, turn NAME to an uppercase identifier.
25457 Factor out 'CODE_FOR_' from ICODE, use 'nothing' instead of '-1'.
25458 Remove ID. Adjust comments.
25459 * config/avr/avr-c.c (avr_builtin_name): Remove.
25460 (avr_cpu_cpp_builtins): Use DEF_BUILTIN instead of for-loop.
25461 * config/avr/avr.c (avr_tolower): New static function.
25462 (DEF_BUILTIN): Remove parameter ID. Prefix ICODE by 'CODE_FOR_'.
25463 Stringify NAME, prefix it with "__builtin_avr_" and lowercase it.
25464 (avr_expand_builtin): Assert insn_code != CODE_FOR_nothing for
25465 default expansion.
25466
25467 2013-01-17 Jan Hubicka <jh@suse.cz>
25468
25469 PR tree-optimization/55273
25470 * loop-iv.c (iv_number_of_iterations): Consider zero iteration case.
25471
25472 2013-01-17 Uros Bizjak <ubizjak@gmail.com>
25473
25474 PR target/55981
25475 * config/i386/sync.md (atomic_store<mode>): Always generate SWImode
25476 store through atomic_store<mode>_1.
25477 (atomic_store<mode>_1): Macroize insn using SWI mode iterator.
25478
25479 2013-01-17 Martin Jambor <mjambor@suse.cz>
25480
25481 PR tree-optimizations/55264
25482 * ipa-inline-transform.c (can_remove_node_now_p_1): Never return true
25483 for virtual methods.
25484 * ipa.c (symtab_remove_unreachable_nodes): Never return true for
25485 virtual methods before inlining is over.
25486 * cgraph.h (cgraph_only_called_directly_or_aliased_p): Return false for
25487 virtual functions.
25488 * cgraphclones.c (cgraph_create_virtual_clone): Mark clones as
25489 non-virtual.
25490
25491 2013-01-16 Vladimir Makarov <vmakarov@redhat.com>
25492
25493 PR rtl-optimization/56005
25494 * sched-deps.c (sched_analyze_2): Check deps->readonly for adding
25495 pending reads for prefetch.
25496
25497 2013-01-16 Ian Bolton <ian.bolton@arm.com>
25498
25499 * config/aarch64/aarch64.md
25500 (*cstoresi_neg_uxtw): New pattern.
25501 (*cmovsi_insn_uxtw): New pattern.
25502 (*<optab>si3_uxtw): New pattern.
25503 (*<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): New pattern.
25504 (*<optab>si3_insn_uxtw): New pattern.
25505 (*bswapsi2_uxtw): New pattern.
25506
25507 2013-01-16 Richard Biener <rguenther@suse.de>
25508
25509 * tree-inline.c (tree_function_versioning): Remove set but
25510 never used variable.
25511
25512 2013-01-16 Richard Biener <rguenther@suse.de>
25513
25514 PR tree-optimization/55964
25515 * tree-flow.h (rename_variables_in_loop): Remove.
25516 (rename_variables_in_bb): Likewise.
25517 * tree-loop-distribution.c (update_phis_for_loop_copy): Remove.
25518 (copy_loop_before): Adjust and delete update-ssa status.
25519 * tree-vect-loop-manip.c (rename_variables_in_bb): Make static.
25520 (rename_variables_in_bb): Likewise. Properly walk over predecessors.
25521 (rename_variables_in_loop): Remove.
25522 (slpeel_update_phis_for_duplicate_loop): Likewise.
25523 (slpeel_tree_duplicate_loop_to_edge_cfg): Handle nested loops,
25524 use available cfg machinery instead of duplicating it.
25525 Update PHI nodes and perform poor-mans SSA update here.
25526 (slpeel_tree_peel_loop_to_edge): Adjust.
25527
25528 2013-01-16 Richard Biener <rguenther@suse.de>
25529
25530 PR tree-optimization/54767
25531 PR tree-optimization/53465
25532 * tree-vrp.c (vrp_meet_1): Revert original fix for PR53465.
25533 (vrp_visit_phi_node): For PHI arguments coming via backedges
25534 drop all symbolical range information.
25535 (execute_vrp): Compute backedges.
25536
25537 2013-01-16 Richard Biener <rguenther@suse.de>
25538
25539 * doc/install.texi: Update CLooG and ISL requirements to
25540 0.18.0 and 0.11.1.
25541
25542 2013-01-16 Christian Bruel <christian.bruel@st.com>
25543
25544 PR target/55301
25545 * config/sh/sh.c (sh_expand_prologue): Postpone new_stack mem symbol.
25546 (broken_move): Handle UNSPECV_SP_SWITCH_B.
25547 * config/sh/sh.md (sp_switch_1): Use set (reg:SI SP_REG).
25548
25549 2013-01-16 DJ Delorie <dj@redhat.com>
25550
25551 * config/sh/sh.md (UNSPECV_SP_SWITCH_B): New.
25552 (UNSPECV_SP_SWITCH_E): New.
25553 (sp_switch_1): Change to an unspec.
25554 (sp_switch_2): Change to an unspec. Don't use post-inc when we
25555 replace $r15.
25556
25557 2013-01-16 Uros Bizjak <ubizjak@gmail.com>
25558
25559 * emit-rtl.c (need_atomic_barrier_p): Mask memory model argument
25560 with MEMMODEL_MASK before comparing with MEMMODEL_* memory types.
25561 * optabs.c (maybe_emit_sync_lock_test_and_set): Ditto.
25562 (expand_mem_thread_fence): Ditto.
25563 (expand_mem_signal_fence): Ditto.
25564 (expand_atomic_load): Ditto.
25565 (expand_atomic_store): Ditto.
25566
25567 2013-01-16 Alexandre Oliva <aoliva@redhat.com>
25568
25569 PR rtl-optimization/55547
25570 PR rtl-optimization/53827
25571 PR debug/53671
25572 PR debug/49888
25573 * alias.c (memrefs_conflict_p): Set sizes to negative after
25574 AND adjustments.
25575
25576 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25577
25578 PR target/55940
25579 * function.c (thread_prologue_and_epilogue_insns): Always
25580 add crtl->drap_reg to set_up_by_prologue.set, even if
25581 stack_realign_drap is false.
25582
25583 2013-01-15 Jan-Benedict Glaw <jbglaw@lug-owl.de>
25584
25585 * config/vax/vax.md (add<mode>3, sub<mode>3, mul<mode>3, div<mode>3,
25586 and<mode>3, *and<mode>_const_int, ior<mode>3, xor<mode>3, ashrsi3,
25587 *call): Fix indention.
25588
25589 2013-01-15 Tom de Vries <tom@codesourcery.com>
25590
25591 PR target/55876
25592 * optabs.c (widen_operand): Use gen_lowpart instead of gen_rtx_SUBREG.
25593 Update comment.
25594
25595 2013-01-15 Vladimir Makarov <vmakarov@redhat.com>
25596
25597 PR rtl-optimization/55153
25598 * sched-deps.c (sched_analyze_2): Add pending reads for prefetch.
25599
25600 2013-01-15 Martin Jambor <mjambor@suse.cz>
25601
25602 PR tree-optimization/55920
25603 * tree-sra.c (analyze_access_subtree): Do not mark non-removable
25604 accesses as grp_to_be_debug_replaced.
25605
25606 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25607
25608 PR tree-optimization/55920
25609 * tree-sra.c (sra_modify_assign): If for lacc->grp_to_be_debug_replaced
25610 there is non-useless type conversion needed from debug rhs to lhs,
25611 use build_debug_ref_for_model and/or VIEW_CONVERT_EXPR.
25612
25613 2013-01-15 Joseph Myers <joseph@codesourcery.com>
25614 Mikael Pettersson <mikpe@it.uu.se>
25615
25616 PR target/43961
25617 * config/arm/arm.h (ADDR_VEC_ALIGN): Align SImode jump tables for
25618 Thumb.
25619 (ASM_OUTPUT_CASE_LABEL): Remove.
25620 (ASM_OUTPUT_BEFORE_CASE_LABEL): Define to empty.
25621 * final.c (shorten_branches): Update alignment of labels before
25622 jump tables if CASE_VECTOR_SHORTEN_MODE.
25623
25624 2013-01-15 Richard Biener <rguenther@suse.de>
25625
25626 PR bootstrap/55961
25627 * system.h: Do not include gmp.h for building host tools.
25628
25629 2013-01-15 Richard Biener <rguenther@suse.de>
25630
25631 PR middle-end/55882
25632 * emit-rtl.c (set_mem_attributes_minus_bitpos): Correctly
25633 account for bitpos when computing alignment.
25634
25635 2013-01-15 Vladimir Yakovlev <vladimir.b.yakovlev@intel.com>
25636
25637 * config/i386/i386-c.c (ix86_target_macros_internal): New case.
25638 (ix86_target_macros_internal): Likewise.
25639
25640 * config/i386/i386.c (m_CORE2I7): Removed.
25641 (m_CORE_HASWELL): New macro.
25642 (m_CORE_ALL): Likewise.
25643 (initial_ix86_tune_features): m_CORE2I7 is replaced by m_CORE_ALL.
25644 (initial_ix86_arch_features): Likewise.
25645 (processor_target_table): Initializations for Core avx2.
25646 (cpu_names): New names "core-avx2".
25647 (ix86_option_override_internal): Changed PROCESSOR_COREI7 by
25648 PROCESSOR_CORE_HASWELL.
25649 (ix86_issue_rate): New case.
25650 (ia32_multipass_dfa_lookahead): Likewise.
25651 (ix86_sched_init_global): Likewise.
25652
25653 * config/i386/i386.h (TARGET_HASWELL): New macro.
25654 (target_cpu_default): New TARGET_CPU_DEFAULT_haswell.
25655 (processor_type): New PROCESSOR_HASWELL.
25656
25657 2013-01-15 Jakub Jelinek <jakub@redhat.com>
25658
25659 PR tree-optimization/55955
25660 * tree-vect-loop.c (vectorizable_reduction): Give up early on
25661 *SHIFT_EXPR and *ROTATE_EXPR codes.
25662
25663 PR tree-optimization/48766
25664 * opts.c (common_handle_option): For -fwrapv disable -ftrapv, for
25665 -ftrapv disable -fwrapv.
25666
25667 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25668
25669 PR target/55974
25670 * config/avr/avr-c.c (avr_cpu_cpp_builtins): Define __FLASH
25671 etc. to 1 and not to __flash.
25672 Use LL suffix for __INT24_MAX__ with -mint8.
25673 Use ULL suffix for __UINT24_MAX__ with -mint8.
25674
25675 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25676
25677 * config/avr/avr-arch.h
25678 (struct base_arch_s): Use typedef avr_arch_t instead.
25679 (struct arch_info_s): Use typedef avr_arch_info_t instead.
25680 (struct mcu_type_s): Use typedef avr_mcu_t instead.
25681 * config/avr/avr.c: Same.
25682 * config/avr/avr-devices.c: Same.
25683 * config/avr/driver-avr.c: Same.
25684 * config/avr/gen-avr-mmcu-texi.c: Same.
25685 * config/avr/avr-mcus.def: Adjust comment.
25686
25687 2013-01-14 Tejas Belagod <tejas.belagod@arm.com>
25688
25689 * config/aarch64/aarch64-simd.md (*aarch64_simd_ld1r<mode>): New.
25690 * config/aarch64/iterators.md (VALLDI): New.
25691
25692 2013-01-14 Uros Bizjak <ubizjak@gmail.com>
25693 Andi Kleen <ak@linux.intel.com>
25694
25695 PR target/55948
25696 * config/i386/sync.md (atomic_store<mode>_1): New pattern.
25697 (atomic_store<mode>): Call atomic_store<mode>_1 for IX86_HLE_RELEASE
25698 memmodel flag.
25699
25700 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25701
25702 * config/avr/avr-stdint.h: Remove trailing blanks.
25703 * config/avr/avr-log.h: Same.
25704 * config/avr/avr-arch.h: Same.
25705 * config/avr/avr-devices.c: Same.
25706 * config/avr/avr-dimode.md: Same.
25707 * config/avr/predicates.md: Same.
25708 * config/avr/avr-c.c: Same. And fix typo.
25709
25710 * config/avr/avr-protos.h: Same. And:
25711 (function_arg_regno_p): Rename to avr_function_arg_regno_p.
25712 (init_cumulative_args): Rename to avr_init_cumulative_args.
25713 (expand_prologue): Rename to avr_expand_prologue.
25714 (expand_epilogue): Rename to avr_expand_epilogue.
25715 (adjust_insn_length): Rename to avr_adjust_insn_length.
25716 (notice_update_cc): Rename to avr_notice_update_cc.
25717 (final_prescan_insn): Rename to avr_final_prescan_insn.
25718 * config/avr/avr.c: Same.
25719 * config/avr/avr.h: Same.
25720 * config/avr/avr.md: Remove trailing blanks.
25721 (prologue): Use avr_expand_prologue.
25722 (epilogue, sibcall_epilogue): Use avr_expand_epilogue.
25723
25724 2013-01-14 Richard Biener <rguenther@suse.de>
25725
25726 * tree-cfg.c (verify_expr_location, verify_expr_location_1,
25727 verify_location, collect_subblocks): New functions.
25728 (verify_gimple_in_cfg): Verify that locations only reference
25729 BLOCKs in the functions BLOCK tree.
25730
25731 2013-01-14 Richard Biener <rguenther@suse.de>
25732
25733 * tree-cfgcleanup.c (remove_forwarder_block): Unshare propagated
25734 PHI argument.
25735 * graphite-sese-to-poly.c (insert_out_of_ssa_copy): Properly
25736 unshare reference.
25737 (insert_out_of_ssa_copy_on_edge): Likewise.
25738 (rewrite_close_phi_out_of_ssa): Likewise.
25739 * tree-ssa.c (insert_debug_temp_for_var_def): Properly unshare
25740 debug expressions.
25741 * tree-ssa-pre.c (insert_into_preds_of_block): Properly unshare
25742 propagated constants.
25743 * tree-cfg.c (tree_node_can_be_shared): Handled component-refs
25744 can not be shared.
25745
25746 2013-01-14 Georg-Johann Lay <avr@gjlay.de>
25747
25748 * config/avr/avr-modes.def: Add GPL copyright notice.
25749
25750 2013-01-13 Uros Bizjak <ubizjak@gmail.com>
25751
25752 * config/i386/sync.md (mem_thread_fence): Mask operands[0] with
25753 MEMMODEL_MASK to determine memory model.
25754 (atomic_store<mode>): Ditto from operands[2].
25755 * config/i386/i386.c (ix86_memmodel_check): Declare "strong" as bool.
25756
25757 2013-01-13 Jakub Jelinek <jakub@redhat.com>
25758
25759 PR fortran/55935
25760 * gimple-fold.c (get_symbol_constant_value): Call unshare_expr.
25761 (fold_gimple_assign): Don't call unshare_expr here.
25762 (fold_ctor_reference): Call unshare_expr.
25763
25764 2013-01-13 Terry Guo <terry.guo@arm.com>
25765
25766 * Makefile.in (s-mlib): New argument MULTILIB_REUSE.
25767 * doc/fragments.texi: Document MULTILIB_REUSE.
25768 * gcc.c (multilib_reuse): New internal spec.
25769 (set_multilib_dir): Also search multilib from multilib_reuse.
25770 * genmultilib (tmpmultilib3): Refactor code.
25771 (tmpmultilib4): Ditto.
25772 (multilib_reuse): New multilib argument.
25773
25774 2013-01-13 Richard Sandiford <rdsandiford@googlemail.com>
25775
25776 * Makefile.in: Update copyright.
25777
25778 2013-01-12 Tom de Vries <tom@codesourcery.com>
25779
25780 PR middle-end/55890
25781 * calls.c (expand_call): Check if arg_nr is valid.
25782
25783 2013-01-11 Michael Meissner <meissner@linux.vnet.ibm.com>
25784
25785 * doc/extend.texi (X86 Built-in Functions): Add whitespace in
25786 __builtin_ia32_paddb256 and __builtin_ia32_pavgb256
25787 documentation. Add missing '__' in front of
25788 __builtin_ia32_packssdw256.
25789
25790 2013-01-11 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25791
25792 PR target/55719
25793 * config/s390/s390.c (s390_preferred_reload_class): Do not return
25794 NO_REGS for larl operands.
25795 (s390_reload_larl_operand): Use s390_load_address instead of
25796 emit_move_insn.
25797
25798 2013-01-11 Richard Biener <rguenther@suse.de>
25799
25800 * tree-cfg.c (verify_node_sharing_1): Split out from ...
25801 (verify_node_sharing): ... here.
25802 (verify_gimple_in_cfg): Use verify_node_sharing_1 for walk_tree.
25803
25804 2013-01-11 Eric Botcazou <ebotcazou@adacore.com>
25805
25806 * configure.ac (Tree checking): Set TREECHECKING to yes if enabled.
25807 Substitute TREECHECKING.
25808 * configure: Regenerate.
25809 * Makefile.in (TREECHECKING): New.
25810
25811 2013-01-11 Richard Guenther <rguenther@suse.de>
25812
25813 PR tree-optimization/44061
25814 * tree-vrp.c (extract_range_basic): Compute zero as
25815 value-range for __builtin_constant_p of function parameters.
25816
25817 2013-01-10 Richard Sandiford <rdsandiford@googlemail.com>
25818
25819 Update copyright years.
25820
25821 2013-01-10 Vladimir Makarov <vmakarov@redhat.com>
25822
25823 PR rtl-optimization/55672
25824 * lra-eliminations.c (mark_not_eliminable): Permit addition with
25825 const to be eliminable.
25826
25827 2013-01-10 David Edelsohn <dje.gcc@gmail.com>
25828
25829 * configure.ac (HAVE_AS_TLS): Add check for powerpc-ibm-aix.
25830 * configure: Regenerate.
25831
25832 2013-01-10 Richard Biener <rguenther@suse.de>
25833
25834 * builtins.c (expand_builtin_init_trampoline): Use set_mem_attributes.
25835
25836 2013-01-10 Richard Biener <rguenther@suse.de>
25837
25838 PR bootstrap/55792
25839 * tree-into-ssa.c (rewrite_add_phi_arguments): Do not set
25840 locations for virtual PHI arguments.
25841 (rewrite_update_phi_arguments): Likewise.
25842
25843 2013-01-10 Joel Sherrill <joel.sherrill@OARcorp.com>
25844
25845 * config/v850/rtems.h (ASM_SPEC): Pass -m8byte-align and -mgcc-abi
25846 on to assembler.
25847
25848 2013-01-10 Jakub Jelinek <jakub@redhat.com>
25849
25850 PR tree-optimization/55921
25851 * tree-complex.c (expand_complex_asm): New function.
25852 (expand_complex_operations_1): Call it for GIMPLE_ASM.
25853
25854 2013-01-10 Andreas Krebbel <Andreas.Krebbel@de.ibm.com>
25855
25856 PR target/55718
25857 * config/s390/s390.c (s390_symref_operand_p)
25858 (s390_loadrelative_operand_p): Merge the two functions.
25859 (s390_check_qrst_address, print_operand_address): Add parameters
25860 to s390_loadrelative_operand_p invokation.
25861 (s390_check_symref_alignment): Use s390_loadrelative_operand_p.
25862 (s390_reload_larl_operand, s390_secondary_reload): Use
25863 s390_loadrelative_operand_p instead of s390_symref_operand_p.
25864 (legitimize_pic_address): Handle @GOTENT and @PLT + addend.
25865
25866 2013-01-09 Mike Stump <mikestump@comcast.net>
25867
25868 * dse.c (record_store): Remove unnecessary assert.
25869
25870 2013-01-09 Jan Hubicka <jh@suse.cz>
25871
25872 PR tree-optimization/55569
25873 * cfgloopmanip.c (scale_loop_profile): Make ITERATION_BOUND gcov_type.
25874 * cfgloop.h (scale_loop_profile): Likewise.
25875
25876 2013-01-09 Jan Hubicka <jh@suse.cz>
25877
25878 PR lto/45375
25879 * ipa-inline.c (ipa_inline): Remove extern inlines and virtual
25880 functions.
25881 * cgraphclones.c (cgraph_clone_node): Cpoy also LTO file data.
25882
25883 2013-01-09 Richard Sandiford <rdsandiford@googlemail.com>
25884
25885 PR middle-end/55114
25886 * expr.h (maybe_emit_group_store): Declare.
25887 * expr.c (maybe_emit_group_store): New function.
25888 * builtins.c (expand_builtin_int_roundingfn): Call it.
25889 (expand_builtin_int_roundingfn_2): Likewise.
25890
25891 2013-01-09 Vladimir Makarov <vmakarov@redhat.com>
25892
25893 PR rtl-optimization/55829
25894 * lra-constraints.c (match_reload): Add code for absent output.
25895 (curr_insn_transform): Add code for reloads of matched inputs
25896 without output.
25897
25898 2013-01-09 Uros Bizjak <ubizjak@gmail.com>
25899
25900 * config/i386/sse.md (*vec_interleave_highv2df): Change mode
25901 attribute of movddup insn to DF.
25902 (*vec_interleave_lowv2df): Ditto.
25903 (vec_dupv2df): Ditto.
25904
25905 2013-01-09 Jan Hubicka <jh@suse.cz>
25906
25907 PR tree-optimiation/55875
25908 * tree-ssa-loop-niter.c (number_of_iterations_cond): Add
25909 EVERY_ITERATION parameter.
25910 (number_of_iterations_exit): Check if exit is executed every iteration.
25911 (idx_infer_loop_bounds): Similarly here.
25912 (n_of_executions_at_most): Simplify
25913 to only test for cases where statement is dominated by the
25914 particular bound; handle correctly the "postdominance" test.
25915 (scev_probably_wraps_p): Use max loop iterations info
25916 as a global bound first.
25917
25918 2013-01-09 Nguyen Duy Dat <dat.nguyen.yn@rvc.renesas.com>
25919 Nick Clifton <nickc@redhat.com>
25920
25921 * config/v850/v850.md (cbranchsf4): New pattern.
25922 (cstoresf4): New pattern.
25923 (cbranchdf4): New pattern.
25924 (cstoredf4): New pattern.
25925 (movsicc): Disallow floating point comparisons.
25926 (cmpsf_le_insn): Fix order of operators.
25927 (cmpsf_lt_insn): Likewise.
25928 (cmpsf_eq_insn): Likewise.
25929 (cmpdf_le_insn): Likewise.
25930 (cmpdf_lt_insn): Likewise.
25931 (cmpdf_eq_insn): Likewise.
25932 (cmpsf_ge_insn): Use LE comparison.
25933 (cmpdf_ge_insn): Likewise.
25934 (cmpsf_gt_insn): Use LT comparison.
25935 (cmpdf_gt_insn): Likewise.
25936 (cmpsf_ne_insn): Delete pattern.
25937 (cmpdf_ne_insn): Delete pattern.
25938 * config/v850/v850.c (v850_gen_float_compare): Use
25939 gen_cmpdf_eq_insn for NE comparison.
25940 (v850_float_z_comparison_operator)
25941 (v850_float_nz_comparison_operator): Move from here ...
25942 * config/v850/predicates.md: ... to here. Move GT and GE
25943 comparisons into v850_float_z_comparison_operator.
25944 * config/v850/v850-protos.h (v850_float_z_comparison_operator):
25945 Delete prototype.
25946 (v850_float_nz_comparison_operator): Likewise.
25947
25948 2013-01-09 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25949
25950 * config/pa/pa.c (pa_emit_move_sequence): Replace calls to gen_insv
25951 with calls to gen_insvsi/gen_insvdi.
25952
25953 2013-01-09 Venkataramanan Kumar <venkataramanan.kumar@amd.com>
25954
25955 * config/i386/i386.c (initial_ix86_tune_features): Set up
25956 X86_TUNE_AVX128_OPTIMAL for m_BTVER2.
25957
25958 2013-01-09 Steven Bosscher <steven@gcc.gnu.org>
25959 Jakub Jelinek <jakub@redhat.com>
25960
25961 PR tree-optimization/48189
25962 * predict.c (predict_loops): If max is 0, don't call compare_tree_int.
25963 If nitercst is 0, don't predict the exit edge.
25964
25965 2013-01-08 Naveen H.S <Naveen.Hurugalawadi@caviumnetworks.com>
25966
25967 * config/aarch64/aarch64.c (aarch64_print_operand): Replace %r
25968 in asm_fprintf with reg_names.
25969 (aarch64_print_operand_address): Likewise.
25970 (aarch64_return_addr): Likewise.
25971 * config/aarch64/aarch64.h (ASM_FPRINTF_EXTENSIONS): Remove.
25972
25973 2013-01-08 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
25974
25975 * config/pa/pa.h (VAL_U6_BITS_P): Define.
25976 (INT_U6_BITS): Likewise.
25977 * config/pa/predicates.md (uint6_operand): New predicate.
25978 (shift5_operand, shift6_operand): Likewise.
25979 * config/pa/pa.md (lshrsi3, rotrsi3): Use shift5_operand instead of
25980 arith32_operand.
25981 (lshrdi3): Use shift6_operand.
25982 (shrpsi4, shrpdi4): New insn patterns.
25983 (extzv): Delete expander.
25984 (extzvsi, extzvdi): New expanders. Use uint5_operand and uint6_operand
25985 predicates in unamed zero extract patterns. Tighten common constraint.
25986 (extv): Delete expander.
25987 (extvsi, extvdi): New expanders. Use uint5_operand and uint6_operand
25988 predicates in unamed sign extract patterns. Tighten common constraint.
25989 (insv): Delete expander.
25990 (insvsi, insvdi): New expanders. Use uint5_operand and uint6_operand
25991 predicates in unamed insert patterns. Tighten common constraint.
25992 Change uint32_operand predicate to uint6_operand predicate in unamed
25993 DImode pattern to insert constant values of type 1...1xxxx.
25994
25995 2013-01-04 Jan Hubicka <jh@suse.cz>
25996
25997 PR tree-optimization/55823
25998 * ipa-prop.c (update_indirect_edges_after_inlining): Fix ordering
25999 issue.
26000
26001 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26002 Uros Bizjak <ubizjak@gmail.com>
26003
26004 PR rtl-optimization/55845
26005 * df-problems.c (can_move_insns_across): Stop scanning at
26006 volatile_insn_p source instruction or give up if
26007 across_from .. across_to range contains any volatile_insn_p
26008 instructions.
26009
26010 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26011
26012 * config/aarch64/aarch64-simd.md (vec_init<mode>): New.
26013 * config/aarch64/aarch64-protos.h (aarch64_expand_vector_init):
26014 Declare.
26015 * config/aarch64/aarch64.c (aarch64_simd_dup_constant,
26016 aarch64_simd_make_constant, aarch64_expand_vector_init): New.
26017
26018 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26019
26020 PR fortran/55341
26021 * asan.c (asan_clear_shadow): New function.
26022 (asan_emit_stack_protection): Use it.
26023
26024 2013-01-08 Tejas Belagod <tejas.belagod@arm.com>
26025
26026 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_<su>mult_lo_<mode>,
26027 aarch64_simd_vec_<su>mult_hi_<mode>): Separate instruction and operand
26028 with tab instead of space.
26029
26030 2013-01-08 Nick Clifton <nickc@redhat.com>
26031
26032 * config/rl78/rl78.c (rl78_expand_prologue): Always select
26033 register bank 0 at the start of an interrupt handler.
26034 * config/rl78/rl78.md (mulsi3_g13): Correct values for MDBL and
26035 MDBH registers.
26036
26037 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26038
26039 * config/aarch64/aarch64-simd.md
26040 (aarch64_simd_bsl<mode>_internal): Add floating-point modes.
26041 (aarch64_simd_bsl): Likewise.
26042 (aarch64_vcond_internal<mode>): Likewise.
26043 (vcond<mode><mode>): Likewise.
26044 (aarch64_cm<cmp><mode>): Fix constraints, add new modes.
26045 * config/aarch64/iterators.md (V_cmp_result): Add V2DF.
26046
26047 2013-01-08 James Greenhalgh <james.greenhalgh@arm.com>
26048
26049 * config/aarch64/aarch64-builtins.c
26050 (aarch64_builtin_vectorized_function): Handle sqrt, sqrtf.
26051
26052 2013-01-08 Martin Jambor <mjambor@suse.cz>
26053
26054 PR debug/55579
26055 * tree-sra.c (analyze_access_subtree): Return true also after
26056 potentially creating a debug-only replacement.
26057
26058 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26059
26060 PR middle-end/55890
26061 * tree-ssa-ccp.c (evaluate_stmt): Use gimple_call_builtin_p.
26062
26063 PR tree-optimization/54120
26064 * tree-vrp.c (range_fits_type_p): Don't allow
26065 src_precision < precision from signed vr to unsigned_p
26066 if vr->min or vr->max is negative.
26067 (simplify_float_conversion_using_ranges): Test can_float_p
26068 against CODE_FOR_nothing.
26069
26070 2013-01-08 Jakub Jelinek <jakub@redhat.com>
26071 Richard Biener <rguenther@suse.de>
26072
26073 PR middle-end/55851
26074 * fold-const.c (int_binop_types_match_p): Allow all INTEGRAL_TYPE_P
26075 types instead of just INTEGER_TYPE types.
26076
26077 2013-01-07 Mark Kettenis <kettenis@openbsd.org>
26078
26079 * config/i386/openbsdelf.h (LIBGCC2_HAS_TF_MODE, LIBGCC2_TF_CEXT,
26080 TF_SIZE): Define.
26081
26082 2013-01-07 Steve Ellcey <sellcey@mips.com>
26083
26084 PR target/42661
26085 * config/mips/mips.opt: Change mad to mmad to match documentation.
26086
26087 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26088
26089 PR target/55897
26090 * doc/extend.texi (AVR Named Address Spaces): __memx goes into
26091 .progmemx.data now.
26092
26093 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26094
26095 PR target/55897
26096 * config/avr/avr.h (ADDR_SPACE_COUNT): New enum.
26097 (avr_addrspace_t): Add .section_name field.
26098 * config/avr/avr.c (progmem_section): Use ADDR_SPACE_COUNT as
26099 array size.
26100 (avr_addrspace): Same. Initialize .section_name. Remove last
26101 NULL entry. Put __memx into .progmemx.data.
26102 (progmem_section_prefix): Remove.
26103 (avr_asm_init_sections): No need to initialize progmem_section.
26104 (avr_asm_named_section): Use avr_addrspace[].section_name to get
26105 section name prefix.
26106 (avr_asm_select_section): Ditto. And use get_unnamed_section to
26107 retrieve the progmem section.
26108 * avr-c.c (avr_cpu_cpp_builtins): Use ADDR_SPACE_COUNT as loop
26109 boundary to run over avr_addrspace[].
26110 (avr_register_target_pragmas): Ditto.
26111
26112 2013-01-06 Jakub Jelinek <jakub@redhat.com>
26113
26114 * varasm.c (output_constant_def_contents): For asan_protect_global
26115 protected strings, adjust DECL_ALIGN if needed, before testing for
26116 anchored symbols.
26117 (place_block_symbol): Adjust size for asan protected STRING_CSTs if
26118 TREE_CONSTANT_POOL_ADDRESS_P. Increase alignment for asan protected
26119 normal decls.
26120 (output_object_block): For asan protected decls, emit asan padding
26121 after their contents.
26122 * asan.c (asan_protect_global): Don't check TREE_ASM_WRITTEN here.
26123 (asan_finish_file): Test it here instead.
26124
26125 2013-01-07 Nick Clifton <nickc@redhat.com>
26126 Matthias Klose <doko@debian.org>
26127 Doug Kwan <dougkwan@google.com>
26128 H.J. Lu <hongjiu.lu@intel.com>
26129
26130 PR driver/55470
26131 * collect2.c (main): Support -fuse-ld=bfd and -fuse-ld=gold.
26132
26133 * common.opt: Add fuse-ld=bfd and fuse-ld=gold.
26134
26135 * gcc.c (LINK_COMMAND_SPEC): Pass -fuse-ld=* to collect2.
26136
26137 * opts.c (comman_handle_option): Ignore -fuse-ld=bfd and -fuse-ld=gold.
26138
26139 * doc/invoke.texi: Document -fuse-ld=bfd and -fuse-ld=gold.
26140
26141 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26142
26143 PR target/54461
26144 * doc/install.texi (Cross-Compiler-Specific Options): Document
26145 --with-avrlibc.
26146
26147 2013-01-07 Tejas Belagod <tejas.belagod@arm.com>
26148
26149 * config/aarch64/arm_neon.h (vmovn_high_is16, vmovn_high_s32,
26150 vmovn_high_s64, vmovn_high_u16, vmovn_high_u32, vmovn_high_u64,
26151 vqmovn_high_s16, vqmovn_high_s32, vqmovn_high_s64, vqmovn_high_u16,
26152 vqmovn_high_u32, vqmovn_high_u64, vqmovun_high_s16, vqmovun_high_s32,
26153 vqmovun_high_s64): Fix source operand number and update copyright.
26154
26155 2013-01-07 Richard Biener <rguenther@suse.de>
26156
26157 PR middle-end/55890
26158 * gimple.h (gimple_call_builtin_p): New overload.
26159 * gimple.c (validate_call): New function.
26160 (gimple_call_builtin_p): Likewise.
26161 * tree-ssa-structalias.c (find_func_aliases_for_builtin_call):
26162 Use gimple_call_builtin_p.
26163 (find_func_clobbers): Likewise.
26164 * tree-ssa-strlen.c (adjust_last_stmt): Likewise.
26165 (strlen_optimize_stmt): Likewise.
26166
26167 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26168
26169 * config/aarch64/arm_neon.h (vld1_dup_*): Make argument const.
26170 (vld1q_dup_*): Likewise.
26171 (vld1_*): Likewise.
26172 (vld1q_*): Likewise.
26173 (vld1_lane_*): Likewise.
26174 (vld1q_lane_*): Likewise.
26175
26176 2013-01-07 Richard Biener <rguenther@suse.de>
26177
26178 * lto-streamer.h (LTO_minor_version): Bump to 2.
26179
26180 2013-01-07 James Greenhalgh <james.greenhalgh@arm.com>
26181
26182 * config/aarch64/aarch64-protos.h
26183 (aarch64_const_double_zero_rtx_p): Rename to...
26184 (aarch64_float_const_zero_rtx_p): ...this.
26185 (aarch64_float_const_representable_p): New.
26186 (aarch64_output_simd_mov_immediate): Likewise.
26187 * config/aarch64/aarch64-simd.md (*aarch64_simd_mov<mode>): Refactor
26188 move immediate case.
26189 * config/aarch64/aarch64.c
26190 (aarch64_const_double_zero_rtx_p): Rename to...
26191 (aarch64_float_const_zero_rtx_p): ...this.
26192 (aarch64_print_operand): Allow printing of new constants.
26193 (aarch64_valid_floating_const): New.
26194 (aarch64_legitimate_constant_p): Check for valid floating-point
26195 constants.
26196 (aarch64_simd_valid_immediate): Likewise.
26197 (aarch64_vect_float_const_representable_p): New.
26198 (aarch64_float_const_representable_p): Likewise.
26199 (aarch64_simd_imm_zero_p): Also allow for floating-point 0.0.
26200 (aarch64_output_simd_mov_immediate): New.
26201 * config/aarch64/aarch64.md (*movsf_aarch64): Add new alternative.
26202 (*movdf_aarch64): Likewise.
26203 * config/aarch64/constraints.md (Ufc): New.
26204 (Y): call aarch64_float_const_zero_rtx.
26205 * config/aarch64/predicates.md (aarch64_fp_compare_operand): New.
26206
26207 2013-01-07 Richard Biener <rguenther@suse.de>
26208
26209 PR tree-optimization/55888
26210 PR tree-optimization/55862
26211 * tree-ssa-pre.c (phi_translate_1): Revert previous change.
26212 (valid_in_sets): Check if a NAME has a leader in AVAIL_OUT,
26213 not if it is contained therein.
26214
26215 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26216
26217 * config/avr/t-avr: Typo.
26218
26219 2013-01-07 Georg-Johann Lay <avr@gjlay.de>
26220
26221 PR55243
26222 * config/avr/t-avr: Don't automatically rebuild
26223 $(srcdir)/config/avr/t-multilib
26224 $(srcdir)/config/avr/avr-tables.opt
26225 $(srcdir)/doc/avr-mmcu.texi
26226 (avr-mcus): New phony target to build them on request.
26227 (s-avr-mlib, s-avr-mmcu-texi): Remove.
26228 * avr/avr-mcus.def: Adjust comments.
26229
26230 2013-01-07 Uros Bizjak <ubizjak@gmail.com>
26231
26232 * config/i386/i386.c (DEFAULT_PCC_STRUCT_RETURN): Remove.
26233
26234 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26235
26236 * file-find.c, file-find.h, realmpfr.c: Add FSF as copyright holder.
26237
26238 2013-01-06 Richard Sandiford <rdsandiford@googlemail.com>
26239
26240 * config/tilepro/gen-mul-tables.cc: Put copyright on one line.
26241
26242 2013-01-05 David Edelsohn <dje.gcc@gmail.com>
26243
26244 * config/rs6000/aix53.h (LIB_SPEC): Add -lpthreads when compiling
26245 to generate profiling.
26246 * config/rs6000/aix64.h (LIB_SPEC): Same.
26247
26248 2013-01-04 Andrew Pinski <apinski@cavium.com>
26249
26250 * config/aarch64/aarch64.c (aarch64_fixed_condition_code_regs):
26251 New function.
26252 (TARGET_FIXED_CONDITION_CODE_REGS): Define.
26253
26254 2013-01-04 Uros Bizjak <ubizjak@gmail.com>
26255
26256 * config/i386/i386.c (ix86_legitimize_address): Call convert_to_mode
26257 unconditionally.
26258 (ix86_expand_move): Ditto.
26259 (ix86_zero_extend_to_Pmode): Ditto.
26260 (ix86_expand_call): Ditto.
26261 (ix86_expand_special_args_builtin): Ditto.
26262 (ix86_expand_builtin): Ditto.
26263
26264 2013-01-04 Richard Biener <rguenther@suse.de>
26265
26266 PR tree-optimization/55862
26267 * tree-ssa-pre.c (phi_translate_1): Valueize SSA names after
26268 translating them through PHI nodes.
26269
26270 2013-01-04 Martin Jambor <mjambor@suse.cz>
26271
26272 PR tree-optimization/55755
26273 * tree-sra.c (sra_modify_assign): Do not check that an access has no
26274 children when trying to avoid producing a VIEW_CONVERT_EXPR.
26275
26276 2013-01-04 Marek Polacek <polacek@redhat.com>
26277
26278 PR middle-end/55859
26279 * opts.c (default_options_optimization): Clarify error message.
26280
26281 2013-01-04 Richard Biener <rguenther@suse.de>
26282
26283 PR middle-end/55863
26284 * fold-const.c (split_tree): Undo -X - 1 to ~X folding for
26285 reassociation.
26286
26287 2013-01-03 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26288
26289 PR target/53789
26290 * config/pa/pa.md (movsi): Revert previous change.
26291 * config/pa/pa.c (pa_legitimate_constant_p): Reject all TLS symbol
26292 references.
26293
26294 2013-01-03 Richard Henderson <rth@redhat.com>
26295
26296 * config/i386/i386.c (ix86_expand_move): Always assign to op1
26297 after eliminating TLS symbols.
26298
26299 2013-01-03 Marc Glisse <marc.glisse@inria.fr>
26300
26301 PR bootstrap/50167
26302 * graphite-interchange.c (pdr_stride_in_loop): Use gmp_fprintf.
26303 * graphite-poly.c (debug_gmp_value): Likewise.
26304
26305 2013-01-03 Uros Bizjak <ubizjak@gmail.com>
26306
26307 PR target/55712
26308 * config/i386/i386-c.c (ix86_target_macros_internal): Depending on
26309 selected code model, define __code_mode_small__, __code_model_medium__,
26310 __code_model_large__, __code_model_32__ or __code_model_kernel__.
26311 * config/i386/cpuid.h (__cpuid, __cpuid_count) [__i386__]: Prefix
26312 xchg temporary register with %k. Declare temporary register as
26313 early clobbered.
26314 [__x86_64__]: For medium and large code models, preserve %rbx register.
26315
26316 2013-01-03 Richard Biener <rguenther@suse.de>
26317
26318 * tree-data-ref.c (dump_conflict_function): Use less vertical spacing.
26319 (dump_subscript): Adjust.
26320 (finalize_ddr_dependent): Do not dump redundant info.
26321 (analyze_siv_subscript): Adjust.
26322 (subscript_dependence_tester): Likewise.
26323 (compute_affine_dependence): Likewise.
26324
26325 2013-01-03 Richard Biener <rguenther@suse.de>
26326
26327 Revert
26328 2013-01-03 Richard Biener <rguenther@suse.de>
26329
26330 PR tree-optimization/55857
26331 * tree-vect-stmts.c (vectorizable_load): Do not setup
26332 re-alignment for invariant loads.
26333
26334 2013-01-02 Richard Biener <rguenther@suse.de>
26335
26336 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26337 invariant load do not generate a vector load from the scalar location.
26338
26339 2013-01-03 Richard Biener <rguenther@suse.de>
26340
26341 * tree-vect-loop.c (vect_analyze_loop_form): Clarify reason
26342 for not vectorizing.
26343 * tree-vect-data-refs.c (vect_create_addr_base_for_vector_ref): Do
26344 not build INDIRECT_REFs, call get_name once only.
26345 (vect_create_data_ref_ptr): Likewise. Dump base object kind
26346 based on DR_BASE_OBJECT, not DR_BASE_ADDRESS.
26347
26348 2013-01-03 Richard Biener <rguenther@suse.de>
26349
26350 PR tree-optimization/55857
26351 * tree-vect-stmts.c (vectorizable_load): Do not setup
26352 re-alignment for invariant loads.
26353
26354 2013-01-03 Richard Biener <rguenther@suse.de>
26355
26356 PR lto/55848
26357 * lto-symtab.c (lto_symtab_merge_decls_1): As last resort, always
26358 prefer a built-in decl.
26359
26360 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26361
26362 * gcc.c (process_command): Update copyright notice dates.
26363 * gcov.c (print_version): Likewise.
26364 * gcov-dump.c (print_version): Likewise.
26365
26366 PR rtl-optimization/55838
26367 * loop-iv.c (iv_number_of_iterations): Call lowpart_subreg on
26368 iv0.step, iv1.step and step.
26369
26370 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26371 Marc Glisse <marc.glisse@inria.fr>
26372
26373 PR tree-optimization/55832
26374 * fold-const.c (fold_binary_loc): For ABS_EXPR<x> >= 0 and
26375 ABS_EXPR<x> < 0 folding use constant_boolean_node instead of
26376 integer_{one,zero}_node.
26377
26378 2013-01-03 Jakub Jelinek <jakub@redhat.com>
26379
26380 PR debug/54402
26381 * params.def (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE): New param.
26382 * var-tracking.c (reverse_op): Don't add reverse ops to
26383 VALUEs that have already
26384 PARAM_VALUE (PARAM_MAX_VARTRACK_REVERSE_OP_SIZE) or longer locs list.
26385
26386 2013-01-02 Gerald Pfeifer <gerald@pfeifer.com>
26387
26388 * doc/contrib.texi: Note years as release manager for Mark Mitchell.
26389
26390 2013-01-02 Teresa Johnson <tejohnson@google.com>
26391
26392 * dumpfile.c (dump_loc): Print filename with location.
26393 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Use
26394 new location_t parameter to emit complete unroll message with
26395 new dump framework.
26396 (canonicalize_loop_induction_variables): Compute loops location
26397 and pass to try_unroll_loop_completely.
26398 * loop-unroll.c (report_unroll_peel): New function.
26399 (peel_loops_completely): Use new dump format with location
26400 for main dumpfile message, and invoke report_unroll_peel on success.
26401 (decide_unrolling_and_peeling): Ditto.
26402 (decide_peel_once_rolling): Remove old dumpfile message subsumed
26403 by report_unroll_peel.
26404 (decide_peel_completely): Ditto.
26405 (decide_unroll_constant_iterations): Ditto.
26406 (decide_unroll_runtime_iterations): Ditto.
26407 (decide_peel_simple): Ditto.
26408 (decide_unroll_stupid): Ditto.
26409 * cfgloop.c (get_loop_location): New function.
26410 * cfgloop.h (get_loop_location): Declare.
26411
26412 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26413
26414 * config/i386/i386.c (fold_builtin_cpu): Remove unnecessary checks for
26415 NULL.
26416
26417 2013-01-02 John David Anglin <dave.anglin@nrc-cnrc.gc.ca>
26418
26419 PR middle-end/55198
26420 * expr.c (expand_expr_real_1): Don't use bitfield extraction for non
26421 BLKmode objects when EXPAND_MEMORY is specified.
26422
26423 2013-01-02 Sriraman Tallam <tmsriram@google.com>
26424
26425 * config/i386/i386.c (ix86_get_function_versions_dispatcher): Fix bug
26426 in loop predicate.
26427 (fold_builtin_cpu): Do not share cpu model decls across statements.
26428
26429 2013-01-02 Jason Merrill <jason@redhat.com>
26430
26431 PR c++/55804
26432 * tree.c (build_array_type_1): Revert earlier change.
26433
26434 2013-01-02 Yufeng Zhang <yufeng.zhang@arm.com>
26435
26436 * config/aarch64/aarch64-cores.def: Add entries for "cortex-a53" and
26437 "cortex-a57".
26438 * config/aarch64/aarch64-tune.md: Re-generate.
26439
26440 2013-01-02 Richard Biener <rguenther@suse.de>
26441
26442 * tree-vect-stmts.c (vectorizable_load): When vectorizing an
26443 invariant load do not generate a vector load from the scalar location.
26444
26445 2013-01-02 Richard Biener <rguenther@suse.de>
26446
26447 PR bootstrap/55784
26448 * configure.ac: Add $GMPINC to CFLAGS/CXXFLAGS.
26449 * configure: Regenerate.
26450
26451 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26452
26453 * builtins.c (expand_builtin_mathfn, expand_builtin_mathfn_2)
26454 (expand_builtin_mathfn_ternary, expand_builtin_mathfn_3)
26455 (expand_builtin_int_roundingfn_2): Keep the original target around
26456 for the fallback case.
26457
26458 2013-01-02 Richard Sandiford <rdsandiford@googlemail.com>
26459
26460 * tree-vrp.c (range_fits_type_p): Require the MSB of the double_int
26461 to be clear for sign changes.
26462
26463 2013-01-01 Jan Hubicka <jh@suse.cz>
26464
26465 * ipa-inline-analysis.c: Fix formatting.
26466
26467 2013-01-01 Jakub Jelinek <jakub@redhat.com>
26468
26469 PR tree-optimization/55831
26470 * tree-vect-loop.c (get_initial_def_for_induction): Use
26471 gsi_after_labels instead of gsi_start_bb.
26472 \f
26473 Copyright (C) 2013 Free Software Foundation, Inc.
26474
26475 Copying and distribution of this file, with or without modification,
26476 are permitted in any medium without royalty provided the copyright
26477 notice and this notice are preserved.