1 2024-03-04 David Faust <david.faust@oracle.com>
3 * config/bpf/bpf-protos.h (bpf_expand_setmem): New prototype.
4 * config/bpf/bpf.cc (bpf_expand_setmem): New.
5 * config/bpf/bpf.md (setmemdi): New define_expand.
7 2024-03-04 Jakub Jelinek <jakub@redhat.com>
9 PR rtl-optimization/113010
10 * combine.cc (simplify_comparison): Guard the
11 WORD_REGISTER_OPERATIONS check on scalar_int_mode of SUBREG_REG
12 and initialize inner_mode.
14 2024-03-04 Andre Vieira <andre.simoesdiasvieira@arm.com>
16 * config/arm/iterators.md (supf): Remove VMLALDAVXQ_U, VMLALDAVXQ_P_U,
18 (VMLALDAVXQ): Remove iterator.
19 (VMLALDAVXQ_P): Likewise.
20 (VMLALDAVAXQ): Likewise.
21 * config/arm/mve.md (mve_vstrwq_p_fv4sf): Replace use of <MVE_VPRED>
22 mode iterator attribute with V4BI mode.
23 * config/arm/unspecs.md (VMLALDAVXQ_U, VMLALDAVXQ_P_U,
24 VMLALDAVAXQ_U): Remove unused unspecs.
26 2024-03-04 Andre Vieira <andre.simoesdiasvieira@arm.com>
28 * config/arm/arm.md (mve_safe_imp_xlane_pred): New attribute.
29 * config/arm/iterators.md (mve_vmaxmin_safe_imp): New iterator
31 * config/arm/mve.md (vaddvq_s, vaddvq_u, vaddlvq_s, vaddlvq_u,
32 vaddvaq_s, vaddvaq_u, vmaxavq_s, vmaxvq_u, vmladavq_s, vmladavq_u,
33 vmladavxq_s, vmlsdavq_s, vmlsdavxq_s, vaddlvaq_s, vaddlvaq_u,
34 vmlaldavq_u, vmlaldavq_s, vmlaldavq_u, vmlaldavxq_s, vmlsldavq_s,
35 vmlsldavxq_s, vrmlaldavhq_u, vrmlaldavhq_s, vrmlaldavhxq_s,
36 vrmlsldavhq_s, vrmlsldavhxq_s, vrmlaldavhaq_s, vrmlaldavhaq_u,
37 vrmlaldavhaxq_s, vrmlsldavhaq_s, vrmlsldavhaxq_s, vabavq_s, vabavq_u,
38 vmladavaq_u, vmladavaq_s, vmladavaxq_s, vmlsdavaq_s, vmlsdavaxq_s,
39 vmlaldavaq_s, vmlaldavaq_u, vmlaldavaxq_s, vmlsldavaq_s,
40 vmlsldavaxq_s): Added mve_safe_imp_xlane_pred.
42 2024-03-04 Stam Markianos-Wright <stam.markianos-wright@arm.com>
44 * config/arm/arm.md (mve_unpredicated_insn): New attribute.
45 * config/arm/arm.h (MVE_VPT_PREDICATED_INSN_P): New define.
46 (MVE_VPT_UNPREDICATED_INSN_P): Likewise.
47 (MVE_VPT_PREDICABLE_INSN_P): Likewise.
48 * config/arm/vec-common.md (mve_vshlq_<supf><mode>): Add attribute.
49 * config/arm/mve.md (arm_vcx1q<a>_p_v16qi): Add attribute.
50 (arm_vcx1q<a>v16qi): Likewise.
51 (arm_vcx1qav16qi): Likewise.
52 (arm_vcx1qv16qi): Likewise.
53 (arm_vcx2q<a>_p_v16qi): Likewise.
54 (arm_vcx2q<a>v16qi): Likewise.
55 (arm_vcx2qav16qi): Likewise.
56 (arm_vcx2qv16qi): Likewise.
57 (arm_vcx3q<a>_p_v16qi): Likewise.
58 (arm_vcx3q<a>v16qi): Likewise.
59 (arm_vcx3qav16qi): Likewise.
60 (arm_vcx3qv16qi): Likewise.
61 (@mve_<mve_insn>q_<supf><mode>): Likewise.
62 (@mve_<mve_insn>q_int_<supf><mode>): Likewise.
63 (@mve_<mve_insn>q_<supf>v4si): Likewise.
64 (@mve_<mve_insn>q_n_<supf><mode>): Likewise.
65 (@mve_<mve_insn>q_r_<supf><mode>): Likewise.
66 (@mve_<mve_insn>q_f<mode>): Likewise.
67 (@mve_<mve_insn>q_m_<supf><mode>): Likewise.
68 (@mve_<mve_insn>q_m_n_<supf><mode>): Likewise.
69 (@mve_<mve_insn>q_m_r_<supf><mode>): Likewise.
70 (@mve_<mve_insn>q_m_f<mode>): Likewise.
71 (@mve_<mve_insn>q_int_m_<supf><mode>): Likewise.
72 (@mve_<mve_insn>q_p_<supf>v4si): Likewise.
73 (@mve_<mve_insn>q_p_<supf><mode>): Likewise.
74 (@mve_<mve_insn>q<mve_rot>_<supf><mode>): Likewise.
75 (@mve_<mve_insn>q<mve_rot>_f<mode>): Likewise.
76 (@mve_<mve_insn>q<mve_rot>_m_<supf><mode>): Likewise.
77 (@mve_<mve_insn>q<mve_rot>_m_f<mode>): Likewise.
78 (mve_v<absneg_str>q_f<mode>): Likewise.
79 (mve_<mve_addsubmul>q<mode>): Likewise.
80 (mve_<mve_addsubmul>q_f<mode>): Likewise.
81 (mve_vadciq_<supf>v4si): Likewise.
82 (mve_vadciq_m_<supf>v4si): Likewise.
83 (mve_vadcq_<supf>v4si): Likewise.
84 (mve_vadcq_m_<supf>v4si): Likewise.
85 (mve_vandq_<supf><mode>): Likewise.
86 (mve_vandq_f<mode>): Likewise.
87 (mve_vandq_m_<supf><mode>): Likewise.
88 (mve_vandq_m_f<mode>): Likewise.
89 (mve_vandq_s<mode>): Likewise.
90 (mve_vandq_u<mode>): Likewise.
91 (mve_vbicq_<supf><mode>): Likewise.
92 (mve_vbicq_f<mode>): Likewise.
93 (mve_vbicq_m_<supf><mode>): Likewise.
94 (mve_vbicq_m_f<mode>): Likewise.
95 (mve_vbicq_m_n_<supf><mode>): Likewise.
96 (mve_vbicq_n_<supf><mode>): Likewise.
97 (mve_vbicq_s<mode>): Likewise.
98 (mve_vbicq_u<mode>): Likewise.
99 (@mve_vclzq_s<mode>): Likewise.
100 (mve_vclzq_u<mode>): Likewise.
101 (@mve_vcmp_<mve_cmp_op>q_<mode>): Likewise.
102 (@mve_vcmp_<mve_cmp_op>q_n_<mode>): Likewise.
103 (@mve_vcmp_<mve_cmp_op>q_f<mode>): Likewise.
104 (@mve_vcmp_<mve_cmp_op>q_n_f<mode>): Likewise.
105 (@mve_vcmp_<mve_cmp_op1>q_m_f<mode>): Likewise.
106 (@mve_vcmp_<mve_cmp_op1>q_m_n_<supf><mode>): Likewise.
107 (@mve_vcmp_<mve_cmp_op1>q_m_<supf><mode>): Likewise.
108 (@mve_vcmp_<mve_cmp_op1>q_m_n_f<mode>): Likewise.
109 (mve_vctp<MVE_vctp>q<MVE_vpred>): Likewise.
110 (mve_vctp<MVE_vctp>q_m<MVE_vpred>): Likewise.
111 (mve_vcvtaq_<supf><mode>): Likewise.
112 (mve_vcvtaq_m_<supf><mode>): Likewise.
113 (mve_vcvtbq_f16_f32v8hf): Likewise.
114 (mve_vcvtbq_f32_f16v4sf): Likewise.
115 (mve_vcvtbq_m_f16_f32v8hf): Likewise.
116 (mve_vcvtbq_m_f32_f16v4sf): Likewise.
117 (mve_vcvtmq_<supf><mode>): Likewise.
118 (mve_vcvtmq_m_<supf><mode>): Likewise.
119 (mve_vcvtnq_<supf><mode>): Likewise.
120 (mve_vcvtnq_m_<supf><mode>): Likewise.
121 (mve_vcvtpq_<supf><mode>): Likewise.
122 (mve_vcvtpq_m_<supf><mode>): Likewise.
123 (mve_vcvtq_from_f_<supf><mode>): Likewise.
124 (mve_vcvtq_m_from_f_<supf><mode>): Likewise.
125 (mve_vcvtq_m_n_from_f_<supf><mode>): Likewise.
126 (mve_vcvtq_m_n_to_f_<supf><mode>): Likewise.
127 (mve_vcvtq_m_to_f_<supf><mode>): Likewise.
128 (mve_vcvtq_n_from_f_<supf><mode>): Likewise.
129 (mve_vcvtq_n_to_f_<supf><mode>): Likewise.
130 (mve_vcvtq_to_f_<supf><mode>): Likewise.
131 (mve_vcvttq_f16_f32v8hf): Likewise.
132 (mve_vcvttq_f32_f16v4sf): Likewise.
133 (mve_vcvttq_m_f16_f32v8hf): Likewise.
134 (mve_vcvttq_m_f32_f16v4sf): Likewise.
135 (mve_vdwdupq_m_wb_u<mode>_insn): Likewise.
136 (mve_vdwdupq_wb_u<mode>_insn): Likewise.
137 (mve_veorq_s><mode>): Likewise.
138 (mve_veorq_u><mode>): Likewise.
139 (mve_veorq_f<mode>): Likewise.
140 (mve_vidupq_m_wb_u<mode>_insn): Likewise.
141 (mve_vidupq_u<mode>_insn): Likewise.
142 (mve_viwdupq_m_wb_u<mode>_insn): Likewise.
143 (mve_viwdupq_wb_u<mode>_insn): Likewise.
144 (mve_vldrbq_<supf><mode>): Likewise.
145 (mve_vldrbq_gather_offset_<supf><mode>): Likewise.
146 (mve_vldrbq_gather_offset_z_<supf><mode>): Likewise.
147 (mve_vldrbq_z_<supf><mode>): Likewise.
148 (mve_vldrdq_gather_base_<supf>v2di): Likewise.
149 (mve_vldrdq_gather_base_wb_<supf>v2di_insn): Likewise.
150 (mve_vldrdq_gather_base_wb_z_<supf>v2di_insn): Likewise.
151 (mve_vldrdq_gather_base_z_<supf>v2di): Likewise.
152 (mve_vldrdq_gather_offset_<supf>v2di): Likewise.
153 (mve_vldrdq_gather_offset_z_<supf>v2di): Likewise.
154 (mve_vldrdq_gather_shifted_offset_<supf>v2di): Likewise.
155 (mve_vldrdq_gather_shifted_offset_z_<supf>v2di): Likewise.
156 (mve_vldrhq_<supf><mode>): Likewise.
157 (mve_vldrhq_fv8hf): Likewise.
158 (mve_vldrhq_gather_offset_<supf><mode>): Likewise.
159 (mve_vldrhq_gather_offset_fv8hf): Likewise.
160 (mve_vldrhq_gather_offset_z_<supf><mode>): Likewise.
161 (mve_vldrhq_gather_offset_z_fv8hf): Likewise.
162 (mve_vldrhq_gather_shifted_offset_<supf><mode>): Likewise.
163 (mve_vldrhq_gather_shifted_offset_fv8hf): Likewise.
164 (mve_vldrhq_gather_shifted_offset_z_<supf><mode>): Likewise.
165 (mve_vldrhq_gather_shifted_offset_z_fv8hf): Likewise.
166 (mve_vldrhq_z_<supf><mode>): Likewise.
167 (mve_vldrhq_z_fv8hf): Likewise.
168 (mve_vldrwq_<supf>v4si): Likewise.
169 (mve_vldrwq_fv4sf): Likewise.
170 (mve_vldrwq_gather_base_<supf>v4si): Likewise.
171 (mve_vldrwq_gather_base_fv4sf): Likewise.
172 (mve_vldrwq_gather_base_wb_<supf>v4si_insn): Likewise.
173 (mve_vldrwq_gather_base_wb_fv4sf_insn): Likewise.
174 (mve_vldrwq_gather_base_wb_z_<supf>v4si_insn): Likewise.
175 (mve_vldrwq_gather_base_wb_z_fv4sf_insn): Likewise.
176 (mve_vldrwq_gather_base_z_<supf>v4si): Likewise.
177 (mve_vldrwq_gather_base_z_fv4sf): Likewise.
178 (mve_vldrwq_gather_offset_<supf>v4si): Likewise.
179 (mve_vldrwq_gather_offset_fv4sf): Likewise.
180 (mve_vldrwq_gather_offset_z_<supf>v4si): Likewise.
181 (mve_vldrwq_gather_offset_z_fv4sf): Likewise.
182 (mve_vldrwq_gather_shifted_offset_<supf>v4si): Likewise.
183 (mve_vldrwq_gather_shifted_offset_fv4sf): Likewise.
184 (mve_vldrwq_gather_shifted_offset_z_<supf>v4si): Likewise.
185 (mve_vldrwq_gather_shifted_offset_z_fv4sf): Likewise.
186 (mve_vldrwq_z_<supf>v4si): Likewise.
187 (mve_vldrwq_z_fv4sf): Likewise.
188 (mve_vmvnq_s<mode>): Likewise.
189 (mve_vmvnq_u<mode>): Likewise.
190 (mve_vornq_<supf><mode>): Likewise.
191 (mve_vornq_f<mode>): Likewise.
192 (mve_vornq_m_<supf><mode>): Likewise.
193 (mve_vornq_m_f<mode>): Likewise.
194 (mve_vornq_s<mode>): Likewise.
195 (mve_vornq_u<mode>): Likewise.
196 (mve_vorrq_<supf><mode>): Likewise.
197 (mve_vorrq_f<mode>): Likewise.
198 (mve_vorrq_m_<supf><mode>): Likewise.
199 (mve_vorrq_m_f<mode>): Likewise.
200 (mve_vorrq_m_n_<supf><mode>): Likewise.
201 (mve_vorrq_n_<supf><mode>): Likewise.
202 (mve_vorrq_s<mode>): Likewise.
203 (mve_vorrq_s<mode>): Likewise.
204 (mve_vsbciq_<supf>v4si): Likewise.
205 (mve_vsbciq_m_<supf>v4si): Likewise.
206 (mve_vsbcq_<supf>v4si): Likewise.
207 (mve_vsbcq_m_<supf>v4si): Likewise.
208 (mve_vshlcq_<supf><mode>): Likewise.
209 (mve_vshlcq_m_<supf><mode>): Likewise.
210 (mve_vshrq_m_n_<supf><mode>): Likewise.
211 (mve_vshrq_n_<supf><mode>): Likewise.
212 (mve_vstrbq_<supf><mode>): Likewise.
213 (mve_vstrbq_p_<supf><mode>): Likewise.
214 (mve_vstrbq_scatter_offset_<supf><mode>_insn): Likewise.
215 (mve_vstrbq_scatter_offset_p_<supf><mode>_insn): Likewise.
216 (mve_vstrdq_scatter_base_<supf>v2di): Likewise.
217 (mve_vstrdq_scatter_base_p_<supf>v2di): Likewise.
218 (mve_vstrdq_scatter_base_wb_<supf>v2di): Likewise.
219 (mve_vstrdq_scatter_base_wb_p_<supf>v2di): Likewise.
220 (mve_vstrdq_scatter_offset_<supf>v2di_insn): Likewise.
221 (mve_vstrdq_scatter_offset_p_<supf>v2di_insn): Likewise.
222 (mve_vstrdq_scatter_shifted_offset_<supf>v2di_insn): Likewise.
223 (mve_vstrdq_scatter_shifted_offset_p_<supf>v2di_insn): Likewise.
224 (mve_vstrhq_<supf><mode>): Likewise.
225 (mve_vstrhq_fv8hf): Likewise.
226 (mve_vstrhq_p_<supf><mode>): Likewise.
227 (mve_vstrhq_p_fv8hf): Likewise.
228 (mve_vstrhq_scatter_offset_<supf><mode>_insn): Likewise.
229 (mve_vstrhq_scatter_offset_fv8hf_insn): Likewise.
230 (mve_vstrhq_scatter_offset_p_<supf><mode>_insn): Likewise.
231 (mve_vstrhq_scatter_offset_p_fv8hf_insn): Likewise.
232 (mve_vstrhq_scatter_shifted_offset_<supf><mode>_insn): Likewise.
233 (mve_vstrhq_scatter_shifted_offset_fv8hf_insn): Likewise.
234 (mve_vstrhq_scatter_shifted_offset_p_<supf><mode>_insn): Likewise.
235 (mve_vstrhq_scatter_shifted_offset_p_fv8hf_insn): Likewise.
236 (mve_vstrwq_<supf>v4si): Likewise.
237 (mve_vstrwq_fv4sf): Likewise.
238 (mve_vstrwq_p_<supf>v4si): Likewise.
239 (mve_vstrwq_p_fv4sf): Likewise.
240 (mve_vstrwq_scatter_base_<supf>v4si): Likewise.
241 (mve_vstrwq_scatter_base_fv4sf): Likewise.
242 (mve_vstrwq_scatter_base_p_<supf>v4si): Likewise.
243 (mve_vstrwq_scatter_base_p_fv4sf): Likewise.
244 (mve_vstrwq_scatter_base_wb_<supf>v4si): Likewise.
245 (mve_vstrwq_scatter_base_wb_fv4sf): Likewise.
246 (mve_vstrwq_scatter_base_wb_p_<supf>v4si): Likewise.
247 (mve_vstrwq_scatter_base_wb_p_fv4sf): Likewise.
248 (mve_vstrwq_scatter_offset_<supf>v4si_insn): Likewise.
249 (mve_vstrwq_scatter_offset_fv4sf_insn): Likewise.
250 (mve_vstrwq_scatter_offset_p_<supf>v4si_insn): Likewise.
251 (mve_vstrwq_scatter_offset_p_fv4sf_insn): Likewise.
252 (mve_vstrwq_scatter_shifted_offset_<supf>v4si_insn): Likewise.
253 (mve_vstrwq_scatter_shifted_offset_fv4sf_insn): Likewise.
254 (mve_vstrwq_scatter_shifted_offset_p_<supf>v4si_insn): Likewise.
255 (mve_vstrwq_scatter_shifted_offset_p_fv4sf_insn): Likewise.
257 2024-03-04 Marek Polacek <polacek@redhat.com>
259 * doc/extend.texi: Update [[gnu::no_dangling]].
261 2024-03-04 Andrew Stubbs <ams@baylibre.com>
263 * dojump.cc (do_compare_and_jump): Use full-width integers for shifts.
264 * expr.cc (store_constructor): Likewise.
265 (do_store_flag): Likewise.
267 2024-03-04 Mark Wielaard <mark@klomp.org>
269 * common.opt.urls: Regenerate.
270 * config/avr/avr.opt.urls: Likewise.
271 * config/i386/i386.opt.urls: Likewise.
272 * config/pru/pru.opt.urls: Likewise.
273 * config/riscv/riscv.opt.urls: Likewise.
274 * config/rs6000/rs6000.opt.urls: Likewise.
276 2024-03-04 Richard Biener <rguenther@suse.de>
278 PR tree-optimization/114197
279 * tree-if-conv.cc (bitfields_to_lower_p): Do not lower if
280 there are volatile bitfield accesses.
281 (pass_if_conversion::execute): Throw away result if the
282 if-converted and original loops are not nested as expected.
284 2024-03-04 Richard Biener <rguenther@suse.de>
286 PR tree-optimization/114164
287 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Fail if
288 the code generated for mask argument setup is not supported.
290 2024-03-04 Richard Biener <rguenther@suse.de>
292 PR tree-optimization/114203
293 * tree-ssa-loop-niter.cc (build_cltz_expr): Apply CTZ->CLZ
294 adjustment before making the result defined at zero.
296 2024-03-04 Richard Biener <rguenther@suse.de>
298 PR tree-optimization/114192
299 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Use the
300 appropriate def for the live out stmt in case of an alternate
303 2024-03-04 Jakub Jelinek <jakub@redhat.com>
306 * gimple-lower-bitint.cc (bitint_large_huge::limb_access): Call
307 unshare_expr when creating a MEM_REF from MEM_REF.
308 (bitint_large_huge::lower_stmt): Call unshare_expr.
310 2024-03-04 Jakub Jelinek <jakub@redhat.com>
313 * config/i386/i386-expand.cc (ix86_expand_move): If XFmode op1
314 is SUBREG of CONSTANT_P, force the SUBREG_REG into memory or
317 2024-03-04 Roger Sayle <roger@nextmovesoftware.com>
320 * simplify-rtx.cc (simplify_context::simplify_subreg): Call
321 lowpart_subreg to perform type conversion, to avoid confusion
322 over the offset to use in the call to simplify_reg_subreg.
324 2024-03-03 Greg McGary <gkm@rivosinc.com>
326 PR rtl-optimization/113010
327 * combine.cc (simplify_comparison): Simplify a SUBREG on
328 WORD_REGISTER_OPERATIONS targets only if it is a zero-extending
331 2024-03-03 Georg-Johann Lay <avr@gjlay.de>
333 * config/avr/avr.cc: Resolve ATTRIBUTE_UNUSED.
334 Use bool in place of int for boolean logic (if possible).
335 Move declarations to definitions (if possible).
336 * config/avr/avr.md: Use C++ comments. Fix some indentation glitches.
337 * config/avr/avr-dimode.md: Same.
338 * config/avr/constraints.md: Same.
339 * config/avr/predicates.md: Same.
341 2024-03-03 Uros Bizjak <ubizjak@gmail.com>
344 * config/alpha/alpha.md (umuldi3_highpart): Remove expander.
345 (*umuldi3_highpart_reg): Rename to umuldi3_highpart and
346 simplify insn RTX using UMUL_HIGHPART rtx_code.
347 (*umuldi3_highpart_const): Remove.
349 2024-03-03 Georg-Johann Lay <avr@gjlay.de>
352 * config/avr/avr-protos.h (_reg_unused_after): Remove proto.
353 * config/avr/avr.cc (_reg_unused_after): Make static. And
354 add 3rd argument to skip the current insn.
355 (reg_unused_after): Adjust call of reg_unused_after.
356 (avr_out_plus_1) [AVR_TINY && -mfuse-add >= 2]: Don't output
357 unneeded frame pointer adjustments.
359 2024-03-03 Georg-Johann Lay <avr@gjlay.de>
362 * config/avr/avr.md (define_attr "cc"): Remove.
363 * config/avr/avr-protos.h (avr_out_plus): Remove pcc argument
365 * config/avr/avr.cc (avr_out_plus_1): Remove pcc argument and
366 its uses. Add insn argument.
367 (avr_out_plus_symbol): Remove pcc argument and its uses.
368 (avr_out_plus): Remove pcc argument and its uses.
369 Adjust calls of avr_out_plus_symbol and avr_out_plus_1.
370 (avr_out_round): Adjust call of avr_out_plus.
372 2024-03-03 Georg-Johann Lay <avr@gjlay.de>
374 * config/avr/avr.cc (avr_init_cumulative_args): Fix a typo
377 2024-03-03 Oleg Endo <olegendo@gcc.gnu.org>
380 * config/sh/sh.cc (sh_is_nott_insn): Handle case where the input
381 is not an insn, but e.g. a code label.
383 2024-03-02 Georg-Johann Lay <avr@gjlay.de>
385 * config/avr/avr.md (REG_0, ... REG_36): New define_constants.
386 * config/avr/avr.cc: Use them instead of magic numbers when it
387 means a register number.
389 2024-03-02 Georg-Johann Lay <avr@gjlay.de>
391 * config/avr/avr.cc: Adjust some comments.
393 2024-03-02 Georg-Johann Lay <avr@gjlay.de>
396 * config/avr/avr.cc (avr_out_plus_1) [-mtiny-stack]: Only adjust
397 the low part of the frame pointer with 8-bit stack pointer.
399 2024-03-01 Patrick Palka <ppalka@redhat.com>
403 * tree-inline.cc (remap_decl): Handle copy_decl returning the
405 (remap_decls): Handle remap_decl returning the original decl.
406 (copy_fn): Adjust copy_decl callback to skip TYPE_DECL and
409 2024-03-01 Jeff Law <jlaw@ventanamicro.com>
411 * config/riscv/riscv.md (zero_extendqi<SUPERQI:mode>2_internal): Fix
413 (extendsidi2_internal, movhf_hardfloat, movhf_softfloat): Likewise.
414 (movdi_32bit, movdi_64bit, movsi_internal): Likewise.
415 (movhi_internal, movqi_internal): Likewise.
416 (movsf_softfloat, movsf_hardfloat): Likewise.
417 (movdf_hardfloat_rv32, movdf_hardfloat_rv64): Likewise.
418 (movdf_softfloat): Likewise.
420 2024-03-01 Marek Polacek <polacek@redhat.com>
424 * doc/extend.texi: Document gnu::no_dangling.
425 * doc/invoke.texi: Mention that gnu::no_dangling disables
426 -Wdangling-reference.
428 2024-03-01 Georg-Johann Lay <avr@gjlay.de>
430 * config/avr/avr.opt: Overhaul help screen.
432 2024-03-01 Jakub Jelinek <jakub@redhat.com>
433 Tobias Burnus <tburnus@baylibre.com>
436 * gimplify.cc (omp_notice_variable): Fix 'shared' arg to
437 lang_hooks.decls.omp_disregard_value_expr for
438 (first)private in target regions.
440 2024-03-01 Jakub Jelinek <jakub@redhat.com>
443 * calls.cc (expand_call): For TYPE_NO_NAMED_ARGS_STDARG_P set
444 n_named_args initially before INIT_CUMULATIVE_ARGS to
445 structure_value_addr_parm rather than 0, after it don't modify
446 it if strict_argument_naming and clear only if
447 !pretend_outgoing_varargs_named.
449 2024-03-01 Jakub Jelinek <jakub@redhat.com>
452 * dwarf2out.cc (should_move_die_to_comdat): Return false for
453 aggregates without DW_AT_byte_size attribute or with non-constant
456 2024-03-01 Georg-Johann Lay <avr@gjlay.de>
458 * doc/invoke.texi (AVR Options) <-mfuse-add=level>: Document
459 valid values for level.
461 2024-03-01 Richard Biener <rguenther@suse.de>
464 * match.pd ((c ? a : b) op d --> c ? (a op d) : (b op d)):
465 Allow the folding if before lowering and the current IL
466 isn't supported with vcond_mask.
468 2024-03-01 xuli <xuli1@eswincomputing.com>
470 * config/riscv/riscv.cc (TARGET_GNU_ATTRIBUTES): Add riscv_vector_cc
471 attribute to riscv_attribute_table.
472 (riscv_vector_cc_function_p): Return true if FUNC is a riscv_vector_cc function.
473 (riscv_fntype_abi): Add riscv_vector_cc attribute check.
474 * doc/extend.texi: Add riscv_vector_cc attribute description.
476 2024-03-01 Pan Li <pan2.li@intel.com>
479 * config/riscv/riscv-avlprop.cc (pass_avlprop::execute): Replace
480 RVV_FIXED_VLMAX to RVV_VECTOR_BITS_ZVL.
481 * config/riscv/riscv-opts.h (enum riscv_autovec_preference_enum): Remove.
482 (enum rvv_vector_bits_enum): New enum for different RVV vector bits.
483 * config/riscv/riscv-selftests.cc (riscv_run_selftests): Update
484 comments for option replacement.
485 * config/riscv/riscv-v.cc (autovec_use_vlmax_p): Replace enum of
486 riscv_autovec_preference to rvv_vector_bits.
487 (vls_mode_valid_p): Ditto.
488 (estimated_poly_value): Ditto.
489 * config/riscv/riscv.cc (riscv_convert_vector_chunks): Rename to
490 vector chunks and honor new option mrvv-vector-bits.
491 (riscv_override_options_internal): Update comments and rename the
493 * config/riscv/riscv.opt: Add option mrvv-vector-bits and remove
494 internal option param=riscv-autovec-preference.
496 2024-03-01 Jakub Jelinek <jakub@redhat.com>
498 * function.cc (assign_parms): Only call assign_parms_setup_varargs
499 early for TYPE_NO_NAMED_ARGS_STDARG_P functions if fnargs is empty.
501 2024-03-01 Jakub Jelinek <jakub@redhat.com>
504 * gimple-lower-bitint.cc (bitint_large_huge::lower_stmt): Allow
505 rhs1 of a VCE to have no underlying variable if it is a load and
508 2024-02-29 David Malcolm <dmalcolm@redhat.com>
511 * function.cc (function_name): Make param const.
512 * function.h (function_name): Likewise.
514 2024-02-29 Georg-Johann Lay <avr@gjlay.de>
517 * doc/invoke.texi (AVR Options) <-mfuse-add>: Document.
518 * config/avr/avr.opt (-mfuse-add=): New target option.
519 * common/config/avr/avr-common.cc (avr_option_optimization_table)
520 [OPT_LEVELS_1_PLUS]: Set -mfuse-add=1.
521 [OPT_LEVELS_2_PLUS]: Set -mfuse-add=2.
522 * config/avr/avr-passes.def (avr_pass_fuse_add): Insert new pass.
523 * config/avr/avr-protos.h (avr_split_tiny_move)
524 (make_avr_pass_fuse_add): New protos.
525 * config/avr/avr.md [AVR_TINY]: New post-reload splitter uses
526 avr_split_tiny_move to split indirect memory accesses.
527 (gen_move_clobbercc): New define_expand helper.
528 * config/avr/avr.cc (avr_pass_data_fuse_add): New pass data.
529 (avr_pass_fuse_add): New class from rtl_opt_pass.
530 (make_avr_pass_fuse_add, avr_split_tiny_move): New functions.
531 (reg_seen_between_p, emit_move_ccc, emit_move_ccc_after): New functions.
532 (avr_legitimate_address_p) [AVR_TINY]: Don't restrict offsets
533 of PLUS addressing for AVR_TINY.
534 (avr_regno_mode_code_ok_for_base_p) [AVR_TINY]: Ignore -mstrict-X.
535 (avr_out_plus_1) [AVR_TINY]: Tweak ++Y and --Y.
536 (avr_mode_code_base_reg_class) [AVR_TINY]: Always return POINTER_REGS.
538 2024-02-29 Georg-Johann Lay <avr@gjlay.de>
541 * config/avr/avr.h (CUMULATIVE_ARGS) <has_stack_args>: New field.
542 * config/avr/avr.cc (avr_init_cumulative_args): Initialize it.
543 (avr_function_arg): Set it.
544 (avr_frame_pointer_required_p): Use it instead of .nregs.
546 2024-02-29 Andrew Pinski <quic_apinski@quicinc.com>
549 * config/aarch64/aarch64-builtins.cc (aarch64_memtag_builtin_data): Make
550 static and mark with GTY.
552 2024-02-29 Xi Ruoyao <xry111@xry111.site>
554 * config/loongarch/loongarch.md
555 (loongarch_<crc>_w_<size>_w_extended): New define_insn.
557 2024-02-29 Xi Ruoyao <xry111@xry111.site>
559 * config/loongarch/loongarch.md (CRC): New define_int_iterator.
560 (crc): New define_int_attr.
561 (loongarch_crc_w_<size>_w, loongarch_crcc_w_<size>_w): Unify
563 (loongarch_<crc>_w_<size>_w): ... here.
565 2024-02-29 Kito Cheng <kito.cheng@sifive.com>
568 * config/riscv/sync.md (atomic_compare_and_swap<mode>): Sign
569 extend the expected value if needed.
571 2024-02-28 Cupertino Miranda <cupertino.miranda@oracle.com>
573 * config.gcc (target_gtfiles): Change coreout to btfext-out.
574 (extra_objs): Change coreout to btfext-out.
575 * config/bpf/coreout.cc: Rename to btfext-out.cc.
576 * config/bpf/btfext-out.cc: Add.
577 * config/bpf/coreout.h: Rename to btfext-out.h.
578 * config/bpf/btfext-out.h: Add.
579 * config/bpf/core-builtins.cc: Change include.
580 * config/bpf/core-builtins.h: Change include.
581 * config/bpf/t-bpf: Accomodate renamed files.
583 2024-02-28 Cupertino Miranda <cupertino.miranda@oracle.com>
586 * config/bpf/bpf.cc (bpf_function_prologue): Define target
588 * config/bpf/coreout.cc (brf_ext_info_section)
589 (btf_ext_info): Move from coreout.h
590 (btf_ext_funcinfo, btf_ext_lineinfo): Add struct.
591 (bpf_core_reloc): Rename to btf_ext_core_reloc.
592 (btf_ext): Add static variable.
593 (btfext_info_sec_find_or_add, SEARCH_NODE_AND_RETURN)
594 (bpf_create_or_find_funcinfo, bpt_create_core_reloc)
595 (btf_ext_add_string, btf_funcinfo_type_callback)
596 (btf_add_func_info_for, btf_validate_funcinfo)
597 (btf_ext_info_len, output_btfext_func_info): Add function.
598 (output_btfext_header, bpf_core_reloc_add)
599 (output_btfext_core_relocs, btf_ext_init, btf_ext_output):
600 Change to support new structs.
601 * config/bpf/coreout.h (btf_ext_funcinfo, btf_ext_lineinfo):
602 Move and change in coreout.cc.
603 (btf_add_func_info_for, btf_ext_add_string): Add prototypes.
605 2024-02-28 Cupertino Miranda <cupertino.miranda@oracle.com>
607 * config/bpf/bpf.cc (bpf_option_override): Make .BTF.ext
608 enabled by default for BPF.
609 (bpf_file_end): Call BTF deallocation.
610 (bpf_asm_init_sections): Correct condition.
611 * dwarf2ctf.cc (ctf_debug_finalize): Conditionally execute BTF
613 (ctf_debuf_finish): Correct condition for calling
616 2024-02-28 Cupertino Miranda <cupertino.miranda@oracle.com>
618 * btfout.cc (output_btf_func_types): Use FOR_EACH_VEC_ELT.
619 (traverse_btf_func_types): Define function.
620 * ctfc.h (funcs_traverse_callback): Typedef for function
622 (traverse_btf_func_types): Add prototype.
624 2024-02-28 Cupertino Miranda <cupertino.miranda@oracle.com>
626 * btfout.cc (btf_collect_dataset): Corrects BTF type id.
628 2024-02-28 Richard Biener <rguenther@suse.de>
630 PR tree-optimization/113831
631 PR tree-optimization/108355
632 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): Revert
635 2024-02-28 Richard Biener <rguenther@suse.de>
637 PR tree-optimization/114121
638 * tree-ssa-sccvn.h (vn_reference_s::offset,
639 vn_reference_s::max_size): New fields.
640 (vn_reference_insert_pieces): Adjust prototype.
641 * tree-ssa-pre.cc (phi_translate_1): Preserve offset/max_size.
642 * tree-ssa-sccvn.cc (vn_reference_eq): Compare offset and
643 size, allow using "don't know" state.
644 (vn_walk_cb_data::finish): Pass along offset/max_size.
645 (vn_reference_lookup_or_insert_for_pieces): Take offset and
646 max_size as argument and use it.
647 (vn_reference_lookup_3): Properly adjust offset and max_size
648 according to the adjusted ao_ref.
649 (vn_reference_lookup_pieces): Initialize offset and max_size.
650 (vn_reference_lookup): Likewise.
651 (vn_reference_lookup_call): Likewise.
652 (vn_reference_insert): Likewise.
653 (visit_reference_op_call): Likewise.
654 (vn_reference_insert_pieces): Take offset and max_size
655 as argument and use it.
657 2024-02-28 Juergen Christ <jchrist@linux.ibm.com>
659 PR tree-optimization/114075
660 * tree-vect-stmts.cc (vectorizable_operation): Don't emulate floating
663 2024-02-28 Jakub Jelinek <jakub@redhat.com>
665 PR tree-optimization/114041
666 * graphite-sese-to-poly.cc (add_conditions_to_domain): Check for
667 INTEGRAL_TYPE_P check rather than INTEGER_TYPE.
669 2024-02-28 Jakub Jelinek <jakub@redhat.com>
671 PR tree-optimization/113988
672 * stor-layout.h (bitwise_mode_for_size): Declare.
673 * stor-layout.cc (bitwise_mode_for_size): New function.
674 * gimple-fold.cc (gimple_fold_builtin_memory_op): Use it.
675 Use bitwise_type_for_mode instead of build_nonstandard_integer_type.
676 Use BITS_PER_UNIT instead of 8.
678 2024-02-27 Uros Bizjak <ubizjak@gmail.com>
681 * config/i386/mmx.md (V248FI): Add V2BF mode.
684 2024-02-27 Eric Botcazou <ebotcazou@adacore.com>
686 * tree-ssa-dse.cc (compute_trims): Fix description. Return early
687 if either ref->offset is not byte aligned or ref->size is not known
688 to be equal to ref->max_size.
689 (maybe_trim_complex_store): Fix description.
690 (maybe_trim_constructor_store): Likewise.
691 (maybe_trim_partially_dead_store): Likewise.
693 2024-02-27 Richard Earnshaw <rearnsha@arm.com>
695 * config/arm/mmintrin.h: Warn if this header is included without
696 defining __ENABLE_DEPRECATED_IWMMXT.
698 2024-02-27 Richard Biener <rguenther@suse.de>
700 PR tree-optimization/114074
701 * tree-chrec.h (chrec_convert_rhs): Default at_stmt arg to NULL.
702 * tree-chrec.cc (chrec_fold_multiply): Canonicalize inputs.
703 Handle poly vs. non-poly multiplication correctly with respect
704 to undefined behavior on overflow.
706 2024-02-27 Jakub Jelinek <jakub@redhat.com>
708 PR rtl-optimization/114044
709 * internal-fn.def (CLRSB, CLZ, CTZ, FFS, PARITY): Use
710 DEF_INTERNAL_INT_EXT_FN macro rather than DEF_INTERNAL_INT_FN.
711 * internal-fn.h (expand_CLRSB, expand_CLZ, expand_CTZ, expand_FFS,
712 expand_PARITY): Declare.
713 * internal-fn.cc (expand_bitquery, expand_CLRSB, expand_CLZ,
714 expand_CTZ, expand_FFS, expand_PARITY): New functions.
715 (expand_POPCOUNT): Use expand_bitquery.
717 2024-02-27 Richard Biener <rguenther@suse.de>
719 PR tree-optimization/114081
720 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
721 Perform manual dominator update for prologue peeling.
722 (vect_do_peeling): Properly update dominators after adding the
723 prologue-around guard.
725 2024-02-26 Georg-Johann Lay <avr@gjlay.de>
727 * config/avr/avr.opt (mcall-prologues, mrelax, maccumulate-args)
728 (mstrict-X): Tag as "Optimization".
730 2024-02-26 Georg-Johann Lay <avr@gjlay.de>
732 * config/avr/avr.cc (avr_out_compare) [AVR_TINY]: Remove code in
733 an "if avr_adiw_reg_p()" block that's dead for AVR_TINY.
735 2024-02-26 Jakub Jelinek <jakub@redhat.com>
736 H.J. Lu <hjl.tools@gmail.com>
738 PR rtl-optimization/113617
739 * varasm.cc (default_elf_select_rtx_section): For
740 references to private symbols in comdat sections
741 use .data.relro.local.pool.<comdat>, .data.relro.pool.<comdat>
742 or .rodata.<comdat> comdat sections.
744 2024-02-26 Richard Biener <rguenther@suse.de>
746 PR tree-optimization/114099
747 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
748 Create and fill in a needed virtual LC PHI for the alternate
749 exits. Remove code dealing with that missing.
751 2024-02-26 Richard Biener <rguenther@suse.de>
753 PR tree-optimization/114068
754 * tree-vect-loop-manip.cc (get_live_virtual_operand_on_edge):
756 (slpeel_tree_duplicate_loop_to_edge_cfg): Add a virtual LC PHI
757 on the main exit if needed. Remove band-aid for the case
760 2024-02-26 H.J. Lu <hjl.tools@gmail.com>
763 * config/i386/i386-options.cc (ix86_set_func_type): Check
764 interrupt instead of noreturn attribute.
766 2024-02-26 Jakub Jelinek <jakub@redhat.com>
768 * config/i386/i386.cc (ix86_bitint_type_info): Add support for
771 2024-02-26 Jakub Jelinek <jakub@redhat.com>
773 PR tree-optimization/114090
774 * match.pd ((x >= 0 ? x : 0) + (x <= 0 ? -x : 0) -> abs x):
775 Restrict pattern to ANY_INTEGRAL_TYPE_P and TYPE_OVERFLOW_UNDEFINED
777 ((x <= 0 ? -x : 0) -> max(-x, 0)): Likewise.
779 2024-02-26 Jakub Jelinek <jakub@redhat.com>
782 * fold-const.cc (fold_binary_loc): Avoid the final associate_trees
783 if all subtrees of var0 come from one of the op0 or op1 operands
784 and all subtrees of con0 come from the other one. Don't clear
785 variables which are never used afterwards.
787 2024-02-26 Richard Biener <rguenther@suse.de>
790 * genmatch.cc (parser::parse_c_expr): Do not record operand
791 lists but only mark operators used.
792 * match.pd ((c ? a : b) op (c ? d : e) --> c ? (a op d) : (b op e)):
793 Properly guard the case of tcc_comparison changing the VEC_COND
796 2024-02-26 Jakub Jelinek <jakub@redhat.com>
799 * config/i386/i386.cc (x86_function_profiler): Add missing new-line
800 to printed instruction.
802 2024-02-26 H.J. Lu <hjl.tools@gmail.com>
805 * config/i386/amxtileintrin.h (_tile_loadconfig): Use
806 __builtin_ia32_ldtilecfg.
807 (_tile_storeconfig): Use __builtin_ia32_sttilecfg.
808 * config/i386/i386-builtin.def (BDESC): Add
809 __builtin_ia32_ldtilecfg and __builtin_ia32_sttilecfg.
810 * config/i386/i386-expand.cc (ix86_expand_builtin): Handle
811 IX86_BUILTIN_LDTILECFG and IX86_BUILTIN_STTILECFG.
812 * config/i386/i386.md (ldtilecfg): New pattern.
813 (sttilecfg): Likewise.
815 2024-02-24 Richard Sandiford <richard.sandiford@arm.com>
817 PR tree-optimization/113205
818 * tree-vect-slp.cc (vect_optimize_slp_pass::forward_cost): Reject
819 the proposed layout if it does not allow a source partition with
820 layout 2 to keep that layout.
822 2024-02-24 Jakub Jelinek <jakub@redhat.com>
824 * builtins.cc (fold_builtin_isascii): Use HOST_WIDE_INT_UC macro.
825 * combine.cc (make_field_assignment): Use HOST_WIDE_INT_1U macro.
826 * double-int.cc (double_int::mask): Use HOST_WIDE_INT_UC macros.
827 * genattrtab.cc (attr_alt_complement): Use HOST_WIDE_INT_1 macro.
828 (mk_attr_alt): Use HOST_WIDE_INT_0 macro.
829 * genautomata.cc (bitmap_set_bit, CLEAR_BIT): Use HOST_WIDE_INT_1
831 * ipa-strub.cc (can_strub_internally_p): Use HOST_WIDE_INT_1 macro.
832 * loop-iv.cc (implies_p): Use HOST_WIDE_INT_1U macro.
833 * pretty-print.cc (test_pp_format): Use HOST_WIDE_INT_C and
834 HOST_WIDE_INT_UC macros.
835 * rtlanal.cc (nonzero_bits1): Use HOST_WIDE_INT_UC macro.
836 * tree.cc (build_replicated_int_cst): Use HOST_WIDE_INT_1U macro.
837 * tree.h (DECL_OFFSET_ALIGN): Use HOST_WIDE_INT_1U macro.
838 * tree-ssa-structalias.cc (dump_varinfo): Use ~HOST_WIDE_INT_0U
840 * wide-int.cc (divmod_internal_2): Use HOST_WIDE_INT_1U macro.
841 * config/i386/constraints.md (define_constraint "L"): Use
842 HOST_WIDE_INT_C macro.
843 * config/i386/i386.md (movabsq split peephole2): Use HOST_WIDE_INT_C
845 (movl + movb peephole2): Likewise.
846 * config/i386/predicates.md (x86_64_zext_immediate_operand): Likewise.
847 (const_32bit_mask): Likewise.
849 2024-02-24 Jakub Jelinek <jakub@redhat.com>
852 * gimple-lower-bitint.cc (bitint_large_huge::lower_stmt): Handle
853 VIEW_CONVERT_EXPRs between large/huge _BitInt and non-integer/pointer
854 types like vector or complex types.
855 (gimple_lower_bitint): Don't merge VIEW_CONVERT_EXPRs to non-integral
856 types. Fix up VIEW_CONVERT_EXPR handling. Allow merging
857 VIEW_CONVERT_EXPR from non-integral/pointer types with a store.
859 2024-02-23 Robin Dapp <rdapp@ventanamicro.com>
862 * config/riscv/riscv-v.cc (rvv_builder::can_duplicate_repeating_sequence_p):
863 Return false if inner mode is already Pmode.
864 (rvv_builder::is_all_same_sequence): New function.
865 (expand_vec_init): Emit broadcast if sequence is all same.
867 2024-02-23 Richard Sandiford <richard.sandiford@arm.com>
870 * config/aarch64/aarch64-early-ra.cc
871 (early_ra::m_current_region): New member variable.
872 (early_ra::m_fpr_recency): Likewise.
873 (early_ra::start_new_region): Bump m_current_region.
874 (early_ra::allocate_colors): Prefer less recently used registers
875 in the event of a tie. Add a comment to explain why we prefer(ed)
876 higher-numbered registers.
877 (early_ra::find_oldest_color): Prefer less recently used registers
879 (early_ra::finalize_allocation): Update recency information for
881 (early_ra::process_blocks): Initialize m_current_region and
884 2024-02-23 Richard Sandiford <richard.sandiford@arm.com>
887 * config/aarch64/aarch64-early-ra.cc
888 (early_ra::test_strictness): New enum.
889 (early_ra::is_chain_candidate): Add a strictness parameter to
890 control whether only correctness matters, or whether both correctness
891 and heuristics should be used. Handle multiple levels of equivalence.
892 (early_ra::find_related_start): Update call accordingly.
893 (early_ra::strided_polarity_pref): Likewise.
894 (early_ra::form_chains): Likewise.
895 (early_ra::try_to_chain_allocnos): Use is_chain_candidate in
896 correctness mode rather than trying to inline the test.
898 2024-02-23 Richard Sandiford <richard.sandiford@arm.com>
901 * config/aarch64/aarch64-early-ra.cc
902 (early_ra::find_related_start): Account for definitions by shared
903 registers when testing for a single register definition.
904 (early_ra::accumulate_defs): New function.
905 (early_ra::record_copy): If A shares B's register, fold A's
906 definition information into B's. Fold A's use information into B's.
908 2024-02-23 H.J. Lu <hjl.tools@gmail.com>
910 * configure.ac (HAVE_AS_R_X86_64_CODE_6_GOTTPOFF): Defined as 1
911 if R_X86_64_CODE_6_GOTTPOFF is supported.
912 * config.in: Regenerated.
913 * configure: Likewise.
914 * config/i386/predicates.md (apx_ndd_add_memory_operand): Allow
915 UNSPEC_GOTNTPOFF if R_X86_64_CODE_6_GOTTPOFF is supported.
917 2024-02-23 Richard Earnshaw <rearnsha@arm.com>
920 * config/arm/neon.md (div<VCVTF:mode>3): Rename from div<mode>3.
921 Gate with ARM_HAVE_NEON_<MODE>_ARITH.
923 2024-02-23 Jakub Jelinek <jakub@redhat.com>
925 PR rtl-optimization/114054
926 * expr.cc (expand_expr_real_2) <case MULT_EXPR>: Use
927 temp variable instead of target parameter for result.
929 2024-02-23 Jakub Jelinek <jakub@redhat.com>
931 PR tree-optimization/114040
932 * gimple-lower-bitint.cc (bitint_large_huge::lower_addsub_overflow):
933 Use EQ_EXPR rather than LT_EXPR for g2 condition and change its
934 probability from likely to unlikely. When handling the true true
935 store, first cast to limb_access_type and then to l's type.
937 2024-02-23 Richard Biener <rguenther@suse.de>
940 * config.gcc: Add ia64*-*-* to the list of obsoleted targets.
942 2024-02-23 Palmer Dabbelt <palmer@rivosinc.com>
945 * config/riscv/arch-canonicalize: Move to python3
946 * config/riscv/multilib-generator: Likewise
948 2024-02-23 Palmer Dabbelt <palmer@rivosinc.com>
950 * doc/invoke.texi: Document -mcpu.
952 2024-02-23 Lulu Cheng <chenglulu@loongson.cn>
954 * configure: Regenerate.
955 * configure.ac: Add parameter "--fatal-warnings" to assemble
956 when checking whether the assemble support conditional branch
959 2024-02-22 Jakub Jelinek <jakub@redhat.com>
962 * doc/extend.texi: (__extension__): Remove comments about scope
963 tokens vs. two colons.
965 2024-02-22 Andrew Pinski <quic_apinski@quicinc.com>
967 PR tree-optimization/109804
968 * gimple-ssa-warn-access.cc (new_delete_mismatch_p): Handle
969 DEMANGLE_COMPONENT_UNNAMED_TYPE.
971 2024-02-22 Richard Biener <rguenther@suse.de>
973 PR tree-optimization/114048
974 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): MEM_REF
975 can also produce -1 off.
977 2024-02-22 Richard Biener <rguenther@suse.de>
979 PR tree-optimization/114027
980 * tree-vect-loop.cc (vecctorizable_reduction): Use optimized
981 condition reduction classification only for single-element
984 2024-02-22 Jakub Jelinek <jakub@redhat.com>
987 * profile-count.h (profile_count::dump): Remove overload with
988 char * first argument.
989 * profile-count.cc (profile_count::dump): Change overload with char *
990 first argument which uses sprintf into the overfload with FILE *
991 first argument and use fprintf instead. Remove overload which wrapped
994 2024-02-22 Jakub Jelinek <jakub@redhat.com>
996 PR tree-optimization/113993
997 * tree-call-cdce.cc (get_no_error_domain): Handle
998 BUILT_IN_{COSH,SINH,EXP{,M1,2}}{F32X,F64X}. Handle
999 BUILT_IN_{COSH,SINH,EXP{,M1,2}}L for
1000 REAL_MODE_FORMAT (TYPE_MODE (long_double_type_node))->emax == 16384
1001 the as the F128 suffixed cases, otherwise as non-suffixed ones.
1002 Handle BUILT_IN_{EXP,POW}10L for
1003 REAL_MODE_FORMAT (TYPE_MODE (long_double_type_node))->emax == 16384
1006 2024-02-22 Jakub Jelinek <jakub@redhat.com>
1008 PR tree-optimization/114038
1009 * gimple-lower-bitint.cc (bitint_large_huge::lower_mul_overflow): Fix
1010 loop exit condition if end is divisible by limb_prec.
1012 2024-02-22 YunQiang Su <syq@gcc.gnu.org>
1014 * doc/invoke.texi(MIPS Options): Fix skipping UrlSuffix
1015 problem of mabi=, mno-flush-func, mexplicit-relocs;
1016 add missing leading - of mbranch-cost option.
1017 * config/mips/mips.opt.urls: Regenerate.
1019 2024-02-22 Kewen Lin <linkw@linux.ibm.com>
1022 * config/rs6000/constraints.md (we): Update internal doc without
1023 referring to option -mpower9-vector.
1024 * config/rs6000/driver-rs6000.cc (asm_names): Remove mpower9-vector
1026 * config/rs6000/rs6000-cpus.def (OTHER_P9_VECTOR_MASKS,
1027 OTHER_P8_VECTOR_MASKS): Merge to ...
1028 (OTHER_VSX_VECTOR_MASKS): ... here.
1029 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Remove
1030 some error message handlings and explicit option mask adjustments on
1031 explicit option power{8,9}-vector conflicting with other options.
1032 (rs6000_print_isa_options): Update comments.
1033 (rs6000_disable_incompatible_switches): Remove power{8,9}-vector
1034 related array items and handlings.
1035 * config/rs6000/rs6000.h (ASM_CPU_SPEC): Remove mpower9-vector
1037 * config/rs6000/rs6000.opt: Make option power{8,9}-vector as
1039 * doc/extend.texi: Remove documentation referring to option
1041 * doc/invoke.texi: Remove documentation for option
1042 -mpower{8,9}-vector and adjust some documentation referring to them.
1043 * doc/md.texi: Update documentation for constraint we.
1044 * doc/sourcebuild.texi: Remove documentation for powerpc_p8vector_ok.
1046 2024-02-22 Pan Li <pan2.li@intel.com>
1049 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins): Upgrade
1050 the version to 0.12.
1052 2024-02-21 Edwin Lu <ewlu@rivosinc.com>
1054 * config/riscv/riscv.cc (riscv_sched_variable_issue): Enable assert
1056 2024-02-21 Edwin Lu <ewlu@rivosinc.com>
1057 Robin Dapp <rdapp.gcc@gmail.com>
1059 * config/riscv/generic-ooo.md (generic_ooo): Move reservation
1060 (generic_ooo_vec_load): Ditto
1061 (generic_ooo_vec_store): Ditto
1062 (generic_ooo_vec_loadstore_seg): Ditto
1063 (generic_ooo_vec_alu): Ditto
1064 (generic_ooo_vec_fcmp): Ditto
1065 (generic_ooo_vec_imul): Ditto
1066 (generic_ooo_vec_fadd): Ditto
1067 (generic_ooo_vec_fmul): Ditto
1068 (generic_ooo_crypto): Ditto
1069 (generic_ooo_perm): Ditto
1070 (generic_ooo_vec_reduction): Ditto
1071 (generic_ooo_vec_ordered_reduction): Ditto
1072 (generic_ooo_vec_idiv): Ditto
1073 (generic_ooo_vec_float_divsqrt): Ditto
1074 (generic_ooo_vec_mask): Ditto
1075 (generic_ooo_vec_vesetvl): Ditto
1076 (generic_ooo_vec_setrm): Ditto
1077 (generic_ooo_vec_readlen): Ditto
1078 * config/riscv/riscv.md: Include generic-vector-ooo
1079 * config/riscv/generic-vector-ooo.md: New file. To here
1081 2024-02-21 Edwin Lu <ewlu@rivosinc.com>
1083 * config/riscv/generic-ooo.md (generic_ooo_sfb_alu): Add reservation
1084 (generic_ooo_branch): Ditto
1085 * config/riscv/generic.md (generic_sfb_alu): Ditto
1086 (generic_fmul_half): Ditto
1087 * config/riscv/riscv.md: Remove cbo, pushpop, and rdfrm types
1088 * config/riscv/sifive-7.md (sifive_7_hfma): Add reservation
1089 (sifive_7_popcount): Ditto
1090 * config/riscv/sifive-p400.md (sifive_p400_clmul): Ditto
1091 * config/riscv/sifive-p600.md (sifive_p600_clmul): Ditto
1092 * config/riscv/vector.md: Change rdfrm to fmove
1093 * config/riscv/zc.md: Change pushpop to load/store
1095 2024-02-21 Jonathan Wakely <jwakely@redhat.com>
1097 * doc/invoke.texi (Warning Options): Fix typos.
1099 2024-02-21 David Faust <david.faust@oracle.com>
1101 * config/bpf/bpf-protos.h (bpf_expand_cpymem): New.
1102 * config/bpf/bpf.cc: (emit_move_loop, bpf_expand_cpymem): New.
1103 * config/bpf/bpf.md: (cpymemdi, movmemdi): New define_expands.
1105 2024-02-21 Martin Jambor <mjambor@suse.cz>
1108 * ipa-prop.h (ipa_node_params): Convert lattices to a vector, adjust
1109 initializers in the contructor.
1110 (ipa_node_params::~ipa_node_params): Release lattices as a vector.
1111 * ipa-cp.h: New file.
1112 * ipa-cp.cc: Include sreal.h and ipa-cp.h.
1113 (ipcp_value_source): Move to ipa-cp.h.
1114 (ipcp_value_base): Likewise.
1115 (ipcp_value): Likewise.
1116 (ipcp_lattice): Likewise.
1117 (ipcp_agg_lattice): Likewise.
1118 (ipcp_bits_lattice): Likewise.
1119 (ipcp_vr_lattice): Likewise.
1120 (ipcp_param_lattices): Likewise.
1121 (ipa_get_parm_lattices): Remove assert latticess is non-NULL.
1122 (ipa_value_from_jfunc): Adjust a check for empty lattices.
1123 (ipa_context_from_jfunc): Likewise.
1124 (ipa_agg_value_from_jfunc): Likewise.
1125 (merge_agg_lats_step): Do not memset new aggregate lattices to zero.
1126 (ipcp_propagate_stage): Allocate lattices in a vector as opposed to
1127 just in contiguous memory.
1128 (ipcp_store_vr_results): Adjust a check for empty lattices.
1129 * auto-profile.cc: Include sreal.h and ipa-cp.h.
1130 * cgraph.cc: Likewise.
1131 * cgraphclones.cc: Likewise.
1132 * cgraphunit.cc: Likewise.
1133 * config/aarch64/aarch64.cc: Likewise.
1134 * config/i386/i386-builtins.cc: Likewise.
1135 * config/i386/i386-expand.cc: Likewise.
1136 * config/i386/i386-features.cc: Likewise.
1137 * config/i386/i386-options.cc: Likewise.
1138 * config/i386/i386.cc: Likewise.
1139 * config/rs6000/rs6000.cc: Likewise.
1140 * config/s390/s390.cc: Likewise.
1141 * gengtype.cc (open_base_files): Added sreal.h and ipa-cp.h to the
1142 files to be included in gtype-desc.cc.
1143 * gimple-range-fold.cc: Include sreal.h and ipa-cp.h.
1144 * ipa-devirt.cc: Likewise.
1145 * ipa-fnsummary.cc: Likewise.
1146 * ipa-icf.cc: Likewise.
1147 * ipa-inline-analysis.cc: Likewise.
1148 * ipa-inline-transform.cc: Likewise.
1149 * ipa-inline.cc: Include ipa-cp.h, move inclusion of sreal.h higher.
1150 * ipa-modref.cc: Include sreal.h and ipa-cp.h.
1151 * ipa-param-manipulation.cc: Likewise.
1152 * ipa-predicate.cc: Likewise.
1153 * ipa-profile.cc: Likewise.
1154 * ipa-prop.cc: Likewise.
1155 (ipa_node_params_t::duplicate): Assert new lattices remain empty
1156 instead of setting them to NULL.
1157 * ipa-pure-const.cc: Include sreal.h and ipa-cp.h.
1158 * ipa-split.cc: Likewise.
1159 * ipa-sra.cc: Likewise.
1160 * ipa-strub.cc: Likewise.
1161 * ipa-utils.cc: Likewise.
1163 * toplev.cc: Likewise.
1164 * tree-ssa-ccp.cc: Likewise.
1165 * tree-ssa-sccvn.cc: Likewise.
1166 * tree-vrp.cc: Likewise.
1168 2024-02-21 Tamar Christina <tamar.christina@arm.com>
1170 * config/aarch64/aarch64-arches.def (AARCH64_ARCH): Remove LS64 from
1173 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1175 * config/aarch64/aarch64.cc (aarch64_mode_emit_local_sme_state):
1176 Use aarch64_gen_compare_zero_and_branch rather than emitting
1179 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1181 * config/aarch64/aarch64.cc (aarch64_option_valid_attribute_p):
1182 Remove duplicated call.
1184 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1186 * config/aarch64/aarch64.cc (aarch64_function_ok_for_sibcall):
1187 Check that each individual piece of state is shared in the same
1188 way, rather than using an aggregate check for PSTATE.ZA.
1190 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1192 * config/aarch64/aarch64.cc (aarch64_mode_emit_local_sme_state):
1193 In the code that commits a lazy save, only zero ZA if the function
1194 has ZA state. Similarly zero ZT0 if the function has ZT0 state.
1196 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1198 * config/aarch64/aarch64-sme.md (aarch64_commit_lazy_save): Remove,
1199 directly inserting the associated sequence
1200 * config/aarch64/aarch64.cc (aarch64_mode_emit_local_sme_state):
1203 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1206 * config/aarch64/aarch64.cc (aarch64_expand_prologue): Don't
1207 fold the SVE allocation into the initial allocation if the
1208 initial allocation includes a VG save.
1210 2024-02-21 Richard Sandiford <richard.sandiford@arm.com>
1213 * cfgrtl.cc (commit_one_edge_insertion): Handle sequences that
1214 contain jumps even if called after initial RTL expansion.
1215 * mode-switching.cc: Include cfgbuild.h.
1216 (optimize_mode_switching): Allow the sequence returned by the
1217 emit hook to contain internal jumps. Record which blocks
1218 contain such jumps and split the blocks at the end.
1219 * config/aarch64/aarch64.cc (aarch64_mode_emit): Check for
1220 non-debug insns when scanning the sequence.
1222 2024-02-21 Tobias Burnus <tburnus@baylibre.com>
1224 * config/nvptx/gen-omp-device-properties.sh: Add 'nvptx64' to arch.
1225 * config/nvptx/nvptx.cc (nvptx_omp_device_kind_arch_isa): Likewise.
1227 2024-02-21 Dimitar Dimitrov <dimitar@dinux.eu>
1229 * doc/invoke.texi (-mmcu): Add information about MCU specs.
1231 2024-02-21 Dimitar Dimitrov <dimitar@dinux.eu>
1233 * doc/invoke.texi (-minrt): Clarify that main
1234 must take no arguments.
1236 2024-02-20 Georg-Johann Lay <avr@gjlay.de>
1238 * config/avr/builtins.def: Use function prototypes of given size
1240 * config/avr/avr.cc (avr_init_builtins): Adjust types required
1242 * doc/extend.texi (AVR Built-in Functions): Adjust accordingly.
1244 2024-02-20 Georg-Johann Lay <avr@gjlay.de>
1246 * doc/extend.texi (AVR Built-in Functions): Use @defbuiltin
1249 2024-02-20 Will Hawkins <hawkinsw@obs.cr>
1251 * config/bpf/bpf.opt: Add help information for -mcpu.
1253 2024-02-20 Richard Sandiford <richard.sandiford@arm.com>
1256 * config/aarch64/aarch64-passes.def (pass_late_track_speculation):
1258 * config/aarch64/aarch64-protos.h (make_pass_late_track_speculation):
1260 * config/aarch64/aarch64.md (is_call): New attribute.
1261 (*and<mode>3nr_compare0): Rename to...
1262 (@aarch64_and<mode>3nr_compare0): ...this.
1263 * config/aarch64/aarch64-sme.md (aarch64_get_sme_state)
1264 (aarch64_tpidr2_save, aarch64_tpidr2_restore): Add is_call attributes.
1265 * config/aarch64/aarch64-speculation.cc: Update file comment to
1266 describe the new late pass.
1267 (aarch64_do_track_speculation): Handle is_call insns like other calls.
1268 (pass_track_speculation): Add an is_late member variable.
1269 (pass_track_speculation::gate): Run the late pass for streaming-
1270 compatible functions and the early pass for other functions.
1271 (make_pass_track_speculation): Update accordingly.
1272 (make_pass_late_track_speculation): New function.
1273 * config/aarch64/aarch64.cc (aarch64_gen_test_and_branch): New
1275 (aarch64_guard_switch_pstate_sm): Use it.
1277 2024-02-19 Iain Sandoe <iain@sandoe.co.uk>
1279 * config/aarch64/aarch64-builtins.cc (aarch64_init_rng_builtins):
1280 Register these builtins with a pointer to uint64_t rather than unsigned
1283 2024-02-19 Thomas Schwinge <tschwinge@baylibre.com>
1286 * config/gcn/gcn-valu.md (define_expand "reduc_<fexpander>_scal_<mode>"):
1287 Conditionalize on '!TARGET_RDNA2_PLUS'.
1288 * config/gcn/gcn.cc (gcn_expand_dpp_shr_insn)
1289 (gcn_expand_reduc_scalar):
1290 'gcc_checking_assert (!TARGET_RDNA2_PLUS);'.
1292 2024-02-19 Thomas Schwinge <tschwinge@baylibre.com>
1294 * config/gcn/gcn.h (TARGET_CPU_CPP_BUILTINS): Restore lost
1295 '__gfx90a__' target CPU definition. Add some safeguards for the future.
1297 2024-02-19 Richard Biener <rguenther@suse.de>
1299 PR rtl-optimization/54052
1300 * rtl-ssa/blocks.cc (function_info::place_phis): Filter
1301 local defs by LR_OUT.
1303 2024-02-19 Jakub Jelinek <jakub@redhat.com>
1305 PR tree-optimization/113967
1306 * match.pd (bit_insert @0 (BIT_FIELD_REF @1 ..) ..): Require
1307 in condition that @rpos is multiple of vector element size.
1309 2024-02-19 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1312 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::earliest_fuse_vsetvl_info):
1313 Suppress vsetvl fusion.
1315 2024-02-18 H.J. Lu <hjl.tools@gmail.com>
1318 * config/i386/i386.cc (ix86_can_use_push2pop2): New.
1319 (ix86_pro_and_epilogue_can_use_push2pop2): Use it.
1320 (ix86_emit_save_regs): Don't generate push2 if
1321 ix86_can_use_push2pop2 return false.
1322 (ix86_expand_epilogue): Don't generate pop2 if
1323 ix86_can_use_push2pop2 return false.
1325 2024-02-18 Georg-Johann Lay <avr@gjlay.de>
1327 * doc/invoke.texi (AVR Options) <-mmcu>: Remove "Atmel".
1328 Note on complete device support.
1330 2024-02-18 Georg-Johann Lay <avr@gjlay.de>
1332 * doc/extend.texi (AVR Function Attributes): Fuse description
1333 of "signal" and "interrupt" attribute. Link pseudo instruction.
1335 2024-02-18 Lulu Cheng <chenglulu@loongson.cn>
1337 * config/loongarch/larchintrin.h (__movgr2fcsr): Remove redundant
1338 symbol type conversions.
1339 (__cacop_d): Likewise.
1340 (__cpucfg): Likewise.
1341 (__asrtle_d): Likewise.
1342 (__asrtgt_d): Likewise.
1343 (__lddir_d): Likewise.
1344 (__ldpte_d): Likewise.
1345 (__crc_w_b_w): Likewise.
1346 (__crc_w_h_w): Likewise.
1347 (__crc_w_w_w): Likewise.
1348 (__crc_w_d_w): Likewise.
1349 (__crcc_w_b_w): Likewise.
1350 (__crcc_w_h_w): Likewise.
1351 (__crcc_w_w_w): Likewise.
1352 (__crcc_w_d_w): Likewise.
1353 (__csrrd_w): Likewise.
1354 (__csrwr_w): Likewise.
1355 (__csrxchg_w): Likewise.
1356 (__csrrd_d): Likewise.
1357 (__csrwr_d): Likewise.
1358 (__csrxchg_d): Likewise.
1359 (__iocsrrd_b): Likewise.
1360 (__iocsrrd_h): Likewise.
1361 (__iocsrrd_w): Likewise.
1362 (__iocsrrd_d): Likewise.
1363 (__iocsrwr_b): Likewise.
1364 (__iocsrwr_h): Likewise.
1365 (__iocsrwr_w): Likewise.
1366 (__iocsrwr_d): Likewise.
1367 (__frecipe_s): Likewise.
1368 (__frecipe_d): Likewise.
1369 (__frsqrte_s): Likewise.
1370 (__frsqrte_d): Likewise.
1372 2024-02-18 Lulu Cheng <chenglulu@loongson.cn>
1374 * config/loongarch/larchintrin.h (__iocsrrd_h): Modify the
1375 function return value type to unsigned short.
1377 2024-02-16 Edwin Lu <ewlu@rivosinc.com>
1379 * doc/sourcebuild.texi: add scan-assembler-bound
1381 2024-02-16 Jason Merrill <jason@redhat.com>
1383 * gdbhooks.py: Fix regex syntax.
1385 2024-02-16 Richard Biener <rguenther@suse.de>
1387 PR tree-optimization/113895
1388 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): Disable
1389 consistency checking when there are out-of-bound array
1390 accesses. Allow -1 off when from an array reference with
1393 2024-02-16 Kito Cheng <kito.cheng@sifive.com>
1396 * config/riscv/riscv.md (*sge<u>_<X:mode><GPR:mode>): Fix asm
1399 2024-02-16 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
1401 * doc/sourcebuild.texi (Effective-Target Keywords, Other
1402 attribugs): Document linker_plugin.
1403 (Require Support): Document dg-require-linker-plugin.
1405 2024-02-16 Kito Cheng <kito.cheng@sifive.com>
1408 * common/config/riscv/riscv-common.cc (riscv_arch_help): New.
1409 * config/riscv/riscv-protos.h (RISCV_MAJOR_VERSION_BASE): New.
1410 (RISCV_MINOR_VERSION_BASE): Ditto.
1411 (RISCV_REVISION_VERSION_BASE): Ditto.
1412 * config/riscv/riscv-c.cc (riscv_ext_version_value): Use enum
1413 rather than magic number.
1414 * config/riscv/riscv.h (riscv_arch_help): New.
1415 (EXTRA_SPEC_FUNCTIONS): Add riscv_arch_help.
1416 (DRIVER_SELF_SPECS): Handle -march=help, -print-supported-extensions and
1417 --print-supported-extensions.
1418 * config/riscv/riscv.opt (march=help): New.
1419 (print-supported-extensions): New.
1420 (-print-supported-extensions): New.
1421 * doc/invoke.texi (RISC-V Options): Document -march=help.
1423 2024-02-16 Tejas Belagod <tejas.belagod@arm.com>
1426 * config/arm/arm.cc (arm_function_ok_for_sibcall): Don't allow tailcalls
1427 for indirect calls with 4 or more arguments in pac-enabled functions.
1429 2024-02-15 David Faust <david.faust@oracle.com>
1431 * config/bpf/bpf.md (zero_extendqidi2): Correct asm template to
1432 use ldxb instead of ldxh.
1434 2024-02-15 Jakub Jelinek <jakub@redhat.com>
1436 PR middle-end/113921
1437 * cfgrtl.h (prepend_insn_to_edge): New declaration.
1438 * cfgrtl.cc (insert_insn_on_edge): Clarify behavior in function
1440 (prepend_insn_to_edge): New function.
1441 * cfgexpand.cc (expand_asm_stmt): Use prepend_insn_to_edge instead of
1442 insert_insn_on_edge.
1444 2024-02-15 Richard Biener <rguenther@suse.de>
1446 PR tree-optimization/111156
1447 * tree-vect-loop.cc (vect_dissolve_slp_only_groups): Look
1448 at the pattern stmt if any.
1450 2024-02-15 Georg-Johann Lay <avr@gjlay.de>
1453 * config/avr/avr.h (AVR_HAVE_ADIW): New macro.
1454 * config/avr/avr-protos.h (avr_adiw_reg_p): New proto.
1455 * config/avr/avr.cc (avr_adiw_reg_p): New function.
1456 (avr_conditional_register_usage) [AVR_TINY]: Don't clear ADDW_REGS.
1457 Replace test_hard_reg_class (ADDW_REGS, ...) with calls to
1458 * config/avr/avr.md: Same.
1459 (attr "isa") <tiny, no_tiny>: Remove.
1460 <adiw, no_adiw>: Add.
1461 (define_insn, define_insn_and_split): When an alternative has
1462 constraint "w", then set attribute "isa" to "adiw".
1463 * config/avr/avr-c.cc (avr_cpu_cpp_builtins) [AVR_HAVE_ADIW]:
1464 Built-in define __AVR_HAVE_ADIW__.
1465 * doc/invoke.texi (AVR Options): Document it.
1467 2024-02-15 Andrew Stubbs <ams@baylibre.com>
1469 * config/gcn/gcn-valu.md
1470 (vec_extract<V_MOV:mode><V_MOV_ALT:mode>): Add conditions for RDNA.
1471 * config/gcn/gcn.cc (gcn_vectorize_vec_perm_const): Check permutation
1472 details are supported on RDNA devices.
1474 2024-02-15 Andrew Pinski <quic_apinski@quicinc.com>
1476 PR middle-end/113508
1477 * doc/md.texi (sdot_prod@var{m}, udot_prod@var{m},
1478 usdot_prod@var{m}, ssad@var{m}, usad@var{m}, widen_usum@var{m}3,
1479 smulhs@var{m}3, umulhs@var{m}3, smulhrs@var{m}3, umulhrs@var{m}3):
1480 Add sentence about what the mode m is.
1482 2024-02-15 Andrew Pinski <quic_apinski@quicinc.com>
1484 * doc/md.texi (widen_ssum, widen_usum, smulhs, umulhs,
1485 smulhrs, umulhrs, sdiv_pow2): Move the 3 outside of the
1488 2024-02-15 Richard Biener <rguenther@suse.de>
1490 * tree-ssa-tail-merge.cc (same_succ_hash): Skip debug
1493 2024-02-15 Jakub Jelinek <jakub@redhat.com>
1495 PR tree-optimization/113567
1496 * gimple-lower-bitint.cc (gimple_lower_bitint): For large/huge
1497 _BitInt multiplication, division or modulo with
1498 SSA_NAME_OCCURS_IN_ABNORMAL_PHI lhs and at least one of rhs1 and rhs2
1499 force the affected inputs into a new SSA_NAME.
1501 2024-02-14 Uros Bizjak <ubizjak@gmail.com>
1504 * config/i386/mmx.md (V248FI): New mode iterator.
1506 (vec_shl_<V248FI:mode>): New expander.
1507 (vec_shl_<V24FI_32:mode>): Ditto.
1508 (vec_shr_<V248FI:mode>): Ditto.
1509 (vec_shr_<V24FI_32:mode>): Ditto.
1510 * config/i386/sse.md (vec_shl_<V_128:mode>): Simplify expander.
1511 (vec_shr_<V248FI:mode>): Ditto.
1513 2024-02-14 Jan Hubicka <jh@suse.cz>
1515 PR tree-optimization/111054
1516 * tree-ssa-loop-split.cc (split_loop): Check for profile being present.
1518 2024-02-14 Tamar Christina <tamar.christina@arm.com>
1520 * tree-cfg.cc (replace_loop_annotate): Inspect loop edges for annotations.
1522 2024-02-14 Richard Biener <rguenther@suse.de>
1524 PR tree-optimization/113910
1525 * bitmap.cc (bitmap_hash): Mix the full element "hash" to
1528 2024-02-14 Jakub Jelinek <jakub@redhat.com>
1530 * pretty-print.cc (PTRDIFF_MAX): Define if not yet defined.
1531 (pp_integer_with_precision): For unsigned ptrdiff_t printing
1532 with u, o or x print ptrdiff_t argument converted to
1533 unsigned long long and masked with 2ULL * PTRDIFF_MAX + 1.
1535 2024-02-14 Richard Biener <rguenther@suse.de>
1537 PR middle-end/113576
1538 * expr.cc (do_store_flag): For vector bool compares of vectors
1539 with padding zero that.
1540 * dojump.cc (do_compare_and_jump): Likewise.
1542 2024-02-14 Gerald Pfeifer <gerald@pfeifer.com>
1544 * doc/install.texi (Prerequisites): Update gettext link.
1546 2024-02-13 H.J. Lu <hjl.tools@gmail.com>
1549 * config/i386/i386.cc (ix86_pro_and_epilogue_can_use_push2pop2):
1550 Return false if the incoming stack isn't 16-byte aligned.
1552 2024-02-13 Tobias Burnus <tburnus@baylibre.com>
1554 PR middle-end/113904
1555 * omp-general.cc (struct omp_ts_info): Update for splitting of
1556 OMP_TRAIT_PROPERTY_EXPR into OMP_TRAIT_PROPERTY_{DEV_NUM,BOOL}_EXPR.
1557 * omp-selectors.h (enum omp_tp_type): Replace
1558 OMP_TRAIT_PROPERTY_EXPR by OMP_TRAIT_PROPERTY_{DEV_NUM,BOOL}_EXPR.
1560 2024-02-13 Monk Chiang <monk.chiang@sifive.com>
1563 * config/riscv/riscv.cc (riscv_macro_fusion_pair_p): Fix
1564 recognizes UNSPEC_AUIPC for RISCV_FUSE_LUI_ADDI.
1566 2024-02-13 Richard Biener <rguenther@suse.de>
1568 PR tree-optimization/113895
1569 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): Track
1570 offset to discover constant array indices in bits, handle
1571 COMPONENT_REF to bitfields.
1573 2024-02-13 Richard Biener <rguenther@suse.de>
1575 PR tree-optimization/113831
1576 * tree-ssa-sccvn.cc (ao_ref_init_from_vn_reference): Fix
1579 2024-02-13 Richard Biener <rguenther@suse.de>
1581 PR tree-optimization/113902
1582 * tree-vect-loop.cc (move_early_exit_stmts): Track
1583 last_seen_vuse for VUSE updating.
1585 2024-02-13 Tamar Christina <tamar.christina@arm.com>
1587 PR tree-optimization/113734
1588 * tree-vect-loop.cc (vect_transform_loop): Treat the final iteration of
1589 an early break loop as partial.
1591 2024-02-13 Richard Biener <rguenther@suse.de>
1593 PR tree-optimization/113898
1594 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): Add
1595 missing accumulated off adjustment.
1597 2024-02-13 Jakub Jelinek <jakub@redhat.com>
1599 * hwint.h (GCC_PRISZ, fmt_size_t): Fix preprocessor conditions,
1600 instead of comparing SIZE_MAX against INT_MAX and LONG_MAX compare
1601 it against UINT_MAX and ULONG_MAX.
1603 2024-02-13 David Malcolm <dmalcolm@redhat.com>
1605 * diagnostic-core.h (emit_diagnostic_valist): Rename overload
1607 (emit_diagnostic_valist_meta): ...this.
1608 * diagnostic.cc (emit_diagnostic_valist): Likewise, to...
1609 (emit_diagnostic_valist_meta): ...this.
1611 2024-02-12 Jakub Jelinek <jakub@redhat.com>
1613 PR tree-optimization/113849
1614 * gimple-lower-bitint.cc (bitint_large_huge::handle_cast): Don't use
1615 fast path for widening casts where !m_upwards_2limb and lhs_type
1616 has precision which is a multiple of limb_prec.
1618 2024-02-12 Jakub Jelinek <jakub@redhat.com>
1621 * attribs.cc (extract_attribute_substring): Remove.
1622 (lookup_scoped_attribute_spec): Don't call it.
1624 2024-02-12 Jakub Jelinek <jakub@redhat.com>
1626 * gengtype.cc (adjust_field_rtx_def): Use HOST_SIZE_T_PRINT_UNSIGNED
1627 and cast to fmt_size_t instead of %lu and cast to unsigned long.
1629 2024-02-12 Christophe Lyon <christophe.lyon@linaro.org>
1631 * Makefile.in: Add no-info dependency.
1632 * configure.ac: Set BUILD_INFO=no-info if makeinfo is not
1634 * configure: Regenerate.
1636 2024-02-12 Iain Sandoe <iain@sandoe.co.uk>
1639 * config/i386/darwin.h (DARWIN_HEAP_T_LIB): Moved to be
1640 available to all sub-targets.
1641 * config/i386/darwin32-biarch.h (DARWIN_HEAP_T_LIB): Delete.
1642 * config/i386/darwin64-biarch.h (DARWIN_HEAP_T_LIB): Delete.
1644 2024-02-12 Richard Biener <rguenther@suse.de>
1646 PR tree-optimization/113831
1647 PR tree-optimization/108355
1648 * tree-ssa-sccvn.cc (copy_reference_ops_from_ref): When
1649 we see variable array indices and get_ref_base_and_extent
1650 can resolve those to constants fix up the ops to constants
1652 (ao_ref_init_from_vn_reference): Use 'off' member for
1653 ARRAY_REF and ARRAY_RANGE_REF instead of recomputing it.
1654 (valueize_refs_1): Also fixup 'off' of ARRAY_RANGE_REF.
1656 2024-02-12 Pan Li <pan2.li@intel.com>
1658 * config/riscv/riscv-vector-builtins.cc (resolve_overloaded_builtin):
1659 Replace args to arguments for misspelled term.
1661 2024-02-12 Georg-Johann Lay <avr@gjlay.de>
1664 * config/avr/gen-avr-mmcu-specs.cc (print_mcu) [have_flmap]:
1665 <*link_rodata_in_ram>: Spec undefs symbol __do_flmap_init
1666 when not linked with -mrodata-in-ram.
1668 2024-02-12 Richard Biener <rguenther@suse.de>
1670 PR tree-optimization/113863
1671 * tree-vect-data-refs.cc (vect_analyze_early_break_dependences):
1672 Record crossed virtual PHIs.
1673 * tree-vect-loop.cc (move_early_exit_stmts): Elide crossed
1676 2024-02-10 Marek Polacek <polacek@redhat.com>
1681 * doc/invoke.texi: Document -Wtemplate-id-cdtor.
1683 2024-02-10 Jakub Jelinek <jakub@redhat.com>
1685 * gimple-lower-bitint.cc (itint_large_huge::lower_addsub_overflow): Fix
1686 computation of idx for i == 4 of bitint_prec_huge.
1688 2024-02-10 Jakub Jelinek <jakub@redhat.com>
1690 PR middle-end/110754
1691 * gimple-low.cc (assumption_copy_decl): For TREE_THIS_VOLATILE
1692 decls create PARM_DECL with pointer to original type, set
1693 TREE_READONLY and keep TREE_THIS_VOLATILE, TREE_ADDRESSABLE,
1694 DECL_NOT_GIMPLE_REG_P and DECL_BY_REFERENCE cleared.
1695 (adjust_assumption_stmt_op): For remapped TREE_THIS_VOLATILE decls
1696 wrap PARM_DECL into a simple TREE_THIS_NO_TRAP MEM_REF.
1697 (lower_assumption): For TREE_THIS_VOLATILE vars pass ADDR_EXPR
1698 of the var as argument.
1700 2024-02-10 Jakub Jelinek <jakub@redhat.com>
1702 * pretty-print.cc (pp_integer_with_precision): Handle precision 3 for
1703 size_t and precision 4 for ptrdiff_t. Formatting fix.
1704 (pp_format): Document %{t,z}{d,i,u,o,x}. Implement t and z modifiers.
1706 (test_pp_format): Test t and z modifiers.
1707 * gcc.cc (read_specs): Use %td instead of %ld and casts to long.
1709 2024-02-10 Jakub Jelinek <jakub@redhat.com>
1711 * ipa-icf.cc (sem_item_optimizer::process_cong_reduction,
1712 sem_item_optimizer::dump_cong_classes): Use HOST_SIZE_T_PRINT_UNSIGNED
1713 and casts to fmt_size_t instead of "%lu" and casts to unsigned long.
1714 * tree.cc (print_debug_expr_statistics): Use HOST_SIZE_T_PRINT_DEC
1715 and casts to fmt_size_t instead of "%ld" and casts to long.
1716 (print_value_expr_statistics, print_type_hash_statistics): Likewise.
1717 * dwarf2out.cc (output_macinfo_op): Use HOST_WIDE_INT_PRINT_UNSIGNED
1718 instead of "%lu" and casts to unsigned long.
1719 * gcov-dump.cc (dump_gcov_file): Use %u instead of %lu and casts to
1721 * tree-ssa-dom.cc (htab_statistics): Use HOST_SIZE_T_PRINT_DEC
1722 and casts to fmt_size_t instead of "%ld" and casts to long.
1723 * cfgexpand.cc (dump_stack_var_partition): Use
1724 HOST_SIZE_T_PRINT_UNSIGNED and casts to fmt_size_t instead of "%lu"
1725 and casts to unsigned long.
1726 * gengtype.cc (adjust_field_rtx_def): Likewise.
1727 * tree-into-ssa.cc (htab_statistics): Use HOST_SIZE_T_PRINT_DEC
1728 and casts to fmt_size_t instead of "%ld" and casts to long.
1729 * postreload-gcse.cc (dump_hash_table): Likewise.
1730 * ggc-page.cc (alloc_page): Use HOST_SIZE_T_PRINT_UNSIGNED
1731 and casts to fmt_size_t instead of "%lu" and casts to unsigned long.
1732 (ggc_internal_alloc, ggc_free): Likewise.
1733 * genpreds.cc (write_lookup_constraint_1): Likewise.
1734 (write_insn_constraint_len): Likewise.
1735 * tree-dfa.cc (dump_dfa_stats): Use HOST_SIZE_T_PRINT_DEC
1736 and casts to fmt_size_t instead of "%ld" and casts to long.
1737 * varasm.cc (output_constant_pool_contents): Use
1738 HOST_WIDE_INT_PRINT_DEC instead of "%ld" and casts to long.
1739 * var-tracking.cc (dump_var): Likewise.
1741 2024-02-09 Jakub Jelinek <jakub@redhat.com>
1743 PR tree-optimization/113783
1744 * gimple-lower-bitint.cc (bitint_large_huge::lower_stmt): Look
1745 through VIEW_CONVERT_EXPR for final cast checks. Handle
1746 VIEW_CONVERT_EXPRs from large/huge _BitInt to > MAX_FIXED_MODE_SIZE
1748 (gimple_lower_bitint): Don't merge mergeable operations or other
1749 casts with VIEW_CONVERT_EXPRs to > MAX_FIXED_MODE_SIZE INTEGER_TYPEs.
1750 * expr.cc (expand_expr_real_1): Don't use convert_modes if either
1753 2024-02-09 Jakub Jelinek <jakub@redhat.com>
1755 * hwint.h (GCC_PRISZ, fmt_size_t, HOST_SIZE_T_PRINT_DEC,
1756 HOST_SIZE_T_PRINT_UNSIGNED, HOST_SIZE_T_PRINT_HEX,
1757 HOST_SIZE_T_PRINT_HEX_PURE): Define.
1758 * ira-conflicts.cc (build_conflict_bit_table): Use it. Formatting
1761 2024-02-09 Jakub Jelinek <jakub@redhat.com>
1763 PR middle-end/113415
1764 * cfgexpand.cc (expand_asm_stmt): For asm goto, use
1765 duplicate_insn_chain to duplicate after_rtl_seq sequence instead
1766 of hand written loop with emit_insn of copy_insn and emit original
1767 after_rtl_seq on the last edge.
1769 2024-02-09 Jakub Jelinek <jakub@redhat.com>
1771 PR tree-optimization/113818
1772 * gimple-lower-bitint.cc (add_eh_edge): New function.
1773 (bitint_large_huge::handle_load,
1774 bitint_large_huge::lower_mergeable_stmt,
1775 bitint_large_huge::lower_muldiv_stmt): Use it.
1777 2024-02-09 Jakub Jelinek <jakub@redhat.com>
1779 PR tree-optimization/113774
1780 * gimple-lower-bitint.cc (bitint_large_huge::handle_cast): Don't
1781 emit any comparison if m_first and low + 1 is equal to
1782 m_upwards_2limb, simplify condition for that. If not
1783 single_comparison, not m_first and we can prove that the idx <= low
1784 comparison will be always true, emit instead of idx <= low
1785 comparison low <= low such that cfg cleanup will optimize it at
1786 the end of the pass.
1788 2024-02-08 Aldy Hernandez <aldyh@redhat.com>
1790 PR tree-optimization/113735
1791 * value-relation.cc (equiv_oracle::add_equiv_to_block): Call
1794 2024-02-08 Georg-Johann Lay <avr@gjlay.de>
1796 * config/avr/gen-avr-mmcu-specs.cc (struct McuInfo): New.
1797 (main, print_mcu, diagnose_mrodata_in_ram): Pass it down.
1799 2024-02-08 H.J. Lu <hjl.tools@gmail.com>
1803 * config/i386/constraints.md: List all constraints with j prefix.
1804 (j>): Change auto-dec to auto-inc in documentation.
1805 (je): Changed to a memory constraint with APX NDD TLS operand
1807 (jM): New memory constraint for APX NDD instructions.
1809 * config/i386/i386-protos.h (x86_poff_operand_p): Removed.
1810 * config/i386/i386.cc (x86_poff_operand_p): Likewise.
1811 * config/i386/i386.md (*add<dwi>3_doubleword): Use rjO.
1812 (*add<mode>_1[SWI48]): Use je and jM.
1813 (addsi_1_zext): Use jM.
1814 (*addv<dwi>4_doubleword_1[DWI]): Likewise.
1815 (*sub<mode>_1[SWI]): Use jM.
1816 (@add<mode>3_cc_overflow_1[SWI]): Likewise.
1817 (*add<dwi>3_doubleword_cc_overflow_1): Use rjO.
1818 (*and<dwi>3_doubleword): Likewise.
1820 (*andsi_1_zext): Likewise.
1821 (*and<mode>_1[SWI24]): Likewise.
1822 (*<code><dwi>3_doubleword[any_or]): Use rjO
1823 (*code<mode>_1[any_or SWI248]): Use jM.
1824 (*<code>si_1_zext[zero_extend + any_or]): Likewise.
1825 * config/i386/predicates.md (apx_ndd_memory_operand): New.
1826 (apx_ndd_add_memory_operand): Likewise.
1828 2024-02-08 Georg-Johann Lay <avr@gjlay.de>
1831 * config/avr/avr-mcus.def (ata5797): Move from avr5 to avr4.
1832 * doc/avr-mmcu.texi: Rebuild.
1834 2024-02-08 Tamar Christina <tamar.christina@arm.com>
1836 PR tree-optimization/113808
1837 * tree-vect-loop.cc (vectorizable_live_operation): Don't cache the
1838 value cross iterations.
1840 2024-02-08 Georg-Johann Lay <avr@gjlay.de>
1842 * config/avr/gen-avr-mmcu-specs.cc (print_mcu) <*cpp_mcu>: Spec always
1843 defines __AVR_PM_BASE_ADDRESS__ if the core has it.
1845 2024-02-08 Richard Biener <rguenther@suse.de>
1847 * tree-vect-data-refs.cc (vect_analyze_early_break_dependences):
1848 Revert last change to dr_may_alias_p.
1850 2024-02-08 Georg-Johann Lay <avr@gjlay.de>
1852 * config/avr/gen-avr-mmcu-specs.cc: Rename spec cc1_misc to
1853 cc1_rodata_in_ram. Rename spec link_misc to link_rodata_in_ram.
1854 Remove spec asm_misc.
1855 * config/avr/specs.h: Same.
1857 2024-02-08 Pan Li <pan2.li@intel.com>
1860 * config/riscv/riscv-vector-builtins-shapes.cc (struct alu_def): Make
1861 sure the c.arg_num is >= 2 before checking.
1862 (struct build_frm_base): Ditto.
1863 (struct narrow_alu_def): Ditto.
1865 2024-02-07 Richard Biener <rguenther@suse.de>
1867 PR tree-optimization/113796
1868 * tree-if-conv.cc (combine_blocks): Wipe range-info before
1869 replacing PHIs and inserting predicates.
1871 2024-02-07 Roger Sayle <roger@nextmovesoftware.com>
1872 Uros Bizjak <ubizjak@gmail.com>
1875 * config/i386/i386-features.cc (timode_convert_cst): New helper
1876 function to convert a TImode CONST_SCALAR_INT_P to a V1TImode
1878 (timode_scalar_chain::convert_op): Use timode_convert_cst.
1879 (timode_scalar_chain::convert_insn): Delete REG_EQUAL notes.
1880 Use timode_convert_cst.
1882 2024-02-07 Victor Do Nascimento <victor.donascimento@arm.com>
1884 * config/aarch64/aarch64-sys-regs.def: Copy from Binutils.
1885 * config/aarch64/aarch64.h (AARCH64_FL_AIE): New.
1886 (AARCH64_FL_DEBUGv8p9): Likewise.
1887 (AARCH64_FL_FGT2): Likewise.Likewise.
1888 (AARCH64_FL_ITE): Likewise.
1889 (AARCH64_FL_PFAR): Likewise.
1890 (AARCH64_FL_PMUv3_ICNTR): Likewise.
1891 (AARCH64_FL_PMUv3_SS): Likewise.
1892 (AARCH64_FL_PMUv3p9): Likewise.
1893 (AARCH64_FL_RASv2): Likewise.
1894 (AARCH64_FL_S1PIE): Likewise.
1895 (AARCH64_FL_S1POE): Likewise.
1896 (AARCH64_FL_S2PIE): Likewise.
1897 (AARCH64_FL_S2POE): Likewise.
1898 (AARCH64_FL_SCTLR2): Likewise.
1899 (AARCH64_FL_SEBEP): Likewise.
1900 (AARCH64_FL_SPE_FDS): Likewise.
1901 (AARCH64_FL_TCR2): Likewise.
1903 2024-02-07 Richard Biener <rguenther@suse.de>
1905 * tree-vect-data-refs.cc (vect_analyze_early_break_dependences):
1906 Only check whether reads are in-bound in places that are not safe.
1907 Fix dependence check. Add missing newline. Clarify comments.
1909 2024-02-07 Tamar Christina <tamar.christina@arm.com>
1911 PR tree-optimization/113750
1912 * tree-vect-data-refs.cc (vect_analyze_early_break_dependences): Check
1913 for single predecessor when doing early break vect.
1914 * tree-vect-loop.cc (move_early_exit_stmts): Get gsi at the start but
1917 2024-02-07 Tamar Christina <tamar.christina@arm.com>
1919 PR tree-optimization/113731
1920 * gimple-iterator.cc (gsi_move_before): Take new parameter for update
1922 * gimple-iterator.h (gsi_move_before): Default new param to
1924 * tree-vect-loop.cc (move_early_exit_stmts): Call gsi_move_before with
1927 2024-02-07 Jakub Jelinek <jakub@redhat.com>
1929 PR tree-optimization/113756
1930 * range-op.cc (update_known_bitmask): For GIMPLE_UNARY_RHS,
1931 use TYPE_SIGN (lh.type ()) instead of sign for widest_int::from
1932 of lh_bits value and mask.
1934 2024-02-07 Jakub Jelinek <jakub@redhat.com>
1936 PR tree-optimization/113753
1937 * wide-int.cc (wi::mul_internal): Unpack op1val and op2val with
1938 UNSIGNED rather than SIGNED. If high or needs_overflow and prec is
1939 not a multiple of HOST_BITS_PER_WIDE_INT, shift left bits above prec
1940 so that they start with r[half_blocks_needed] lowest bit. Fix up
1941 computation of top mask for SIGNED.
1943 2024-02-07 Pan Li <pan2.li@intel.com>
1946 * config/riscv/riscv-protos.h (resolve_overloaded_builtin): Adjust
1947 the signature of func.
1948 * config/riscv/riscv-c.cc (riscv_resolve_overloaded_builtin): Ditto.
1949 * config/riscv/riscv-vector-builtins.cc (resolve_overloaded_builtin): Make
1950 overloaded func with empty args error.
1952 2024-02-06 H.J. Lu <hjl.tools@gmail.com>
1955 * config/i386/i386.cc (x86_64_select_profile_regnum): Return
1956 R10_REG after sorry.
1958 2024-02-06 Andrew Carlotti <andrew.carlotti@arm.com>
1960 * config/aarch64/aarch64.cc (aarch64_mangle_decl_assembler_name):
1961 Move before new caller, and add ".default" suffix.
1962 (get_suffixed_assembler_name): New.
1963 (make_resolver_func): Use get_suffixed_assembler_name.
1964 (aarch64_generate_version_dispatcher_body): Redo name mangling.
1966 2024-02-06 Jakub Jelinek <jakub@redhat.com>
1969 * config/aarch64/aarch64.cc (aarch64_output_sme_zero_za): Change tiles
1970 element from std::pair<unsigned int, char> to an unnamed struct.
1971 Adjust uses of tile range variable.
1973 2024-02-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
1975 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::emit_vsetvl): Fix inifinite compilation.
1976 (pre_vsetvl::remove_vsetvl_pre_insns): Ditto.
1978 2024-02-06 Jakub Jelinek <jakub@redhat.com>
1981 * gimple-fold.cc (gimple_fold_builtin_strlen): For -fsanitize=address
1982 reset maxlen to sizetype maximum.
1984 2024-02-06 Jakub Jelinek <jakub@redhat.com>
1986 PR tree-optimization/113736
1987 * gimple-lower-bitint.cc (bitint_large_huge::limb_access): Use
1988 var's address space for MEM_REF or VIEW_CONVERT_EXPRs.
1990 2024-02-06 Jakub Jelinek <jakub@redhat.com>
1992 PR tree-optimization/113759
1993 * tree-ssa-math-opts.cc (convert_mult_to_widen): If actual_precision
1994 or from_unsignedN differs from properties of typeN, update typeN
1995 to build_nonstandard_integer_type. If TREE_TYPE (rhsN) is not
1996 uselessly convertible to typeN, convert it using fold_convert or
1997 build_and_insert_cast depending on if rhsN is INTEGER_CST or not.
1998 (convert_plusminus_to_widen): Likewise.
2000 2024-02-06 Tejas Belagod <tejas.belagod@arm.com>
2003 * config/aarch64/aarch64.cc (aarch64_class_max_nregs): Handle 64-bit
2004 vector structure modes correctly.
2006 2024-02-05 Christoph Müllner <christoph.muellner@vrull.eu>
2008 * config/riscv/thead.cc (th_print_operand_address): Fix compiler
2011 2024-02-05 H.J. Lu <hjl.tools@gmail.com>
2014 * config/i386/i386.cc (x86_64_select_profile_regnum): New.
2015 (x86_function_profiler): Call x86_64_select_profile_regnum to
2016 get a scratch register for large model profiling.
2018 2024-02-05 Richard Ball <richard.ball@arm.com>
2020 * config/arm/arm.cc (arm_output_mi_thunk): Emit
2021 insn for bti_c when bti is enabled.
2023 2024-02-05 Xi Ruoyao <xry111@xry111.site>
2025 * config/mips/mips-msa.md (neg<mode:MSA>2): Add missing mode for
2028 2024-02-05 Xi Ruoyao <xry111@xry111.site>
2030 * config/mips/mips-msa.md (elmsgnbit): New define_mode_attr.
2031 (neg<mode>2): Change the mode iterator from MSA to IMSA because
2032 in FP arithmetic we cannot use (0 - x) for -x.
2033 (neg<mode>2): New define_insn to implement FP vector negation,
2034 using a bnegi instruction to negate the sign bit.
2036 2024-02-05 Richard Biener <rguenther@suse.de>
2038 PR tree-optimization/113707
2039 * tree-ssa-sccvn.cc (rpo_elim::eliminate_avail): After
2040 checking the avail set treat out-of-region defines as
2043 2024-02-05 Richard Biener <rguenther@suse.de>
2045 * tree-vect-data-refs.cc (vect_create_data_ref_ptr): Use
2046 the default mode when building a pointer.
2048 2024-02-05 Jakub Jelinek <jakub@redhat.com>
2050 PR tree-optimization/113737
2051 * gimple-lower-bitint.cc (gimple_lower_bitint): If GIMPLE_SWITCH
2052 has just a single label, remove it and make single successor edge
2055 2024-02-05 Jakub Jelinek <jakub@redhat.com>
2058 * config/i386/i386-features.cc (rest_of_handle_insert_vzeroupper):
2059 Remove REG_DEAD/REG_UNUSED notes at the end of the pass before
2062 2024-02-05 Richard Biener <rguenther@suse.de>
2065 * config/i386/i386-expand.cc
2066 (expand_set_or_cpymem_prologue_epilogue_by_misaligned_moves):
2067 Use a new pseudo for the skipped number of bytes.
2069 2024-02-05 Monk Chiang <monk.chiang@sifive.com>
2071 * config/riscv/riscv-cores.def: Add sifive-p450, sifive-p670.
2072 * doc/invoke.texi (RISC-V Options): Add sifive-p450,
2075 2024-02-05 Monk Chiang <monk.chiang@sifive.com>
2077 * config/riscv/riscv.md: Include sifive-p400.md.
2078 * config/riscv/sifive-p400.md: New file.
2079 * config/riscv/riscv-cores.def (RISCV_TUNE): Add parameter.
2080 * config/riscv/riscv-opts.h (enum riscv_microarchitecture_type):
2082 * config/riscv/riscv.cc (sifive_p400_tune_info): New.
2083 * config/riscv/riscv.h (TARGET_SFB_ALU): Update.
2084 * doc/invoke.texi (RISC-V Options): Add sifive-p400-series
2086 2024-02-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
2088 * config/xtensa/xtensa.md (*eqne_zero_masked_bits):
2089 Add missing ":SI" to the match_operator.
2091 2024-02-04 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
2093 * config/xtensa/xtensa.md (SHI): New mode iterator.
2094 (2 split patterns related to constsynth):
2095 Change to also accept HImode operands.
2097 2024-02-04 Jeff Law <jlaw@ventanamicro.com>
2099 * config/riscv/riscv.cc (riscv_rtx_costs): Handle SUBREG and REG
2102 2024-02-04 Xi Ruoyao <xry111@xry111.site>
2104 * config/loongarch/lsx.md (neg<mode:FLSX>2): Remove the
2106 * config/loongarch/simd.md (simdfmt_as_i): New define_mode_attr.
2107 (elmsgnbit): Likewise.
2108 (neg<mode:FVEC>2): New define_insn.
2109 * config/loongarch/lasx.md (negv4df2, negv8sf2): Remove as they
2110 are now instantiated in simd.md.
2112 2024-02-04 Xi Ruoyao <xry111@xry111.site>
2114 * config/loongarch/loongarch.cc (loongarch_symbol_insns): Do not
2115 use LSX_SUPPORTED_MODE_P or LASX_SUPPORTED_MODE_P if mode is
2118 2024-02-04 Li Wei <liwei@loongson.cn>
2120 * config/loongarch/loongarch.cc (loongarch_expand_vselect): Adjust.
2121 (loongarch_expand_vselect_vconcat): Ditto.
2122 (loongarch_try_expand_lsx_vshuf_const): New, use vshuf to implement
2123 all 128-bit constant permutation situations.
2124 (loongarch_expand_lsx_shuffle): Adjust and rename function name.
2125 (loongarch_is_imm_set_shuffle): Renamed function name.
2126 (loongarch_expand_vec_perm_even_odd): Function forward declaration.
2127 (loongarch_expand_vec_perm_even_odd_1): Add implement for 128-bit
2128 extract-even and extract-odd permutations.
2129 (loongarch_is_odd_extraction): Delete.
2130 (loongarch_is_even_extraction): Ditto.
2131 (loongarch_expand_vec_perm_const): Adjust.
2133 2024-02-03 Jakub Jelinek <jakub@redhat.com>
2135 PR middle-end/113722
2136 * wide-int.cc (wi::bswap_large): Rename third argument from
2137 len to xlen and adjust use in safe_uhwi. Add len variable, set
2138 it to BLOCKS_NEEDED (precision) and use it for clearing of val
2139 and as canonize argument. Clear val using memset instead of
2142 2024-02-03 Jakub Jelinek <jakub@redhat.com>
2144 * ggc-common.cc (gt_pch_save): Allow addr to be equal to
2145 mmi.preferred_base + mmi.size - sizeof (void *).
2147 2024-02-03 Xi Ruoyao <xry111@xry111.site>
2149 * config/loongarch/loongarch-def.h (abi_minimal_isa): Declare.
2150 * config/loongarch/loongarch-opts.cc (abi_minimal_isa): Remove
2151 the ODR-violating locale declaration.
2153 2024-02-02 Tamar Christina <tamar.christina@arm.com>
2155 PR tree-optimization/113588
2156 PR tree-optimization/113467
2157 * tree-vect-data-refs.cc
2158 (vect_analyze_data_ref_dependence): Choose correct dest and fix checks.
2159 (vect_analyze_early_break_dependences): Update comments.
2161 2024-02-02 John David Anglin <danglin@gcc.gnu.org>
2164 * config/pa/pa.cc (enum pa_builtins): Add PA_BUILTIN_GET_FPSR
2165 and PA_BUILTIN_SET_FPSR builtins.
2166 * (pa_builtins_icode): Declare.
2167 * (def_builtin, pa_fpu_init_builtins): New.
2168 * (pa_init_builtins): Initialize FPU builtins.
2169 * (pa_builtin_decl, pa_expand_builtin_1): New.
2170 * (pa_expand_builtin): Handle PA_BUILTIN_GET_FPSR and
2171 PA_BUILTIN_SET_FPSR builtins.
2172 * (pa_atomic_assign_expand_fenv): New.
2173 * config/pa/pa.md (UNSPECV_GET_FPSR, UNSPECV_SET_FPSR): New
2175 (get_fpsr, put_fpsr): New expanders.
2176 (get_fpsr_32, get_fpsr_64, set_fpsr_32, set_fpsr_64): New
2179 2024-02-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2182 * config/riscv/riscv-v.cc (expand_reduction): Pass VLMAX avl to scalar move.
2184 2024-02-02 Jonathan Wakely <jwakely@redhat.com>
2186 * doc/extend.texi (Common Type Attributes): Fix typo in
2187 description of hardbool.
2189 2024-02-02 Jakub Jelinek <jakub@redhat.com>
2191 PR tree-optimization/113692
2192 * gimple-lower-bitint.cc (bitint_large_huge::lower_stmt): Handle casts
2193 from large/huge BITINT_TYPEs to POINTER_TYPE/REFERENCE_TYPE as
2196 2024-02-02 Jakub Jelinek <jakub@redhat.com>
2198 PR middle-end/113699
2199 * gimple-lower-bitint.cc (bitint_large_huge::lower_asm): Handle
2200 uninitialized large/huge _BitInt SSA_NAME inputs.
2202 2024-02-02 Jakub Jelinek <jakub@redhat.com>
2204 PR middle-end/113705
2205 * tree-ssa-math-opts.cc (is_widening_mult_rhs_p): Use wide_int_from
2206 around wi::to_wide in order to compare value in prec precision.
2208 2024-02-02 Lehua Ding <lehua.ding@rivai.ai>
2211 2024-02-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2213 * config/riscv/riscv.cc (riscv_legitimize_move): Fix poly_int dest generation.
2215 2024-02-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2217 * config/riscv/riscv.cc (riscv_legitimize_move): Fix poly_int dest generation.
2219 2024-02-02 Pan Li <pan2.li@intel.com>
2221 * config/riscv/riscv.cc (riscv_get_arg_info): Cleanup comments.
2222 (riscv_pass_by_reference): Ditto.
2223 (riscv_fntype_abi): Ditto.
2225 2024-02-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2227 * config/riscv/riscv-vsetvl.cc (vsetvl_pre_insn_p): New function.
2228 (pre_vsetvl::cleaup): Remove vsetvl_pre.
2229 (pre_vsetvl::remove_vsetvl_pre_insns): New function.
2231 2024-02-02 Jiahao Xu <xujiahao@loongson.cn>
2233 * config/loongarch/larchintrin.h
2234 (__frecipe_s): Update function return type.
2235 (__frecipe_d): Ditto.
2236 (__frsqrte_s): Ditto.
2237 (__frsqrte_d): Ditto.
2239 2024-02-02 Li Wei <liwei@loongson.cn>
2241 * config/loongarch/loongarch.cc (loongarch_multiply_add_p): New.
2242 (loongarch_vector_costs::add_stmt_cost): Adjust.
2244 2024-02-02 Xi Ruoyao <xry111@xry111.site>
2246 * config/loongarch/loongarch.md (unspec): Add
2247 UNSPEC_LA_PCREL_64_PART1 and UNSPEC_LA_PCREL_64_PART2.
2248 (la_pcrel64_two_parts): New define_insn.
2249 * config/loongarch/loongarch.cc (loongarch_tls_symbol): Fix a
2250 typo in the comment.
2251 (loongarch_call_tls_get_addr): If -mcmodel=extreme
2252 -mexplicit-relocs={always,auto}, use la_pcrel64_two_parts for
2253 addressing the TLS symbol and __tls_get_addr. Emit an REG_EQUAL
2254 note to allow CSE addressing __tls_get_addr.
2255 (loongarch_legitimize_tls_address): If -mcmodel=extreme
2256 -mexplicit-relocs={always,auto}, address TLS IE symbols with
2257 la_pcrel64_two_parts.
2258 (loongarch_split_symbol): If -mcmodel=extreme
2259 -mexplicit-relocs={always,auto}, address symbols with
2260 la_pcrel64_two_parts.
2261 (loongarch_output_mi_thunk): Clean up unreachable code. If
2262 -mcmodel=extreme -mexplicit-relocs={always,auto}, address the MI
2263 thunks with la_pcrel64_two_parts.
2265 2024-02-02 Lulu Cheng <chenglulu@loongson.cn>
2267 * config/loongarch/loongarch.cc (loongarch_call_tls_get_addr):
2268 Add support for call36.
2270 2024-02-02 Lulu Cheng <chenglulu@loongson.cn>
2272 * config/loongarch/loongarch.cc (loongarch_explicit_relocs_p):
2273 When the code model of the symbol is extreme and -mexplicit-relocs=auto,
2274 the macro instruction loading symbol address is not applicable.
2275 (loongarch_call_tls_get_addr): Adjust code.
2276 (loongarch_legitimize_tls_address): Likewise.
2278 2024-02-02 Lulu Cheng <chenglulu@loongson.cn>
2280 * config/loongarch/loongarch-protos.h (loongarch_symbol_extreme_p):
2281 Add function declaration.
2282 * config/loongarch/loongarch.cc (loongarch_symbolic_constant_p):
2283 For SYMBOL_PCREL64, non-zero addend of "la.local $rd,$rt,sym+addend"
2285 (loongarch_load_tls): Added macro support in extreme mode.
2286 (loongarch_call_tls_get_addr): Likewise.
2287 (loongarch_legitimize_tls_address): Likewise.
2288 (loongarch_force_address): Likewise.
2289 (loongarch_legitimize_move): Likewise.
2290 (loongarch_output_mi_thunk): Likewise.
2291 (loongarch_option_override_internal): Remove the code that detects
2292 explicit relocs status.
2293 (loongarch_handle_model_attribute): Likewise.
2294 * config/loongarch/loongarch.md (movdi_symbolic_off64): New template.
2295 * config/loongarch/predicates.md (symbolic_off64_operand): New predicate.
2296 (symbolic_off64_or_reg_operand): Likewise.
2298 2024-02-02 Lulu Cheng <chenglulu@loongson.cn>
2300 * config/loongarch/loongarch.cc (loongarch_load_tls):
2301 Load all types of tls symbols through one function.
2302 (loongarch_got_load_tls_gd): Delete.
2303 (loongarch_got_load_tls_ld): Delete.
2304 (loongarch_got_load_tls_ie): Delete.
2305 (loongarch_got_load_tls_le): Delete.
2306 (loongarch_call_tls_get_addr): Modify the called function name.
2307 (loongarch_legitimize_tls_address): Likewise.
2308 * config/loongarch/loongarch.md (@got_load_tls_gd<mode>): Delete.
2309 (@load_tls<mode>): New template.
2310 (@got_load_tls_ld<mode>): Delete.
2311 (@got_load_tls_le<mode>): Delete.
2312 (@got_load_tls_ie<mode>): Delete.
2314 2024-02-02 Lulu Cheng <chenglulu@loongson.cn>
2316 * config/loongarch/loongarch.cc (mem_shadd_or_shadd_rtx_p): New function.
2317 (loongarch_legitimize_address): Add logical transformation code.
2319 2024-02-01 Marek Polacek <polacek@redhat.com>
2321 * doc/invoke.texi: Update -Wdangling-reference documentation.
2323 2024-02-01 Uros Bizjak <ubizjak@gmail.com>
2326 * config/i386/i386.md (*cmp<dwi>_doubleword):
2327 Do not force SUBREG pieces to pseudos.
2329 2024-02-01 John David Anglin <danglin@gcc.gnu.org>
2331 * config/pa/pa.md (atomic_storedi_1): Fix bug in
2334 2024-02-01 Georg-Johann Lay <avr@gjlay.de>
2336 * config/avr/avr.cc: Tabify.
2338 2024-02-01 Richard Ball <richard.ball@arm.com>
2340 PR tree-optimization/111268
2341 * tree-vect-slp.cc (vectorizable_slp_permutation_1):
2342 Add variable-length check for vector input arguments
2345 2024-02-01 Thomas Schwinge <tschwinge@baylibre.com>
2347 * config/gcn/gcn.cc (gcn_hsa_declare_function_name): Don't
2348 hard-code number of SGPR/VGPR/AVGPR registers.
2349 * config/gcn/gcn.h: Add a 'STATIC_ASSERT's for number of
2350 SGPR/VGPR/AVGPR registers.
2352 2024-02-01 Monk Chiang <monk.chiang@sifive.com>
2354 * config/riscv/riscv.md: Add "fcvt_i2f", "fcvt_f2i" type
2355 attribute, and include sifive-p600.md.
2356 * config/riscv/generic-ooo.md: Update type attribute.
2357 * config/riscv/generic.md: Update type attribute.
2358 * config/riscv/sifive-7.md: Update type attribute.
2359 * config/riscv/sifive-p600.md: New file.
2360 * config/riscv/riscv-cores.def (RISCV_TUNE): Add parameter.
2361 * config/riscv/riscv-opts.h (enum riscv_microarchitecture_type):
2363 * config/riscv/riscv.cc (sifive_p600_tune_info): New.
2364 * config/riscv/riscv.h (TARGET_SFB_ALU): Update.
2365 * doc/invoke.texi (RISC-V Options): Add sifive-p600-series
2367 2024-02-01 Monk Chiang <monk.chiang@sifive.com>
2369 * common/config/riscv/riscv-common.cc: Add Za64rs, Za128rs,
2370 Ziccif, Ziccrse, Ziccamoa, Zicclsm, Zic64b items.
2371 * config/riscv/riscv.opt: New macro for 7 new unprivileged
2373 * doc/invoke.texi (RISC-V Options): Add Za64rs, Za128rs,
2374 Ziccif, Ziccrse, Ziccamoa, Zicclsm, Zic64b extensions.
2376 2024-02-01 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
2378 * config/sol2.h (LIBASAN_EARLY_SPEC): Add -z now unless
2379 -static-libasan. Add missing whitespace.
2381 2024-02-01 Thomas Schwinge <tschwinge@baylibre.com>
2383 * config/gcn/gcn.md (FIRST_SGPR_REG, LAST_SGPR_REG)
2384 (FIRST_VGPR_REG, LAST_VGPR_REG, FIRST_AVGPR_REG, LAST_AVGPR_REG):
2385 Don't 'define_constants'.
2387 2024-02-01 Thomas Schwinge <tschwinge@baylibre.com>
2389 * config/gcn/gcn.h (SGPR_OR_VGPR_REGNO_P): Remove.
2391 2024-02-01 Thomas Schwinge <tschwinge@baylibre.com>
2393 * config/gcn/gcn.md (sync_compare_and_swap<mode>_lds_insn)
2394 [TARGET_RDNA3]: Adjust.
2396 2024-02-01 Richard Biener <rguenther@suse.de>
2398 PR tree-optimization/113693
2399 * tree-ssa-sccvn.cc (rpo_elim::eliminate_avail): Honor avail
2400 data when available.
2402 2024-02-01 Jakub Jelinek <jakub@redhat.com>
2403 Jason Merrill <jason@redhat.com>
2406 * gimple-low.cc (lower_stmt): Remove .ASAN_MARK calls
2407 on variables which were promoted to TREE_STATIC.
2409 2024-02-01 Roger Sayle <roger@nextmovesoftware.com>
2410 Richard Biener <rguenther@suse.de>
2413 * tree-ssa-math-opts.cc (is_widening_mult_rhs_p): Use range
2414 information via tree_non_zero_bits to check if this operand
2415 is suitably extended for a widening (or highpart) multiplication.
2416 (convert_mult_to_widen): Insert explicit casts if the RHS or LHS
2417 isn't already of the claimed type.
2419 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2422 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2424 * config/riscv/generic-ooo.md (generic_ooo_sfb_alu): Add reservation
2425 (generic_ooo_branch): ditto
2426 * config/riscv/generic.md (generic_sfb_alu): ditto
2427 (generic_fmul_half): ditto
2428 * config/riscv/riscv.md: Remove cbo, pushpop, and rdfrm types
2429 * config/riscv/sifive-7.md (sifive_7_hfma):Add reservation
2430 (sifive_7_popcount): ditto
2431 * config/riscv/vector.md: change rdfrm to fmove
2432 * config/riscv/zc.md: change pushpop to load/store
2434 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2437 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2438 Robin Dapp <rdapp.gcc@gmail.com>
2440 * config/riscv/generic-ooo.md (generic_ooo): Move reservation
2441 (generic_ooo_vec_load): ditto
2442 (generic_ooo_vec_store): ditto
2443 (generic_ooo_vec_loadstore_seg): ditto
2444 (generic_ooo_vec_alu): ditto
2445 (generic_ooo_vec_fcmp): ditto
2446 (generic_ooo_vec_imul): ditto
2447 (generic_ooo_vec_fadd): ditto
2448 (generic_ooo_vec_fmul): ditto
2449 (generic_ooo_crypto): ditto
2450 (generic_ooo_perm): ditto
2451 (generic_ooo_vec_reduction): ditto
2452 (generic_ooo_vec_ordered_reduction): ditto
2453 (generic_ooo_vec_idiv): ditto
2454 (generic_ooo_vec_float_divsqrt): ditto
2455 (generic_ooo_vec_mask): ditto
2456 (generic_ooo_vec_vesetvl): ditto
2457 (generic_ooo_vec_setrm): ditto
2458 (generic_ooo_vec_readlen): ditto
2459 * config/riscv/riscv.md: include generic-vector-ooo
2460 * config/riscv/generic-vector-ooo.md: New file. to here
2462 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2465 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2467 * config/riscv/riscv.cc (riscv_sched_variable_issue): enable assert
2469 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2471 * config/riscv/riscv.cc (riscv_sched_variable_issue): enable assert
2473 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2474 Robin Dapp <rdapp.gcc@gmail.com>
2476 * config/riscv/generic-ooo.md (generic_ooo): Move reservation
2477 (generic_ooo_vec_load): ditto
2478 (generic_ooo_vec_store): ditto
2479 (generic_ooo_vec_loadstore_seg): ditto
2480 (generic_ooo_vec_alu): ditto
2481 (generic_ooo_vec_fcmp): ditto
2482 (generic_ooo_vec_imul): ditto
2483 (generic_ooo_vec_fadd): ditto
2484 (generic_ooo_vec_fmul): ditto
2485 (generic_ooo_crypto): ditto
2486 (generic_ooo_perm): ditto
2487 (generic_ooo_vec_reduction): ditto
2488 (generic_ooo_vec_ordered_reduction): ditto
2489 (generic_ooo_vec_idiv): ditto
2490 (generic_ooo_vec_float_divsqrt): ditto
2491 (generic_ooo_vec_mask): ditto
2492 (generic_ooo_vec_vesetvl): ditto
2493 (generic_ooo_vec_setrm): ditto
2494 (generic_ooo_vec_readlen): ditto
2495 * config/riscv/riscv.md: include generic-vector-ooo
2496 * config/riscv/generic-vector-ooo.md: New file. to here
2498 2024-02-01 Edwin Lu <ewlu@rivosinc.com>
2500 * config/riscv/generic-ooo.md (generic_ooo_sfb_alu): Add reservation
2501 (generic_ooo_branch): ditto
2502 * config/riscv/generic.md (generic_sfb_alu): ditto
2503 (generic_fmul_half): ditto
2504 * config/riscv/riscv.md: Remove cbo, pushpop, and rdfrm types
2505 * config/riscv/sifive-7.md (sifive_7_hfma):Add reservation
2506 (sifive_7_popcount): ditto
2507 * config/riscv/vector.md: change rdfrm to fmove
2508 * config/riscv/zc.md: change pushpop to load/store
2510 2024-02-01 Andrew Pinski <quic_apinski@quicinc.com>
2513 * config/aarch64/aarch64-simd.md (split for movv8di):
2514 For strict aligned mode, use DImode instead of TImode.
2516 2024-01-31 Robin Dapp <rdapp@ventanamicro.com>
2518 PR middle-end/113607
2519 * match.pd: Make sure else values match when folding a
2520 vec_cond into a conditional operation.
2522 2024-01-31 Marek Polacek <polacek@redhat.com>
2524 * doc/invoke.texi: Mention that -fconcepts-ts was deprecated in GCC 14.
2526 2024-01-31 Tamar Christina <tamar.christina@arm.com>
2527 Matthew Malcomson <matthew.malcomson@arm.com>
2530 * asan.h (asan_intercepted_p): Incercept memset, memmove, memcpy and
2532 * builtins.cc (expand_builtin): Include HWASAN when checking for
2535 2024-01-31 Richard Biener <rguenther@suse.de>
2537 PR middle-end/110176
2538 * match.pd (zext (bool) <= (int) 4294967295u): Make sure
2539 to match INTEGER_CST only without outstanding conversion.
2541 2024-01-31 Alex Coplan <alex.coplan@arm.com>
2544 * config/aarch64/aarch64.cc (aarch64_reg_save_mode): Use
2545 V16QImode for the full 16-byte FPR saves in the vector PCS case.
2547 2024-01-31 Richard Biener <rguenther@suse.de>
2549 PR tree-optimization/111444
2550 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Do not use
2551 vn_reference_lookup_2 when optimistically skipping may-defs.
2553 2024-01-31 Richard Biener <rguenther@suse.de>
2555 PR tree-optimization/113630
2556 * tree-ssa-pre.cc (compute_avail): Avoid registering a
2557 reference with a representation with not matching base
2560 2024-01-31 Jakub Jelinek <jakub@redhat.com>
2562 PR rtl-optimization/113656
2563 * simplify-rtx.cc (simplify_context::simplify_unary_operation_1)
2564 <case FLOAT_TRUNCATE>: Fix up last argument to simplify_gen_unary.
2566 2024-01-31 Jakub Jelinek <jakub@redhat.com>
2569 * dwarf2out.cc (loc_list_from_tree_1): Assume integral types
2570 with BLKmode are larger than DWARF2_ADDR_SIZE.
2572 2024-01-31 Jakub Jelinek <jakub@redhat.com>
2574 PR tree-optimization/113639
2575 * gimple-lower-bitint.cc (bitint_large_huge::handle_operand_addr):
2576 For VIEW_CONVERT_EXPR set rhs1 to its operand.
2578 2024-01-31 Richard Biener <rguenther@suse.de>
2580 PR tree-optimization/113670
2581 * tree-vect-data-refs.cc (vect_check_gather_scatter):
2582 Make sure we can take the address of the reference base.
2584 2024-01-31 Georg-Johann Lay <avr@gjlay.de>
2586 * config/avr/avr-mcus.def: Add AVR64DU28, AVR64DU32, ATA5787,
2587 ATA5835, ATtiny64AUTO, ATA5700M322.
2588 * doc/avr-mmcu.texi: Rebuild.
2590 2024-01-31 Alexandre Oliva <oliva@adacore.com>
2593 * ipa-strub.cc (build_ref_type_for): Drop nonaliased. Adjust
2596 2024-01-31 Alexandre Oliva <oliva@adacore.com>
2598 PR middle-end/112917
2599 PR middle-end/113100
2600 * builtins.cc (expand_builtin_stack_address): Use
2601 STACK_ADDRESS_OFFSET.
2602 * doc/extend.texi (__builtin_stack_address): Adjust.
2603 * config/sparc/sparc.h (STACK_ADDRESS_OFFSET): Define.
2604 * doc/tm.texi.in (STACK_ADDRESS_OFFSET): Document.
2605 * doc/tm.texi: Rebuilt.
2607 2024-01-31 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2610 * config/riscv/riscv-vsetvl.cc (extract_single_source): Remove.
2611 (pre_vsetvl::compute_vsetvl_def_data): Fix compile time issue.
2612 (pre_vsetvl::compute_transparent): New function.
2613 (pre_vsetvl::compute_lcm_local_properties): Fix compile time time issue.
2615 2024-01-30 Fangrui Song <maskray@google.com>
2618 * config/i386/constraints.md: Define constraint "Ws".
2619 * doc/md.texi: Document it.
2621 2024-01-30 Marek Polacek <polacek@redhat.com>
2625 * doc/invoke.texi: Update -Wdangling-reference description.
2627 2024-01-30 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
2629 * config/xtensa/constraints.md (R, T, U):
2630 Change define_constraint to define_memory_constraint.
2631 * config/xtensa/predicates.md (move_operand): Don't check that a
2632 constant pool operand size is a multiple of UNITS_PER_WORD.
2633 * config/xtensa/xtensa.cc
2634 (xtensa_lra_p, TARGET_LRA_P): Remove.
2635 (xtensa_emit_move_sequence): Remove "if (reload_in_progress)"
2636 clause as it can no longer be true.
2637 (fixup_subreg_mem): Drop function.
2638 (xtensa_output_integer_literal_parts): Consider 16-bit wide
2640 (xtensa_legitimate_constant_p): Add short-circuit path for
2641 integer load instructions. Don't check that mode size is
2642 at least UNITS_PER_WORD.
2643 * config/xtensa/xtensa.md (movsf): Use can_create_pseudo_p()
2644 rather reload_in_progress and reload_completed.
2645 (doloop_end): Drop operand 2.
2646 (movhi_internal): Add alternative loading constant from a
2648 (define_split for DI register_operand): Don't limit to
2649 !TARGET_AUTO_LITPOOLS.
2650 * config/xtensa/xtensa.opt (mlra): Change to no effect.
2652 2024-01-30 Pan Li <pan2.li@intel.com>
2654 * config/riscv/riscv.cc (riscv_v_vls_mode_aggregate_gpr_count): New function to
2655 calculate the gpr count required by vls mode.
2656 (riscv_v_vls_to_gpr_mode): New function convert vls mode to gpr mode.
2657 (riscv_pass_vls_aggregate_in_gpr): New function to return the rtx of gpr
2659 (riscv_get_arg_info): Add vls mode handling.
2660 (riscv_pass_by_reference): Return false if arg info has no zero gpr count.
2662 2024-01-30 Richard Biener <rguenther@suse.de>
2664 PR tree-optimization/113659
2665 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
2666 Handle main exit without virtual use.
2668 2024-01-30 Christoph Müllner <christoph.muellner@vrull.eu>
2670 * config/riscv/riscv.md: Move UNSPEC_XTHEADFMV* to unspec enum.
2672 2024-01-30 Iain Sandoe <iain@sandoe.co.uk>
2675 * config/darwin.h (DARWIN_SHARED_WEAK_ADDS, DARWIN_WEAK_CRTS): New.
2676 (REAL_LIBGCC_SPEC): Move weak CRT handling to separate spec.
2677 * config/i386/darwin.h (DARWIN_HEAP_T_LIB): New.
2678 * config/i386/darwin32-biarch.h (DARWIN_HEAP_T_LIB): New.
2679 * config/i386/darwin64-biarch.h (DARWIN_HEAP_T_LIB): New.
2680 * config/rs6000/darwin.h (DARWIN_HEAP_T_LIB): New.
2682 2024-01-30 Richard Sandiford <richard.sandiford@arm.com>
2685 * config/aarch64/aarch64-early-ra.cc (early_ra::preprocess_insns):
2686 Mark all registers that occur in addresses as needing a GPR.
2688 2024-01-30 Richard Sandiford <richard.sandiford@arm.com>
2691 * config/aarch64/aarch64-early-ra.cc (early_ra::replace_regs): Take
2692 the containing insn as an extra parameter. Reset debug instructions
2693 if they reference a register that is no longer used by real insns.
2694 (early_ra::apply_allocation): Update calls accordingly.
2696 2024-01-30 Jakub Jelinek <jakub@redhat.com>
2698 PR tree-optimization/113603
2699 * tree-ssa-strlen.cc (strlen_pass::handle_store): After
2700 count_nonzero_bytes call refetch si using get_strinfo in case it
2701 has been unshared in the meantime.
2703 2024-01-30 Jakub Jelinek <jakub@redhat.com>
2705 PR middle-end/101195
2706 * except.cc (expand_builtin_eh_return_data_regno): If which doesn't
2707 fit into unsigned HOST_WIDE_INT, return constm1_rtx.
2709 2024-01-30 Jin Ma <jinma@linux.alibaba.com>
2711 * config/riscv/thead.cc (th_print_operand_address): Change %ld
2714 2024-01-29 Manos Anagnostakis <manos.anagnostakis@vrull.eu>
2715 Manolis Tsamis <manolis.tsamis@vrull.eu>
2716 Philipp Tomsich <philipp.tomsich@vrull.eu>
2718 * config/aarch64/aarch64-ldpstp.md: Remove unused mode.
2719 * config/aarch64/aarch64-protos.h (aarch64_operands_ok_for_ldpstp):
2721 * config/aarch64/aarch64.cc (aarch64_operands_ok_for_ldpstp):
2722 Call on framework moved later.
2724 2024-01-29 Jose E. Marchesi <jose.marchesi@oracle.com>
2726 * config/bpf/bpf.cc (bpf_expand_epilogue): Do not emit a return
2727 instruction in naked function epilogues.
2729 2024-01-29 YunQiang Su <syq@gcc.gnu.org>
2732 * configure.ac: Fix typo gcc_cv_as_mips_explicit should be
2733 gcc_cv_as_mips_explicit_relocs.
2734 * configure: Regnerated.
2736 2024-01-29 Matthieu Longo <matthieu.longo@arm.com>
2739 * config/arm/arm.md (arm_rev16si2): Convert to define_insn.
2740 Correct generated RTL.
2741 (arm_rev16si2_alt1): Correctly handle conditional execution.
2742 (arm_rev16si2_alt2): Likewise.
2744 2024-01-29 Richard Biener <rguenther@suse.de>
2746 PR middle-end/113622
2747 * expr.cc (expand_assignment): Spill hard registers if
2748 we index them with a variable offset.
2750 2024-01-29 Richard Biener <rguenther@suse.de>
2752 PR middle-end/113622
2753 * gimple-isel.cc (gimple_expand_vec_set_extract_expr):
2754 Also allow DECL_HARD_REGISTER variables.
2756 2024-01-29 Alex Coplan <alex.coplan@arm.com>
2759 * config/aarch64/aarch64-ldp-fusion.cc (fixup_debug_uses_trailing_add):
2760 Use iterate_safely when iterating over debug uses.
2761 (fixup_debug_uses): Likewise.
2762 (ldp_bb_info::cleanup_tombstones): Use iterate_safely to iterate
2763 over nondebug insns instead of manually maintaining the next insn.
2764 * iterator-utils.h (class safe_iterator): New.
2765 (iterate_safely): New.
2767 2024-01-29 H.J. Lu <hjl.tools@gmail.com>
2770 * config/i386/i386-options.cc (ix86_set_func_type): Save
2771 callee-saved registers in noreturn functions for -O0/-Og.
2773 2024-01-29 Tobias Burnus <tburnus@baylibre.com>
2776 * config/gcn/gcn-valu.md (fold_left_plus_<mode>): Only
2777 define for !TARGET_RDNA2_PLUS.
2779 2024-01-29 Richard Sandiford <richard.sandiford@arm.com>
2782 * tree-vect-patterns.cc (vect_recog_over_widening_pattern): Remove
2783 workaround for right shifts.
2784 (vect_truncatable_operation_p): Handle NEGATE_EXPR and BIT_NOT_EXPR.
2785 (vect_determine_precisions_from_range): Be more selective about
2786 which codes can be narrowed based on their input and output ranges.
2787 For shifts, require at least one more bit of precision than the
2788 maximum shift amount.
2790 2024-01-29 Tobias Burnus <tburnus@baylibre.com>
2792 * config/nvptx/nvptx.opt (march-map=): Add sm_89 and sm_90a.
2794 2024-01-29 Tobias Burnus <tburnus@baylibre.com>
2796 * doc/install.texi (amdgcn): Recommend LLVM 15+ and newlib 4.4+,
2797 but keep requiring only newlib 4.3+ and, if gfx1100 is disabled,
2800 2024-01-29 Tobias Burnus <tburnus@baylibre.com>
2803 * config/gcn/mkoffload.cc (SET_XNACK_UNSET, TEST_SRAM_ECC_UNSET): New.
2804 (SET_SRAM_ECC_UNSUPPORTED): Renamed to ...
2805 (SET_SRAM_ECC_UNSET): ... this.
2806 (copy_early_debug_info): Remove gfx900 special case, now handled as
2807 part of the generic handling.
2808 (main): Update SRAM_ECC and XNACK for the -march as done in gcn-hsa.h.
2810 2024-01-29 Jakub Jelinek <jakub@redhat.com>
2812 PR tree-optimization/110603
2813 * tree-ssa-strlen.cc (get_range_strlen_dynamic): Remove incorrect
2814 setting of pdata->maxlen to vr.upper_bound (which is unconditionally
2815 overwritten anyway). Avoid creating invalid range with minlen
2816 larger than maxlen. Formatting fix.
2818 2024-01-29 Richard Biener <rguenther@suse.de>
2821 * tree-inline.cc (initialize_inlined_parameters): Reverse
2822 the decl chain of inlined parameters.
2824 2024-01-28 Iain Sandoe <iain@sandoe.co.uk>
2826 * config/darwin.cc (darwin_build_constant_cfstring): Prevent over-
2827 alignment of CFString constants by setting DECL_USER_ALIGN.
2829 2024-01-28 Iain Sandoe <iain@sandoe.co.uk>
2830 Jakub Jelinek <jakub@redhat.com>
2833 * builtins.cc (expand_builtin): Handle BUILT_IN_GCC_NESTED_PTR_CREATED
2834 and BUILT_IN_GCC_NESTED_PTR_DELETED.
2835 * builtins.def (BUILT_IN_GCC_NESTED_PTR_CREATED,
2836 BUILT_IN_GCC_NESTED_PTR_DELETED): Make these builtins LIB-EXT and
2837 rename the library fallbacks to __gcc_nested_func_ptr_created and
2838 __gcc_nested_func_ptr_deleted.
2839 * doc/invoke.texi: Rename these to __gcc_nested_func_ptr_created
2840 and __gcc_nested_func_ptr_deleted.
2841 * tree-nested.cc (finalize_nesting_tree_1): Use builtin_explicit for
2842 BUILT_IN_GCC_NESTED_PTR_CREATED and BUILT_IN_GCC_NESTED_PTR_DELETED.
2843 * tree.cc (build_common_builtin_nodes): Build the
2844 BUILT_IN_GCC_NESTED_PTR_CREATED and BUILT_IN_GCC_NESTED_PTR_DELETED local
2845 builtins only for non-explicit.
2847 2024-01-28 YunQiang Su <syq@gcc.gnu.org>
2849 * doc/invoke.texi: Remove duplicate MIPS explicit-relocs option.
2851 2024-01-27 H.J. Lu <hjl.tools@gmail.com>
2854 * config/i386/i386-options.cc (ix86_set_func_type): Don't
2855 save and restore callee saved registers for a noreturn function
2856 with nothrow or compiled with -fno-exceptions.
2858 2024-01-27 H.J. Lu <hjl.tools@gmail.com>
2862 * config/i386/i386-expand.cc (ix86_expand_call): Replace
2863 no_caller_saved_registers check with call_saved_registers check.
2864 Clobber all registers that are not used by the callee with
2865 no_callee_saved_registers attribute.
2866 * config/i386/i386-options.cc (ix86_set_func_type): Set
2867 call_saved_registers to TYPE_NO_CALLEE_SAVED_REGISTERS for
2868 noreturn function. Disallow no_callee_saved_registers with
2869 interrupt or no_caller_saved_registers attributes together.
2870 (ix86_set_current_function): Replace no_caller_saved_registers
2871 check with call_saved_registers check.
2872 (ix86_handle_no_caller_saved_registers_attribute): Renamed to ...
2873 (ix86_handle_call_saved_registers_attribute): This.
2874 (ix86_gnu_attributes): Add
2875 ix86_handle_call_saved_registers_attribute.
2876 * config/i386/i386.cc (ix86_conditional_register_usage): Replace
2877 no_caller_saved_registers check with call_saved_registers check.
2878 (ix86_function_ok_for_sibcall): Don't allow callee with
2879 no_callee_saved_registers attribute when the calling function
2880 has callee-saved registers.
2881 (ix86_comp_type_attributes): Also check
2882 no_callee_saved_registers.
2883 (ix86_epilogue_uses): Replace no_caller_saved_registers check
2884 with call_saved_registers check.
2885 (ix86_hard_regno_scratch_ok): Likewise.
2886 (ix86_save_reg): Replace no_caller_saved_registers check with
2887 call_saved_registers check. Don't save any registers for
2888 TYPE_NO_CALLEE_SAVED_REGISTERS. Save all registers with
2889 TYPE_DEFAULT_CALL_SAVED_REGISTERS if function with
2890 no_callee_saved_registers attribute is called.
2891 (find_drap_reg): Replace no_caller_saved_registers check with
2892 call_saved_registers check.
2893 * config/i386/i386.h (call_saved_registers_type): New enum.
2894 (machine_function): Replace no_caller_saved_registers with
2895 call_saved_registers.
2896 * doc/extend.texi: Document no_callee_saved_registers attribute.
2898 2024-01-27 Jakub Jelinek <jakub@redhat.com>
2900 PR tree-optimization/113614
2901 * gimple-lower-bitint.cc (gimple_lower_bitint): Don't merge
2902 widening casts from signed to unsigned types with TRUNC_DIV_EXPR,
2903 TRUNC_MOD_EXPR or FLOAT_EXPR uses.
2905 2024-01-27 Jakub Jelinek <jakub@redhat.com>
2907 PR tree-optimization/113568
2908 * gimple-lower-bitint.cc (bitint_large_huge::lower_mergeable_stmt):
2909 For VIEW_CONVERT_EXPR use first operand of rhs1 instead of rhs1
2910 in the widening extension checks.
2912 2024-01-27 Jakub Jelinek <jakub@redhat.com>
2914 * gimple-lower-bitint.cc (gimple_lower_bitint): For
2915 TDF_DETAILS dump mapping of SSA_NAMEs to decls.
2917 2024-01-26 Hans-Peter Nilsson <hp@axis.com>
2919 * cgraphunit.cc (process_function_and_variable_attributes): Tweak
2920 the warning for an attribute-always_inline without inline declaration.
2922 2024-01-26 Robin Dapp <rdapp@ventanamicro.com>
2925 * genopinit.cc (main): Split init_all_optabs into functions
2926 of 1000 patterns each.
2928 2024-01-26 Tobias Burnus <tburnus@baylibre.com>
2930 * config.gcc (amdgcn-*-*): Add gfx1030 and gfx1100 to
2932 * doc/install.texi (Configuration amdgcn-*-*): Mention gfx1030/gfx1100.
2933 * doc/invoke.texi (AMD GCN Options): Add gfx1030 and gfx1100 to
2936 2024-01-26 Andrew Stubbs <ams@baylibre.com>
2938 * config/gcn/gcn-opts.h (TARGET_PACKED_WORK_ITEMS): Add TARGET_RDNA3.
2939 * config/gcn/gcn-valu.md (all_convert): New iterator.
2940 (<convop><V_INT_1REG_ALT:mode><V_INT_1REG:mode>2<exec>): New
2941 define_expand, and rename the old one to ...
2942 (*<convop><V_INT_1REG_ALT:mode><V_INT_1REG:mode>_sdwa<exec>): ... this.
2943 (extend<V_INT_1REG_ALT:mode><V_INT_1REG:mode>2<exec>): Likewise, to ...
2944 (extend<V_INT_1REG_ALT:mode><V_INT_1REG:mode>_sdwa<exec>): .. this.
2945 (*<convop><V_INT_1REG_ALT:mode><V_INT_1REG:mode>_shift<exec>): New.
2946 * config/gcn/gcn.cc (gcn_global_address_p): Use "offsetbits" correctly.
2947 (gcn_hsa_declare_function_name): Update the vgpr counting for gfx1100.
2948 * config/gcn/gcn.md (<u>mulhisi3): Disable on RDNA3.
2949 (<u>mulqihi3_scalar): Likewise.
2951 2024-01-26 Richard Biener <rguenther@suse.de>
2953 PR tree-optimization/113602
2954 * tree-data-ref.cc (dr_analyze_innermost): Fail when
2955 the base object isn't addressable.
2957 2024-01-26 Tobias Burnus <tburnus@baylibre.com>
2959 * config/gcn/gcn-hsa.h (ABI_VERSION_SPEC): New; creates the
2960 "--amdhsa-code-object-version=" argument.
2961 (ASM_SPEC): Use it; replace previous version of it.
2963 2024-01-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2965 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::earliest_fuse_vsetvl_info): Refine some codes.
2966 (pre_vsetvl::emit_vsetvl): Ditto.
2968 2024-01-26 Jiahao Xu <xujiahao@loongson.cn>
2970 * config/loongarch/lasx.md (vec_extract<mode>_0):
2971 New define_insn_and_split patten.
2973 2024-01-26 Jiahao Xu <xujiahao@loongson.cn>
2975 * config/loongarch/loongarch.h (LOGICAL_OP_NON_SHORT_CIRCUIT): Define.
2977 2024-01-26 Li Wei <liwei@loongson.cn>
2979 * config/loongarch/loongarch.cc (loongarch_emit_swdivsf): Adjust.
2981 2024-01-26 Juzhe-Zhong <juzhe.zhong@rivai.ai>
2984 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::compute_lcm_local_properties): Fix bug.
2986 2024-01-26 Andrew Pinski <quic_apinski@quicinc.com>
2989 * config/aarch64/aarch64.cc (aarch64_classify_index): Avoid
2990 undefined shift after the call to exact_log2.
2992 2024-01-25 Andrew Pinski <quic_apinski@quicinc.com>
2995 * config/aarch64/constraints.md (J): Cast to `unsigned HOST_WIDE_INT`
2996 before taking the negative of it.
2998 2024-01-25 Vladimir N. Makarov <vmakarov@redhat.com>
3001 * lra-constraints.cc (curr_insn_transform): Change class even for
3002 spilled pseudo successfully matched with with NO_REGS.
3004 2024-01-25 Georg-Johann Lay <avr@gjlay.de>
3007 * config/avr/avr-mcus.def (atmega3208, atmega3209): Fix data_section_start.
3009 2024-01-25 Szabolcs Nagy <szabolcs.nagy@arm.com>
3012 * config/aarch64/aarch64.cc (aarch64_gen_compare_zero_and_branch): New.
3013 (aarch64_expand_epilogue): Use the new function.
3014 (aarch64_split_compare_and_swap): Likewise.
3015 (aarch64_split_atomic_op): Likewise.
3017 2024-01-25 Robin Dapp <rdapp.gcc@gmail.com>
3019 PR middle-end/112971
3020 * fold-const.cc (simplify_const_binop): New function for binop
3021 simplification of two constant vectors when element-wise
3022 handling is not necessary.
3023 (const_binop): Call new function.
3025 2024-01-25 Mary Bennett <mary.bennett@embecosm.com>
3027 * common/config/riscv/riscv-common.cc: Add XCVbitmanip.
3028 * config/riscv/constraints.md: Likewise.
3029 * config/riscv/corev.def: Likewise.
3030 * config/riscv/corev.md: Likewise.
3031 * config/riscv/predicates.md: Likewise.
3032 * config/riscv/riscv-builtins.cc (AVAIL): Likewise.
3033 * config/riscv/riscv-ftypes.def: Likewise.
3034 * config/riscv/riscv.opt: Likewise.
3035 * config/riscv/riscv.cc (riscv_print_operand): Add new operand 'Y'.
3036 * doc/extend.texi: Add XCVbitmanip builtin documentation.
3037 * doc/sourcebuild.texi: Likewise.
3039 2024-01-25 Tobias Burnus <tburnus@baylibre.com>
3041 * config/gcn/gcn-hsa.h (ASM_SPEC): Add space after -mxnack= argument.
3043 2024-01-25 Yanzhang Wang <yanzhang.wang@intel.com>
3046 * config/riscv/riscv.cc (riscv_get_arg_info): Remove the flag.
3047 (riscv_fntype_abi): Ditto.
3048 * config/riscv/riscv.opt: Ditto.
3050 2024-01-25 Jakub Jelinek <jakub@redhat.com>
3052 PR middle-end/113574
3053 * convert.cc (convert_to_integer_1) <case LSHIFT_EXPR>: Compare shift
3054 count against TYPE_PRECISION rather than TYPE_SIZE.
3056 2024-01-25 Richard Sandiford <richard.sandiford@arm.com>
3059 * config/aarch64/aarch64-sve-builtins.cc (vector_cst_all_same):
3060 Check VECTOR_CST_ELT instead of VECTOR_CST_ENCODED_ELT
3062 2024-01-25 Richard Sandiford <richard.sandiford@arm.com>
3065 * config/aarch64/aarch64-simd.md: In the movv8di splitter, check
3066 whether each split instruction is a load that clobbers the source
3067 address. Emit that instruction last if so.
3069 2024-01-25 Richard Sandiford <richard.sandiford@arm.com>
3072 * config/aarch64/aarch64-simd.md (aarch64_zip1<mode>_low): New
3074 (<optab><Vnarrowq><mode>2): Use it instead of generating a
3075 paradoxical subreg for the input.
3077 2024-01-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3079 * config/riscv/riscv-vsetvl.cc (get_all_predecessors): New function.
3080 (pre_vsetvl::pre_global_vsetvl_info): Add LCM delete block all
3081 predecessors dump information.
3083 2024-01-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3085 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::compute_vsetvl_def_data): Remove
3086 redundant full available computation.
3087 (pre_vsetvl::pre_global_vsetvl_info): Ditto.
3089 2024-01-25 Jakub Jelinek <jakub@redhat.com>
3091 * doc/generic.texi (VECTOR_CST): Fix typo - petterns -> patterns.
3092 * doc/rtl.texi (CONST_VECTOR): Likewise.
3094 2024-01-25 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3096 * config/riscv/riscv-opts.h (enum vsetvl_strategy_enum): Add optim-no-fusion option.
3097 * config/riscv/riscv-vsetvl.cc (pass_vsetvl::lazy_vsetvl): Ditto.
3098 (pass_vsetvl::execute): Ditto.
3099 * config/riscv/riscv.opt: Ditto.
3101 2024-01-25 Jiahao Xu <xujiahao@loongson.cn>
3103 * config/loongarch/lasx.md (@vec_concatz<mode>): Remove this define_insn pattern.
3104 * config/loongarch/loongarch.cc (loongarch_expand_vector_group_init): Use vec_concat<mode>.
3106 2024-01-25 Richard Biener <rguenther@suse.de>
3108 PR tree-optimization/113576
3109 * tree-vect-loop.cc (vec_init_loop_exit_info): Only allow
3110 exits with may_be_zero niters when its the last one.
3112 2024-01-25 Lulu Cheng <chenglulu@loongson.cn>
3114 * config/loongarch/loongarch.cc (loongarch_symbolic_constant_p):
3115 For symbols of type tls, non-zero Offset is not generated.
3117 2024-01-25 Haochen Gui <guihaoc@gcc.gnu.org>
3119 * config/rs6000/rs6000-string.cc (expand_block_compare): Enable
3120 P9 with m32 and mpowerpc64.
3122 2024-01-25 liuhongt <hongtao.liu@intel.com>
3124 * config/i386/i386-options.cc (ix86_option_override_internal):
3125 Enable -mlam=u57 by default when compiled with
3126 -fsanitize=hwaddress.
3128 2024-01-25 Palmer Dabbelt <palmer@rivosinc.com>
3130 * common/config/riscv/riscv-common.cc (riscv_implied_info):
3131 Remove {"ztso", "a"}.
3133 2024-01-24 Martin Jambor <mjambor@suse.cz>
3137 * cgraph.h (cgraph_edge): Add a parameter to
3138 redirect_call_stmt_to_callee.
3139 * ipa-param-manipulation.h (ipa_param_adjustments): Add a
3140 parameter to modify_call.
3141 (ipa_release_ssas_in_hash): Declare.
3142 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee): New
3143 parameter killed_ssas, pass it to padjs->modify_call.
3144 * ipa-param-manipulation.cc (purge_all_uses): New function.
3145 (ipa_param_adjustments::modify_call): New parameter killed_ssas.
3146 Instead of substituting uses, invoke purge_all_uses. If
3147 hash of killed SSAs has not been provided, create a temporary one
3148 and release SSAs that have been added to it.
3149 (compare_ssa_versions): New function.
3150 (ipa_release_ssas_in_hash): Likewise.
3151 * tree-inline.cc (redirect_all_calls): Create
3152 id->killed_new_ssa_names earlier, pass it to edge redirection,
3154 (copy_body): Release SSAs in id->killed_new_ssa_names.
3156 2024-01-24 Andrew Pinski <quic_apinski@quicinc.com>
3159 * config/aarch64/aarch64.cc (aarch64_get_reg_raw_mode): For
3160 TARGET_GENERAL_REGS_ONLY, return VOIDmode for non-GP_REGNUM_P regno.
3162 2024-01-24 Monk Chiang <monk.chiang@sifive.com>
3165 * config/riscv/sfb.md: New splitters to rewrite single bit
3166 sign extension as the condition to SFB instructions.
3168 2024-01-24 Jan Hubicka <jh@suse.cz>
3171 * common.opt: (flimit-function-alignment): Reorder alphabeticaly
3172 (fmin-function-alignment): New parameter.
3173 * doc/invoke.texi: (-fmin-function-alignment): Document.
3174 (-falign-functions,-falign-loops,-falign-labels): Mention that
3175 aglinments are ignored in cold code.
3176 * varasm.cc (assemble_start_function): Handle min-function-alignment.
3178 2024-01-24 Tamar Christina <tamar.christina@arm.com>
3181 * config/aarch64/aarch64-simd.md (<su_optab>div<mode>3,
3183 * config/aarch64/iterators.md (VQDIV): Remove.
3184 (SVE_FULL_SDI_SIMD, SVE_FULL_HSDI_SIMD_DI,
3185 SVE_I_SIMD_DI): New.
3186 (VPRED, sve_lane_con): Add V4SI and V2DI.
3187 * config/aarch64/aarch64-sve.md (<optab><mode>3,
3188 @aarch64_pred_<optab><mode>): Support Advanced SIMD types.
3189 (mul<mode>3): New, split from <optab><mode>3.
3190 (@aarch64_pred_<optab><mode>, *post_ra_<optab><mode>3): New.
3191 * config/aarch64/aarch64-sve2.md (@aarch64_mul_lane_<mode>,
3192 *aarch64_mul_unpredicated_<mode>): Change SVE_FULL_HSDI to
3193 SVE_FULL_HSDI_SIMD_DI.
3195 2024-01-24 Tamar Christina <tamar.christina@arm.com>
3197 PR tree-optimization/113552
3198 * config/aarch64/aarch64.cc
3199 (aarch64_simd_clone_compute_vecsize_and_simdlen): Block simdlen 1.
3201 2024-01-24 Martin Jambor <mjambor@suse.cz>
3204 * ipa-cp.cc (ipcp_lattice<valtype>::add_value): Bail out if value
3205 count is equal or greater than the limit. Use the limit from the
3208 2024-01-24 YunQiang Su <syq@gcc.gnu.org>
3210 * configure.ac: Detect the explicit relocs support for
3211 mips, and define C macro MIPS_EXPLICIT_RELOCS.
3212 * config.in: Regenerated.
3213 * configure: Regenerated.
3214 * doc/invoke.texi(MIPS Options): Add -mexplicit-relocs.
3215 * config/mips/mips-opts.h: Define enum mips_explicit_relocs.
3216 * config/mips/mips.cc(mips_set_compression_mode): Sorry if
3217 !TARGET_EXPLICIT_RELOCS instead of just set it.
3218 * config/mips/mips.h: Define TARGET_EXPLICIT_RELOCS and
3219 TARGET_EXPLICIT_RELOCS_PCREL with mips_opt_explicit_relocs.
3220 * config/mips/mips.opt: Introduce -mexplicit-relocs= option
3221 and define -m(no-)explicit-relocs as aliases.
3223 2024-01-24 Alex Coplan <alex.coplan@arm.com>
3225 * config/aarch64/aarch64.opt (-mearly-ldp-fusion): Set default
3227 (-mlate-ldp-fusion): Likewise.
3229 2024-01-24 Tamar Christina <tamar.christina@arm.com>
3231 * tree-vect-loop.cc (vect_get_vect_def,
3232 vect_create_epilog_for_reduction): Rename main_exit_p to
3235 2024-01-24 Tamar Christina <tamar.christina@arm.com>
3237 PR tree-optimization/113364
3238 * tree-vect-loop.cc (vect_create_epilog_for_reduction): If all exits all
3239 early exits then we must reduce from the first offset for all of them.
3241 2024-01-24 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3244 * config/riscv/riscv-vsetvl.cc (get_expr_id): Remove.
3246 (get_bb_index): Ditto.
3247 (pre_vsetvl::compute_avl_def_data): Ditto.
3248 (pre_vsetvl::earliest_fuse_vsetvl_info): Fix large memory usage.
3249 (pre_vsetvl::pre_global_vsetvl_info): Ditto.
3251 2024-01-23 Andrew Pinski <quic_apinski@quicinc.com>
3252 Richard Sandiford <richard.sandiford@arm.com>
3255 * ccmp.cc (ccmp_candidate_p): Add outer argument.
3256 Allow if the outer is true and the lhs is used more
3258 (expand_ccmp_expr): Update call to ccmp_candidate_p.
3259 * expr.h (expand_expr_real_gassign): Declare.
3260 * expr.cc (expand_expr_real_gassign): New function, split out from...
3261 (expand_expr_real_1): ...here.
3262 * cfgexpand.cc (expand_gimple_stmt_1): Use expand_expr_real_gassign.
3264 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3267 * config/aarch64/aarch64-ldp-fusion.cc (reset_debug_use): New.
3268 (fixup_debug_use): New.
3269 (fixup_debug_uses_trailing_add): New.
3270 (fixup_debug_uses): New. Use it ...
3271 (ldp_bb_info::fuse_pair): ... here.
3272 (try_promote_writeback): Call fixup_debug_uses_trailing_add to
3273 fix up debug uses of the base register that are affected by
3274 folding in the trailing add insn.
3276 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3279 * config/aarch64/aarch64-ldp-fusion.cc (ldp_bb_info::fuse_pair):
3280 Update trailing nondebug uses of the base register in the case
3281 of cancelling writeback.
3283 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3286 * rtl-ssa/accesses.h (use_info::next_debug_insn_use): New.
3287 (debug_insn_use_iterator): New.
3288 (set_info::first_debug_insn_use): New.
3289 (set_info::debug_insn_uses): New.
3290 * rtl-ssa/member-fns.inl (use_info::next_debug_insn_use): New.
3291 (set_info::first_debug_insn_use): New.
3292 (set_info::debug_insn_uses): New.
3294 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3297 * config/aarch64/aarch64-ldp-fusion.cc (ldp_bb_info::try_fuse_pair):
3298 Don't record hazards against the opposite insn in the pair.
3300 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3303 * config/aarch64/aarch64-ldp-fusion.cc
3304 (struct stp_change_builder): New.
3305 (decide_stp_strategy): Reanme to ...
3306 (try_repurpose_store): ... this.
3307 (ldp_bb_info::fuse_pair): Refactor to use stp_change_builder to
3308 construct stp changes. Fix up uses when inserting new stp insns.
3310 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3313 * rtl-ssa.h: Include hash-set.h.
3314 * rtl-ssa/changes.cc (function_info::finalize_new_accesses): Add
3315 new_sets parameter and use it to keep track of new user-created sets.
3316 (function_info::apply_changes_to_insn): Also call add_def on new sets.
3317 (function_info::change_insns): Add hash_set to keep track of new
3318 user-created defs. Plumb it through.
3319 * rtl-ssa/functions.h: Add hash_set parameter to finalize_new_accesses and
3320 apply_changes_to_insn.
3322 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3325 * rtl-ssa/accesses.cc (function_info::create_use): New.
3326 * rtl-ssa/changes.cc (function_info::finalize_new_accesses):
3327 Ensure new uses end up referring to permanent defs.
3328 * rtl-ssa/functions.h (function_info::create_use): Declare.
3330 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3333 * rtl-ssa/changes.cc (function_info::change_insns): Split out the call
3334 to finalize_new_accesses from the backwards placement loop, run it
3335 forwards in a separate loop.
3337 2024-01-23 Richard Biener <rguenther@suse.de>
3339 PR tree-optimization/113552
3340 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Use
3341 floor_log2 instead of exact_log2 on the number of calls.
3343 2024-01-23 Jeff Law <jlaw@ventanamicro.com>
3344 Jakub Jelinek <jakub@redhat.com>
3346 * config/ia64/ia64.cc (ia64_start_function): Add ATTRIBUTE_UNUSED to
3349 2024-01-23 Richard Biener <rguenther@suse.de>
3351 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
3352 Separate single and multi-exit case when creating PHIs between
3353 the main and epilogue.
3355 2024-01-23 Richard Sandiford <richard.sandiford@arm.com>
3358 * config/aarch64/aarch64-sve-builtins-shapes.cc (build_one): Skip
3359 MODE_single variants of functions that don't take tuple arguments.
3361 2024-01-23 Alex Coplan <alex.coplan@arm.com>
3364 * config/aarch64/aarch64-ldp-fusion.cc (try_promote_writeback):
3365 Don't assert recog success, just punt if the writeback pair
3368 2024-01-23 Jakub Jelinek <jakub@redhat.com>
3370 * config/gcn/gcn.cc (gcn_hsa_declare_function_name): Add
3371 ATTRIBUTE_UNUSED to decl.
3373 2024-01-23 Richard Biener <rguenther@suse.de>
3376 * dwarf2out.cc (dwarf2out_die_ref_for_decl): Gracefully
3377 handle unexpected but bogus DIE contexts when not checking
3380 2024-01-23 Jakub Jelinek <jakub@redhat.com>
3382 PR tree-optimization/113462
3383 * fold-const.cc (native_interpret_int): Don't punt if total_bytes
3384 is larger than HOST_BITS_PER_DOUBLE_INT / BITS_PER_UNIT.
3385 (fold_view_convert_expr): Use XALLOCAVEC buffers for types with
3386 sizes between 129 and 8192 bytes.
3388 2024-01-23 Xi Ruoyao <xry111@xry111.site>
3390 * config/loongarch/loongarch.cc (loongarch_explicit_relocs_p):
3391 If la_opt_explicit_relocs is EXPLICIT_RELOCS_AUTO, return false
3392 for SYMBOL_TLS_LDM and SYMBOL_TLS_GD.
3393 (loongarch_call_tls_get_addr): Do not split symbols of
3394 SYMBOL_TLS_LDM or SYMBOL_TLS_GD if la_opt_explicit_relocs is
3395 EXPLICIT_RELOCS_AUTO.
3397 2024-01-23 Richard Biener <rguenther@suse.de>
3399 * alias.cc (known_base_value_p): Remove.
3400 (find_base_value): Remove PLUS/MINUS handling
3401 when both operands are not CONST_INT_P.
3403 2024-01-23 Richard Biener <rguenther@suse.de>
3405 PR rtl-optimization/113255
3406 * alias.cc (find_base_term): Remove PLUS/MINUS handling
3407 when both operands are not CONST_INT_P.
3409 2024-01-23 Richard Biener <rguenther@suse.de>
3412 * dwarf2out.cc (dwarf2out_finish): Reset all type units
3413 for the fat part of an LTO compile.
3415 2024-01-23 chenxiaolong <chenxiaolong@loongson.cn>
3417 * doc/sourcebuild.texi: Add attributes for keywords.
3419 2024-01-23 Sandra Loosemore <sandra@codesourcery.com>
3422 * doc/invoke.texi (Warning Options): Correct lists of options
3423 enabled by -Wall and -Wextra by checking against common.opt
3426 2024-01-22 Andrew Pinski <quic_apinski@quicinc.com>
3429 * config/arm/parsecpu.awk (check_cpu): Use cpu_opt_alias
3430 instead of cpu_optaliases.
3431 (check_arch): Use arch_opt_alias instead of arch_optaliases.
3433 2024-01-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3435 * config/riscv/riscv-protos.h (splat_to_scalar_move_p): New function.
3436 * config/riscv/riscv-v.cc (splat_to_scalar_move_p): Ditto.
3437 * config/riscv/vector.md: Simplify vmv.v.x. into vmv.s.x.
3439 2024-01-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3442 * config/riscv/riscv.md: Use reg instead of subreg.
3444 2024-01-22 Tobias Burnus <tburnus@baylibre.com>
3447 * config/gcn/mkoffload.cc (elf_arch): Change default to gfx900
3448 to match the compiler default.
3449 (simple_object_copy_lto_debug_sections): Never unlink the outfile
3450 on error as the caller does so.
3451 (maybe_unlink, compile_native): Use %<...%> and %qs in fatal_error.
3452 (main): Likewise. Fix 'mkoffload.dbg.o' cleanup.
3454 2024-01-22 Richard Biener <rguenther@suse.de>
3456 PR tree-optimization/113373
3457 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
3458 Create LC PHIs in the exit blocks where necessary.
3459 * tree-vect-loop.cc (vectorizable_live_operation): Do not try
3460 to handle missing LC PHIs.
3461 (find_connected_edge): Remove.
3462 (vect_create_epilog_for_reduction): Cleanup use of auto_vec.
3464 2024-01-22 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3466 * config/riscv/vector.md: Fix vfirst/vmsbf/vmsof ratio attributes.
3468 2024-01-22 xuli <xuli1@eswincomputing.com>
3471 * config/riscv/riscv-vector-builtins.cc (has_vxrm_or_frm_p):remove.
3472 (registered_function::overloaded_hash):refactor.
3473 (resolve_overloaded_builtin):avoid internal ICE.
3475 2024-01-21 Mikael Pettersson <mikpelinux@gmail.com>
3479 * calls.cc (emit_library_call_value_1): Pass valid TYPE
3481 * expr.cc (emit_push_insn): Likewise.
3483 2024-01-21 Jeff Law <jlaw@ventanamicro.com>
3485 * config/riscv/riscv.cc (riscv_init_cumulative_args): Install
3486 correcction version of last change.
3488 2024-01-21 Jeff Law <jlaw@ventanamicro.com>
3490 * config/riscv/riscv.cc (riscv_init_cumulative_args): Update and
3491 fix bugs in signature.
3493 2024-01-21 Roger Sayle <roger@nextmovesoftware.com>
3494 Richard Biener <rguenther@suse.de>
3496 PR rtl-optimization/111267
3497 * fwprop.cc (fwprop_propagation::profitabe_p): Rename
3498 profitable_p method to likely_profitable_p.
3499 (try_fwprop_subst_node): Update call to likely_profitable_p.
3500 Only bail-out early when !prop.likely_profitable_p for instructions
3501 that are not single sets. When comparing costs, bail-out if the
3502 cost is unchanged and !prop.likely_profitable_p.
3504 2024-01-21 Sandra Loosemore <sandra@codesourcery.com>
3507 * doc/invoke.texi (Warning Options): Document that -Wunused-parameter
3508 isn't enabled by -Wunused unless -Wextra is provided, and that
3509 -Wunused does enable -Wunused-const-variable=1 for C. Clarify that
3510 -Wunused doesn't enable -Wunused-* options documented as behaving
3511 otherwise, and list them explicitly.
3513 2024-01-21 Sandra Loosemore <sandra@codesourcery.com>
3516 * doc/invoke.texi (Warning Options): Fix broken example and
3517 clean up/reorganize the others. Also describe what the short-form
3520 2024-01-20 Sandra Loosemore <sandra@codesourcery.com>
3523 * doc/invoke.texi (Option Summary): Add -Warray-parameter.
3524 (Warning Options): Correct/edit discussion of -Warray-parameter
3525 to make the first example less confusing, and fill in missing info.
3527 2024-01-20 Jakub Jelinek <jakub@redhat.com>
3529 PR tree-optimization/113462
3530 * gimple-lower-bitint.cc (bitint_large_huge::handle_cast):
3531 Handle rhs1 INTEGER_CST like SSA_NAME.
3533 2024-01-20 Jakub Jelinek <jakub@redhat.com>
3535 PR tree-optimization/113491
3536 * tree-switch-conversion.cc (switch_conversion::build_constructors):
3537 If elt.index has precision higher than sizetype, fold_convert it to
3539 (switch_conversion::array_value_type): Return type if type is
3540 BITINT_TYPE with precision above MAX_FIXED_MODE_SIZE or with BLKmode.
3541 (switch_conversion::build_arrays): Use unsigned_type_for rather than
3542 lang_hooks.types.type_for_mode if utype is BITINT_TYPE with precision
3543 above MAX_FIXED_MODE_SIZE or with BLKmode. If utype has precision
3544 higher than sizetype, use sizetype as tidx type and fold_convert the
3545 subtraction to sizetype.
3547 2024-01-20 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3549 * config/riscv/riscv.cc (riscv_init_cumulative_args): Suppress warning.
3550 (riscv_vector_mode_supported_any_target_p): Ditto.
3552 2024-01-19 Mikael Pettersson <mikpelinux@gmail.com>
3555 * config/m68k/m68k.cc (m68k_zero_call_used_regs): New function.
3556 (TARGET_ZERO_CALL_USED_REGS): Define.
3558 2024-01-19 Mikael Pettersson <mikpelinux@gmail.com>
3561 * config/m68k/m68k.cc (output_andsi3): Use QImode for
3562 address adjusted for 1-byte RMW access.
3563 (output_iorsi3): Likewise.
3564 (output_xorsi3): Likewise.
3566 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3568 * doc/invoke.texi (RISC-V Options): Add list of supported
3571 2024-01-19 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3574 * config/riscv/riscv-protos.h (RVV_VLMAX): Change to regno_reg_rtx[X0_REGNUM].
3575 (RVV_VUNDEF): Ditto.
3576 * config/riscv/riscv-vsetvl.cc: Add timevar.
3578 2024-01-19 Richard Biener <rguenther@suse.de>
3581 * lto-streamer-in.cc (lto_read_tree_1): When there isn't
3582 an early DIE but there should be, do not pretend there is.
3584 2024-01-19 Richard Biener <rguenther@suse.de>
3586 PR tree-optimization/113494
3587 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
3588 Handle endless loop on exit. Handle re-allocated PHI.
3590 2024-01-19 Jakub Jelinek <jakub@redhat.com>
3592 PR tree-optimization/113464
3593 * gimple-lower-bitint.cc (gimple_lower_bitint): Don't try to
3594 optimize loads into GIMPLE_ASM stmts.
3596 2024-01-19 Jakub Jelinek <jakub@redhat.com>
3598 PR tree-optimization/113463
3599 * gimple-ssa-warn-restrict.cc (builtin_memref::extend_offset_range):
3600 Only look through NOP_EXPRs if rhs1 doesn't have wider type than
3603 2024-01-19 Jakub Jelinek <jakub@redhat.com>
3605 PR tree-optimization/113459
3606 * tree-ssa-sccvn.cc (vn_walk_cb_data::push_partial_def): Use
3607 TREE_INT_CST_LOW of TYPE_SIZE_UNIT rather than GET_MODE_SIZE
3608 of SCALAR_INT_TYPE_MODE if type has BLKmode.
3609 (vn_reference_lookup_3): Likewise. Formatting fix.
3611 2024-01-19 Jakub Jelinek <jakub@redhat.com>
3612 Richard Biener <rguenther@suse.de>
3614 * cfgexpand.cc (discover_nonconstant_array_refs_r): Force non-BLKmode
3615 VAR_DECLs referenced in BLKmode VIEW_CONVERT_EXPRs into memory.
3616 * expr.cc (expand_expr_real_1) <case VIEW_CONVERT_EXPR>: Do nothing
3617 but adjust_address also for BLKmode mode and MEM op0.
3619 2024-01-19 Palmer Dabbelt <palmer@rivosinc.com>
3621 * common/config/riscv/riscv-common.cc: Add Zihpm and Zicnttr
3624 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3626 * doc/invoke.texi (RISC-V Options): Document the syntax of -march.
3628 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3630 * common/config/riscv/riscv-common.cc
3631 (riscv_subset_list::parse_std_ext): Remove.
3632 (riscv_subset_list::parse_multiletter_ext): Remove.
3633 * config/riscv/riscv-subset.h
3634 (riscv_subset_list::parse_std_ext): Remove.
3635 (riscv_subset_list::parse_multiletter_ext): Remove.
3637 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3639 * common/config/riscv/riscv-common.cc
3640 (riscv_subset_list::parse_single_std_ext): New parameter.
3641 (riscv_subset_list::parse_single_multiletter_ext): Ditto.
3642 (riscv_subset_list::parse_single_ext): Ditto.
3643 (riscv_subset_list::parse): Relax the order for the input of ISA
3645 * config/riscv/riscv-subset.h
3646 (riscv_subset_list::parse_single_std_ext): New parameter.
3647 (riscv_subset_list::parse_single_multiletter_ext): Ditto.
3648 (riscv_subset_list::parse_single_ext): Ditto.
3650 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3652 * common/config/riscv/riscv-common.cc
3653 (riscv_subset_list::parse_base_ext): New.
3654 (riscv_subset_list::parse): Extract part of logic into
3655 riscv_subset_list::parse_base_ext.
3656 * config/riscv/riscv-subset.h (riscv_subset_list::parse_base_ext):
3659 2024-01-19 Kito Cheng <kito.cheng@sifive.com>
3661 * config/riscv/riscv.cc (riscv_override_options_internal): Tweak
3664 2024-01-19 Kuan-Lin Chen <rufus@andestech.com>
3666 * config/riscv/vector-crypto.md (UNSPEC_CLMUL): Rename to
3669 2024-01-19 Sandra Loosemore <sandra@codesourcery.com>
3672 * doc/extend.texi (Common Variable Attributes): Explain what
3673 happens when multiple variables with cleanups are in the same scope.
3675 2024-01-18 Sandra Loosemore <sandra@codesourcery.com>
3678 * doc/extend.texi (Common Function Attributes): Document that
3679 noinline also disables some interprocedural optimizations and
3680 improve flow to the part about using inline asm instead to
3681 disable calls from being optimized away completely. Remove the
3682 sentence that says noipa is mainly for internal compiler testing.
3684 2024-01-18 John David Anglin <danglin@gcc.gnu.org>
3686 PR tree-optimization/69807
3687 * config/pa/pa.cc (pa_option_override): Set flag_pie on TARGET_64BIT.
3689 2024-01-18 Brian Inglis <Brian.Inglis@Shaw.ca>
3692 * doc/invoke.texi (Option Summary): Remove -mcygwin and -mno-cygwin
3693 from x86 Windows Options.
3695 2024-01-18 Sandra Loosemore <sandra@codesourcery.com>
3698 * doc/extend.texi (C Extensions): Add new section to menu.
3699 (Function Attributes): Move dangling index entries to....
3700 (Const and Volatile Functions): New section.
3702 2024-01-18 David Malcolm <dmalcolm@redhat.com>
3704 PR middle-end/112684
3705 * toplev.cc (toplev::main): Don't ICE in
3706 -fdiagnostics-generate-patch when exiting after options,
3707 since no edit context will have been created.
3709 2024-01-18 Richard Biener <rguenther@suse.de>
3711 * tree-vect-stmts.cc (vectorizable_store): Do not pre-allocate
3714 2024-01-18 Iain Sandoe <iain@sandoe.co.uk>
3716 * Makefile.in: Emit ENABLE_DARWIN_AT_RPATH into site.exp
3717 when ENABLE_DARWIN_AT_RPATH_TRUE is not '#'.
3719 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3720 Jin Ma <jinma@linux.alibaba.com>
3721 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3722 Christoph Müllner <christoph.muellner@vrull.eu>
3724 * config/riscv/thead.cc
3725 (th_asm_output_opcode): Rewrite some instructions.
3727 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3728 Jin Ma <jinma@linux.alibaba.com>
3729 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3730 Christoph Müllner <christoph.muellner@vrull.eu>
3732 * config/riscv/riscv.md (none,thv,rvv): New attribute.
3733 (no,yes): Add an attribute to disable alternative
3734 for xtheadvector or RVV1.0.
3735 * config/riscv/vector.md:
3736 Disable alternatives that destination register overlaps
3737 source register group for xtheadvector.
3739 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3740 Jin Ma <jinma@linux.alibaba.com>
3741 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3742 Christoph Müllner <christoph.muellner@vrull.eu>
3744 * config/riscv/riscv-vector-builtins-bases.cc
3745 (class th_loadstore_width): Define new builtin bases.
3746 (class th_extract): Define new builtin bases.
3747 (BASE): Define new builtin bases.
3748 * config/riscv/riscv-vector-builtins-bases.h:
3749 Define new builtin class.
3750 * config/riscv/riscv-vector-builtins-shapes.cc
3751 (struct th_loadstore_width_def): Define new builtin shapes.
3752 (struct th_indexed_loadstore_width_def):
3753 Define new builtin shapes.
3754 (struct th_extract_def): Define new builtin shapes.
3755 (SHAPE): Define new builtin shapes.
3756 * config/riscv/riscv-vector-builtins-shapes.h:
3757 Define new builtin shapes.
3758 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_FUNCTION):
3759 Redefine DEF_RVV_FUNCTION for XTheadVector special intrinsics.
3760 * config/riscv/riscv-vector-builtins.h
3761 (enum required_ext): Add new XTheadVector member.
3762 (struct function_group_info): Likewise.
3763 * config/riscv/t-riscv:
3764 Add thead-vector-builtins-functions.def
3765 * config/riscv/thead-vector.md
3766 (@pred_mov_width<vlmem_op_attr><mode>): Add new patterns.
3767 (*pred_mov_width<vlmem_op_attr><mode>): Likewise.
3768 (@pred_store_width<vlmem_op_attr><mode>): Likewise.
3769 (@pred_strided_load_width<vlmem_op_attr><mode>): Likewise.
3770 (@pred_strided_store_width<vlmem_op_attr><mode>): Likewise.
3771 (@pred_indexed_load_width<vlmem_op_attr><mode>): Likewise.
3772 (@pred_th_extract<mode>): Likewise.
3773 (*pred_th_extract<mode>): Likewise.
3774 * config/riscv/thead-vector-builtins-functions.def: New file.
3776 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3777 Jin Ma <jinma@linux.alibaba.com>
3778 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3779 Christoph Müllner <christoph.muellner@vrull.eu>
3781 * config.gcc: Add files for XTheadVector intrinsics.
3782 * config/riscv/autovec.md: Guard XTheadVector.
3783 * config/riscv/predicates.md: Disable immediate vl
3785 * config/riscv/riscv-c.cc (riscv_pragma_intrinsic):
3786 Add pragma for XTheadVector.
3787 * config/riscv/riscv-string.cc (riscv_expand_block_move):
3789 * config/riscv/riscv-v.cc (vls_mode_valid_p):
3791 * config/riscv/riscv-vector-builtins-bases.cc:
3792 Do not normalize vsetvl instructions for XTheadVector.
3793 * config/riscv/riscv-vector-builtins-shapes.cc (check_type):
3794 New check type function.
3795 (build_one): Adjust for XTheadVector.
3796 * config/riscv/riscv-vector-switch.def (ENTRY):
3797 Disable fractional mode for the XTheadVector extension.
3798 (TUPLE_ENTRY): Likewise.
3799 * config/riscv/riscv.cc (riscv_v_adjust_bytesize):
3801 (riscv_preferred_simd_mode): Likewsie.
3802 (riscv_autovectorize_vector_modes): Likewise.
3803 (riscv_vector_mode_supported_any_target_p): Likewise.
3804 (TARGET_VECTOR_MODE_SUPPORTED_ANY_TARGET_P): Likewise.
3805 * config/riscv/thead.cc (th_asm_output_opcode):
3806 Rewrite vsetvl instructions.
3807 * config/riscv/vector.md:
3808 Include thead-vector.md and change fractional LMUL
3810 * config/riscv/riscv_th_vector.h: New file.
3811 * config/riscv/thead-vector.md: New file.
3813 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3814 Jin Ma <jinma@linux.alibaba.com>
3815 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3816 Christoph Müllner <christoph.muellner@vrull.eu>
3818 * config/riscv/riscv-protos.h (riscv_asm_output_opcode):
3819 Add new function to add assembler insn code prefix/suffix.
3820 (th_asm_output_opcode):
3821 Add Thead function to add assembler insn code prefix/suffix.
3822 * config/riscv/riscv.cc (riscv_asm_output_opcode):
3823 Implement function to add assembler insn code prefix/suffix.
3824 * config/riscv/riscv.h (ASM_OUTPUT_OPCODE):
3825 Add new function to add assembler insn code prefix/suffix.
3826 * config/riscv/thead.cc (th_asm_output_opcode):
3827 Implement Thead function to add assembler insn code
3830 2024-01-18 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
3831 Jin Ma <jinma@linux.alibaba.com>
3832 Xianmiao Qu <cooper.qu@linux.alibaba.com>
3833 Christoph Müllner <christoph.muellner@vrull.eu>
3835 * common/config/riscv/riscv-common.cc
3836 (riscv_subset_list::parse): Add new vendor extension.
3837 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins):
3839 * config/riscv/riscv.opt: Add new mask.
3841 2024-01-18 Iain Sandoe <iain@sandoe.co.uk>
3843 * config/darwin.h (DARWIN_RPATH_SPEC): Arrange for the %P spec
3844 to be conditional on macosx-version-min.
3846 2024-01-18 Iain Sandoe <iain@sandoe.co.uk>
3848 * config/darwin.cc (darwin_objc1_section): Use the correct
3849 meta-data version for constant strings.
3850 (machopic_select_section): Assert if we fail to handle CFString
3851 sections as Obejctive-C meta-data or drectly.
3853 2024-01-18 Iain Sandoe <iain@sandoe.co.uk>
3855 * lto-section-names.h (OFFLOAD_SECTION_NAME_PREFIX,
3856 OFFLOAD_VAR_TABLE_SECTION_NAME, OFFLOAD_FUNC_TABLE_SECTION_NAME,
3857 OFFLOAD_IND_FUNC_TABLE_SECTION_NAME): Provide Mach-O syntax
3858 versions when the object format is Mach-O.
3860 2024-01-18 Iain Sandoe <iain@sandoe.co.uk>
3863 * config/darwin.cc (machopic_select_section): Handle C and C++
3865 (darwin_rename_builtins): Move this out of the CFString code.
3866 (darwin_libc_has_function): Likewise.
3867 (darwin_build_constant_cfstring): Create an anonymous var to
3869 * config/darwin.h (ASM_OUTPUT_LABELREF): Handle constant
3872 2024-01-18 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
3875 * haifa-sched.cc (dep_list_size): Make global.
3876 * sched-deps.cc (find_inc): Use instead of sd_lists_size().
3877 * sched-int.h (dep_list_size): Declare.
3879 2024-01-18 Martin Jambor <mjambor@suse.cz>
3881 PR tree-optimization/110422
3882 * tree-sra.cc (scan_function): Disqualify bases of operands of asm
3885 2024-01-18 Richard Biener <rguenther@suse.de>
3887 PR tree-optimization/113475
3888 * gimple-range-phi.h (phi_analyzer::m_phi_groups): New.
3889 * gimple-range-phi.cc (phi_analyzer::phi_analyzer): Initialize.
3890 (phi_analyzer::~phi_analyzer): Deallocate and free collected
3892 (phi_analyzer::process_phi): Record allocated phi_groups.
3894 2024-01-18 Richard Biener <rguenther@suse.de>
3896 * tree-vect-stmts.cc (vectorizable_store): Do not allocate
3897 storage for gvec_oprnds elements.
3899 2024-01-18 Richard Biener <rguenther@suse.de>
3901 * tree-vect-loop.cc (vec_init_loop_exit_info): Adjust comment,
3902 prefer all later exits we can handle.
3903 (vect_analyze_loop_form): Free the allocated loop body.
3906 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3908 * config/avr/avr-log.cc: Tabify.
3910 2024-01-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3912 * config/riscv/autovec.md: Support vi variant.
3914 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3916 * config/avr/avr-devices.cc: Tabify.
3918 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3920 * config/avr/avr-c.cc: Tabify.
3922 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3924 * config/avr/driver-avr.cc: Tabify.
3926 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3928 * config/avr/gen-avr-mmcu-texi.cc: Tabify.
3930 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3932 * config/avr/gen-avr-mmcu-specs.cc: Tabify.
3934 2024-01-18 Jakub Jelinek <jakub@redhat.com>
3936 * config/riscv/riscv.opt (mshorten-memrefs, mrelax, mcsr-check,
3937 minline-strcmp, minline-strncmp, minline-strlen,
3938 -param=riscv-vector-abi): Remove Bool keywords.
3940 2024-01-18 Jakub Jelinek <jakub@redhat.com>
3943 * config/i386/i386.cc (x86_function_profiler): Add -masm=intel
3944 support. Add missing space after , in emitted assembly in some
3945 cases. Formatting fixes.
3947 2024-01-18 Xi Ruoyao <xry111@xry111.site>
3949 * config/loongarch/loongarch.md (movsi_internal): Remove
3952 2024-01-18 Georg-Johann Lay <avr@gjlay.de>
3954 * config/avr/gen-avr-mmcu-specs.cc (diagnose_rodata_in_ram): Fix typo
3955 in the diagnostic, and capitalize the device name.
3956 (print_mcu): Generate specs such that:
3957 <*check_rodata_in_ram>: New.
3958 <*cc1_misc>: Use check_rodata_in_ram instead of cc1_rodata_in_ram.
3959 <*link_misc>: Use check_rodata_in_ram instead of link_rodata_in_ram.
3960 <*cc1_rodata_in_ram, *link_rodata_in_ram>: Remove.
3962 2024-01-18 Jakub Jelinek <jakub@redhat.com>
3965 * common.opt (ffold-mem-offsets): Remove Target and Bool keywords, add
3966 Common and Optimization.
3968 2024-01-18 Richard Biener <rguenther@suse.de>
3970 PR tree-optimization/113431
3971 * tree-vect-data-refs.cc (vect_preserves_scalar_order_p):
3972 When there is an invariant load we might not preserve
3975 2024-01-18 Richard Biener <rguenther@suse.de>
3977 PR tree-optimization/113374
3978 * tree-ssa-operands.h (SET_PHI_ARG_DEF_ON_EDGE): New.
3979 * tree-vect-loop.cc (move_early_exit_stmts): Update
3981 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
3982 Refactor. Preserve virtual LC PHIs on all exits.
3984 2024-01-18 Lulu Cheng <chenglulu@loongson.cn>
3986 * config/loongarch/loongarch.cc (loongarch_split_symbol):
3987 Assign the '/u' attribute to the mem.
3989 2024-01-18 Sandra Loosemore <sandra@codesourcery.com>
3991 PR middle-end/110847
3992 * doc/invoke.texi (Option Summary): Document negative forms of
3993 -Wtsan and -Wxor-used-as-pow.
3994 (Warning Options): Likewise.
3996 2024-01-18 Juzhe-Zhong <juzhe.zhong@rivai.ai>
3999 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::earliest_fuse_vsetvl_info): Fix bug.
4001 2024-01-18 Sandra Loosemore <sandra@codesourcery.com>
4003 * doc/extend.texi (Common Function Attributes): Re-alphabetize
4005 (Common Variable Attributes): Likewise.
4006 (Common Type Attributes): Likewise.
4008 2024-01-17 Sandra Loosemore <sandra@codesourcery.com>
4010 PR middle-end/111659
4011 * doc/extend.texi (Common Variable Attributes): Fix long lines
4012 in documentation of strict_flex_array + other minor copy-editing.
4013 Add a cross-reference to -Wstrict-flex-arrays.
4014 * doc/invoke.texi (Option Summary): Fix whitespace in tables
4015 before -fstrict-flex-arrays and -Wstrict-flex-arrays.
4016 (C Dialect Options): Combine the docs for the two
4017 -fstrict-flex-arrays forms into a single entry. Note this option
4018 is for C/C++ only. Add a cross-reference to -Wstrict-flex-arrays.
4019 (Warning Options): Note -Wstrict-flex-arrays is for C/C++ only.
4020 Minor copy-editing. Add cross references to the strict_flex_array
4021 attribute and -fstrict-flex-arrays option. Add note that this
4022 option depends on -ftree-vrp.
4024 2024-01-17 Andrew Pinski <quic_apinski@quicinc.com>
4027 * config/aarch64/predicates.md (aarch64_ldp_reg_operand): For subreg,
4028 only allow REG operands instead of allowing all.
4030 2024-01-17 Vineet Gupta <vineetg@rivosinc.com>
4032 * config/riscv/riscv-vsetvl.cc (earliest_fuse_vsetvl_info):
4033 Remove redundant checks in else condition for readablity.
4034 (earliest_fuse_vsetvl_info) Print iteration count in debug
4036 (earliest_fuse_vsetvl_info) Fix misleading vsetvl info
4037 dump details in certain cases.
4039 2024-01-17 Vineet Gupta <vineetg@rivosinc.com>
4041 * config/riscv/riscv.opt: New -param=vsetvl-strategy.
4042 * config/riscv/riscv-opts.h: New enum vsetvl_strategy_enum.
4043 * config/riscv/riscv-vsetvl.cc
4044 (pre_vsetvl::pre_global_vsetvl_info): Use vsetvl_strategy.
4045 (pass_vsetvl::execute): Use vsetvl_strategy.
4047 2024-01-17 Jan Hubicka <jh@suse.cz>
4049 * ipa-polymorphic-call.cc (ipa_polymorphic_call_context::set_by_invariant): Remove
4050 accidental hack reseting offset.
4052 2024-01-17 Jan Hubicka <jh@suse.cz>
4054 * config/i386/i386-options.cc (ix86_option_override_internal): Fix
4055 handling of X86_TUNE_AVOID_512FMA_CHAINS.
4057 2024-01-17 Jan Hubicka <jh@suse.cz>
4058 Jakub Jelinek <jakub@redhat.com>
4060 PR tree-optimization/110852
4061 * predict.cc (expr_expected_value_1): Fix profile merging of PHI and
4063 (get_predictor_value): Handle PRED_COMBINED_VALUE_PREDICTIONS and
4064 PRED_COMBINED_VALUE_PREDICTIONS_PHI
4065 * predict.def (PRED_COMBINED_VALUE_PREDICTIONS): New predictor.
4066 (PRED_COMBINED_VALUE_PREDICTIONS_PHI): New predictor.
4068 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4070 PR tree-optimization/113421
4071 * gimple-lower-bitint.cc (stmt_needs_operand_addr): Adjust function
4073 (bitint_dom_walker::before_dom_children): Add g temporary to simplify
4074 formatting. Start at vop rather than cvop even if stmt is a store
4075 and needs_operand_addr.
4077 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4079 PR middle-end/113410
4080 * gimple-ssa-warn-access.cc (pass_waccess::maybe_check_access_sizes):
4081 If access_nelts is integral with larger precision than sizetype,
4082 fold_convert it to sizetype.
4084 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4086 PR tree-optimization/113408
4087 * gimple-lower-bitint.cc (bitint_large_huge::handle_stmt): For
4088 VIEW_CONVERT_EXPR, pass TREE_OPERAND (rhs1, 0) rather than rhs1
4091 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4093 PR middle-end/113406
4094 * ipa-strub.cc (pass_ipa_strub::execute): Check aggregate_value_p
4095 regardless of whether is_gimple_reg_type (restype) or not.
4097 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4099 * tree-into-ssa.cc (pass_build_ssa::gate): Fix comment typo,
4100 funcions -> functions, and use were instead of was.
4101 * gengtype.cc (dump_typekind): Fix comment typos, funcion -> function
4102 and guaranteee -> guarantee.
4103 * attribs.h (struct attr_access): Fix comment typo funcion -> function.
4105 2024-01-17 Jakub Jelinek <jakub@redhat.com>
4107 PR middle-end/113409
4108 * omp-general.cc (omp_adjust_for_condition): Handle BITINT_TYPE like
4110 (omp_extract_for_data): Use build_bitint_type rather than
4111 build_nonstandard_integer_type if either iter_type or loop->v type
4113 * omp-expand.cc (expand_omp_for_generic,
4114 expand_omp_taskloop_for_outer, expand_omp_taskloop_for_inner): Handle
4115 BITINT_TYPE like INTEGER_TYPE.
4117 2024-01-17 Richard Biener <rguenther@suse.de>
4119 PR tree-optimization/113371
4120 * tree-vect-data-refs.cc (vect_enhance_data_refs_alignment):
4121 Do not peel when LOOP_VINFO_EARLY_BREAKS_VECT_PEELED.
4122 * tree-vect-loop-manip.cc (vect_do_peeling): Assert we do
4123 not perform prologue peeling when LOOP_VINFO_EARLY_BREAKS_VECT_PEELED.
4125 2024-01-17 Maxim Kuvyrkov <maxim.kuvyrkov@linaro.org>
4127 PR rtl-optimization/96388
4128 PR rtl-optimization/111554
4129 * sched-deps.cc (find_inc): Avoid exponential behavior.
4131 2024-01-17 Sandra Loosemore <sandra@codesourcery.com>
4134 * doc/invoke.texi (Option Summary): Move -Wuseless-cast
4135 from C++ Language Options to Warning Options. Add entry for
4137 (C++ Dialect Options): Move -Wuse-after-free and -Wuseless-cast
4139 (Warning Options): ...to here. Minor copy-editing to fix typo
4142 2024-01-17 YunQiang Su <syq@gcc.gnu.org>
4144 * config/mips/mips.cc (mips_compute_frame_info): If another
4145 register is used as global_pointer, mark $GP live false.
4147 2024-01-17 Sandra Loosemore <sandra@codesourcery.com>
4150 * doc/extend.texi (BPF Built-in Functions): Wrap long lines and
4151 give the section a light copy-editing pass.
4153 2024-01-16 Wilco Dijkstra <wilco.dijkstra@arm.com>
4155 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Add 'cobalt-100' CPU.
4156 * config/aarch64/aarch64-tune.md: Regenerated.
4157 * doc/invoke.texi (-mcpu): Add cobalt-100 core.
4159 2024-01-16 Wilco Dijkstra <wilco.dijkstra@arm.com>
4162 * config/aarch64/aarch64.cc (aarch64_legitimize_address): Reassociate
4163 badly formed CONST expressions.
4165 2024-01-16 Daniel Cederman <cederman@gaisler.com>
4167 * config/sparc/sparc.cc (next_active_non_empty_insn): Length 0 treated as empty
4169 2024-01-16 Daniel Cederman <cederman@gaisler.com>
4171 * config/sparc/sparc.cc (atomic_insn_for_leon3_p): Treat membar_storeload as atomic
4172 * config/sparc/sync.md (membar_storeload): Turn into named insn
4173 and add GR712RC errata workaround.
4174 (membar_v8): Add GR712RC errata workaround.
4176 2024-01-16 Andreas Larsson <andreas@gaisler.com>
4178 * config/sparc/sync.md (*membar_storeload_leon3): Remove
4179 (*membar_storeload): Enable for LEON
4181 2024-01-16 Jakub Jelinek <jakub@redhat.com>
4183 PR tree-optimization/113372
4185 PR middle-end/110115
4186 PR middle-end/111422
4187 * cfgexpand.cc (add_scope_conflicts_2): New function.
4188 (add_scope_conflicts_1): Use it.
4190 2024-01-16 Georg-Johann Lay <avr@gjlay.de>
4192 * config/avr/avr-mcus.def (avr16eb14, avr16eb20, avr16eb28, avr16eb32)
4193 (avr16ea28, avr16ea32, avr16ea48, avr32ea28, avr32ea32, avr32ea48): Add.
4194 * doc/avr-mmcu.texi: Regenerate.
4196 2024-01-16 Feng Xue <fxue@os.amperecomputing.com>
4198 PR tree-optimization/113091
4199 * tree-vect-slp.cc (vect_slp_has_scalar_use): New function.
4200 (vect_bb_slp_mark_live_stmts): New parameter scalar_use_map, check
4201 scalar use with new function.
4202 (vect_bb_slp_mark_live_stmts): New function as entry to existing
4203 overriden functions with same name.
4204 (vect_slp_analyze_operations): Call new entry function to mark
4207 2024-01-16 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4210 * config/riscv/riscv.cc (riscv_override_options_internal): Report sorry
4211 for RVV in big-endian mode.
4213 2024-01-16 Yanzhang Wang <yanzhang.wang@intel.com>
4215 * config/riscv/riscv.cc (riscv_arg_has_vector): Delete.
4216 (riscv_pass_in_vector_p): Delete.
4217 (riscv_init_cumulative_args): Delete the checking.
4218 (riscv_get_arg_info): Delete the checking.
4219 (riscv_function_value): Delete the checking.
4220 * config/riscv/riscv.h: Delete the member for checking.
4222 2024-01-15 Georg-Johann Lay <avr@gjlay.de>
4224 * doc/invoke.texi (AVR Options) [-mskip-bug]: Add documentation.
4226 2024-01-15 Liao Shihua <shihua@iscas.ac.cn>
4228 * config.gcc: Include riscv_bitmanip.h.
4229 * config/riscv/bitmanip.md: Changed mode form X to GPR in orcb and clmul pattern.
4230 * config/riscv/crypto.md: Changed mode form X to GPR in brev8 pattern.
4231 * config/riscv/riscv-builtins.cc (AVAIL): Adding new bitmanip builtins.
4232 (RISCV_BUILTIN_NO_PREFIX): New helper macro.
4233 * config/riscv/riscv-cmo.def (RISCV_BUILTIN): Add '_32'/'_64' postfix to builtins.
4234 * config/riscv/riscv-ftypes.def (2): New ftypes.
4235 * config/riscv/riscv-scalar-crypto.def (RISCV_BUILTIN): New builtins.
4236 (RISCV_BUILTIN_NO_PREFIX): Likewise.
4237 * config/riscv/riscv_bitmanip.h: New file.
4239 2024-01-15 Liao Shihua <shihua@iscas.ac.cn>
4241 * config.gcc: Include riscv_crypto.h.
4242 * config/riscv/riscv_crypto.h: New file.
4244 2024-01-15 Vladimir N. Makarov <vmakarov@redhat.com>
4246 PR middle-end/113354
4247 * lra-constraints.cc (curr_insn_transform): Spill pseudo only used
4248 in the insn if the corresponding operand does not require hard
4251 2024-01-15 Georg-Johann Lay <avr@gjlay.de>
4254 * config/avr/avr.h (EXTRA_SPEC_FUNCTIONS): Add no-devlib, avr_no_devlib.
4255 * config/avr/driver-avr.cc (avr_no_devlib): New function.
4256 (avr_devicespecs_file): Use it to remove -nodevicelib from the
4257 options for cores only.
4258 * config/avr/avr-arch.h (avr_get_parch): New prototype.
4259 * config/avr/avr-devices.cc (avr_get_parch): New function.
4261 2024-01-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4264 * config/riscv/riscv-protos.h (struct regmove_vector_cost): Add vector to scalar regmove.
4265 * config/riscv/riscv-vector-costs.cc (adjust_stmt_cost): Ditto.
4266 * config/riscv/riscv.cc (riscv_builtin_vectorization_cost): Adjust vec_construct cost.
4268 2024-01-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4271 * config/riscv/riscv-vector-costs.cc (costs::adjust_vect_cost_per_loop): New function.
4272 (costs::finish_cost): Adjust cost for LOOP LEN with NITERS < VF.
4273 * config/riscv/riscv-vector-costs.h: New function.
4275 2024-01-15 Richard Biener <rguenther@suse.de>
4277 PR tree-optimization/113385
4278 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
4279 First redirect, then split the exit edge.
4281 2024-01-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4283 * config/riscv/riscv-vector-costs.cc (costs::analyze_loop_vinfo):
4284 Remove m_num_vector_iterations.
4285 * config/riscv/riscv-vector-costs.h: Ditto.
4287 2024-01-15 Andrew Pinski <quic_apinski@quicinc.com>
4290 * config/avr/avr.opt (-mdouble, -mlong-double): Add "Save" flag.
4291 (-mbranch-cost): Set "Optimization" flag.
4293 2024-01-15 Jakub Jelinek <jakub@redhat.com>
4295 PR tree-optimization/113370
4296 * gimple-lower-bitint.cc (bitint_large_huge::handle_operand): Only
4297 set rem to prec % (2 * limb_prec) if m_upwards_2limb, otherwise
4298 set it to just prec % limb_prec.
4300 2024-01-15 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4303 * config/riscv/vector.md: Fix ternary attributes.
4305 2024-01-14 Georg-Johann Lay <avr@gjlay.de>
4308 * configure.ac [target=avr]: Check availability of emulations
4309 avrxmega2_flmap and avrxmega4_flmap, resulting in new config vars
4310 HAVE_LD_AVR_AVRXMEGA2_FLMAP and HAVE_LD_AVR_AVRXMEGA4_FLMAP.
4311 * configure: Regenerate.
4312 * config.in: Regenerate.
4313 * doc/invoke.texi (AVR Options): Document -mflmap, -mrodata-in-ram,
4314 __AVR_HAVE_FLMAP__, __AVR_RODATA_IN_RAM__.
4315 * config/avr/avr.opt (-mflmap, -mrodata-in-ram): New options.
4316 * config/avr/avr-arch.h (enum avr_device_specific_features):
4318 * config/avr/avr-mcus.def (AVR_MCU) [avr64*, avr128*]: Set isa flag
4320 * config/avr/avr.cc (avr_arch_index, avr_has_rodata_p): New vars.
4321 (avr_set_core_architecture): Set avr_arch_index.
4322 (have_avrxmega2_flmap, have_avrxmega4_flmap)
4323 (have_avrxmega3_rodata_in_flash): Set new static const bool according
4324 to configure results.
4325 (avr_rodata_in_flash_p): New function using them.
4326 (avr_asm_init_sections): Let readonly_data_section->unnamed.callback
4327 track avr_need_copy_data_p only if not avr_rodata_in_flash_p().
4328 (avr_asm_named_section): Track avr_has_rodata_p.
4329 (avr_file_end): Emit __do_copy_data also when avr_has_rodata_p
4330 and not avr_rodata_in_flash_p ().
4331 * config/avr/specs.h (CC1_SPEC): Add %(cc1_rodata_in_ram).
4332 (LINK_SPEC): Add %(link_rodata_in_ram).
4333 (LINK_ARCH_SPEC): Remove.
4334 * config/avr/gen-avr-mmcu-specs.cc (have_avrxmega3_rodata_in_flash)
4335 (have_avrxmega2_flmap, have_avrxmega4_flmap): Set new static
4336 const bool according to configure results.
4337 (diagnose_mrodata_in_ram): New function.
4338 (print_mcu): Generate specs with the following changes:
4339 <*cc1_misc, *asm_misc, *link_misc>: New specs so that we don't
4340 need to extend avr/specs.h each time we add a new bell or whistle.
4341 <*cc1_rodata_in_ram, *link_rodata_in_ram>: New specs to diagnose
4342 -m[no-]rodata-in-ram.
4343 <*cpp_rodata_in_ram>: New. Does -D__AVR_RODATA_IN_RAM__=0/1.
4344 <*cpp_mcu>: Add -D__AVR_AVR_FLMAP__ if it applies.
4345 <*cpp>: Add %(cpp_rodata_in_ram).
4346 <*link_arch>: Use emulation avrxmega2_flmap, avrxmega4_flmap as
4348 <*self_spec>: Add -mflmap or %<mflmap as needed.
4350 2024-01-14 Jeff Law <jlaw@ventanamicro.com>
4352 * config/mips/mips.md (ior<mode>3_mips16_asmacro): Use SImode,
4353 not the GPR iterator. Adjust pattern name and mode attribute
4356 2024-01-13 Jakub Jelinek <jakub@redhat.com>
4358 PR tree-optimization/113361
4359 * gimple-lower-bitint.cc (bitint_large_huge::handle_operand_addr):
4360 Fix up determination of the type for > limb_prec constants.
4362 2024-01-12 Georg-Johann Lay <avr@gjlay.de>
4364 * doc/extend.texi (AVR Named Address Spaces, Limitations and Caveats):
4365 Add web-link to the avr-gcc wiki.
4367 2024-01-12 Georg-Johann Lay <avr@gjlay.de>
4369 * doc/extend.texi (AVR Variable Attributes) [address]: Remove
4370 documentation for a version without argument, which is not supported.
4372 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4374 * config/arm/arm_neon.h
4375 (vld1_u8_x4, vld1_u16_x4, vld1_u32_x4, vld1_u64_x4): New.
4376 (vld1_s8_x4, vld1_s16_x4, vld1_s32_x4, vld1_s64_x4): New.
4377 (vld1_f16_x4, vld1_f32_x4): New.
4378 (vld1_p8_x4, vld1_p16_x4, vld1_p64_x4): New.
4379 (vld1_bf16_x4): New.
4380 (vld1q_types_x4): Updated to use vld1q_x4
4381 from arm_neon_builtins.def
4382 * config/arm/arm_neon_builtins.def
4383 (vld1_x4): Updated entries.
4384 (vld1q_x4): New entries, but comes from the old vld1_x4
4385 * config/arm/neon.md
4386 (neon_vld1q_x4<mode>): Updated from neon_vld1_x4<mode>.
4388 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4390 * config/arm/arm_neon.h
4391 (vld1_u8_x3, vld1_u16_x3, vld1_u32_x3, vld1_u64_x3): New.
4392 (vld1_s8_x3, vld1_s16_x3, vld1_s32_x3, vld1_s64_x3): New.
4393 (vld1_f16_x3, vld1_f32_x3): New.
4394 (vld1_p8_x3, vld1_p16_x3, vld1_p64_x3): New.
4395 (vld1_bf16_x3): New.
4396 (vld1q_types_x3): Updated to use vld1q_x3 from
4397 arm_neon_builtins.def
4398 * config/arm/arm_neon_builtins.def
4399 (vld1_x3): Updated entries.
4400 (vld1q_x3): New entries, but comes from the old vld1_x2
4401 * config/arm/neon.md
4402 (neon_vld1q_x3<mode>): Updated from neon_vld1_x3<mode>.
4404 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4406 * config/arm/arm_neon.h
4407 (vld1_u8_x2, vld1_u16_x2, vld1_u32_x2, vld1_u64_x2): New.
4408 (vld1_s8_x2, vld1_s16_x2, vld1_s32_x2, vld1_s64_x2): New.
4409 (vld1_f16_x2, vld1_f32_x2): New.
4410 (vld1_p8_x2, vld1_p16_x2, vld1_p64_x2): New.
4411 (vld1_bf16_x2): New.
4412 (vld1q_types_x2): Updated to use vld1q_x2 from
4413 arm_neon_builtins.def
4414 * config/arm/arm_neon_builtins.def
4415 (vld1_x2): Updated entries.
4416 (vld1q_x2): New entries, but comes from the old vld1_x2
4417 * config/arm/neon.md
4418 (neon_vld1<VMEMX2_q>_x2<VDQX:mode>): Updated from
4421 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4423 * config/arm/arm_neon.h
4424 (vst1q_u8_x4, vst1q_u16_x4, vst1q_u32_x4, vst1q_u64_x4): New.
4425 (vst1q_s8_x4, vst1q_s16_x4, vst1q_s32_x4, vst1q_s64_x4): New.
4426 (vst1q_f16_x4, vst1q_f32_x4): New.
4427 (vst1q_p8_x4, vst1q_p16_x4, vst1q_p64_x4): New.
4428 (vst1q_bf16_x4): New.
4429 * config/arm/arm_neon_builtins.def (vst1q_x4): New entries.
4430 * config/arm/neon.md
4431 (neon_vst1q_x4<mode>): New.
4432 (neon_vst1x4qa<mode>, neon_vst1x4qb<mode>): New.
4433 * config/arm/unspecs.md
4434 (UNSPEC_VST1X4A, UNSPEC_VST1X4B): New.
4436 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4438 * config/arm/arm_neon.h
4439 (vst1q_u8_x3, vst1q_u16_x3, vst1q_u32_x3, vst1q_u64_x3): New.
4440 (vst1q_s8_x3, vst1q_s16_x3, vst1q_s32_x3, vst1q_s64_x3): New.
4441 (vst1q_f16_x3, vst1q_f32_x3): New.
4442 (vst1q_p8_x3, vst1q_p16_x3, vst1q_p64_x3): New.
4443 (vst1q_bf16_x3): New.
4444 * config/arm/arm_neon_builtins.def (vst1q_x3): New entries.
4445 * config/arm/neon.md
4446 (neon_vst1q_x3<mode>): New.
4447 (neon_vld1x3qa<mode>, neon_vst1x3qb<mode>): New.
4448 * config/arm/unspecs.md
4449 (UNSPEC_VST1X3A, UNSPEC_VST1X3B): New.
4451 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4453 * config/arm/arm_neon.h
4454 (vst1q_u8_x2, vst1q_u16_x2, vst1q_u32_x2, vst1q_u64_x2): New.
4455 (vst1q_s8_x2, vst1q_s16_x2, vst1q_s32_x2, vst1q_s64_x2): New.
4456 (vst1q_f16_x2, vst1q_f32_x2): New.
4457 (vst1q_p8_x2, vst1q_p16_x2, vst1q_p64_x2): New.
4458 (vst1q_bf16_x2): New.
4459 * config/arm/arm_neon_builtins.def (vst1<_x2): New entries.
4460 * config/arm/neon.md
4461 (neon_vst1<VMEMX2_q>_x2<VDQX:mode>): Updated from
4463 * config/arm/iterators.md
4464 (VMEMX2): New mode iterator.
4465 (VMEMX2_q): New mode attribute.
4467 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4469 * config/arm/arm_neon.h
4470 (vst1_u8_x4, vst1_u16_x4, vst1_u32_x4, vst1_u64_x4): New.
4471 (vst1_s8_x4, vst1_s16_x4, vst1_s32_x4, vst1_s64_x4): New.
4472 (vst1_f16_x4, vst1_f32_x4): New.
4473 (vst1_p8_x4, vst1_p16_x4, vst1_p64_x4): New.
4474 (vst1_bf16_x4): New.
4475 * config/arm/arm_neon_builtins.def (vst1_x4): New entries.
4476 * config/arm/neon.md (vst1_x4<mode>): New.
4478 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4480 * config/arm/arm_neon.h
4481 (vst1_u8_x3, vst1_u16_x3, vst1_u32_x3, vst1_u64_x3): New.
4482 (vst1_s8_x3, vst1_s16_x3, vst1_s32_x3, vst1_s64_x3): New.
4483 (vst1_f16_x3, vst1_f32_x3): New.
4484 (vst1_p8_x3, vst1_p16_x3, vst1_p64_x3): New.
4485 (vst1_bf16_x3): New.
4486 * config/arm/arm_neon_builtins.def (vst1_x3): New entries.
4487 * config/arm/neon.md (vst1_x3<mode>): New.
4489 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4491 * config/arm/arm_neon.h
4492 (vst1_u8_x2, vst1_u16_x2, vst1_u32_x2, vst1_u64_x2): New.
4493 (vst1_s8_x2, vst1_s16_x2, vst1_s32_x2, vst1_s64_x2): New.
4494 (vst1_f16_x2, vst1_f32_x2): New.
4495 (vst1_p8_x2, vst1_p16_x2, vst1_p64_x2): New.
4496 (vst1_bf16_x2): New.
4497 * config/arm/arm_neon_builtins.def (vst1_x2): New entries.
4498 * config/arm/neon.md (vst1_x2<mode>): New.
4500 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4502 * config/arm/arm_neon.h
4503 (vld1q_u8_x4, vld1q_u16_x4, vld1q_u32_x4, vld1q_u64_x4): New.
4504 (vld1q_s8_x4, vld1q_s16_x4, vld1q_s32_x4, vld1q_s64_x4): New.
4505 (vld1q_f16_x4, vld1q_f32_x4): New.
4506 (vld1q_p8_x4, vld1q_p16_x4, vld1q_p64_x4): New.
4507 (vld1q_bf16_x4): New.
4508 * config/arm/arm_neon_builtins.def (vld1_x4): New entries.
4509 * config/arm/neon.md
4510 (neon_vld1_x4<mode>): New.
4511 (neon_vld1x4qa<mode>, neon_vld1x4qb<mode>): New
4512 * config/arm/unspecs.md
4513 (UNSPEC_VLD1X4A, UNSPEC_VLD1X4B): New.
4515 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4517 * config/arm/arm_neon.h
4518 (vld1q_u8_x3, vld1q_u16_x3, vld1q_u32_x3, vld1q_u64_x3): New.
4519 (vld1q_s8_x3, vld1q_s16_x3, vld1q_s32_x3, vld1q_s64_x3): New.
4520 (vld1q_f16_x3, vld1q_f32_x3): New.
4521 (vld1q_p8_x3, vld1q_p16_x3, vld1q_p64_x3): New.
4522 (vld1q_bf16_x3): New.
4523 * config/arm/arm_neon_builtins.def (vld1_x3): New entries.
4524 * config/arm/neon.md
4525 (neon_vld1_x3<mode>): New.
4526 (neon_vld1x3qa<mode>, neon_vld1x3qb<mode>): New.
4527 * config/arm/unspecs.md
4528 (UNSPEC_VLD1X3A, UNSPEC_VLD1X3B): New.
4530 2024-01-12 Ezra Sitorus <ezra.sitorus@arm.com>
4532 * config/arm/arm_neon.h
4533 (vld1q_u8_x2, vld1q_u16_x2, vld1q_u32_x2, vld1q_u64_x2): New.
4534 (vld1q_s8_x2, vld1q_s16_x2, vld1q_s32_x2, vld1q_s64_x2): New.
4535 (vld1q_f16_x2, vld1q_f32_x2): New.
4536 (vld1q_p8_x2, vld1q_p16_x2, vld1q_p64_x2): New.
4537 (vld1q_bf16_x2): New.
4538 * config/arm/arm_neon_builtins.def (vld1_x2): New entries.
4539 * config/arm/neon.md (vld1_x2<mode>): New.
4541 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4543 PR tree-optimization/113287
4544 * doc/sourcebuild.texi (check_effective_target_bitint65535): New.
4546 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4548 * tree-vect-loop-manip.cc (vect_loop_versioning): Replace single_exit.
4549 * tree-vect-loop.cc (vect_transform_loop): Likewise.
4551 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4553 PR tree-optimization/113178
4554 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Fill in all
4557 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4559 PR tree-optimization/113237
4560 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg): Use
4561 existing LCSSA variable for exit when all exits are early break.
4563 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4565 PR tree-optimization/113137
4566 PR tree-optimization/113136
4567 PR tree-optimization/113172
4568 PR tree-optimization/113178
4569 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
4570 Maintain PHIs on inverted loops.
4571 (vect_do_peeling): Maintain virtual PHIs on inverted loops.
4572 * tree-vect-loop.cc (vec_init_loop_exit_info): Pick exit closes to
4574 (vect_create_loop_vinfo): Record all conds instead of only alt ones.
4576 2024-01-12 Tamar Christina <tamar.christina@arm.com>
4578 PR tree-optimization/113135
4579 * tree-vect-data-refs.cc (vect_analyze_early_break_dependences): Rework
4580 dependency analysis.
4582 2024-01-12 Iain Sandoe <iain@sandoe.co.uk>
4584 * config/rs6000/host-darwin.cc (segv_handler): Use the revised
4585 diagnostics class member name for abort of error.
4587 2024-01-12 Georg-Johann Lay <avr@gjlay.de>
4589 * config/avr/avr.cc (avr_handle_addr_attribute): Move "..." from
4590 format string to %s argument.
4592 2024-01-12 John David Anglin <danglin@gcc.gnu.org>
4593 Jakub Jelinek <jakub@redhat.com>
4595 PR middle-end/113182
4596 * varasm.cc (process_pending_assemble_externals,
4597 assemble_external_libcall): Use targetm.strip_name_encoding
4598 before calling get_identifier.
4600 2024-01-12 Richard Sandiford <richard.sandiford@arm.com>
4603 * config/aarch64/aarch64.h (machine_function::advsimd_zero_insn):
4604 New member variable.
4605 * config/aarch64/aarch64-protos.h (aarch64_split_simd_shift_p):
4607 * config/aarch64/iterators.md (Vnarrowq2): New mode attribute.
4608 * config/aarch64/aarch64-simd.md
4609 (vec_unpacku_hi_<mode>, vec_unpacks_hi_<mode>): Recombine into...
4610 (vec_unpack<su>_hi_<mode>): ...this. Move the generation of
4611 zip2 for zero-extends to...
4612 (aarch64_simd_vec_unpack<su>_hi_<mode>): ...a split of this
4613 instruction. Fix big-endian handling.
4614 (vec_unpacku_lo_<mode>, vec_unpacks_lo_<mode>): Recombine into...
4615 (vec_unpack<su>_lo_<mode>): ...this. Move the generation of
4616 zip1 for zero-extends to...
4617 (<optab><Vnarrowq><mode>2): ...a split of this instruction.
4618 Fix big-endian handling.
4619 (*aarch64_zip1_uxtl): New pattern.
4620 (aarch64_usubw<mode>_lo_zip, aarch64_uaddw<mode>_lo_zip): Delete
4621 (aarch64_usubw<mode>_hi_zip, aarch64_uaddw<mode>_hi_zip): Likewise.
4622 * config/aarch64/aarch64.cc (aarch64_get_shareable_reg): New function.
4623 (aarch64_gen_shareable_zero): Use it.
4624 (aarch64_split_simd_shift_p): New function.
4626 2024-01-12 Richard Sandiford <richard.sandiford@arm.com>
4628 * emit-rtl.h (rtl_data::x_function_beg_note): New member variable.
4629 (function_beg_insn): New macro.
4630 * function.cc (expand_function_start): Initialize function_beg_insn.
4632 2024-01-12 Richard Sandiford <richard.sandiford@arm.com>
4635 * config/aarch64/aarch64-sve-builtins.h
4636 (function_builder::m_overload_names): Replace with...
4637 * config/aarch64/aarch64-sve-builtins.cc (overload_names): ...this
4639 (add_overloaded_function): Update accordingly, using get_identifier
4640 to get a GGC-friendly record of the name.
4642 2024-01-12 Richard Sandiford <richard.sandiford@arm.com>
4645 * config/aarch64/aarch64-sve-builtins.def: Don't include
4646 aarch64-sve-builtins-sme.def.
4647 (DEF_SME_ZA_FUNCTION_GS, DEF_SME_ZA_FUNCTION): Move to...
4648 * config/aarch64/aarch64-sve-builtins-sme.def: ...here.
4649 (DEF_SME_FUNCTION): New macro. Use it and DEF_SME_FUNCTION_GS
4650 instead of DEF_SVE_*. Add AARCH64_FL_SME to anything that
4651 requires AARCH64_FL_SME2.
4652 * config/aarch64/aarch64-sve-builtins-sve2.def: Make same
4653 AARCH64_FL_SME adjustment here.
4654 * config/aarch64/aarch64-sve-builtins.cc (function_groups): Don't
4655 include SME intrinsics.
4656 (sme_function_groups): New array.
4657 (handle_arm_sve_h): Remove check for AARCH64_FL_SME.
4658 (handle_arm_sme_h): Use sme_function_groups instead of function_groups.
4660 2024-01-12 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4663 * config/riscv/riscv-protos.h (struct regmove_vector_cost): New struct.
4664 (struct cpu_vector_cost): Add regmove struct.
4665 (get_vector_costs): Export as global.
4666 * config/riscv/riscv-vector-costs.cc (adjust_stmt_cost): Adjust scalar_to_vec cost.
4667 (costs::add_stmt_cost): Ditto.
4668 * config/riscv/riscv.cc (get_common_costs): Export global function.
4670 2024-01-12 Jakub Jelinek <jakub@redhat.com>
4672 PR tree-optimization/113334
4673 * gimple-lower-bitint.cc (bitint_large_huge::handle_operand): Use
4674 wi::neg_p (wi::to_wide (op)) instead of tree_int_cst_sgn (op) < 0
4675 to determine if number should be extended by all ones rather than zero
4678 2024-01-12 Jakub Jelinek <jakub@redhat.com>
4680 PR tree-optimization/113330
4681 * tree-sra.cc (create_access): Punt for BITINT_TYPE accesses with
4684 2024-01-12 Jakub Jelinek <jakub@redhat.com>
4686 PR tree-optimization/113323
4687 * gimple-lower-bitint.cc (bitint_dom_walker::before_dom_children): Fix
4688 check for lhs being large/huge _BitInt not in m_names.
4690 2024-01-12 Jakub Jelinek <jakub@redhat.com>
4692 PR tree-optimization/113316
4693 * gimple-lower-bitint.cc (bitint_large_huge::lower_call): Handle
4694 uninitialized large/huge _BitInt arguments to calls.
4696 2024-01-12 Jakub Jelinek <jakub@redhat.com>
4698 * gimple-lower-bitint.cc (mergeable_op): Instead of comparing
4699 TYPE_SIZE (t) of large/huge BITINT_TYPEs, compare
4700 CEIL (TYPE_PRECISION (t), limb_prec).
4701 (bitint_large_huge::handle_cast): Likewise.
4703 2024-01-12 Ilya Leoshkevich <iii@linux.ibm.com>
4706 * config/rs6000/rs6000.cc (rs6000_elf_declare_function_name):
4707 Use assemble_function_label_final () for Power ELF V1 ABI.
4708 * output.h (assemble_function_label_final): New function.
4709 * varasm.cc (assemble_function_label_raw): Use
4710 assemble_function_label_final ().
4711 (assemble_function_label_final): New function.
4713 2024-01-12 Richard Biener <rguenther@suse.de>
4715 PR middle-end/113344
4716 * match.pd ((double)float CMP (double)float -> float CMP float):
4717 Perform result type check only for vectors.
4718 * fold-const.cc (fold_binary_loc): Likewise.
4720 2024-01-12 Haochen Jiang <haochen.jiang@intel.com>
4722 * config/i386/sse.md (sdot_prod<mode>): Remove redundant SET.
4723 (usdot_prod<mode>): Ditto.
4724 (sdot_prod<mode>): Ditto.
4725 (udot_prod<mode>): Ditto.
4727 2024-01-12 Haochen Jiang <haochen.jiang@intel.com>
4730 * config/i386/i386-c.cc (ix86_target_macros_internal):
4731 Add __AVX10_1__, __AVX10_1_256__ and __AVX10_1_512__.
4733 2024-01-12 Richard Biener <rguenther@suse.de>
4736 * config/s390/s390.cc (expand_perm_as_a_vlbr_vstbr_candidate):
4737 Do not generate code when d.testing_p.
4739 2024-01-12 liuhongt <hongtao.liu@intel.com>
4742 * doc/invoke.texi (fcf-protection=): Update documents.
4744 2024-01-12 Pan Li <pan2.li@intel.com>
4746 * config/riscv/riscv.cc (riscv_v_ext_mode_p): Update the
4747 comments of predicate func riscv_v_ext_mode_p.
4749 2024-01-12 Feng Wang <wangfeng@eswincomputing.com>
4751 * config/riscv/riscv-vector-builtins.def (vfloat16m8_t):
4752 Modify ABI-name length of vfloat16m8_t
4754 2024-01-12 Li Wei <liwei@loongson.cn>
4756 * config/loongarch/loongarch.cc (loongarch_expand_conditional_move):
4759 2024-01-12 Li Wei <liwei@loongson.cn>
4761 * config/loongarch/loongarch.md (add<mode>3): Removed.
4765 (*addsi3_extended): Removed.
4766 (addsi3_extended): New.
4768 2024-01-11 Jin Ma <jinma@linux.alibaba.com>
4770 * config/riscv/thead.md: Add limits for splits.
4772 2024-01-11 Andrew Pinski <quic_apinski@quicinc.com>
4774 PR middle-end/113322
4775 * expr.cc (do_store_flag): Don't try single bit tests with
4776 comparison on vector types.
4778 2024-01-11 Andrew Pinski <quic_apinski@quicinc.com>
4780 PR tree-optimization/113301
4781 * match.pd (`1/x`): Delay signed case until late.
4783 2024-01-11 Georg-Johann Lay <avr@gjlay.de>
4785 * doc/invoke.texi (AVR Options): Move -mrmw, -mn-flash, -mshort-calls
4787 (AVR Internal Options): ...this new @subsubsection.
4789 2024-01-11 Vladimir N. Makarov <vmakarov@redhat.com>
4791 PR rtl-optimization/112918
4792 * lra-constraints.cc (SMALL_REGISTER_CLASS_P): Move before in_class_p.
4793 (in_class_p): Restrict condition for narrowing class in case of
4794 allow_all_reload_class_changes_p.
4795 (process_alt_operands): Try to match operand without and with
4796 narrowing reg class. Discourage narrowing the class. Finish insn
4797 matching only if there is no class narrowing.
4798 (curr_insn_transform): Pass true to in_class_p for reg operand win.
4800 2024-01-11 Richard Biener <rguenther@suse.de>
4802 PR tree-optimization/112505
4803 * tree-vect-loop.cc (vectorizable_induction): Reject
4804 bit-precision induction.
4806 2024-01-11 Richard Biener <rguenther@suse.de>
4808 PR tree-optimization/113126
4809 * match.pd ((double)float CMP (double)float -> float CMP float):
4810 Make sure the boolean type is the same.
4811 * fold-const.cc (fold_binary_loc): Likewise.
4813 2024-01-11 Richard Biener <rguenther@suse.de>
4815 PR tree-optimization/112636
4816 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Call
4817 estimate_numbers_of_iterations before querying
4818 get_max_loop_iterations_int.
4819 (pass_ch::execute): Initialize SCEV and loops appropriately.
4821 2024-01-11 Georg-Johann Lay <avr@gjlay.de>
4823 * config/avr/avr-devices.cc (avr_texinfo): Adjust documentation for
4825 * config/avr/gen-avr-mmcu-texi.cc (main): Add @anchor for each core.
4826 * doc/extend.texi (AVR Variable Attributes): Improve documentation
4827 of io, io_low and address attributes.
4828 * doc/invoke.texi (AVR Options): Add some anchors for external refs.
4829 * doc/avr-mmcu.texi: Rebuild.
4831 2024-01-11 Yang Yujie <yangyujie@loongson.cn>
4834 * config/loongarch/genopts/loongarch.opt.in: Mark options with
4835 the "Save" property.
4836 * config/loongarch/loongarch.opt: Same.
4837 * config/loongarch/loongarch-opts.cc: Refresh -mcmodel= state
4838 according to la_target.
4839 * config/loongarch/loongarch.cc: Implement TARGET_OPTION_{SAVE,
4840 RESTORE} for the la_target structure; Rename option conditions
4841 to have the same "la_" prefix.
4842 * config/loongarch/loongarch.h: Same.
4844 2024-01-11 Pan Li <pan2.li@intel.com>
4846 * loop-unroll.cc (insert_var_expansion_initialization): Leverage
4847 MODE_HAS_SIGNED_ZEROS for expansion variable initialization.
4849 2024-01-11 Alex Coplan <alex.coplan@arm.com>
4852 * config/aarch64/aarch64-ldp-fusion.cc (filter_notes): Add
4853 fr_expr param to extract REG_FRAME_RELATED_EXPR notes.
4854 (combine_reg_notes): Handle REG_FRAME_RELATED_EXPR notes, and
4855 synthesize these if needed. Update caller ...
4856 (ldp_bb_info::fuse_pair): ... here.
4857 (ldp_bb_info::try_fuse_pair): Punt if either insn has writeback
4858 and either insn is frame-related.
4859 (find_trailing_add): Punt on frame-related insns.
4860 * config/aarch64/aarch64.cc (aarch64_save_callee_saves): Use
4861 REG_FRAME_RELATED_EXPR instead of REG_CFA_OFFSET.
4863 2024-01-11 YunQiang Su <syq@gcc.gnu.org>
4865 * config/mips/mips.cc (mips_start_function_definition):
4866 Add ATTRIBUTE_UNUSED.
4868 2024-01-11 Richard Biener <rguenther@suse.de>
4870 PR middle-end/112740
4871 * expr.cc (store_constructor): Check the integer vector
4872 mask has a single bit per element before using sign-extension
4873 to expand an uniform vector.
4875 2024-01-11 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4877 * config/riscv/riscv-vector-costs.cc (costs::better_main_loop_than_p): VLA
4878 preempt VLS on unknown NITERS loop.
4880 2024-01-11 Haochen Jiang <haochen.jiang@intel.com>
4882 * doc/invoke.texi: Add -mevex512.
4884 2024-01-11 Lulu Cheng <chenglulu@loongson.cn>
4886 * config/loongarch/loongarch.md (one_cmpl<mode>2): Replace GPR with X.
4887 (*nor<mode>3): Likewise.
4888 (nor<mode>3): Likewise.
4889 (*negsi2_extended): New template.
4890 (*<optab>si3_internal): Likewise.
4891 (*one_cmplsi2_internal): Likewise.
4892 (*norsi3_internal): Likewise.
4893 (*<optab>nsi_internal): Likewise.
4894 (bytepick_w_<bytepick_imm>_extend): Modify this template according to the
4895 modified bit operation to make the optimization work.
4897 2024-01-11 liuhongt <hongtao.liu@intel.com>
4900 * match.pd (VEC_COND_EXPR: A < B ? A : B -> MIN_EXPR): New patten match.
4902 2024-01-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4904 * config/riscv/riscv.cc (get_common_costs): Switch RVV cost model.
4905 (get_vector_costs): Ditto.
4906 (riscv_builtin_vectorization_cost): Ditto.
4908 2024-01-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
4910 * config/riscv/riscv-vector-costs.cc (costs::better_main_loop_than_p): Minior tweak.
4912 2024-01-10 Antoni Boucher <bouanto@zoho.com>
4915 * ipa-fnsummary.cc (ipa_fnsummary_cc_finalize): Call
4916 ipa_free_size_summary.
4917 * ipa-icf.cc (ipa_icf_cc_finalize): New function.
4918 * ipa-profile.cc (ipa_profile_cc_finalize): New function.
4919 * ipa-prop.cc (ipa_prop_cc_finalize): New function.
4920 * ipa-prop.h (ipa_prop_cc_finalize): New function.
4921 * ipa-sra.cc (ipa_sra_cc_finalize): New function.
4922 * ipa-utils.h (ipa_profile_cc_finalize, ipa_icf_cc_finalize,
4923 ipa_sra_cc_finalize): New functions.
4924 * toplev.cc (toplev::finalize): Call ipa_icf_cc_finalize,
4925 ipa_prop_cc_finalize, ipa_profile_cc_finalize and
4927 Include ipa-utils.h.
4929 2024-01-10 Jin Ma <jinma@linux.alibaba.com>
4931 * config/riscv/riscv-protos.h (th_int_get_mask): New prototype.
4932 (th_int_get_save_adjustment): Likewise.
4933 (th_int_adjust_cfi_prologue): Likewise.
4934 * config/riscv/riscv.cc (BITSET_P): Moved away from here.
4935 (TH_INT_INTERRUPT): New macro.
4936 (riscv_expand_prologue): Add the processing of XTheadInt.
4937 (riscv_expand_epilogue): Likewise.
4938 * config/riscv/riscv.h (BITSET_P): Moved to here.
4939 * config/riscv/riscv.md: New unspec.
4940 * config/riscv/thead.cc (th_int_get_mask): New function.
4941 (th_int_get_save_adjustment): Likewise.
4942 (th_int_adjust_cfi_prologue): Likewise.
4943 * config/riscv/thead.md (th_int_push): New pattern.
4944 (th_int_pop): new pattern.
4946 2024-01-10 Tamar Christina <tamar.christina@arm.com>
4948 PR tree-optimization/112468
4949 * doc/sourcebuild.texi: Document ifn_copysign.
4950 * match.pd: Only apply transformation if target supports the IFN.
4952 2024-01-10 Andrew Pinski <quic_apinski@quicinc.com>
4954 PR tree-optimization/112581
4955 * gimple-if-to-switch.cc (pass_if_to_switch::execute): Call
4956 mark_ssa_maybe_undefs.
4957 * tree-ssa-reassoc.cc (can_reassociate_op_p): Uninitialized
4958 variables can not be reassociated.
4959 (init_range_entry): Check for uninitialized variables too.
4960 (init_reassoc): Call mark_ssa_maybe_undefs.
4962 2024-01-10 Maciej W. Rozycki <macro@embecosm.com>
4964 * config/riscv/riscv.cc (riscv_noce_conversion_profitable_p):
4965 Also handle sign extension.
4967 2024-01-10 Alex Coplan <alex.coplan@arm.com>
4969 * config/aarch64/aarch64.opt (-mearly-ldp-fusion): Set default
4971 (-mlate-ldp-fusion): Likewise.
4973 2024-01-10 Tamar Christina <tamar.christina@arm.com>
4975 PR tree-optimization/113287
4976 * tree-vect-stmts.cc (vectorizable_early_exit): Check the flags on edge
4977 instead of using BRANCH_EDGE to determine true edge.
4979 2024-01-10 Richard Biener <rguenther@suse.de>
4981 PR tree-optimization/113078
4982 * tree-vect-loop.cc (check_reduction_path): Canonicalize
4983 .COND_SUB to .COND_ADD.
4985 2024-01-10 David Malcolm <dmalcolm@redhat.com>
4987 * gcc-urlifier.cc (gcc_urlifier::get_url_suffix_for_option):
4988 Handle prefix mappings before calling find_opt.
4989 (selftest::gcc_urlifier_cc_tests): Add example of urlifying a
4990 "-fno-"-prefixed command-line option.
4991 * opts-common.cc (get_option_prefix_remapping): New.
4992 * opts.h (get_option_prefix_remapping): New decl.
4994 2024-01-10 David Malcolm <dmalcolm@redhat.com>
4996 * diagnostic.cc (diagnostic_context::report_diagnostic): Pass
4997 m_urlifier to pp_output_formatted_text.
4998 * pretty-print.cc: Add #define of INCLUDE_VECTOR.
4999 (obstack_append_string): New overload, taking a length.
5000 (urlify_quoted_string): Pass in an obstack ptr, rather than using
5001 that of the pp's buffer. Generalize to handle trailing text in
5002 the buffer beyond the run of quoted text.
5003 (class quoting_info): New.
5004 (on_begin_quote): New.
5005 (on_end_quote): New.
5006 (pp_format): Refactor phase 1 and phase 2 quoting support, moving
5007 it to calls to on_begin_quote and on_end_quote.
5008 (struct auto_obstack): New.
5009 (quoting_info::handle_phase_3): New.
5010 (pp_output_formatted_text): Add urlifier param. Use it if there
5011 is deferred urlification. Delete m_quotes.
5012 (selftest::pp_printf_with_urlifier): Pass urlifier to
5013 pp_output_formatted_text.
5014 (selftest::test_urlification): Update results for the existing
5015 case of quoted text stradding chunks; add more such test cases.
5016 * pretty-print.h (class quoting_info): New forward decl.
5017 (chunk_info::m_quotes): New field.
5018 (pp_output_formatted_text): Add optional urlifier param.
5020 2024-01-10 David Malcolm <dmalcolm@redhat.com>
5022 * pretty-print.cc (selftest::test_pp_format): Add selftest
5023 coverage for numbered args.
5025 2024-01-10 Tamar Christina <tamar.christina@arm.com>
5027 PR tree-optimization/113144
5028 PR tree-optimization/113145
5029 * tree-vect-loop-manip.cc (slpeel_tree_duplicate_loop_to_edge_cfg):
5030 Update all BB that the original exits dominated.
5032 2024-01-10 Eric Botcazou <ebotcazou@adacore.com>
5034 * dwarf2out.cc (modified_type_die): Extend the support of reverse
5035 storage order to enumeration types if -gstrict-dwarf is not passed.
5036 (gen_enumeration_type_die): Add REVERSE parameter and generate the
5037 DIE immediately after the existing one if it is true.
5038 (gen_tagged_type_die): Add REVERSE parameter and pass it in the
5039 call to gen_enumeration_type_die.
5040 (gen_type_die_with_usage): Add REVERSE parameter and pass it in the
5041 first recursive call as well as the call to gen_tagged_type_die.
5042 (gen_type_die): Add REVERSE parameter and pass it in the call to
5043 gen_type_die_with_usage.
5045 2024-01-10 Jakub Jelinek <jakub@redhat.com>
5047 PR tree-optimization/113120
5048 * tree-sra.cc (analyze_access_subtree): For BITINT_TYPE
5049 with root->size TYPE_PRECISION don't build anything new.
5050 Otherwise, if root->type is a BITINT_TYPE, use build_bitint_type
5051 rather than build_nonstandard_integer_type.
5053 2024-01-10 Hongyu Wang <hongyu.wang@intel.com>
5055 * config/i386/i386.opt: Adjust document.
5056 * doc/invoke.texi: Add description for
5057 -mapx-inline-asm-use-gpr32.
5059 2024-01-10 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5061 * config/riscv/autovec.md (<u>avg<v_double_trunc>3_floor): Remove.
5062 (avg<v_double_trunc>3_floor): New pattern.
5063 (<u>avg<v_double_trunc>3_ceil): Remove.
5064 (avg<v_double_trunc>3_ceil): New pattern.
5065 (uavg<mode>3_floor): Ditto.
5066 (uavg<mode>3_ceil): Ditto.
5067 * config/riscv/riscv-protos.h (enum insn_flags): Add for average addition.
5068 (enum insn_type): Ditto.
5069 * config/riscv/riscv-v.cc: Ditto.
5070 * config/riscv/vector-iterators.md (ashiftrt): Remove.
5072 * config/riscv/vector.md: Add VLS modes.
5074 2024-01-10 Kewen Lin <linkw@linux.ibm.com>
5077 * config/rs6000/vsx.md (VCZLSBB): New int iterator.
5078 (vczlsbb_char): New int attribute.
5079 (vclzlsbb_<mode>, vctzlsbb_<mode>): Merge to ...
5080 (vc<vczlsbb_char>zlsbb_<mode>): ... this.
5081 (*vctzlsbb_zext_<mode>): Rename to ...
5082 (*vc<vczlsbb_char>zlsbb_zext_<mode>): ... this, and extend it to
5085 2024-01-10 Kewen Lin <linkw@linux.ibm.com>
5088 * config/rs6000/rs6000.md (copysign<mode>3 IEEE128): Change predicate
5089 of the last argument from altivec_register_operand to any_operand. If
5090 operands[2] is CONST_DOUBLE, emit abs or neg abs depending on its sign
5091 otherwise if it doesn't satisfy altivec_register_operand, force it to
5092 REG using copy_to_mode_reg.
5094 2024-01-10 Kewen Lin <linkw@linux.ibm.com>
5096 PR middle-end/113100
5097 * builtins.cc (expand_builtin_stack_address): Guard stack point
5098 adjustment with SPARC_STACK_BOUNDARY_HACK.
5100 2024-01-10 Yang Yujie <yangyujie@loongson.cn>
5102 * config/loongarch/genopts/loongarch-strings: Remove explicit-reloc
5103 argument string definitions.
5104 * config/loongarch/loongarch-str.h: Same.
5105 * config/loongarch/genopts/loongarch.opt.in: Mark -m[no-]explicit-relocs
5106 as aliases to -mexplicit-relocs={always,none}
5107 * config/loongarch/loongarch.opt: Regenerate.
5108 * config/loongarch/loongarch.cc: Same.
5110 2024-01-10 Yang Yujie <yangyujie@loongson.cn>
5112 * config/loongarch/loongarch-def.h: Define constants with
5113 enums instead of Macros.
5115 2024-01-10 Yang Yujie <yangyujie@loongson.cn>
5117 * config/loongarch/genopts/loongarch-strings: Rename.
5118 * config/loongarch/genopts/loongarch.opt.in: Same.
5119 * config/loongarch/loongarch-cpu.cc: Same.
5120 * config/loongarch/loongarch-def.cc: Same.
5121 * config/loongarch/loongarch-def.h: Same.
5122 * config/loongarch/loongarch-opts.cc: Same.
5123 * config/loongarch/loongarch-opts.h: Same.
5124 * config/loongarch/loongarch-str.h: Same.
5125 * config/loongarch/loongarch.opt: Same.
5127 2024-01-10 Yang Yujie <yangyujie@loongson.cn>
5129 * config/loongarch/genopts/genstr.sh: Prepend the isa_evolution
5130 variable with the common la_ prefix.
5131 * config/loongarch/genopts/loongarch.opt.in: Mark ISA evolution
5132 flags as saved using TargetVariable.
5133 * config/loongarch/loongarch.opt: Same.
5134 * config/loongarch/loongarch-def.h: Define evolution_set to
5135 mark changes to the -march default.
5136 * config/loongarch/loongarch-driver.cc: Same.
5137 * config/loongarch/loongarch-opts.cc: Same.
5138 * config/loongarch/loongarch-opts.h: Define and use ISA evolution
5139 conditions around the la_target structure.
5140 * config/loongarch/loongarch.cc: Same.
5141 * config/loongarch/loongarch.md: Same.
5142 * config/loongarch/loongarch-builtins.cc: Same.
5143 * config/loongarch/loongarch-c.cc: Same.
5144 * config/loongarch/lasx.md: Same.
5145 * config/loongarch/lsx.md: Same.
5146 * config/loongarch/sync.md: Same.
5148 2024-01-09 Jeff Law <jlaw@ventanamicro.com>
5150 * config/epiphany/constraints.md (Car): Allow -1024..1023, no more,
5153 2024-01-09 Richard Sandiford <richard.sandiford@arm.com>
5155 * config/mn10300/mn10300.md (subdi3_degenerate): Add isa attribute.
5157 2024-01-09 Tamar Christina <tamar.christina@arm.com>
5159 * tree-vect-loop.cc (vectorizable_live_operation_1): Drop unused
5161 (vectorizable_live_operation): Likewise.
5163 2024-01-09 Tamar Christina <tamar.christina@arm.com>
5165 PR tree-optimization/113199
5166 * tree-vect-loop.cc (vectorizable_live_operation_1): Use
5169 2024-01-09 Jakub Jelinek <jakub@redhat.com>
5172 * config.gcc (aarch64*-*-*): Add aarch64-builtins.h to target_gtfiles.
5173 * config/aarch64/aarch64-builtins.cc (aarch64_simd_types): Add extern
5174 GTY(()) declaration before the definition, drop GTY(()) drom the
5177 2024-01-09 Richard Biener <rguenther@suse.de>
5179 PR tree-optimization/113026
5180 * tree-vect-loop-manip.cc (vect_do_peeling): Remove
5181 redundant and wrong niter bound setting. Move niter
5182 bound adjustment down.
5184 2024-01-09 Tamar Christina <tamar.christina@arm.com>
5186 PR middle-end/113163
5187 * tree-vect-loop-manip.cc (vect_can_peel_nonlinear_iv_p):
5188 Reject non-linear inductions that aren't supported.
5190 2024-01-09 Roger Sayle <roger@nextmovesoftware.com>
5192 * config/arc/arc.cc (arc_shift_alg): New enumerated type for
5193 left shift implementation strategies.
5194 (arc_shift_info): Type for each entry of the shift strategy table.
5195 (arc_shift_context_idx): Return a integer value for each code
5196 generation context, used as an index
5197 (arc_ashl_alg): Table indexed by context and shifted bit count.
5198 (arc_split_ashl): Use the arc_ashl_alg table to select SImode
5199 left shift implementation.
5200 (arc_rtx_costs) <case ASHIFT>: Use the arc_ashl_alg table to
5201 provide accurate costs, when optimizing for speed or size.
5203 2024-01-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5205 * config/riscv/riscv-vector-costs.cc (loop_invariant_op_p): Fix loop invariant check.
5207 2024-01-09 Julian Brown <julian@codesourcery.com>
5209 * gimplify.cc (gimplify_expr): Ensure OMP_ARRAY_SECTION has been
5210 processed out before gimplification.
5211 * tree-pretty-print.cc (dump_generic_node): Support OMP_ARRAY_SECTION.
5212 * tree.def (OMP_ARRAY_SECTION): New tree code.
5214 2024-01-09 Jakub Jelinek <jakub@redhat.com>
5216 PR tree-optimization/113210
5217 * tree-vect-loop.cc (vect_get_loop_niters): If non-INTEGER_CST
5218 value in *number_of_iterationsm1 PLUS_EXPR 1 is folded into
5219 INTEGER_CST, recompute *number_of_iterationsm1 as the INTEGER_CST
5222 2024-01-09 Eric Botcazou <ebotcazou@adacore.com>
5224 PR rtl-optimization/113140
5225 * reorg.cc (fill_slots_from_thread): If we are to branch after the
5226 last instruction of the function, create an end label.
5228 2024-01-09 Roger Sayle <roger@nextmovesoftware.com>
5229 Hongtao Liu <hongtao.liu@intel.com>
5232 * config/i386/i386-expand.cc
5233 (ix86_convert_const_wide_int_to_broadcast): Allow call to
5234 ix86_expand_vector_init_duplicate to fail, and return NULL_RTX.
5235 (ix86_broadcast_from_constant): Revert recent change; Return a
5236 suitable MEMREF independently of mode/target combinations.
5237 (ix86_expand_vector_move): Allow ix86_expand_vector_init_duplicate
5238 to decide whether expansion is possible/preferrable. Only try
5239 forcing DImode constants to memory (and trying again) if calling
5240 ix86_expand_vector_init_duplicate fails with an DImode immediate
5242 (ix86_expand_vector_init_duplicate) <case E_V2DImode>: Try using
5243 V4SImode for suitable immediate constants.
5244 <case E_V4DImode>: Try using V8SImode for suitable constants.
5245 <case E_V4HImode>: Fail for CONST_INT_P, i.e. use constant pool.
5246 <case E_V2HImode>: Likewise.
5247 <case E_V8HImode>: For CONST_INT_P try using V4SImode via widen.
5248 <case E_V16QImode>: For CONT_INT_P try using V8HImode via widen.
5249 <label widen>: Handle CONT_INTs via simplify_binary_operation.
5250 Allow recursive calls to ix86_expand_vector_init_duplicate to fail.
5251 <case E_V16HImode>: For CONST_INT_P try V8SImode via widen.
5252 <case E_V32QImode>: For CONST_INT_P try V16HImode via widen.
5253 (ix86_expand_vector_init): Move try using a broadcast for all_same
5254 with ix86_expand_vector_init_duplicate before using constant pool.
5256 2024-01-09 Chung-Ju Wu <jasonwucj@gmail.com>
5258 * doc/invoke.texi (Arm Options): Document Cortex-M52 options.
5260 2024-01-09 Chung-Ju Wu <jasonwucj@gmail.com>
5262 * config/arm/arm-cpus.in (cortex-m52): New cpu.
5263 * config/arm/arm-tables.opt: Regenerate.
5264 * config/arm/arm-tune.md: Regenerate.
5266 2024-01-09 Jiahao Xu <xujiahao@loongson.cn>
5268 * config/loongarch/lasx.md (vec_initv32qiv16qi): Rename to ..
5269 (vec_init<mode><lasxhalf>): .. this, and extend to mode.
5270 (@vec_concatz<mode>): New insn pattern.
5271 * config/loongarch/loongarch.cc (loongarch_expand_vector_group_init):
5272 Handle VALS containing two vectors.
5274 2024-01-09 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5276 * config/riscv/riscv-vector-builtins-functions.def (vleff): Move comments.
5277 (vundefined): Ditto.
5279 2024-01-09 Feng Wang <wangfeng@eswincomputing.com>
5281 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
5282 Add new function_base for crypto vector.
5283 (class bitmanip): Ditto.
5284 (class b_reverse):Ditto.
5285 (class vwsll): Ditto.
5286 (class clmul): Ditto.
5287 (class vg_nhab): Ditto.
5288 (class crypto_vv):Ditto.
5289 (class crypto_vi):Ditto.
5290 (class vaeskf2_vsm3c):Ditto.
5291 (class vsm3me): Ditto.
5292 (BASE): Add BASE declaration for crypto vector.
5293 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
5294 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
5295 Add crypto vector intrinsic definition.
5323 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
5324 Add new function_shape for crypto vector.
5325 (struct crypto_vi_def): Ditto.
5326 (struct crypto_vv_no_op_type_def): Ditto.
5327 (SHAPE): Add SHAPE declaration of crypto vector.
5328 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
5329 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
5330 Add new data type for crypto vector.
5331 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5332 (vuint32mf2_t): Ditto.
5333 (vuint32m1_t): Ditto.
5334 (vuint32m2_t): Ditto.
5335 (vuint32m4_t): Ditto.
5336 (vuint32m8_t): Ditto.
5337 (vuint64m1_t): Ditto.
5338 (vuint64m2_t): Ditto.
5339 (vuint64m4_t): Ditto.
5340 (vuint64m8_t): Ditto.
5341 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
5342 Add new data struct for crypto vector.
5343 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5344 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
5345 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
5347 2024-01-08 Ilya Leoshkevich <iii@linux.ibm.com>
5350 * varasm.cc (assemble_function_label_raw): Do not call
5351 asan_function_start () without the current function.
5353 2024-01-08 Cupertino Miranda <cupertino.miranda@oracle.com>
5356 * btfout.cc (btf_collect_datasec): Skip creating BTF info for
5357 extern and kernel_helper attributed function decls.
5359 2024-01-08 Cupertino Miranda <cupertino.miranda@oracle.com>
5361 * btfout.cc (output_btf_strs): Changed.
5363 2024-01-08 Tobias Burnus <tobias@codesourcery.com>
5365 * config/gcn/mkoffload.cc (main): Handle gfx1100
5366 when setting the default XNACK.
5368 2024-01-08 Tobias Burnus <tobias@codesourcery.com>
5370 * config.gcc (amdgcn-*-amdhsa): Accept --with-arch=gfx1100.
5371 * config/gcn/gcn-hsa.h (NO_XNACK): Add gfx1100:
5372 (ASM_SPEC): Handle gfx1100.
5373 * config/gcn/gcn-opts.h (enum processor_type): Add PROCESSOR_GFX1100.
5374 (enum gcn_isa): Add ISA_RDNA3.
5375 (TARGET_GFX1100, TARGET_RDNA2_PLUS, TARGET_RDNA3): Define.
5376 * config/gcn/gcn-valu.md: Change TARGET_RDNA2 to TARGET_RDNA2_PLUS.
5377 * config/gcn/gcn.cc (gcn_option_override,
5378 gcn_omp_device_kind_arch_isa, output_file_start): Handle gfx1100.
5379 (gcn_global_address_p, gcn_addr_space_legitimate_address_p): Change
5380 TARGET_RDNA2 to TARGET_RDNA2_PLUS.
5381 (gcn_hsa_declare_function_name): Don't use '.amdhsa_reserve_flat_scratch'
5383 * config/gcn/gcn.h (ASSEMBLER_DIALECT): Likewise.
5384 (TARGET_CPU_CPP_BUILTINS): Define __RDNA3__, __gfx1030__ and
5386 * config/gcn/gcn.md: Change TARGET_RDNA2 to TARGET_RDNA2_PLUS.
5387 * config/gcn/gcn.opt (Enum gpu_type): Add gfx1100.
5388 * config/gcn/mkoffload.cc (EF_AMDGPU_MACH_AMDGCN_GFX1100): Define.
5389 (isa_has_combined_avgprs, main): Handle gfx1100.
5390 * config/gcn/t-omp-device (isa): Add gfx1100.
5392 2024-01-08 Richard Biener <rguenther@suse.de>
5394 * doc/invoke.texi (-mmovbe): Clarify.
5396 2024-01-08 Richard Biener <rguenther@suse.de>
5398 PR tree-optimization/113026
5399 * tree-vect-loop.cc (vect_need_peeling_or_partial_vectors_p):
5400 Avoid an epilog in more cases.
5401 * tree-vect-loop-manip.cc (vect_do_peeling): Adjust the
5402 epilogues niter upper bounds and estimates.
5404 2024-01-08 Jakub Jelinek <jakub@redhat.com>
5406 PR tree-optimization/113228
5407 * gimplify.cc (recalculate_side_effects): Do nothing for SSA_NAMEs.
5409 2024-01-08 Jakub Jelinek <jakub@redhat.com>
5411 PR tree-optimization/113120
5412 * gimple-lower-bitint.cc (gimple_lower_bitint): Fix handling of very
5413 large _BitInt zero INTEGER_CST PHI argument.
5415 2024-01-08 Jakub Jelinek <jakub@redhat.com>
5417 PR tree-optimization/113119
5418 * gimple-lower-bitint.cc (optimizable_arith_overflow): Punt if
5419 both REALPART_EXPR and cast from IMAGPART_EXPR appear, but cast
5420 is before REALPART_EXPR.
5422 2024-01-08 Georg-Johann Lay <avr@gjlay.de>
5425 * config/avr/avr.cc (avr_handle_addr_attribute): Also print valid
5426 range when diagnosing attribute "io" and "io_low" are out of range.
5427 (avr_eval_addr_attrib): Don't ICE on empty address at that place.
5428 (avr_insert_attributes): Reject if attribute "address", "io" or "io_low"
5429 in contexts other than static storage.
5430 (avr_asm_output_aligned_decl_common): Move output of decls with
5431 attribute "address", "io", and "io_low" to...
5432 (avr_output_addr_attrib): ...this new function.
5433 (avr_asm_asm_output_aligned_bss): Remove output for decls with
5434 attribute "address", "io", and "io_low".
5435 (avr_encode_section_info): Rectify handling of decls with attribute
5436 "address", "io", and "io_low".
5438 2024-01-08 Andrew Stubbs <ams@codesourcery.com>
5440 * config/gcn/mkoffload.cc (TEST_XNACK_UNSET): New.
5441 (elf_flags): Remove XNACK from the default value.
5442 (main): Set a default XNACK according to the arch.
5444 2024-01-08 Andrew Stubbs <ams@codesourcery.com>
5446 * config/gcn/mkoffload.cc (isa_has_combined_avgprs): Delete.
5447 (process_asm): Don't count avgprs.
5449 2024-01-08 Hongyu Wang <hongyu.wang@intel.com>
5451 * config/i386/i386.opt: Add supported sub-features.
5452 * doc/extend.texi: Add description for target attribute.
5454 2024-01-08 Feng Wang <wangfeng@eswincomputing.com>
5456 * config/riscv/vector.md: Modify avl_type operand index of zvbc ins.
5458 2024-01-07 Roger Sayle <roger@nextmovesoftware.com>
5459 Uros Bizjak <ubizjak@gmail.com>
5462 * config/i386/i386-features.cc (compute_convert_gain): Include
5463 the overhead of explicit load and store (movd) instructions when
5464 converting non-store scalar operations with memory destinations.
5465 Various indentation whitespace fixes.
5467 2024-01-07 Tamar Christina <tamar.christina@arm.com>
5469 * config/arm/neon.md (cbranch<mode>4): New.
5471 2024-01-07 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5473 * config/riscv/riscv-vsetvl.cc: replace std::max by MAX.
5475 2024-01-06 Jiahao Xu <xujiahao@loongson.cn>
5477 * config/loongarch/lasx.md: Set the unused bits in operand[3] to 0.
5479 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5482 * config/riscv/riscv-vsetvl.cc (pre_vsetvl::fuse_local_vsetvl_info):
5485 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5487 * config/riscv/riscv-vector-costs.cc (loop_invariant_op_p): New function.
5488 (variable_vectorized_p): Teach loop invariant.
5489 (has_unexpected_spills_p): Ditto.
5491 2024-01-06 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5493 * config/riscv/riscv-protos.h (whole_reg_to_reg_move_p): New function.
5494 * config/riscv/riscv-v.cc (whole_reg_to_reg_move_p): Ditto.
5495 * config/riscv/vector.md: Allow non-vlmax with len = NUNITS simplification.
5497 2024-01-05 Richard Sandiford <richard.sandiford@arm.com>
5500 * doc/invoke.texi (aarch64-sve-compare-costs): Replace with...
5501 (aarch64-vect-compare-costs): ...this.
5502 * config/aarch64/aarch64.opt (-param=aarch64-sve-compare-costs=):
5504 (-param=aarch64-vect-compare-costs=): ...this new param.
5505 * config/aarch64/aarch64.cc (aarch64_override_options_internal):
5506 Don't disable it when vectorizing for Advanced SIMD only.
5507 (aarch64_autovectorize_vector_modes): Apply VECT_COMPARE_COSTS
5508 whenever aarch64_vect_compare_costs is true.
5510 2024-01-05 Lulu Cheng <chenglulu@loongson.cn>
5512 * config/loongarch/lasx.md (lasx_mxld_<lasxfmt_f>):
5513 Modify the method of determining the memory offset of [x]vld/[x]vst.
5514 (lasx_mxst_<lasxfmt_f>): Likewise.
5515 * config/loongarch/loongarch.cc (loongarch_valid_offset_p): Delete.
5516 (loongarch_address_insns): Likewise.
5517 * config/loongarch/lsx.md (lsx_ld_<lsxfmt_f>): Likewise.
5518 (lsx_st_<lsxfmt_f>): Likewise.
5519 * config/loongarch/predicates.md (aq10b_operand): Likewise.
5520 (aq10h_operand): Likewise.
5521 (aq10w_operand): Likewise.
5522 (aq10d_operand): Likewise.
5524 2024-01-05 Alex Coplan <alex.coplan@arm.com>
5527 * config/aarch64/aarch64-ldp-fusion.cc
5528 (ldp_bb_info::try_fuse_pair): If the second access can throw,
5529 narrow the move range to exactly that insn.
5531 2024-01-05 Ilya Leoshkevich <iii@linux.ibm.com>
5533 * asan.cc (asan_function_start): Drop switch_to_section ().
5534 (asan_emit_stack_protection): Set .LASANPC alignment.
5535 * config/i386/i386.cc: Use assemble_function_label_raw ()
5536 instead of ASM_OUTPUT_LABEL ().
5537 * config/s390/s390.cc (s390_asm_output_function_label):
5539 * defaults.h (ASM_OUTPUT_FUNCTION_LABEL): Likewise.
5540 * final.cc (final_start_function_1): Drop
5541 asan_function_start ().
5542 * output.h (assemble_function_label_raw): New function.
5543 * varasm.cc (assemble_function_label_raw): Likewise.
5545 2024-01-05 Ilya Leoshkevich <iii@linux.ibm.com>
5547 * config/aarch64/aarch64.cc (aarch64_declare_function_name):
5548 Use ASM_OUTPUT_FUNCTION_LABEL ().
5549 * config/alpha/alpha.cc (alpha_start_function): Likewise.
5550 * config/arm/aout.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
5551 * config/arm/arm.cc (arm_asm_declare_function_name): Likewise.
5552 * config/bfin/bfin.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
5553 * config/c6x/c6x.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
5554 * config/gcn/gcn.cc (gcn_hsa_declare_function_name): Likewise.
5555 * config/h8300/h8300.h (ASM_DECLARE_FUNCTION_NAME): Likewise.
5556 * config/ia64/ia64.cc (ia64_start_function): Likewise.
5557 * config/mcore/mcore-elf.h (ASM_DECLARE_FUNCTION_NAME):
5559 * config/microblaze/microblaze.cc (microblaze_function_prologue):
5561 * config/mips/mips.cc (mips_start_unique_function): Return the
5563 (mips_start_function_definition): Use
5564 ASM_OUTPUT_FUNCTION_LABEL ().
5565 (mips_finish_stub): Pass the tree to
5566 mips_start_function_definition ().
5567 (mips16_build_function_stub): Likewise.
5568 (mips16_build_call_stub): Likewise.
5569 (mips_output_function_prologue): Likewise.
5570 * config/pa/pa.cc (pa_output_function_label): Use
5571 ASM_OUTPUT_FUNCTION_LABEL ().
5572 * config/riscv/riscv.cc (riscv_declare_function_name): Likewise.
5573 * config/rs6000/rs6000.cc (rs6000_elf_declare_function_name):
5575 (rs6000_xcoff_declare_function_name): Likewise.
5577 2024-01-05 Jakub Jelinek <jakub@redhat.com>
5579 PR tree-optimization/113201
5580 * tree-scalar-evolution.cc (final_value_replacement_loop): Don't call
5581 replace_uses_by on SSA_NAME_OCCURS_IN_ABNORMAL_PHI rslt.
5583 2024-01-05 Jakub Jelinek <jakub@redhat.com>
5585 PR tree-optimization/90693
5586 * tree-ssa-math-opts.cc (match_single_bit_test): If
5587 tree_expr_nonzero_p (arg), remember it in the second argument to
5588 IFN_POPCOUNT or lower it as arg & (arg - 1) == 0 rather than
5589 arg ^ (arg - 1) > arg - 1.
5590 * internal-fn.cc (expand_POPCOUNT): If second argument to
5591 IFN_POPCOUNT suggests arg is non-zero, try to expand it as
5592 arg & (arg - 1) == 0 rather than arg ^ (arg - 1) > arg - 1.
5594 2024-01-05 Kito Cheng <kito.cheng@sifive.com>
5596 * config/riscv/riscv-v.cc (expand_load_store):
5598 (expand_cond_len_op): Ditto.
5599 (expand_gather_scatter): Ditto.
5600 (expand_lanes_load_store): Ditto.
5601 (expand_fold_extract_last): Ditto.
5603 2024-01-05 Pan Li <pan2.li@intel.com>
5606 2024-01-05 Feng Wang <wangfeng@eswincomputing.com>
5608 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
5609 Add new function_base for crypto vector.
5610 (class bitmanip): Ditto.
5611 (class b_reverse):Ditto.
5612 (class vwsll): Ditto.
5613 (class clmul): Ditto.
5614 (class vg_nhab): Ditto.
5615 (class crypto_vv):Ditto.
5616 (class crypto_vi):Ditto.
5617 (class vaeskf2_vsm3c):Ditto.
5618 (class vsm3me): Ditto.
5619 (BASE): Add BASE declaration for crypto vector.
5620 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
5621 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
5622 Add crypto vector intrinsic definition.
5650 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
5651 Add new function_shape for crypto vector.
5652 (struct crypto_vi_def): Ditto.
5653 (struct crypto_vv_no_op_type_def): Ditto.
5654 (SHAPE): Add SHAPE declaration of crypto vector.
5655 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
5656 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
5657 Add new data type for crypto vector.
5658 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5659 (vuint32mf2_t): Ditto.
5660 (vuint32m1_t): Ditto.
5661 (vuint32m2_t): Ditto.
5662 (vuint32m4_t): Ditto.
5663 (vuint32m8_t): Ditto.
5664 (vuint64m1_t): Ditto.
5665 (vuint64m2_t): Ditto.
5666 (vuint64m4_t): Ditto.
5667 (vuint64m8_t): Ditto.
5668 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
5669 Add new data struct for crypto vector.
5670 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5671 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
5672 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
5674 2024-01-05 Feng Wang <wangfeng@eswincomputing.com>
5676 * config/riscv/riscv-vector-builtins-bases.cc (class vandn):
5677 Add new function_base for crypto vector.
5678 (class bitmanip): Ditto.
5679 (class b_reverse):Ditto.
5680 (class vwsll): Ditto.
5681 (class clmul): Ditto.
5682 (class vg_nhab): Ditto.
5683 (class crypto_vv):Ditto.
5684 (class crypto_vi):Ditto.
5685 (class vaeskf2_vsm3c):Ditto.
5686 (class vsm3me): Ditto.
5687 (BASE): Add BASE declaration for crypto vector.
5688 * config/riscv/riscv-vector-builtins-bases.h: Ditto.
5689 * config/riscv/riscv-vector-builtins-functions.def (REQUIRED_EXTENSIONS):
5690 Add crypto vector intrinsic definition.
5718 * config/riscv/riscv-vector-builtins-shapes.cc (struct crypto_vv_def):
5719 Add new function_shape for crypto vector.
5720 (struct crypto_vi_def): Ditto.
5721 (struct crypto_vv_no_op_type_def): Ditto.
5722 (SHAPE): Add SHAPE declaration of crypto vector.
5723 * config/riscv/riscv-vector-builtins-shapes.h: Ditto.
5724 * config/riscv/riscv-vector-builtins-types.def (DEF_RVV_CRYPTO_SEW32_OPS):
5725 Add new data type for crypto vector.
5726 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5727 (vuint32mf2_t): Ditto.
5728 (vuint32m1_t): Ditto.
5729 (vuint32m2_t): Ditto.
5730 (vuint32m4_t): Ditto.
5731 (vuint32m8_t): Ditto.
5732 (vuint64m1_t): Ditto.
5733 (vuint64m2_t): Ditto.
5734 (vuint64m4_t): Ditto.
5735 (vuint64m8_t): Ditto.
5736 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_CRYPTO_SEW32_OPS):
5737 Add new data struct for crypto vector.
5738 (DEF_RVV_CRYPTO_SEW64_OPS): Ditto.
5739 (registered_function::overloaded_hash): Processing size_t uimm for C overloaded func.
5740 * config/riscv/riscv-vector-builtins.def (vi): Add vi OP_TYPE.
5742 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5744 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
5746 2024-01-04 Andrew Pinski <quic_apinski@quicinc.com>
5748 PR tree-optimization/113186
5749 * gimple-match-head.cc (gimple_bitwise_inverted_equal_p):
5750 Match `^` with the `==` for 1bit integral types.
5751 * match.pd (maybe_cmp): Allow for bit_xor for 1bit
5754 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5756 * toplev.cc (general_init): Pass lang_mask to urlifier.
5758 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5760 * diagnostic.h (diagnostic_make_option_url_cb): Add lang_mask
5762 (diagnostic_context::make_option_url): Update for lang_mask param.
5763 * gcc-urlifier.cc: Include "opts.h" and "options.h".
5764 (gcc_urlifier::gcc_urlifier): Add lang_mask param.
5765 (gcc_urlifier::m_lang_mask): New field.
5766 (doc_urls): Make static.
5767 (gcc_urlifier::get_url_for_quoted_text): Use label_text.
5768 (gcc_urlifier::get_url_suffix_for_quoted_text): Use label_text.
5769 Look for an option by name before trying a binary search in
5771 (gcc_urlifier::get_url_suffix_for_quoted_text): Use label_text.
5772 (gcc_urlifier::get_url_suffix_for_option): New.
5773 (make_gcc_urlifier): Add lang_mask param.
5774 (selftest::gcc_urlifier_cc_tests): Update for above changes.
5775 Verify that a URL is found for "-fpack-struct".
5776 * gcc-urlifier.def: Drop options "--version" and "-fpack-struct".
5777 * gcc-urlifier.h (make_gcc_urlifier): Add lang_mask param.
5778 * gcc.cc (driver::global_initializations): Pass 0 for lang_mask
5779 to make_gcc_urlifier.
5780 * opts-diagnostic.h (get_option_url): Add lang_mask param.
5781 * opts.cc (get_option_html_page): Remove special-casing for
5783 (get_option_url_suffix): New.
5784 (get_option_url): Reimplement.
5785 (selftest::test_get_option_html_page): Rename to...
5786 (selftest::test_get_option_url_suffix): ...this and update for
5788 (selftest::opts_cc_tests): Update for renaming.
5789 * opts.h: Include "rich-location.h".
5790 (get_option_url_suffix): New decl.
5792 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5794 * Makefile.in (ALL_OPT_URL_FILES): New.
5795 (GCC_OBJS): Add options-urls.o.
5797 (OBJS-libcommon): Likewise.
5798 (s-options): Depend on $(ALL_OPT_URL_FILES), and add this to
5799 inputs to opt-gather.awk.
5800 (options-urls.cc): New Makefile target.
5801 * opt-functions.awk (url_suffix): New function.
5802 (lang_url_suffix): New function.
5803 * options-urls-cc-gen.awk: New file.
5804 * opts.h (get_opt_url_suffix): New decl.
5806 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5808 * params.opt.urls: New file, autogenerated by
5809 regenerate-opt-urls.py.
5811 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5813 * common.opt.urls: New file, autogenerated by
5814 regenerate-opt-urls.py.
5815 * config/aarch64/aarch64.opt.urls: Likewise.
5816 * config/alpha/alpha.opt.urls: Likewise.
5817 * config/alpha/elf.opt.urls: Likewise.
5818 * config/arc/arc-tables.opt.urls: Likewise.
5819 * config/arc/arc.opt.urls: Likewise.
5820 * config/arm/arm-tables.opt.urls: Likewise.
5821 * config/arm/arm.opt.urls: Likewise.
5822 * config/arm/vxworks.opt.urls: Likewise.
5823 * config/avr/avr.opt.urls: Likewise.
5824 * config/bpf/bpf.opt.urls: Likewise.
5825 * config/c6x/c6x-tables.opt.urls: Likewise.
5826 * config/c6x/c6x.opt.urls: Likewise.
5827 * config/cris/cris.opt.urls: Likewise.
5828 * config/cris/elf.opt.urls: Likewise.
5829 * config/csky/csky.opt.urls: Likewise.
5830 * config/csky/csky_tables.opt.urls: Likewise.
5831 * config/darwin.opt.urls: Likewise.
5832 * config/dragonfly.opt.urls: Likewise.
5833 * config/epiphany/epiphany.opt.urls: Likewise.
5834 * config/fr30/fr30.opt.urls: Likewise.
5835 * config/freebsd.opt.urls: Likewise.
5836 * config/frv/frv.opt.urls: Likewise.
5837 * config/ft32/ft32.opt.urls: Likewise.
5838 * config/fused-madd.opt.urls: Likewise.
5839 * config/g.opt.urls: Likewise.
5840 * config/gcn/gcn.opt.urls: Likewise.
5841 * config/gnu-user.opt.urls: Likewise.
5842 * config/h8300/h8300.opt.urls: Likewise.
5843 * config/hpux11.opt.urls: Likewise.
5844 * config/i386/cygming.opt.urls: Likewise.
5845 * config/i386/cygwin.opt.urls: Likewise.
5846 * config/i386/djgpp.opt.urls: Likewise.
5847 * config/i386/i386.opt.urls: Likewise.
5848 * config/i386/mingw-w64.opt.urls: Likewise.
5849 * config/i386/mingw.opt.urls: Likewise.
5850 * config/i386/nto.opt.urls: Likewise.
5851 * config/ia64/ia64.opt.urls: Likewise.
5852 * config/ia64/ilp32.opt.urls: Likewise.
5853 * config/ia64/vms.opt.urls: Likewise.
5854 * config/iq2000/iq2000.opt.urls: Likewise.
5855 * config/linux-android.opt.urls: Likewise.
5856 * config/linux.opt.urls: Likewise.
5857 * config/lm32/lm32.opt.urls: Likewise.
5858 * config/loongarch/loongarch.opt.urls: Likewise.
5859 * config/lynx.opt.urls: Likewise.
5860 * config/m32c/m32c.opt.urls: Likewise.
5861 * config/m32r/m32r.opt.urls: Likewise.
5862 * config/m68k/ieee.opt.urls: Likewise.
5863 * config/m68k/m68k-tables.opt.urls: Likewise.
5864 * config/m68k/m68k.opt.urls: Likewise.
5865 * config/m68k/uclinux.opt.urls: Likewise.
5866 * config/mcore/mcore.opt.urls: Likewise.
5867 * config/microblaze/microblaze.opt.urls: Likewise.
5868 * config/mips/mips-tables.opt.urls: Likewise.
5869 * config/mips/mips.opt.urls: Likewise.
5870 * config/mips/sde.opt.urls: Likewise.
5871 * config/mmix/mmix.opt.urls: Likewise.
5872 * config/mn10300/mn10300.opt.urls: Likewise.
5873 * config/moxie/moxie.opt.urls: Likewise.
5874 * config/msp430/msp430.opt.urls: Likewise.
5875 * config/nds32/nds32-elf.opt.urls: Likewise.
5876 * config/nds32/nds32-linux.opt.urls: Likewise.
5877 * config/nds32/nds32.opt.urls: Likewise.
5878 * config/netbsd-elf.opt.urls: Likewise.
5879 * config/netbsd.opt.urls: Likewise.
5880 * config/nios2/elf.opt.urls: Likewise.
5881 * config/nios2/nios2.opt.urls: Likewise.
5882 * config/nvptx/nvptx-gen.opt.urls: Likewise.
5883 * config/nvptx/nvptx.opt.urls: Likewise.
5884 * config/openbsd.opt.urls: Likewise.
5885 * config/or1k/elf.opt.urls: Likewise.
5886 * config/or1k/or1k.opt.urls: Likewise.
5887 * config/pa/pa-hpux.opt.urls: Likewise.
5888 * config/pa/pa-hpux1010.opt.urls: Likewise.
5889 * config/pa/pa-hpux1111.opt.urls: Likewise.
5890 * config/pa/pa-hpux1131.opt.urls: Likewise.
5891 * config/pa/pa.opt.urls: Likewise.
5892 * config/pa/pa64-hpux.opt.urls: Likewise.
5893 * config/pdp11/pdp11.opt.urls: Likewise.
5894 * config/pru/pru.opt.urls: Likewise.
5895 * config/riscv/riscv.opt.urls: Likewise.
5896 * config/rl78/rl78.opt.urls: Likewise.
5897 * config/rpath.opt.urls: Likewise.
5898 * config/rs6000/476.opt.urls: Likewise.
5899 * config/rs6000/aix64.opt.urls: Likewise.
5900 * config/rs6000/darwin.opt.urls: Likewise.
5901 * config/rs6000/linux64.opt.urls: Likewise.
5902 * config/rs6000/rs6000-tables.opt.urls: Likewise.
5903 * config/rs6000/rs6000.opt.urls: Likewise.
5904 * config/rs6000/sysv4.opt.urls: Likewise.
5905 * config/rtems.opt.urls: Likewise.
5906 * config/rx/elf.opt.urls: Likewise.
5907 * config/rx/rx.opt.urls: Likewise.
5908 * config/s390/s390.opt.urls: Likewise.
5909 * config/s390/tpf.opt.urls: Likewise.
5910 * config/sh/sh.opt.urls: Likewise.
5911 * config/sh/superh.opt.urls: Likewise.
5912 * config/sol2.opt.urls: Likewise.
5913 * config/sparc/long-double-switch.opt.urls: Likewise.
5914 * config/sparc/sparc.opt.urls: Likewise.
5915 * config/stormy16/stormy16.opt.urls: Likewise.
5916 * config/v850/v850.opt.urls: Likewise.
5917 * config/vax/elf.opt.urls: Likewise.
5918 * config/vax/vax.opt.urls: Likewise.
5919 * config/visium/visium.opt.urls: Likewise.
5920 * config/vms/vms.opt.urls: Likewise.
5921 * config/vxworks-smp.opt.urls: Likewise.
5922 * config/vxworks.opt.urls: Likewise.
5923 * config/xtensa/elf.opt.urls: Likewise.
5924 * config/xtensa/uclinux.opt.urls: Likewise.
5925 * config/xtensa/xtensa.opt.urls: Likewise.
5926 * config/bfin/bfin.opt.urls: New file.
5928 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5930 * Makefile.in (OPT_URLS_HTML_DEPS): New.
5931 (regenerate-opt-urls): New target.
5932 (regenerate-opt-urls-unit-test): New target.
5933 * doc/options.texi (Option properties): Add UrlSuffix and
5934 description of regenerate-opt-urls.py. Add LangUrlSuffix_*.
5935 * doc/sourcebuild.texi (Anatomy of a Language Front End): Add
5936 reference to regenerate-opt-urls.py's PER_LANGUAGE_OPTION_INDEXES
5937 and Makefile.in's OPT_URLS_HTML_DEPS.
5938 (Anatomy of a Target Back End): Add
5939 reference to regenerate-opt-urls.py's TARGET_SPECIFIC_PAGES.
5940 * regenerate-opt-urls.py: New file.
5942 2024-01-04 David Malcolm <dmalcolm@redhat.com>
5944 * diagnostic-format-sarif.cc
5945 (sarif_builder::make_logical_location_object): Convert to...
5946 (make_sarif_logical_location_object): ...this.
5947 (sarif_builder::set_any_logical_locs_arr): Update for above
5949 (sarif_builder::make_thread_flow_location_object): Call
5950 maybe_add_sarif_properties on each diagnostic_event.
5951 * diagnostic-format-sarif.h (class logical_location): New forward
5953 (make_sarif_logical_location_object): New decl.
5954 * diagnostic-path.h (class sarif_object): New forward decl.
5955 (diagnostic_event::maybe_add_sarif_properties): New vfunc.
5957 2024-01-04 Kuan-Lin Chen <rufus@andestech.com>
5958 Patrick Lin <patrick@andestech.com>
5959 Rufus Chen <rufus@andestech.com>
5960 Monk Chiang <monk.chiang@sifive.com>
5962 * config/riscv/riscv.cc (riscv_legitimize_move): Expand movfh
5963 with Nan-boxing value.
5964 * config/riscv/riscv.md (*movhf_softfloat_unspec): New pattern.
5966 2024-01-04 Roger Sayle <roger@nextmovesoftware.com>
5967 Jeff Law <jlaw@ventanamicro.com>
5969 PR rtl-optimization/104914
5970 * expr.cc (expand_assignment): When target is SUBREG_PROMOTED_VAR_P
5971 a sign or zero extension is only required if the modified field
5972 overlaps the SUBREG's most significant bit. On MODE_REP_EXTENDED
5973 targets, don't refer to the temporarily incorrectly extended value
5974 using a SUBREG, but instead generate an explicit TRUNCATE rtx.
5976 2024-01-04 Pan Li <pan2.li@intel.com>
5979 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5981 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
5983 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5985 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): Teach vi variant.
5987 2024-01-04 Kito Cheng <kito.cheng@sifive.com>
5989 * config/riscv/riscv.cc (riscv_for_each_saved_reg): Adjust the
5992 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
5994 * config/riscv/riscv-vector-costs.cc (variable_vectorized_p): New function.
5995 (compute_nregs_for_mode): Refine LMUL.
5996 (max_number_of_live_regs): Ditto.
5997 (compute_estimated_lmul): Ditto.
5998 (has_unexpected_spills_p): Ditto.
6000 2024-01-04 Li Wei <liwei@loongson.cn>
6002 * config/loongarch/loongarch.cc (loongarch_is_odd_extraction):
6003 Remove useless forward declaration.
6004 (loongarch_is_even_extraction): Remove useless forward declaration.
6005 (loongarch_try_expand_lsx_vshuf_const): Removed.
6006 (loongarch_expand_vec_perm_const_1): Merged.
6007 (loongarch_is_double_duplicate): Removed.
6008 (loongarch_is_center_extraction): Ditto.
6009 (loongarch_is_reversing_permutation): Ditto.
6010 (loongarch_is_di_misalign_extract): Ditto.
6011 (loongarch_is_si_misalign_extract): Ditto.
6012 (loongarch_is_lasx_lowpart_extract): Ditto.
6013 (loongarch_is_op_reverse_perm): Ditto.
6014 (loongarch_is_single_op_perm): Ditto.
6015 (loongarch_is_divisible_perm): Ditto.
6016 (loongarch_is_triple_stride_extract): Ditto.
6017 (loongarch_expand_vec_perm_const_2): Merged.
6018 (loongarch_expand_vec_perm_const): New.
6019 (loongarch_vectorize_vec_perm_const): Adjust.
6021 2024-01-04 Sandra Loosemore <sandra@codesourcery.com>
6023 * omp-general.cc: Fix comment typos and misplaced/confusing
6024 comments. Delete redundant include of omp-general.h.
6026 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
6028 PR rtl-optimization/104914
6029 * config/mips/mips.md (insqisi_extended): New patterns.
6030 (inshisi_extended): Ditto.
6032 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
6034 * config/mips/mips.cc (mips_insn_cost): New function.
6036 2024-01-04 YunQiang Su <syq@gcc.gnu.org>
6038 * config/mips/mips.md (perf_ratio): New attribute.
6040 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6044 * config/riscv/riscv-vsetvl.cc (invalid_opt_bb_p): New function.
6045 (pre_vsetvl::compute_lcm_local_properties): Disable earliest fusion on
6046 blocks belong to infinite loop.
6047 (pre_vsetvl::emit_vsetvl): Remove fake edges.
6048 * config/riscv/t-riscv: Add a new include file.
6050 2024-01-04 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6052 * config/riscv/vector.md: Fix indent.
6054 2024-01-03 Kwok Cheung Yeung <kcy@codesourcery.com>
6056 * tree-core.h (enum omp_clause_code): Move OMP_CLAUSE_INDIRECT to before
6057 OMP_CLAUSE__SIMDUID_.
6058 * tree.cc (omp_clause_num_ops): Update position of entry for
6059 OMP_CLAUSE_INDIRECT to correspond with omp_clause_code.
6060 (omp_clause_code_name): Likewise.
6062 2024-01-03 Kwok Cheung Yeung <kcy@codesourcery.com>
6064 * config/nvptx/nvptx.cc (nvptx_record_offload_symbol): Restucture
6065 printing of FUNC_MAP/IND_FUNC_MAP labels.
6067 2024-01-03 Jakub Jelinek <jakub@redhat.com>
6069 * gcc.cc (process_command): Update copyright notice dates.
6070 * gcov-dump.cc (print_version): Ditto.
6071 * gcov.cc (print_version): Ditto.
6072 * gcov-tool.cc (print_version): Ditto.
6073 * gengtype.cc (create_file): Ditto.
6074 * doc/cpp.texi: Bump @copying's copyright year.
6075 * doc/cppinternals.texi: Ditto.
6076 * doc/gcc.texi: Ditto.
6077 * doc/gccint.texi: Ditto.
6078 * doc/gcov.texi: Ditto.
6079 * doc/install.texi: Ditto.
6080 * doc/invoke.texi: Ditto.
6082 2024-01-03 Xi Ruoyao <xry111@xry111.site>
6084 * config/loongarch/simd.md (fmax<mode>3): New define_insn.
6085 (fmin<mode>3): Likewise.
6086 (reduc_fmax_scal_<mode>3): New define_expand.
6087 (reduc_fmin_scal_<mode>3): Likewise.
6089 2024-01-03 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6092 * config/riscv/riscv-vector-costs.cc (compute_nregs_for_mode): Add rgroup info.
6093 (max_number_of_live_regs): Ditto.
6094 (has_unexpected_spills_p): Ditto.
6096 2024-01-02 Jun Sha (Joshua) <cooper.joshua@linux.alibaba.com>
6097 Jin Ma <jinma@linux.alibaba.com>
6098 Xianmiao Qu <cooper.qu@linux.alibaba.com>
6099 Christoph Müllner <christoph.muellner@vrull.eu>
6101 * config/riscv/vector.md:
6102 Use vector_length_operand for vsetvl patterns.
6104 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6106 * config/riscv/riscv-v.cc (is_vlmax_len_p): Remove satisfies_constraint_K.
6107 (expand_cond_len_op): Add simplification of dummy len and dummy mask.
6109 2024-01-02 Di Zhao <dizhao@os.amperecomputing.com>
6111 * config/aarch64/aarch64-tuning-flags.def
6112 (AARCH64_EXTRA_TUNING_OPTION): New tuning option
6113 AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA.
6114 * config/aarch64/aarch64.cc
6115 (aarch64_override_options_internal): Set
6116 param_fully_pipelined_fma according to tuning option.
6117 * config/aarch64/tuning_models/ampere1.h: Add
6118 AARCH64_EXTRA_TUNE_FULLY_PIPELINED_FMA to tune_flags.
6119 * config/aarch64/tuning_models/ampere1a.h: Likewise.
6120 * config/aarch64/tuning_models/ampere1b.h: Likewise.
6122 2024-01-02 Feng Wang <wangfeng@eswincomputing.com>
6124 * config/riscv/vector-crypto.md: Modify copyright year.
6126 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6128 * config/riscv/riscv-vector-costs.cc: Move STMT_VINFO_TYPE (...) to local.
6130 2024-01-02 Lulu Cheng <chenglulu@loongson.cn>
6132 * config.in: Regenerate.
6133 * config/loongarch/loongarch-opts.h (HAVE_AS_TLS_LE_RELAXATION): Define.
6134 * config/loongarch/loongarch.cc (loongarch_legitimize_tls_address):
6135 Added TLS Le Relax support.
6136 (loongarch_print_operand_reloc): Add the output string of TLS Le Relax.
6137 * config/loongarch/loongarch.md (@add_tls_le_relax<mode>): New template.
6138 * configure: Regenerate.
6139 * configure.ac: Check if binutils supports TLS le relax.
6141 2024-01-02 Feng Wang <wangfeng@eswincomputing.com>
6143 * config/riscv/iterators.md: Add rotate insn name.
6144 * config/riscv/riscv.md: Add new insns name for crypto vector.
6145 * config/riscv/vector-iterators.md: Add new iterators for crypto vector.
6146 * config/riscv/vector.md: Add the corresponding attr for crypto vector.
6147 * config/riscv/vector-crypto.md: New file.The machine descriptions for crypto vector.
6149 2024-01-02 Juzhe-Zhong <juzhe.zhong@rivai.ai>
6152 * config/riscv/riscv-vector-costs.cc (compute_nregs_for_mode): Fix
6153 pointer type liveness count.
6155 Copyright (C) 2024 Free Software Foundation, Inc.
6157 Copying and distribution of this file, with or without modification,
6158 are permitted in any medium without royalty provided the copyright
6159 notice and this notice are preserved.