1 2022-10-31 Guillermo E. Martinez <guillermo.e.martinez@oracle.com>
3 * btfout.cc (btf_calc_num_vbytes): Compute enumeration size depending of
4 enumerator type btf_enum{,64}.
5 (btf_asm_type): Update btf_kflag according to enumeration type sign
6 using dtd_enum_unsigned field for both: BTF_KIND_ENUM{,64}.
7 (btf_asm_enum_const): New argument to represent the size of
8 the BTF enum type, writing the enumerator constant value for
9 32 bits, if it's 64 bits then explicitly writes lower 32-bits
10 value and higher 32-bits value.
11 (output_asm_btf_enum_list): Add enumeration size argument.
12 * ctfc.cc (ctf_add_enum): New argument to represent CTF enum
14 (ctf_add_generic): Use of ei_{name. size, unsigned} to build the
15 dtd structure containing enumeration information.
16 (ctf_add_enumerator): Update comment mention support for BTF
17 enumeration in 64-bits.
18 * dwarf2ctf.cc (gen_ctf_enumeration_type): Extract signedness
19 for enumeration type and use it in ctf_add_enum.
20 * ctfc.h (ctf_dmdef): Update dmd_value to HOST_WIDE_INT to allow
21 use 32/64 bits enumerators.
23 (ctf_dtdef): New field to describe enum signedness.
25 2022-10-31 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
27 * config/riscv/riscv-vector-builtins-bases.cc: Change constexpr back to CONSTEXPR.
28 * config/riscv/riscv-vector-builtins-shapes.cc (SHAPE): Ditto.
29 * config/riscv/riscv-vector-builtins.cc (struct registered_function_hasher): Ditto.
30 * config/riscv/riscv-vector-builtins.h (struct rvv_arg_type_info): Ditto.
32 2022-10-31 Andrew Stubbs <ams@codesourcery.com>
34 * config/gcn/gcn-valu.md (fminmaxop): New iterator.
35 (<fexpander><mode>3): New define_expand.
36 (<fexpander><mode>3<exec>): Likewise.
37 (reduc_<fexpander>_scal_<mode>): Likewise.
38 * config/gcn/gcn.md (fexpander): New attribute.
40 2022-10-31 Andrew Stubbs <ams@codesourcery.com>
42 * config/gcn/gcn-valu.md (V64_SI): Delete iterator.
45 (V64_INT_1REG): Likewise.
49 (reduc_<reduc_op>_scal_<mode>): Use V_ALL. Use gen_vec_extract.
50 (fold_left_plus_<mode>): Use V_FP.
51 (*<reduc_op>_dpp_shr_<mode>): Use V_1REG.
52 (*<reduc_op>_dpp_shr_<mode>): Use V_DI.
53 (*plus_carry_dpp_shr_<mode>): Use V_INT_1REG.
54 (*plus_carry_in_dpp_shr_<mode>): Use V_SI.
55 (*plus_carry_dpp_shr_<mode>): Use V_DI.
56 (mov_from_lane63_<mode>): Delete.
57 (mov_from_lane63_<mode>): Delete.
58 * config/gcn/gcn.cc (gcn_expand_reduc_scalar): Support partial vectors.
59 * config/gcn/gcn.md (unspec): Remove UNSPEC_MOV_FROM_LANE63.
61 2022-10-31 Andrew Stubbs <ams@codesourcery.com>
63 * config/gcn/gcn.cc (gcn_simd_clone_compute_vecsize_and_simdlen):
64 Set base_type as ARG_UNUSED.
66 2022-10-31 Jakub Jelinek <jakub@redhat.com>
68 * builtin-types.def (BT_COMPLEX_FLOAT16, BT_COMPLEX_FLOAT32,
69 BT_COMPLEX_FLOAT64, BT_COMPLEX_FLOAT128, BT_COMPLEX_FLOAT32X,
70 BT_COMPLEX_FLOAT64X, BT_COMPLEX_FLOAT128X,
71 BT_FN_COMPLEX_FLOAT16_COMPLEX_FLOAT16,
72 BT_FN_COMPLEX_FLOAT32_COMPLEX_FLOAT32,
73 BT_FN_COMPLEX_FLOAT64_COMPLEX_FLOAT64,
74 BT_FN_COMPLEX_FLOAT128_COMPLEX_FLOAT128,
75 BT_FN_COMPLEX_FLOAT32X_COMPLEX_FLOAT32X,
76 BT_FN_COMPLEX_FLOAT64X_COMPLEX_FLOAT64X,
77 BT_FN_COMPLEX_FLOAT128X_COMPLEX_FLOAT128X,
78 BT_FN_FLOAT16_COMPLEX_FLOAT16, BT_FN_FLOAT32_COMPLEX_FLOAT32,
79 BT_FN_FLOAT64_COMPLEX_FLOAT64, BT_FN_FLOAT128_COMPLEX_FLOAT128,
80 BT_FN_FLOAT32X_COMPLEX_FLOAT32X, BT_FN_FLOAT64X_COMPLEX_FLOAT64X,
81 BT_FN_FLOAT128X_COMPLEX_FLOAT128X,
82 BT_FN_COMPLEX_FLOAT16_COMPLEX_FLOAT16_COMPLEX_FLOAT16,
83 BT_FN_COMPLEX_FLOAT32_COMPLEX_FLOAT32_COMPLEX_FLOAT32,
84 BT_FN_COMPLEX_FLOAT64_COMPLEX_FLOAT64_COMPLEX_FLOAT64,
85 BT_FN_COMPLEX_FLOAT128_COMPLEX_FLOAT128_COMPLEX_FLOAT128,
86 BT_FN_COMPLEX_FLOAT32X_COMPLEX_FLOAT32X_COMPLEX_FLOAT32X,
87 BT_FN_COMPLEX_FLOAT64X_COMPLEX_FLOAT64X_COMPLEX_FLOAT64X,
88 BT_FN_COMPLEX_FLOAT128X_COMPLEX_FLOAT128X_COMPLEX_FLOAT128X): New.
89 * builtins.def (CABS_TYPE, CACOSH_TYPE, CARG_TYPE, CASINH_TYPE,
90 CPOW_TYPE, CPROJ_TYPE): Define and undefine later.
91 (BUILT_IN_CABS, BUILT_IN_CACOSH, BUILT_IN_CACOS, BUILT_IN_CARG,
92 BUILT_IN_CASINH, BUILT_IN_CASIN, BUILT_IN_CATANH, BUILT_IN_CATAN,
93 BUILT_IN_CCOSH, BUILT_IN_CCOS, BUILT_IN_CEXP, BUILT_IN_CLOG,
94 BUILT_IN_CPOW, BUILT_IN_CPROJ, BUILT_IN_CSINH, BUILT_IN_CSIN,
95 BUILT_IN_CSQRT, BUILT_IN_CTANH, BUILT_IN_CTAN): Add
96 DEF_EXT_LIB_FLOATN_NX_BUILTINS.
97 * fold-const-call.cc (fold_const_call_sc, fold_const_call_cc,
98 fold_const_call_ccc): Add various CASE_CFN_*_FN: cases when
99 CASE_CFN_* is present.
100 * gimple-ssa-backprop.cc (backprop::process_builtin_call_use):
102 * builtins.cc (expand_builtin, fold_builtin_1): Likewise.
103 * fold-const.cc (negate_mathfn_p, tree_expr_finite_p,
104 tree_expr_maybe_signaling_nan_p, tree_expr_maybe_nan_p,
105 tree_expr_maybe_real_minus_zero_p, tree_call_nonnegative_warnv_p):
108 2022-10-31 Jakub Jelinek <jakub@redhat.com>
110 * builtin-types.def (BT_FN_BFLOAT16_BFLOAT16_BFLOAT16): New.
111 * builtins.def (BUILT_IN_NEXTAFTERF16B): New builtin.
112 * fold-const-call.cc (fold_const_call_sss): Handle
113 CFN_BUILT_IN_NEXTAFTERF16B.
115 2022-10-31 Jakub Jelinek <jakub@redhat.com>
117 * builtin-types.def (BT_FLOAT16_PTR, BT_FLOAT32_PTR, BT_FLOAT64_PTR,
118 BT_FLOAT128_PTR, BT_FLOAT32X_PTR, BT_FLOAT64X_PTR, BT_FLOAT128X_PTR):
119 New DEF_PRIMITIVE_TYPE.
120 (BT_FN_INT_FLOAT16, BT_FN_INT_FLOAT32, BT_FN_INT_FLOAT64,
121 BT_FN_INT_FLOAT128, BT_FN_INT_FLOAT32X, BT_FN_INT_FLOAT64X,
122 BT_FN_INT_FLOAT128X, BT_FN_LONG_FLOAT16, BT_FN_LONG_FLOAT32,
123 BT_FN_LONG_FLOAT64, BT_FN_LONG_FLOAT128, BT_FN_LONG_FLOAT32X,
124 BT_FN_LONG_FLOAT64X, BT_FN_LONG_FLOAT128X, BT_FN_LONGLONG_FLOAT16,
125 BT_FN_LONGLONG_FLOAT32, BT_FN_LONGLONG_FLOAT64,
126 BT_FN_LONGLONG_FLOAT128, BT_FN_LONGLONG_FLOAT32X,
127 BT_FN_LONGLONG_FLOAT64X, BT_FN_LONGLONG_FLOAT128X): New
129 (BT_FN_FLOAT16_FLOAT16_FLOAT16PTR, BT_FN_FLOAT32_FLOAT32_FLOAT32PTR,
130 BT_FN_FLOAT64_FLOAT64_FLOAT64PTR, BT_FN_FLOAT128_FLOAT128_FLOAT128PTR,
131 BT_FN_FLOAT32X_FLOAT32X_FLOAT32XPTR,
132 BT_FN_FLOAT64X_FLOAT64X_FLOAT64XPTR,
133 BT_FN_FLOAT128X_FLOAT128X_FLOAT128XPTR, BT_FN_FLOAT16_FLOAT16_INT,
134 BT_FN_FLOAT32_FLOAT32_INT, BT_FN_FLOAT64_FLOAT64_INT,
135 BT_FN_FLOAT128_FLOAT128_INT, BT_FN_FLOAT32X_FLOAT32X_INT,
136 BT_FN_FLOAT64X_FLOAT64X_INT, BT_FN_FLOAT128X_FLOAT128X_INT,
137 BT_FN_FLOAT16_FLOAT16_INTPTR, BT_FN_FLOAT32_FLOAT32_INTPTR,
138 BT_FN_FLOAT64_FLOAT64_INTPTR, BT_FN_FLOAT128_FLOAT128_INTPTR,
139 BT_FN_FLOAT32X_FLOAT32X_INTPTR, BT_FN_FLOAT64X_FLOAT64X_INTPTR,
140 BT_FN_FLOAT128X_FLOAT128X_INTPTR, BT_FN_FLOAT16_FLOAT16_LONG,
141 BT_FN_FLOAT32_FLOAT32_LONG, BT_FN_FLOAT64_FLOAT64_LONG,
142 BT_FN_FLOAT128_FLOAT128_LONG, BT_FN_FLOAT32X_FLOAT32X_LONG,
143 BT_FN_FLOAT64X_FLOAT64X_LONG, BT_FN_FLOAT128X_FLOAT128X_LONG): New
145 (BT_FN_FLOAT16_FLOAT16_FLOAT16_INTPTR,
146 BT_FN_FLOAT32_FLOAT32_FLOAT32_INTPTR,
147 BT_FN_FLOAT64_FLOAT64_FLOAT64_INTPTR,
148 BT_FN_FLOAT128_FLOAT128_FLOAT128_INTPTR,
149 BT_FN_FLOAT32X_FLOAT32X_FLOAT32X_INTPTR,
150 BT_FN_FLOAT64X_FLOAT64X_FLOAT64X_INTPTR,
151 BT_FN_FLOAT128X_FLOAT128X_FLOAT128X_INTPTR): New DEF_FUNCTION_TYPE_3.
152 * builtins.def (ACOSH_TYPE, ATAN2_TYPE, ATANH_TYPE, COSH_TYPE,
153 FDIM_TYPE, HUGE_VAL_TYPE, HYPOT_TYPE, ILOGB_TYPE, LDEXP_TYPE,
154 LGAMMA_TYPE, LLRINT_TYPE, LOG10_TYPE, LRINT_TYPE, MODF_TYPE,
155 NEXTAFTER_TYPE, REMQUO_TYPE, SCALBLN_TYPE, SCALBN_TYPE, SINH_TYPE):
156 Define and undefine later.
157 (FMIN_TYPE, SQRT_TYPE): Undefine at a later line.
158 (INF_TYPE): Define at a later line.
159 (BUILT_IN_ACOSH, BUILT_IN_ACOS, BUILT_IN_ASINH, BUILT_IN_ASIN,
160 BUILT_IN_ATAN2, BUILT_IN_ATANH, BUILT_IN_ATAN, BUILT_IN_CBRT,
161 BUILT_IN_COSH, BUILT_IN_COS, BUILT_IN_ERFC, BUILT_IN_ERF,
162 BUILT_IN_EXP2, BUILT_IN_EXP, BUILT_IN_EXPM1, BUILT_IN_FDIM,
163 BUILT_IN_FMOD, BUILT_IN_FREXP, BUILT_IN_HYPOT, BUILT_IN_ILOGB,
164 BUILT_IN_LDEXP, BUILT_IN_LGAMMA, BUILT_IN_LLRINT, BUILT_IN_LLROUND,
165 BUILT_IN_LOG10, BUILT_IN_LOG1P, BUILT_IN_LOG2, BUILT_IN_LOGB,
166 BUILT_IN_LOG, BUILT_IN_LRINT, BUILT_IN_LROUND, BUILT_IN_MODF,
167 BUILT_IN_NEXTAFTER, BUILT_IN_POW, BUILT_IN_REMAINDER, BUILT_IN_REMQUO,
168 BUILT_IN_SCALBLN, BUILT_IN_SCALBN, BUILT_IN_SINH, BUILT_IN_SIN,
169 BUILT_IN_TANH, BUILT_IN_TAN, BUILT_IN_TGAMMA): Add
170 DEF_EXT_LIB_FLOATN_NX_BUILTINS.
171 (BUILT_IN_HUGE_VAL): Use HUGE_VAL_TYPE instead of INF_TYPE in
172 DEF_GCC_FLOATN_NX_BUILTINS.
173 * fold-const-call.cc (fold_const_call_ss): Add various CASE_CFN_*_FN:
174 cases when CASE_CFN_* is present.
175 (fold_const_call_sss): Likewise.
176 * builtins.cc (mathfn_built_in_2): Use CASE_MATHFN_FLOATN instead of
177 CASE_MATHFN for various builtins in SEQ_OF_CASE_MATHFN macro.
178 (builtin_with_linkage_p): Add CASE_FLT_FN_FLOATN_NX for various
179 builtins next to CASE_FLT_FN.
180 * fold-const.cc (tree_call_nonnegative_warnv_p): Add CASE_CFN_*_FN:
181 next to CASE_CFN_*: for various builtins.
182 * tree-call-cdce.cc (can_test_argument_range): Add
183 CASE_FLT_FN_FLOATN_NX next to CASE_FLT_FN for various builtins.
184 (edom_only_function): Likewise.
186 2022-10-31 konglin1 <lingling.kong@intel.com>
188 * common/config/i386/i386-common.cc
189 (OPTION_MASK_ISA2_AVXNECONVERT_SET,
190 OPTION_MASK_ISA2_AVXNECONVERT_UNSET): New.
191 (ix86_handle_option): Handle -mavxneconvert, unset
192 avxneconvert when avx2 is disabled.
193 * common/config/i386/i386-cpuinfo.h (processor_types): Add
194 FEATURE_AVXNECONVERT.
195 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
197 * common/config/i386/cpuinfo.h (get_available_features):
199 * config.gcc: Add avxneconvertintrin.h
200 * config/i386/avxneconvertintrin.h: New.
201 * config/i386/avx512bf16vlintrin.h (_mm256_cvtneps_pbh):
202 Unified builtin with avxneconvert.
203 (_mm_cvtneps_pbh): Ditto.
204 * config/i386/cpuid.h (bit_AVXNECONVERT): New.
205 * config/i386/i386-builtin-types.def: Add
206 DEF_POINTER_TYPE (PCV8HF, V8HF, CONST),
207 DEF_POINTER_TYPE (PCV8BF, V8BF, CONST),
208 DEF_POINTER_TYPE (PCV16HF, V16HF, CONST),
209 DEF_POINTER_TYPE (PCV16BF, V16BF, CONST),
210 DEF_FUNCTION_TYPE (V4SF, PCBFLOAT16),
211 DEF_FUNCTION_TYPE (V4SF, PCFLOAT16),
212 DEF_FUNCTION_TYPE (V8SF, PCBFLOAT16),
213 DEF_FUNCTION_TYPE (V8SF, PCFLOAT16),
214 DEF_FUNCTION_TYPE (V4SF, PCV8BF),
215 DEF_FUNCTION_TYPE (V4SF, PCV8HF),
216 DEF_FUNCTION_TYPE (V8SF, PCV16HF),
217 DEF_FUNCTION_TYPE (V8SF, PCV16BF),
218 * config/i386/i386-builtin.def: Add new builtins.
219 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
221 * config/i386/i386-expand.cc (ix86_expand_special_args_builtin):
222 Handle V4SF_FTYPE_PCBFLOAT16,V8SF_FTYPE_PCBFLOAT16, V4SF_FTYPE_PCFLOAT16,
223 V8SF_FTYPE_PCFLOAT16,V4SF_FTYPE_PCV8BF,
224 V4SF_FTYPE_PCV8HF,V8SF_FTYPE_PCV16BF,V8SF_FTYPE_PCV16HF.
225 * config/i386/i386-isa.def : Add DEF_PTA(AVXNECONVERT) New.
226 * config/i386/i386-options.cc (isa2_opts): Add -mavxneconvert.
227 (ix86_valid_target_attribute_inner_p): Handle avxneconvert.
228 * config/i386/i386.md: Add attr avx512bf16vl and avxneconvert.
229 * config/i386/i386.opt: Add option -mavxneconvert.
230 * config/i386/immintrin.h: Inculde avxneconvertintrin.h.
231 * config/i386/sse.md (vbcstnebf162ps_<mode>): New define_insn.
232 (vbcstnesh2ps_<mode>): Ditto.
233 (vcvtnee<bf16_ph>2ps_<mode>):Ditto.
234 (vcvtneo<bf16_ph>2ps_<mode>):Ditto.
235 (vcvtneps2bf16_v4sf): Ditto.
236 (*vcvtneps2bf16_v4sf): Ditto.
237 (vcvtneps2bf16_v8sf): Ditto.
238 * doc/invoke.texi: Document -mavxneconvert.
239 * doc/extend.texi: Document avxneconvert.
240 * doc/sourcebuild.texi: Document target avxneconvert.
242 2022-10-31 konglin1 <lingling.kong@intel.com>
244 * config/i386/avx512bf16intrin.h (__attribute__): Change short to bf16.
245 (_mm_cvtsbh_ss): Ditto.
246 (_mm512_cvtne2ps_pbh): Ditto.
247 (_mm512_mask_cvtne2ps_pbh): Ditto.
248 (_mm512_maskz_cvtne2ps_pbh): Ditto.
249 * config/i386/avx512bf16vlintrin.h (__attribute__): Ditto.
250 (_mm256_cvtne2ps_pbh): Ditto.
251 (_mm256_mask_cvtne2ps_pbh): Ditto.
252 (_mm256_maskz_cvtne2ps_pbh): Ditto.
253 (_mm_cvtne2ps_pbh): Ditto.
254 (_mm_mask_cvtne2ps_pbh): Ditto.
255 (_mm_maskz_cvtne2ps_pbh): Ditto.
256 (_mm_cvtness_sbh): Ditto.
257 * config/i386/i386-builtin-types.def (V8BF): Add new
258 DEF_VECTOR_TYPE for BFmode.
261 * config/i386/i386-builtin.def (BDESC): Fixed builtins.
262 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Changed
263 avx512bf16 ix86_builtin_func_type included HI to BF.
264 * config/i386/immintrin.h: Add SSE2 depend for avx512bf16.
265 * config/i386/sse.md (TARGET_AVX512VL): Changed HI vector to BF
267 (avx512f_cvtneps2bf16_v4sf): New define_expand.
268 (*avx512f_cvtneps2bf16_v4sf): New define_insn.
269 (avx512f_cvtneps2bf16_v4sf_maskz):Ditto.
270 (avx512f_cvtneps2bf16_v4sf_mask): Ditto.
271 (avx512f_cvtneps2bf16_v4sf_mask_1): Ditto.
273 2022-10-31 liuhongt <hongtao.liu@intel.com>
276 * config/i386/i386-modes.def (VECTOR_MODE): Support V2BFmode.
277 * config/i386/i386.cc (classify_argument): Handle V4BFmode and
279 (ix86_convert_const_vector_to_integer): Ditto.
280 * config/i386/i386.h (VALID_AVX512FP16_REG_MODE): Remove
282 (VALID_SSE2_REG_MODE): Add V4BFmode and V2BFmode.
283 (VALID_MMX_REG_MODE): Add V4BFmode.
284 * config/i386/i386.md (mode): Add V4BF and V2BF.
286 * config/i386/mmx.md (MMXMODE) Add V4BF.
288 (V_16_32_64): Add V4BF and V2BF.
289 (mmxinsnmode): Add V4BF and V2BF.
290 (*mov<mode>_internal): Hanlde V4BFmode and V2BFmode.
292 2022-10-29 Eric Botcazou <ebotcazou@adacore.com>
294 * configure.ac (sjlj-exceptions): Restore dropped line.
295 * configure: Regenerate.
297 2022-10-29 Eric Botcazou <ebotcazou@adacore.com>
299 * alias.cc (init_alias_analysis): Do not record sets to the hard
300 frame pointer if the frame pointer has not been eliminated.
302 2022-10-29 Iain Buclaw <ibuclaw@gdcproject.org>
304 * config/darwin-d.cc (TARGET_D_MINFO_START_NAME): Rename to ...
305 (TARGET_D_MINFO_SECTION_START): ...this.
306 (TARGET_D_MINFO_END_NAME): Rename to ...
307 (TARGET_D_MINFO_SECTION_END): ... this.
308 * config/elfos.h (TARGET_D_MINFO_SECTION): Remove.
309 (TARGET_D_MINFO_START_NAME): Remove.
310 (TARGET_D_MINFO_END_NAME): Remove.
311 * config/i386/cygwin-d.cc (TARGET_D_MINFO_SECTION): Remove.
312 (TARGET_D_MINFO_START_NAME): Remove.
313 (TARGET_D_MINFO_END_NAME): Remove.
314 * config/i386/winnt-d.cc (TARGET_D_MINFO_SECTION): Remove.
315 (TARGET_D_MINFO_START_NAME): Remove.
316 (TARGET_D_MINFO_END_NAME): Remove.
317 * doc/tm.texi: Regenerate.
318 * doc/tm.texi.in (TARGET_D_MINFO_START_NAME): Rename to ...
319 (TARGET_D_MINFO_SECTION_START): ...this.
320 (TARGET_D_MINFO_END_NAME): Rename to ...
321 (TARGET_D_MINFO_SECTION_END): ...this.
323 2022-10-29 Iain Buclaw <ibuclaw@gdcproject.org>
325 * config.gcc: Split out glibc-d.o into linux-d.o, kfreebsd-d.o,
326 kopensolaris-d.o, and gnu-d.o. Split out cygwin-d.o from winnt-d.o.
327 * config/arm/linux-eabi.h (EXTRA_TARGET_D_OS_VERSIONS): Remove.
328 * config/gnu.h (GNU_USER_TARGET_D_OS_VERSIONS): Remove.
329 * config/i386/cygwin.h (EXTRA_TARGET_D_OS_VERSIONS): Remove.
330 * config/i386/linux-common.h (EXTRA_TARGET_D_OS_VERSIONS): Remove.
331 * config/i386/mingw32.h (EXTRA_TARGET_D_OS_VERSIONS): Remove.
332 * config/i386/t-cygming: Add cygwin-d.o.
333 * config/i386/winnt-d.cc (winnt_d_os_builtins): Only add
334 MinGW-specific version condition.
335 * config/kfreebsd-gnu.h (GNU_USER_TARGET_D_OS_VERSIONS): Remove.
336 * config/kopensolaris-gnu.h (GNU_USER_TARGET_D_OS_VERSIONS): Remove.
337 * config/linux-android.h (ANDROID_TARGET_D_OS_VERSIONS): Remove.
338 * config/linux.h (GNU_USER_TARGET_D_OS_VERSIONS): Remove.
339 * config/mips/linux-common.h (EXTRA_TARGET_D_OS_VERSIONS): Remove.
340 * config/t-glibc: Remove glibc-d.o, add gnu-d.o, kfreebsd-d.o,
342 * config/t-linux: Add linux-d.o.
343 * config/glibc-d.cc: Remove file.
344 * config/gnu-d.cc: New file.
345 * config/i386/cygwin-d.cc: New file.
346 * config/kfreebsd-d.cc: New file.
347 * config/kopensolaris-d.cc: New file.
348 * config/linux-d.cc: New file.
350 2022-10-29 Jeff Law <jeffreyalaw@gmail.com>
352 * config/h8300/h8300.cc (pre_incdec_with_reg): Make reg argument
354 * config/h8300/h8300-protos.h (pre_incdec_with_reg): Adjust prototype.
356 2022-10-28 Joseph Myers <joseph@codesourcery.com>
358 * config/aarch64/aarch64.cc (aarch64_setup_incoming_varargs):
359 Check TYPE_NO_NAMED_ARGS_STDARG_P.
360 * config/alpha/alpha.cc (alpha_setup_incoming_varargs): Likewise.
361 * config/arc/arc.cc (arc_setup_incoming_varargs): Likewise.
362 * config/arm/arm.cc (arm_setup_incoming_varargs): Likewise.
363 * config/csky/csky.cc (csky_setup_incoming_varargs): Likewise.
364 * config/epiphany/epiphany.cc (epiphany_setup_incoming_varargs):
366 * config/fr30/fr30.cc (fr30_setup_incoming_varargs): Likewise.
367 * config/frv/frv.cc (frv_setup_incoming_varargs): Likewise.
368 * config/ft32/ft32.cc (ft32_setup_incoming_varargs): Likewise.
369 * config/i386/i386.cc (ix86_setup_incoming_varargs): Likewise.
370 * config/ia64/ia64.cc (ia64_setup_incoming_varargs): Likewise.
371 * config/loongarch/loongarch.cc
372 (loongarch_setup_incoming_varargs): Likewise.
373 * config/m32r/m32r.cc (m32r_setup_incoming_varargs): Likewise.
374 * config/mcore/mcore.cc (mcore_setup_incoming_varargs): Likewise.
375 * config/mips/mips.cc (mips_setup_incoming_varargs): Likewise.
376 * config/mmix/mmix.cc (mmix_setup_incoming_varargs): Likewise.
377 * config/nds32/nds32.cc (nds32_setup_incoming_varargs): Likewise.
378 * config/nios2/nios2.cc (nios2_setup_incoming_varargs): Likewise.
379 * config/riscv/riscv.cc (riscv_setup_incoming_varargs): Likewise.
380 * config/rs6000/rs6000-call.cc (setup_incoming_varargs): Likewise.
381 * config/sh/sh.cc (sh_setup_incoming_varargs): Likewise.
382 * config/visium/visium.cc (visium_setup_incoming_varargs):
384 * config/vms/vms-c.cc (vms_c_common_override_options): Do not set
385 flag_allow_parameterless_variadic_functions.
386 * doc/invoke.texi (-fallow-parameterless-variadic-functions): Do
388 * function.cc (assign_parms): Call assign_parms_setup_varargs for
389 TYPE_NO_NAMED_ARGS_STDARG_P case.
390 * ginclude/stdarg.h [__STDC_VERSION__ > 201710L] (va_start): Make
391 variadic macro. Pass second argument of 0 to __builtin_va_start.
392 * target.def (setup_incoming_varargs): Update documentation.
393 * doc/tm.texi: Regenerate.
394 * tree-core.h (struct tree_type_common): Add
395 no_named_args_stdarg_p.
396 * tree-streamer-in.cc (unpack_ts_type_common_value_fields): Unpack
397 TYPE_NO_NAMED_ARGS_STDARG_P.
398 * tree-streamer-out.cc (pack_ts_type_common_value_fields): Pack
399 TYPE_NO_NAMED_ARGS_STDARG_P.
400 * tree.cc (type_cache_hasher::equal): Compare
401 TYPE_NO_NAMED_ARGS_STDARG_P.
402 (build_function_type): Add argument no_named_args_stdarg_p.
403 (build_function_type_list_1, build_function_type_array_1)
404 (reconstruct_complex_type): Update calls to build_function_type.
405 (stdarg_p, prototype_p): Return true for (...) functions.
406 (gimple_canonical_types_compatible_p): Compare
407 TYPE_NO_NAMED_ARGS_STDARG_P.
408 * tree.h (TYPE_NO_NAMED_ARGS_STDARG_P): New.
409 (build_function_type): Update prototype.
411 2022-10-28 Andre Vieira <andre.simoesdiasvieira@arm.com>
413 PR tree-optimization/107346
414 * tree-vect-data-refs.cc (vect_check_gather_scatter): Reject offsets
415 that aren't multiples of BITS_PER_UNIT.
417 2022-10-28 Richard Biener <rguenther@suse.de>
419 PR tree-optimization/107407
420 * tree-ssa-dse.cc (dse_classify_store): Perform backedge
421 varying index check when collecting PHI uses rather than
422 after optimizing processing of the candidate defs.
424 2022-10-28 Richard Biener <rguenther@suse.de>
426 PR tree-optimization/107447
427 * tree-ssa-loop-im.cc (determine_max_movement): Do not
428 hoist returns-twice calls.
430 2022-10-28 Richard Biener <rguenther@suse.de>
432 PR tree-optimization/107435
433 * tree-vect-loop.cc (vectorizable_recurr): Convert initial
434 value to vector component type.
436 2022-10-28 Julian Brown <julian@codesourcery.com>
437 Thomas Schwinge <thomas@codesourcery.com>
440 * omp-low.cc (oacc_privatization_candidate_p): Artificial vars are not
441 privatization candidates.
443 2022-10-28 Martin Liska <mliska@suse.cz>
446 * doc/invoke.texi: Document sanitizers can trigger warnings.
448 2022-10-28 Thomas Schwinge <thomas@codesourcery.com>
450 * doc/makefile.texi (Makefile Targets): Document
451 'distclean-stage[N]'.
453 2022-10-28 Aldy Hernandez <aldyh@redhat.com>
455 * value-range.cc (range_tests_floats): Use HONOR_INFINITIES.
457 2022-10-27 Eric Botcazou <ebotcazou@adacore.com>
459 * config/aarch64/aarch64.h (DONT_USE_BUILTIN_SETJMP): Delete.
461 2022-10-27 H.J. Lu <hjl.tools@gmail.com>
464 * config/i386/i386.md (UNSPEC_CC_NE): New.
465 Replace ne:CCC/ne:CCO with UNSPEC_CC_NE in neg patterns.
467 2022-10-27 Andrew Pinski <apinski@marvell.com>
469 * tree-ssa-phiopt.cc: Include tree-ssa-dce.h
470 (replace_phi_edge_with_variable):
471 New argument, dce_ssa_names. Call simple_dce_from_worklist.
472 (match_simplify_replacement): If we inserted a sequence,
473 mark the lhs of the new sequence to be possible dce.
474 Always move the statement and mark the lhs (if it is a name)
475 as possible to remove.
477 2022-10-27 Richard Sandiford <richard.sandiford@arm.com>
479 * config/aarch64/aarch64-protos.h: Replace constexpr with
481 * config/aarch64/aarch64-sve-builtins-base.cc: Likewise.
482 * config/aarch64/aarch64-sve-builtins-functions.h: Likewise.
483 * config/aarch64/aarch64-sve-builtins-shapes.cc: Likewise.
484 * config/aarch64/aarch64-sve-builtins-sve2.cc: Likewise.
485 * config/aarch64/aarch64-sve-builtins.cc: Likewise.
486 * config/aarch64/aarch64.cc: Likewise.
487 * config/aarch64/driver-aarch64.cc: Likewise
489 2022-10-27 Aldy Hernandez <aldyh@redhat.com>
491 PR tree-optimization/107394
492 * value-range-storage.cc (frange_storage_slot::get_frange): Use
495 2022-10-27 Thomas Schwinge <thomas@codesourcery.com>
497 * optc-save-gen.awk: Clarify 'Init' option property usage for
498 streaming optimization.
500 2022-10-27 Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
501 Yvan ROUX <yvan.roux@foss.st.com>
503 * ira.cc: Resize array after reg number increased.
505 2022-10-27 Jiawei <jiawei@iscas.ac.cn>
506 Sinan Lin <sinan@isrc.iscas.ac.cn>
508 * config/riscv/constraints.md (TARGET_ZFINX ? GR_REGS): Set GPRS
509 use while Zfinx is enable.
510 * config/riscv/riscv.cc (riscv_hard_regno_mode_ok): Limit odd
511 registers use when Zdinx enable in RV32 cases.
512 (riscv_option_override): New target enable MASK_FDIV.
513 (riscv_libgcc_floating_mode_supported_p): New error info when
514 use incompatible arch&abi.
515 (riscv_excess_precision): New target enable FLOAT16.
517 2022-10-27 Jiawei <jiawei@iscas.ac.cn>
519 * config/riscv/iterators.md (TARGET_ZFINX):New target.
520 (TARGET_ZDINX): Ditto.
521 (TARGET_ZHINX): Ditto.
522 * config/riscv/riscv-builtins.cc (AVAIL): Ditto.
523 (riscv_atomic_assign_expand_fenv): Ditto.
524 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins): Ditto.
525 * config/riscv/riscv.md: Ditto.
527 2022-10-27 Jiawei <jiawei@iscas.ac.cn>
528 Sinan Lin <sinan@isrc.iscas.ac.cn>
530 * common/config/riscv/riscv-common.cc: New extensions.
531 * config/riscv/arch-canonicalize: New imply relations.
532 * config/riscv/riscv-opts.h (MASK_ZFINX): New mask.
535 (MASK_ZHINXMIN): Ditto.
536 (TARGET_ZFINX): New target.
537 (TARGET_ZDINX): Ditto.
538 (TARGET_ZHINX): Ditto.
539 (TARGET_ZHINXMIN): Ditto.
540 * config/riscv/riscv.opt: New target variable.
542 2022-10-26 David Faust <david.faust@oracle.com>
544 * config/bpf/bpf.cc: Support __builtin_preserve_field_info.
545 (enum bpf_builtins): Add new builtin.
546 (bpf_init_builtins): Likewise.
547 (bpf_core_field_info): New function.
548 (bpf_expand_builtin): Accomodate new builtin. Refactor adding new
550 (maybe_make_core_relo): ... here. New function.
551 (bpf_resolve_overloaded_builtin): Accomodate new builtin.
552 (bpf_core_newdecl): Likewise.
553 (bpf_core_walk): Likewise.
554 (bpf_core_is_maybe_aggregate_access): Improve logic.
555 (struct core_walk_data): New.
556 * config/bpf/coreout.cc (bpf_core_reloc_add): Allow adding different
558 * config/bpf/coreout.h: Analogous change.
559 * doc/extend.texi: Document BPF __builtin_preserve_field_info.
561 2022-10-26 Marek Polacek <polacek@redhat.com>
564 * doc/invoke.texi: Document -Wdangling-reference.
566 2022-10-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
568 * config/xtensa/xtensa.md (movdi):
569 Copy operands[0...1] to ops[0...3] and then use the latter before
570 calling xtensa_split_DI_reg_imm() and emitting insns.
572 2022-10-26 Alexander Monakov <amonakov@ispras.ru>
575 * ipa-visibility.cc (function_and_variable_visibility):
576 Conditionally upgrade TLS model instead of asserting.
578 2022-10-26 Andrew MacLeod <amacleod@redhat.com>
580 * gimple-range-fold.cc (fold_using_range::fold_stmt): Check if
581 stmt is non-negative and adjust the range.
583 2022-10-26 Martin Liska <mliska@suse.cz>
585 * common/config/i386/cpuinfo.h (has_cpu_feature): Add comment.
586 (reset_cpu_feature): New.
587 (get_zhaoxin_cpu): Use reset_cpu_feature.
589 2022-10-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
591 * config/riscv/riscv.cc (riscv_expand_epilogue): Fix statement.
593 2022-10-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
596 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Set to minimum size.
597 (ADJUST_NUNITS): Adjust according to -march.
598 (ADJUST_BYTESIZE): Ditto.
599 * config/riscv/riscv-protos.h (riscv_v_ext_enabled_vector_mode_p):
601 (riscv_v_ext_vector_mode_p): Change function implementation.
602 * config/riscv/riscv-vector-builtins.cc (rvv_switcher::rvv_switcher):
603 Change to riscv_v_ext_vector_mode_p.
604 (register_builtin_type): Ditto.
605 * config/riscv/riscv.cc (riscv_v_ext_vector_mode_p): Change to enabled
608 (riscv_v_ext_enabled_vector_mode_p): Remove.
609 (riscv_v_adjust_nunits): New function.
610 (riscv_vector_mode_supported_p): Use riscv_v_ext_vector_mode_p instead.
611 * config/riscv/riscv.h (riscv_v_adjust_nunits): New function.
613 2022-10-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
615 * config.gcc (riscv*): Add riscv-v.o to extra_objs.
616 * config/riscv/constraints.md (vu): New constraint.
620 * config/riscv/predicates.md (vector_length_operand): New.
621 (reg_or_mem_operand): Ditto.
622 (vector_move_operand): Ditto.
623 (vector_mask_operand): Ditto.
624 (vector_merge_operand): Ditto.
625 * config/riscv/riscv-protos.h (riscv_regmode_natural_size) New.
626 (riscv_vector::const_vec_all_same_in_range_p): Ditto.
627 (riscv_vector::legitimize_move): Ditto.
628 (tail_policy): Ditto.
629 (mask_policy): Ditto.
630 * config/riscv/riscv-v.cc: New.
631 * config/riscv/riscv-vector-builtins-bases.cc
632 (vsetvl::expand): Refactor how LMUL encoding.
633 * config/riscv/riscv.cc (riscv_print_operand): Update how LMUL
634 print and mask operand print.
635 (riscv_regmode_natural_size): New.
636 * config/riscv/riscv.h (REGMODE_NATURAL_SIZE): New.
637 * config/riscv/riscv.md (mode): Add vector modes.
638 * config/riscv/t-riscv (riscv-v.o) New.
639 * config/riscv/vector-iterators.md: New.
640 * config/riscv/vector.md (vundefined<mode>): New.
643 (@vsetvl<mode>_no_side_effects): New.
644 (@pred_mov<mode>): New.
646 2022-10-26 Monk Chiang <monk.chiang@sifive.com>
648 * common/config/riscv/riscv-common.cc (riscv_ext_version_table):
649 Add svinval and svnapot extension.
650 (riscv_ext_flag_table): Ditto.
651 * config/riscv/riscv-opts.h (MASK_SVINVAL): New.
652 (MASK_SVNAPOT): Ditto.
653 (TARGET_SVINVAL): Ditto.
654 (TARGET_SVNAPOT): Ditto.
655 * config/riscv/riscv.opt (riscv_sv_subext): New.
657 2022-10-26 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
659 * config/riscv/riscv-modes.def: Adjust table indentation in commnet.
661 2022-10-26 Martin Liska <mliska@suse.cz>
663 * configure: Regenerate.
665 2022-10-26 Aldy Hernandez <aldyh@redhat.com>
667 * value-range.cc (frange::set): Use HONOR_*.
668 (frange::verify_range): Same.
669 * value-range.h (frange_val_min): Same.
670 (frange_val_max): Same.
672 2022-10-26 Jiufu Guo <guojiufu@linux.ibm.com>
675 * config/rs6000/rs6000.cc (rs6000_cannot_force_const_mem): Return true
678 2022-10-26 Kito Cheng <kito.cheng@sifive.com>
680 * common/config/riscv/riscv-common.cc (riscv_ext_version_table):
682 (riscv_supported_std_ext): Ditto.
683 (multi_letter_subset_rank): Remove `h`.
684 (riscv_subset_list::parse_std_ext): Handle `h` as single letter
686 (riscv_subset_list::parse): Ditto.
688 2022-10-25 Eugene Rozenfeld <erozen@microsoft.com>
690 * auto-profile.cc (get_combined_location): Include discriminator in the
691 returned combined location.
692 (read_function_instance): Read discriminators from profiles.
694 2022-10-25 H.J. Lu <hjl.tools@gmail.com>
697 * expr.cc (get_inner_reference): Always use TYPE_MODE for vector
698 field with vector raw mode.
700 2022-10-25 Segher Boessenkool <segher@kernel.crashing.org>
702 * config/rs6000/rs6000.md (CCEITHER): Delete.
705 (isel_<un>signed_<GPR:mode>): Rename to...
706 (isel_<CCANY:mode>_<GPR:mode>): ... this. Adjust.
707 (*isel_reversed_<un>signed_<GPR:mode>): Rename to...
708 (*isel_reversed_<CCANY:mode>_<GPR:mode>): ... this. Adjust.
709 (setbc_<un>signed_<GPR:mode>): Rename to...
710 (setbc_<CCANY:mode>_<GPR:mode>C): ... this. Adjust."
711 (*setbcr_<un>signed_<GPR:mode>): Rename to ...
712 (*setbcr_<CCANY:mode>_<GPR:mode>): ... this. Adjust.
713 (*setnbc_<un>signed_<GPR:mode>): Rename to ...
714 (*setnbc_<CCANY:mode>_<GPR:mode>): ... this. Adjust.
715 (*setnbcr_<un>signed_<GPR:mode>): Rename to ...
716 (*setnbcr_<CCANY:mode>_<GPR:mode>): ... this. Adjust.
717 (eq<mode>3 for GPR): Adjust.
718 (ne<mode>3 for GPR): Adjust.
719 * config/rs6000/rs6000-string.cc (do_isel): Adjust.
720 * config/rs6000/rs6000.cc (rs6000_emit_int_cmove): Adjust.
722 2022-10-25 Richard Biener <rguenther@suse.de>
724 PR tree-optimization/107176
725 PR tree-optimization/66375
726 PR tree-optimization/42512
727 * tree-scalar-evolution.cc (follow_ssa_edge_expr): Revert
728 the PR66375 fix, do not not associate PLUS_EXPR to be able
729 to use tail-recursion.
730 (follow_ssa_edge_binary): Likewise.
731 (interpret_loop_phi): Revert PR42512 fix, do not throw
732 away analyze_evolution_in_loop result after the fact.
733 (follow_ssa_edge_expr): When reaching halting_phi initalize
734 the evolution to the symbolic value of the PHI result.
735 (add_to_evolution_1): When adding the first evolution verify
736 we can handle the expression wrapping the symbolic evolution
737 and replace that in full using the initial condition.
738 (class scev_dfs): New, contains ...
739 (follow_ssa_edge_expr, follow_ssa_edge_binary,
740 follow_ssa_edge_in_condition_phi_branch,
741 follow_ssa_edge_in_condition_phi,
742 follow_ssa_edge_inner_loop_phi,
743 add_to_evolution, add_to_evolution_1): ... these with
744 loop and halting_phi arguments in class data.
745 (scev_dfs::get_ev): New toplevel DFS entry, start with
746 a chrec_dont_know evolution.
747 (analyze_evolution_in_loop): Use scev_dfs.
749 2022-10-25 Eric Botcazou <ebotcazou@adacore.com>
751 * profile.cc (branch_prob): Be prepared for ignored functions with
752 DECL_SOURCE_LOCATION set to UNKNOWN_LOCATION.
754 2022-10-25 Richard Biener <rguenther@suse.de>
756 * tree-scalar-evolution.cc (follow_ssa_edge_expr): Move
757 STRIP_USELESS_TYPE_CONVERSIONS to where it matters.
759 2022-10-25 Tejas Joshi <TejasSanjay.Joshi@amd.com>
761 * common/config/i386/i386-common.cc (processor_alias_table): Use
762 CPU_ZNVER3 for znver4.
763 * config/i386/znver.md: Remove znver4 reservations.
765 2022-10-25 Jakub Jelinek <jakub@redhat.com>
767 * gimplify.cc (gimple_boolify): Fix comment typos, prduce -> produce
770 2022-10-25 Jakub Jelinek <jakub@redhat.com>
772 PR tree-optimization/107368
773 * gimplify.cc (gimplify_call_expr): For complex IFN_ASSUME
774 conditions call gimple_boolify on the condition.
776 2022-10-25 YunQiang Su <yunqiang.su@cipunited.com>
778 * config.gcc: add -with-compact-branches=policy build option.
779 * doc/install.texi: Likewise.
780 * config/mips/mips.h: Likewise.
782 2022-10-25 YunQiang Su <yunqiang.su@cipunited.com>
784 * config/mips/mips.cc (mips_option_override): not trigger error
785 for compact-branches=always for pre-R6.
786 * config/mips/mips.h (TARGET_RTP_PIC): not trigger error for
787 compact-branches=always for pre-R6.
788 (TARGET_CB_NEVER): Likewise.
789 (TARGET_CB_ALWAYS): Likewise.
790 (struct mips_cpu_info): define macros for compact branch policy.
791 * doc/invoke.texi: Document "always" with pre-R6.
793 2022-10-25 Jakub Jelinek <jakub@redhat.com>
795 PR tree-optimization/107369
796 * gimplify.cc (gimplify_call_expr): If seen_error, handle complex
797 IFN_ASSUME the same as for -O0.
799 2022-10-25 YunQiang Su <yunqiang.su@cipunited.com>
801 * configure.ac: AC_DEFINE(ENABLE_MULTIARCH, 1)
802 * configure: Regenerated.
803 * config.in: Regenerated.
804 * config/mips/mips.h: don't define STANDARD_STARTFILE_PREFIX_1
805 if ENABLE_MULTIARCH is defined.
806 * config/mips/t-linux64: define correct multiarch path when
807 multiarch is enabled.
809 2022-10-25 Richard Biener <rguenther@suse.de>
811 PR tree-optimization/100756
812 * tree-ssa-loop-niter.cc (expand_simple_operations): Also
813 expand multiplications by invariants.
815 2022-10-25 Kewen Lin <linkw@linux.ibm.com>
817 PR tree-optimization/107338
818 * tree-vect-patterns.cc (vect_recog_bitfield_ref_pattern): Move
819 shfit_n calculation before the adjustments for widening loads.
821 2022-10-25 Martin Liska <mliska@suse.cz>
823 * common/config/riscv/riscv-common.cc
824 (riscv_get_valid_option_values): Get out of ifdef.
826 2022-10-25 Martin Liska <mliska@suse.cz>
829 * common/config/i386/i386-cpuinfo.h (enum processor_vendor):
830 Fix pedantic warning.
832 2022-10-24 Martin Liska <mliska@suse.cz>
835 * diagnostic-format-sarif.cc
836 (sarif_builder::maybe_make_physical_location_object): Gracefully
837 reject locations with NULL filename.
839 2022-10-24 David Malcolm <dmalcolm@redhat.com>
842 * doc/invoke.texi (Static Analyzer Options): Add "pipe" and
843 "pipe2" to the list of functions the analyzer has hardcoded
846 2022-10-24 Jason Merrill <jason@redhat.com>
848 * tree.h (build_string_literal): New one-argument overloads that
849 take tree (identifier) and const char *.
850 * builtins.cc (fold_builtin_FILE)
851 (fold_builtin_FUNCTION)
852 * gimplify.cc (gimple_add_init_for_auto_var)
853 * vtable-verify.cc (verify_bb_vtables): Simplify calls.
855 2022-10-24 Martin Liska <mliska@suse.cz>
858 * common/config/i386/i386-cpuinfo.h (enum processor_vendor):
859 Reorder enum values as BUILTIN_VENDOR_MAX should not point
860 in the middle of the valid enum values.
862 2022-10-24 Marek Polacek <polacek@redhat.com>
865 * tree.cc (maybe_wrap_with_location): Don't create a location wrapper
866 when the type is erroneous.
868 2022-10-24 Wilco Dijkstra <wdijkstr@arm.com>
871 * config/aarch64/aarch64.cc (aarch64_internal_mov_immediate)
872 Add support for a bitmask immediate with 2 MOVKs.
873 (aarch64_check_bitmask): New function after refactorization.
874 (aarch64_bitmask_imm): Simplify replication of small modes.
875 Split function into 64-bit only version for efficiency.
876 (aarch64_move_imm): Move near other immediate functions.
877 (aarch64_uimm12_shift): Likewise.
878 (aarch64_clamp_to_uimm12_shift): Likewise.
879 (aarch64_movk_shift): Likewise.
880 (aarch64_replicate_bitmask_imm): Likewise.
881 (aarch64_and_split_imm1): Likewise.
882 (aarch64_and_split_imm2): Likewise.
883 (aarch64_and_bitmask_imm): Likewise.
884 (aarch64_movw_imm): Likewise.
886 2022-10-24 Aldy Hernandez <aldyh@redhat.com>
888 PR tree-optimization/107355
889 * range-op-float.cc (foperator_abs::op1_range): Handle NAN.
891 2022-10-24 Tobias Burnus <tobias@codesourcery.com>
894 * omp-expand.cc (expand_omp_target): Set calls_declare_variant_alt
895 in DECL_CONTEXT and not to cfun->decl.
896 * cgraphclones.cc (cgraph_node::create_clone): Copy also the
897 node's calls_declare_variant_alt value.
899 2022-10-24 Kito Cheng <kito.cheng@sifive.com>
901 * common/config/riscv/riscv-common.cc (riscv_tunes): New.
902 (riscv_get_valid_option_values): New.
903 (TARGET_GET_VALID_OPTION_VALUES): New.
904 * config/riscv/riscv-cores.def (RISCV_TUNE): New, define options
906 (RISCV_CORE): Fix comment.
907 * config/riscv/riscv.cc (riscv_tune_info_table): Move definition to
910 2022-10-24 Aldy Hernandez <aldyh@redhat.com>
912 PR tree-optimization/107365
913 * value-range.cc (frange::verify_range): Predicate NAN check in
914 VARYING range on HONOR_NANS instead of flag_finite_math_only.
915 (range_tests_floats): Same.
916 (range_tests_floats_various): New.
917 (range_tests): Call range_tests_floats_various.
919 2022-10-24 Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
920 Yvan ROUX <yvan.roux@foss.st.com>
922 * lto-wrapper.cc: Quote paths in makefile.
924 2022-10-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
926 * config/riscv/riscv.cc (riscv_legitimize_move): Support (set (mem) (const_poly_int)).
928 2022-10-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
930 * config/riscv/riscv-vector-builtins-bases.cc: Replace CONSTEXPR
931 with constexpr throughout.
932 * config/riscv/riscv-vector-builtins-shapes.cc (SHAPE): Likewise.
933 * config/riscv/riscv-vector-builtins.cc
934 (struct registered_function_hasher): Likewise.
935 * config/riscv/riscv-vector-builtins.h (struct rvv_arg_type_info):
938 2022-10-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
940 * config/riscv/riscv-vector-switch.def (ENTRY): Remove unused TI/TF vector modes.
942 2022-10-24 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
944 * config/riscv/riscv.h (REG_CLASS_CONTENTS): Fix ALL_REGS.
946 2022-10-22 Michael Eager <eager@eagercon.com>
948 * config/microblaze/microblaze.cc
949 (microblaze_legitimize_address): Initialize 'reg' to NULL, check for NULL.
950 (microblaze_address_insns): Replace abort() with gcc_unreachable().
951 (print_operand_address): Same.
952 (microblaze_expand_move): Initialize 'p1' to NULL, check for NULL.
953 (get_branch_target): Replace abort() with gcc_unreachable().
955 2022-10-22 Aldy Hernandez <aldyh@redhat.com>
957 * value-range.cc (range_tests_floats): Predicate [-Inf, +Inf] test
958 with !flag_finite_math_only.
960 2022-10-22 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
962 * config/xtensa/xtensa.cc (xtensa_conditional_register_usage):
963 Remove register A0 from FIXED_REGS if the CALL0 ABI.
964 (xtensa_expand_epilogue): Change to emit '(use (reg:SI A0_REG))'
965 unconditionally after restoring callee-saved registers for
966 sibling-call functions, in order to prevent misleading that
967 register A0 is free to use.
969 2022-10-21 Jakub Jelinek <jakub@redhat.com>
972 * config/i386/i386-expand.cc (ix86_prepare_fp_compare_args): For
973 BFmode comparisons promote arguments to SFmode and recurse.
974 (ix86_expand_int_movcc, ix86_expand_fp_movcc): Return false early
975 if comparison operands are BFmode and operands[1] is not
976 ix86_fp_comparison_operator.
978 2022-10-21 Tejas Joshi <TejasSanjay.Joshi@amd.com>
980 * common/config/i386/cpuinfo.h (get_amd_cpu): Recognize znver4.
981 * common/config/i386/i386-common.cc (processor_names): Add znver4.
982 (processor_alias_table): Add znver4 and modularize old znvers.
983 * common/config/i386/i386-cpuinfo.h (processor_subtypes):
985 * config.gcc (x86_64-*-* |...): Likewise.
986 * config/i386/driver-i386.cc (host_detect_local_cpu): Let
987 -march=native recognize znver4 cpus.
988 * config/i386/i386-c.cc (ix86_target_macros_internal): Add znver4.
989 * config/i386/i386-options.cc (m_ZNVER4): New definition.
990 (m_ZNVER): Include m_ZNVER4.
991 (processor_cost_table): Add znver4.
992 * config/i386/i386.cc (ix86_reassociation_width): Likewise.
993 * config/i386/i386.h (processor_type): Add PROCESSOR_ZNVER4.
994 (PTA_ZNVER1): New definition.
995 (PTA_ZNVER2): Likewise.
996 (PTA_ZNVER3): Likewise.
997 (PTA_ZNVER4): Likewise.
998 * config/i386/i386.md (define_attr "cpu"): Add znver4 and rename
1000 * config/i386/x86-tune-costs.h (znver4_cost): New definition.
1001 * config/i386/x86-tune-sched.cc (ix86_issue_rate): Add znver4.
1002 (ix86_adjust_cost): Likewise.
1003 * config/i386/znver1.md: Rename to znver.md.
1004 * config/i386/znver.md: Add new reservations for znver4.
1005 * doc/extend.texi: Add details about znver4.
1006 * doc/invoke.texi: Likewise.
1008 2022-10-21 Richard Biener <rguenther@suse.de>
1010 PR tree-optimization/107323
1011 * tree-loop-distribution.cc (pg_unmark_merged_alias_ddrs):
1013 (loop_distribution::break_alias_scc_partitions): Revert
1014 postorder save/restore from the PR94125 fix. Instead
1015 make sure to not ignore edges from SCCs we are going to
1018 2022-10-21 Monk Chiang <monk.chiang@sifive.com>
1020 * config/riscv/riscv.md: Add atomic type attribute.
1021 * config/riscv/sync.md: Add atomic type for atomic instructions.
1023 2022-10-21 Jakub Jelinek <jakub@redhat.com>
1025 PR tree-optimization/54346
1026 * match.pd ((vec_perm (vec_perm@0 @1 @2 VECTOR_CST) @0 VECTOR_CST)):
1027 Optimize nested VEC_PERM_EXPRs even if target can't handle the
1028 new one provided we don't increase number of VEC_PERM_EXPRs the
1029 target can't handle.
1031 2022-10-21 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1033 * config.gcc: Add riscv-vector-builtins-bases.o and riscv-vector-builtins-shapes.o
1034 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_I_OPS): New macro.
1035 (DEF_RVV_FUNCTION): Ditto.
1036 (handle_pragma_vector): Add intrinsic framework.
1037 * config/riscv/riscv.cc (riscv_print_operand): Add operand print for vsetvl/vsetvlmax.
1038 * config/riscv/riscv.md: include vector.md.
1039 * config/riscv/t-riscv: Add riscv-vector-builtins-bases.o and riscv-vector-builtins-shapes.o
1040 * config/riscv/riscv-vector-builtins-bases.cc: New file.
1041 * config/riscv/riscv-vector-builtins-bases.h: New file.
1042 * config/riscv/riscv-vector-builtins-functions.def: New file.
1043 * config/riscv/riscv-vector-builtins-shapes.cc: New file.
1044 * config/riscv/riscv-vector-builtins-shapes.h: New file.
1045 * config/riscv/riscv-vector-builtins-types.def: New file.
1046 * config/riscv/vector.md: New file.
1048 2022-10-21 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1050 * config.gcc: Add gt files since function_instance is GTY ((user)).
1051 * config/riscv/riscv-builtins.cc (riscv_init_builtins): Add RVV intrinsic framework.
1052 (riscv_builtin_decl): Ditto.
1053 (riscv_expand_builtin): Ditto.
1054 * config/riscv/riscv-protos.h (builtin_decl): New function.
1055 (expand_builtin): Ditto.
1056 (enum riscv_builtin_class): New enum to classify RVV intrinsic and RISC-V general built-in.
1057 * config/riscv/riscv-vector-builtins.cc (class GTY): New declaration.
1058 (struct registered_function_hasher): New struct.
1059 (DEF_RVV_OP_TYPE): New macro.
1060 (DEF_RVV_TYPE): Ditto.
1061 (DEF_RVV_PRED_TYPE): Ditto.
1062 (GTY): New declaration.
1063 (add_attribute): New function.
1064 (check_required_extensions): Ditto.
1065 (rvv_arg_type_info::get_tree_type): Ditto.
1066 (function_instance::function_instance): Ditto.
1067 (function_instance::operator==): Ditto.
1068 (function_instance::any_type_float_p): Ditto.
1069 (function_instance::get_return_type): Ditto.
1070 (function_instance::get_arg_type): Ditto.
1071 (function_instance::hash): Ditto.
1072 (function_instance::call_properties): Ditto.
1073 (function_instance::reads_global_state_p): Ditto.
1074 (function_instance::modifies_global_state_p): Ditto.
1075 (function_instance::could_trap_p): Ditto.
1076 (function_builder::function_builder): Ditto.
1077 (function_builder::~function_builder): Ditto.
1078 (function_builder::allocate_argument_types): Ditto.
1079 (function_builder::register_function_group): Ditto.
1080 (function_builder::append_name): Ditto.
1081 (function_builder::finish_name): Ditto.
1082 (function_builder::get_attributes): Ditto.
1083 (function_builder::add_function): Ditto.
1084 (function_builder::add_unique_function): Ditto.
1085 (function_call_info::function_call_info): Ditto.
1086 (function_expander::function_expander): Ditto.
1087 (function_expander::add_input_operand): Ditto.
1088 (function_expander::generate_insn): Ditto.
1089 (registered_function_hasher::hash): Ditto.
1090 (registered_function_hasher::equal): Ditto.
1091 (builtin_decl): Ditto.
1092 (expand_builtin): Ditto.
1093 (gt_ggc_mx): Define for using GCC garbage collect.
1094 (gt_pch_nx): Define for using GCC garbage collect.
1095 * config/riscv/riscv-vector-builtins.def (DEF_RVV_OP_TYPE): New macro.
1096 (DEF_RVV_PRED_TYPE): Ditto.
1097 (vbool64_t): Add suffix.
1104 (vint8mf8_t): Ditto.
1105 (vuint8mf8_t): Ditto.
1106 (vint8mf4_t): Ditto.
1107 (vuint8mf4_t): Ditto.
1108 (vint8mf2_t): Ditto.
1109 (vuint8mf2_t): Ditto.
1111 (vuint8m1_t): Ditto.
1113 (vuint8m2_t): Ditto.
1115 (vuint8m4_t): Ditto.
1117 (vuint8m8_t): Ditto.
1118 (vint16mf4_t): Ditto.
1119 (vuint16mf4_t): Ditto.
1120 (vint16mf2_t): Ditto.
1121 (vuint16mf2_t): Ditto.
1122 (vint16m1_t): Ditto.
1123 (vuint16m1_t): Ditto.
1124 (vint16m2_t): Ditto.
1125 (vuint16m2_t): Ditto.
1126 (vint16m4_t): Ditto.
1127 (vuint16m4_t): Ditto.
1128 (vint16m8_t): Ditto.
1129 (vuint16m8_t): Ditto.
1130 (vint32mf2_t): Ditto.
1131 (vuint32mf2_t): Ditto.
1132 (vint32m1_t): Ditto.
1133 (vuint32m1_t): Ditto.
1134 (vint32m2_t): Ditto.
1135 (vuint32m2_t): Ditto.
1136 (vint32m4_t): Ditto.
1137 (vuint32m4_t): Ditto.
1138 (vint32m8_t): Ditto.
1139 (vuint32m8_t): Ditto.
1140 (vint64m1_t): Ditto.
1141 (vuint64m1_t): Ditto.
1142 (vint64m2_t): Ditto.
1143 (vuint64m2_t): Ditto.
1144 (vint64m4_t): Ditto.
1145 (vuint64m4_t): Ditto.
1146 (vint64m8_t): Ditto.
1147 (vuint64m8_t): Ditto.
1148 (vfloat32mf2_t): Ditto.
1149 (vfloat32m1_t): Ditto.
1150 (vfloat32m2_t): Ditto.
1151 (vfloat32m4_t): Ditto.
1152 (vfloat32m8_t): Ditto.
1153 (vfloat64m1_t): Ditto.
1154 (vfloat64m2_t): Ditto.
1155 (vfloat64m4_t): Ditto.
1156 (vfloat64m8_t): Ditto.
1189 * config/riscv/riscv-vector-builtins.h (GCC_RISCV_VECTOR_BUILTINS_H): New macro.
1190 (RVV_REQUIRE_RV64BIT): Ditto.
1191 (RVV_REQUIRE_ZVE64): Ditto.
1192 (RVV_REQUIRE_ELEN_FP_32): Ditto.
1193 (RVV_REQUIRE_ELEN_FP_64): Ditto.
1194 (enum operand_type_index): New enum.
1195 (DEF_RVV_OP_TYPE): New macro.
1196 (enum predication_type_index): New enum.
1197 (DEF_RVV_PRED_TYPE): New macro.
1198 (enum rvv_base_type): New enum.
1199 (struct rvv_builtin_suffixes): New struct.
1200 (struct rvv_arg_type_info): Ditto.
1201 (struct rvv_type_info): Ditto.
1202 (struct rvv_op_info): Ditto.
1203 (class registered_function): New class.
1204 (class function_base): Ditto.
1205 (class function_shape): Ditto.
1206 (struct function_group_info): New struct.
1207 (class GTY): New class.
1208 (class function_builder): Ditto.
1209 (class function_call_info): Ditto.
1210 (function_call_info::function_returns_void_p): New function.
1211 (class function_expander): New class.
1212 (function_instance::operator!=): New function.
1213 (function_expander::expand): Ditto.
1214 (function_expander::add_input_operand): Ditto.
1215 (function_base::call_properties): Ditto.
1217 2022-10-21 Haochen Jiang <haochen.jiang@intel.com>
1219 * config/i386/sse.md (ssedvecmode): Rename from VI1SI.
1220 (ssedvecmodelower): Rename from vi1si.
1221 (sdot_prod<mode>): New define_expand.
1222 (udot_prod<mode>): Ditto.
1224 2022-10-21 Kong Lingling <lingling.kong@intel.com>
1225 Hongyu Wang <hongyu.wang@intel.com>
1226 Haochen Jiang <haochen.jiang@intel.com>
1228 * common/config/i386/cpuinfo.h (get_available_features): Detect
1230 * common/config/i386/i386-common.cc
1231 (OPTION_MASK_ISA2_AVXVNNIINT8_SET): New.
1232 (OPTION_MASK_ISA2_AVXVNNIINT8_UNSET): Ditto.
1233 (ix86_handle_option): Handle -mavxvnniint8.
1234 * common/config/i386/i386-cpuinfo.h (enum processor_features):
1235 Add FEATURE_AVXVNNIINT8.
1236 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
1238 * config.gcc: Add avxvnniint8intrin.h.
1239 * config/i386/avxvnniint8intrin.h: New file.
1240 * config/i386/cpuid.h (bit_AVXVNNIINT8): New.
1241 * config/i386/i386-builtin.def: Add new builtins.
1242 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
1244 * config/i386/i386-options.cc (isa2_opts): Add -mavxvnniint8.
1245 (ix86_valid_target_attribute_inner_p): Handle avxvnniint8.
1246 * config/i386/i386-isa.def: Add DEF_PTA(AVXVNNIINT8) New..
1247 * config/i386/i386.opt: Add option -mavxvnniint8.
1248 * config/i386/immintrin.h: Include avxvnniint8intrin.h.
1249 * config/i386/sse.md (UNSPEC_VPMADDUBSWACCD
1250 UNSPEC_VPMADDUBSWACCSSD,UNSPEC_VPMADDWDACCD,
1251 UNSPEC_VPMADDWDACCSSD): Rename according to new style.
1252 (vpdp<vpdotprodtype>_<mode>): New define_insn.
1253 * doc/extend.texi: Document avxvnniint8.
1254 * doc/invoke.texi: Document -mavxvnniint8.
1255 * doc/sourcebuild.texi: Document target avxvnniint8.
1257 2022-10-21 Hongyu Wang <hongyu.wang@intel.com>
1259 * common/config/i386/i386-common.cc
1260 (OPTION_MASK_ISA_AVXIFMA_SET, OPTION_MASK_ISA2_AVXIFMA_UNSET,
1261 OPTION_MASK_ISA2_AVX2_UNSET): New macro.
1262 (ix86_handle_option): Handle -mavxifma.
1263 * common/config/i386/i386-cpuinfo.h (processor_types): Add
1265 * common/config/i386/i386-isas.h: Add ISA_NAME_TABLE_ENTRY for
1267 * common/config/i386/cpuinfo.h (get_available_features):
1269 * config.gcc: Add avxifmaintrin.h
1270 * config/i386/avx512ifmavlintrin.h: (_mm_madd52lo_epu64): Change
1272 (_mm_madd52hi_epu64): Likewise.
1273 (_mm256_madd52lo_epu64): Likewise.
1274 (_mm256_madd52hi_epu64): Likewise.
1275 * config/i386/avxifmaintrin.h: New header.
1276 * config/i386/cpuid.h (bit_AVXIFMA): New.
1277 * config/i386/i386-builtin.def: Add new builtins, and correct
1278 pattern names for AVX512IFMA.
1279 * config/i386/i386-builtins.cc (def_builtin): Handle AVX-IFMA
1280 builtins like AVX-VNNI.
1281 * config/i386/i386-c.cc (ix86_target_macros_internal): Define
1283 * config/i386/i386-expand.cc (ix86_check_builtin_isa_match):
1284 Relax ISA masks for AVXIFMA.
1285 * config/i386/i386-isa.def: Add AVXIFMA.
1286 * config/i386/i386-options.cc (isa2_opts): Add -mavxifma.
1287 (ix86_valid_target_attribute_inner_p): Handle avxifma.
1288 * config/i386/i386.md (isa): Add attr avxifma and avxifmavl.
1289 * config/i386/i386.opt: Add option -mavxifma.
1290 * config/i386/immintrin.h: Inculde avxifmaintrin.h.
1291 * config/i386/sse.md (avx_vpmadd52<vpmadd52type>_<mode>):
1293 (vpamdd52<vpmadd52type><mode><sd_maskz_name>): Remove.
1294 (vpamdd52huq<mode>_maskz): Rename to ...
1295 (vpmadd52huq<mode>_maskz): ... this.
1296 (vpamdd52luq<mode>_maskz): Rename to ...
1297 (vpmadd52luq<mode>_maskz): ... this.
1298 (vpmadd52<vpmadd52type><mode>): New define_insn.
1299 (vpmadd52<vpmadd52type>v8di): Likewise.
1300 (vpmadd52<vpmadd52type><mode>_maskz_1): Likewise.
1301 (vpamdd52<vpmadd52type><mode>_mask): Rename to ...
1302 (vpmadd52<vpmadd52type><mode>_mask): ... this.
1303 * doc/invoke.texi: Document -mavxifma.
1304 * doc/extend.texi: Document avxifma.
1305 * doc/sourcebuild.texi: Document target avxifma.
1307 2022-10-20 Aldy Hernandez <aldyh@redhat.com>
1310 * value-query.cc (get_range_global): Handle non integer ranges for
1311 default def SSA names.
1313 2022-10-20 Aldy Hernandez <aldyh@redhat.com>
1315 * range-op-float.cc (foperator_unordered_lt::op1_range): New.
1316 (foperator_unordered_lt::op2_range): New.
1318 2022-10-20 Artem Klimov <jakmobius@gmail.com>
1319 Alexander Monakov <amonakov@gcc.gnu.org>
1322 * ipa-visibility.cc (function_and_variable_visibility): Promote
1323 TLS access model afer visibility optimizations.
1324 * varasm.cc (have_optimized_refs): New helper.
1325 (optimize_dyn_tls_for_decl_p): New helper. Use it ...
1326 (decl_default_tls_model): ... here in place of 'optimize' check.
1328 2022-10-20 Aldy Hernandez <aldyh@redhat.com>
1330 * range-op-float.cc (foperator_unordered_le::op1_range): Adjust
1331 false side with a NAN operand.
1332 (foperator_unordered_le::op2_range): Same.
1333 (foperator_unordered_gt::op1_range): Same.
1334 (foperator_unordered_gt::op2_range): Same.
1335 (foperator_unordered_ge::op1_range): Same.
1336 (foperator_unordered_ge::op2_range): Same.
1337 (foperator_unordered_equal::op1_range): Same.
1339 2022-10-20 Andre Vieira <andre.simoesdiasvieira@arm.com>
1341 PR tree-optimization/107326
1342 * tree-vect-patterns.cc (vect_recog_bitfield_ref_pattern): Change
1343 vectype when widening container.
1345 2022-10-20 Aldy Hernandez <aldyh@redhat.com>
1347 * value-range.h (frange::set_varying): Do not set NAN flags for
1349 * value-range.cc (frange::normalize_kind): Adjust for no NAN when
1351 (frange::verify_range): Same.
1352 * range-op-float.cc (maybe_isnan): Remove flag_finite_math_only check.
1354 2022-10-20 Aldy Hernandez <aldyh@redhat.com>
1356 * range-op-float.cc (finite_operand_p): Remove.
1357 (finite_operands_p): Rename to...
1358 (maybe_isnan): ...this.
1359 (frelop_early_resolve): Use maybe_isnan instead of finite_operands_p.
1360 (foperator_equal::fold_range): Same.
1361 (foperator_equal::op1_range): Same.
1362 (foperator_not_equal::fold_range): Same.
1363 (foperator_lt::fold_range): Same.
1364 (foperator_le::fold_range): Same.
1365 (foperator_gt::fold_range): Same.
1366 (foperator_ge::fold_range): Same.
1368 2022-10-20 Jakub Jelinek <jakub@redhat.com>
1370 * passes.cc (pass_manager::register_pass): Fix a comment
1371 typo - copmilation -> compilation.
1373 2022-10-20 Richard Biener <rguenther@suse.de>
1375 * tree-vect-loop.cc (vect_phi_first_order_recurrence_p):
1376 Disallow latch PHI defs.
1377 (vectorizable_recurr): Revert previous change.
1379 2022-10-20 Julian Brown <julian@codesourcery.com>
1382 * config/gcn/gcn.cc (gcn_detect_incoming_pointer_arg): Any pointer
1383 argument forces FLAT addressing mode, not just
1384 pointer-to-non-aggregate.
1386 2022-10-20 Thomas Schwinge <thomas@codesourcery.com>
1388 * configure.ac (AC_CONFIG_MACRO_DIRS): Instantiate.
1389 * configure: Regenerate.
1391 2022-10-20 Richard Sandiford <richard.sandiford@arm.com>
1393 * config/aarch64/aarch64-sve-builtins.h
1394 (gimple_folder::fold_to_cstu): New member function.
1395 * config/aarch64/aarch64-sve-builtins.cc
1396 (gimple_folder::fold_to_cstu): Define.
1397 * config/aarch64/aarch64-sve-builtins-base.cc
1398 (svcnt_bhwd_impl::fold): Use it.
1400 2022-10-20 Richard Sandiford <richard.sandiford@arm.com>
1402 * config/aarch64/aarch64-sve-builtins-functions.h (quiet)
1403 (rtx_code_function, rtx_code_function_rotated, unspec_based_function)
1404 (unspec_based_function_rotated, unspec_based_function_exact_insn)
1405 (unspec_based_fused_function, unspec_based_fused_lane_function):
1406 Replace constructors with using directives.
1407 * config/aarch64/aarch64-sve-builtins-base.cc (svcnt_bhwd_pat_impl)
1408 (svcreate_impl, svdotprod_lane_impl, svget_impl, svld1_extend_impl)
1409 (svld1_gather_extend_impl, svld234_impl, svldff1_gather_extend)
1410 (svset_impl, svst1_scatter_truncate_impl, svst1_truncate_impl)
1411 (svst234_impl, svundef_impl): Likewise.
1412 * config/aarch64/aarch64-sve-builtins-sve2.cc
1413 (svldnt1_gather_extend_impl, svmovl_lb_impl): Likewise.
1414 (svstnt1_scatter_truncate_impl): Likewise.
1416 2022-10-20 Richard Sandiford <richard.sandiford@arm.com>
1418 * config/aarch64/aarch64-sve-builtins-base.cc: Replace CONSTEXPR
1419 with constexpr throughout.
1420 * config/aarch64/aarch64-sve-builtins-functions.h: Likewise.
1421 * config/aarch64/aarch64-sve-builtins-shapes.cc: Likewise.
1422 * config/aarch64/aarch64-sve-builtins-sve2.cc: Likewise.
1423 * config/aarch64/aarch64-sve-builtins.cc: Likewise.
1425 2022-10-20 Richard Sandiford <richard.sandiford@arm.com>
1427 * config/aarch64/aarch64-sve.md (*aarch64_brk<brk_op>_cc): Remove
1428 merging alternative.
1429 (*aarch64_brk<brk_op>_ptest): Likewise.
1431 2022-10-20 Richard Sandiford <richard.sandiford@arm.com>
1433 * config/aarch64/iterators.md (SVE_BRKP): New iterator.
1434 * config/aarch64/aarch64-sve.md (*aarch64_brkn_cc): New pattern.
1435 (*aarch64_brkn_ptest): Likewise.
1436 (*aarch64_brk<brk_op>_cc): Restrict to SVE_BRKP.
1437 (*aarch64_brk<brk_op>_ptest): Likewise.
1439 2022-10-20 Richard Biener <rguenther@suse.de>
1443 * tree-cfg.h (verify_gimple_in_seq): Add parameter to
1444 indicate whether to emit an ICE. Add return value.
1445 (verify_gimple_in_cfg): Likewise.
1446 * tree-cfg.cc (verify_gimple_in_seq): Likewise.
1447 (verify_gimple_in_cfg): Likewise.
1449 2022-10-20 Richard Biener <rguenther@suse.de>
1451 PR tree-optimization/107240
1452 * tree-vect-patterns.cc (vect_recog_bit_insert_pattern): Attempt to
1453 simplify shifted value first.
1455 2022-10-20 Andrew MacLeod <amacleod@redhat.com>
1457 * gimple-range-gori.h (compute_operand_range): Make public.
1458 * gimple-range-infer.cc (gimple_infer_range::check_assume_func): New.
1459 (gimple_infer_range::gimple_infer_range): Check for assume calls.
1460 * gimple-range-infer.h (check_assume_func): Add prototype.
1461 * gimple-range.cc (assume_query::assume_range_p): New.
1462 (assume_query::range_of_expr): New.
1463 (assume_query::assume_query): New.
1464 (assume_query::calculate_op): New.
1465 (assume_query::calculate_phi): New.
1466 (assume_query::check_taken_edge): New.
1467 (assume_query::calculate_stmt): New.
1468 (assume_query::dump): New.
1469 * gimple-range.h (class assume_query): New.
1470 * tree-vrp.cc (pass_assumptions::execute): Add processing.
1472 2022-10-19 Aldy Hernandez <aldyh@redhat.com>
1474 * range-op-float.cc (build_le): Document result.
1477 (foperator_ge::op2_range): Check result of build_*.
1478 (foperator_unordered_le::op1_range): Same.
1479 (foperator_unordered_le::op2_range): Same.
1480 (foperator_unordered_gt::op1_range): Same.
1481 (foperator_unordered_gt::op2_range): Same.
1482 (foperator_unordered_ge::op1_range): Same.
1483 (foperator_unordered_ge::op2_range): Same.
1485 2022-10-19 Marek Polacek <polacek@redhat.com>
1488 * doc/invoke.texi: Update documentation of -Wuseless-cast.
1490 2022-10-19 Andrew MacLeod <amacleod@redhat.com>
1492 * gimple-range-cache.cc (ranger_cache::range_from_dom): Use
1493 Value_Range not int_range_max.
1495 2022-10-19 Aldy Hernandez <aldyh@redhat.com>
1497 PR tree-optimization/107312
1498 * range.h (range_true_and_false): Special case 1-bit signed types.
1499 * value-range.cc (range_tests_misc): New test.
1501 2022-10-19 LIU Hao <lh_mouse@126.com>
1503 * config/i386/mingw-mcfgthread.h: New file
1504 * config/i386/mingw32.h: Add builtin macro and default libraries
1505 for mcfgthread when thread model is `mcf`
1506 * config.gcc: Include 'i386/mingw-mcfgthread.h' when thread model
1508 * configure.ac: Recognize `mcf` as a valid thread model
1509 * config.in: Regenerate
1510 * configure: Regenerate
1512 2022-10-19 Lewis Hyatt <lhyatt@gmail.com>
1514 * gengtype.cc (output_escaped_param): Add missing const.
1515 (get_string_option): Add missing check for option type.
1516 (walk_type): Support new "string_length" GTY option.
1517 (write_types_process_field): Likewise.
1518 * ggc-common.cc (gt_pch_note_object): Add optional length argument.
1519 * ggc.h (gt_pch_note_object): Adjust prototype for new argument.
1520 (gt_pch_n_S2): Declare...
1521 * stringpool.cc (gt_pch_n_S2): ...new function.
1522 * doc/gty.texi: Document new GTY((string_length)) option.
1524 2022-10-19 Martin Liska <mliska@suse.cz>
1526 * doc/extend.texi: Remove useless @tie{} directives.
1528 2022-10-19 Martin Jambor <mjambor@suse.cz>
1530 PR tree-optimization/107206
1531 * tree-sra.cc (struct access): New field grp_result_of_prop_from_lhs.
1532 (analyze_access_subtree): Do not create replacements for accesses with
1533 this flag when not toally scalarizing.
1534 (propagate_subaccesses_from_lhs): Set the new flag.
1536 2022-10-19 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
1539 * config/s390/s390.cc (s390_call_saved_register_used): For a
1540 parameter with BLKmode fix determining number of consecutive
1543 2022-10-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
1545 * config/xtensa/xtensa-protos.h
1546 (xtensa_split1_finished_p, xtensa_split_DI_reg_imm): New prototypes.
1547 * config/xtensa/xtensa.cc
1548 (xtensa_split1_finished_p, xtensa_split_DI_reg_imm, xtensa_lra_p):
1550 (TARGET_LRA_P): Replace the dummy hook with xtensa_lra_p.
1551 (xt_true_regnum): Rework.
1552 * config/xtensa/xtensa.h (CALL_REALLY_USED_REGISTERS):
1553 Switch from CALL_USED_REGISTERS, and revise the comment.
1554 * config/xtensa/constraints.md (Y):
1555 Use !xtensa_split1_finished_p() instead of can_create_pseudo_p().
1556 * config/xtensa/predicates.md (move_operand): Ditto.
1557 * config/xtensa/xtensa.md: Add two new split patterns:
1558 - splits DImode immediate load into two SImode ones
1559 - puts out-of-constraint SImode constants into the constant pool
1560 * config/xtensa/xtensa.opt (-mlra): New target-specific option
1561 for testing purpose.
1563 2022-10-19 Robin Dapp <rdapp@linux.ibm.com>
1565 * config/s390/s390.md: Move reload_completed and check operands for REG_P.
1567 2022-10-19 Jakub Jelinek <jakub@redhat.com>
1569 PR middle-end/107262
1570 * expr.cc (convert_mode_scalar): For BFmode -> SFmode conversions
1571 of constants, use simplify_unary_operation if fromi has VOIDmode
1572 instead of recursive convert_mode_scalar.
1574 2022-10-19 Jakub Jelinek <jakub@redhat.com>
1576 PR tree-optimization/106990
1577 * match.pd ((~X - ~Y) -> Y - X, -x & 1 -> x & 1): Guard with
1578 !TYPE_OVERFLOW_SANITIZED (type).
1580 2022-10-19 Jakub Jelinek <jakub@redhat.com>
1582 * config/i386/i386-builtins.cc (ix86_bf16_ptr_type_node): Remove.
1583 (ix86_bf16_type_node): New variable.
1584 (ix86_register_bf16_builtin_type): If bfloat16_type_node is NULL
1585 from generic code, set only ix86_bf16_type_node to a new REAL_TYPE
1586 rather than bfloat16_type_node, otherwise set ix86_bf16_type_node
1587 to bfloat16_type_node. Register __bf16 on ix86_bf16_type_node
1588 rather than bfloat16_type_node. Don't initialize unused
1589 ix86_bf16_ptr_type_node.
1590 * config/i386/i386-builtin-types.def (BFLOAT16): Use
1591 ix86_bf16_type_node rather than bfloat16_type_node.
1593 2022-10-19 Richard Biener <rguenther@suse.de>
1595 PR tree-optimization/106781
1596 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee): Copy
1597 LHS removal from fixup_noreturn_call.
1599 2022-10-19 liuhongt <hongtao.liu@intel.com>
1602 * config/i386/i386-expand.cc (ix86_vec_perm_index_canon): New.
1603 (expand_vec_perm_shufps_shufps): Call
1604 ix86_vec_perm_index_canon
1606 2022-10-18 Martin Jambor <mjambor@suse.cz>
1608 * ipa-prop.h (ipa_agg_value): Remove type.
1609 (ipa_agg_value_set): Likewise.
1610 (ipa_copy_agg_values): Remove function.
1611 (ipa_release_agg_values): Likewise.
1612 (ipa_auto_call_arg_values) Add a forward declaration.
1613 (ipa_call_arg_values): Likewise.
1614 (class ipa_argagg_value_list): New constructors, added member function
1616 (class ipa_auto_call_arg_values): Removed the destructor and member
1617 function safe_aggval_at. Use ipa_argagg_values for m_known_aggs.
1618 (class ipa_call_arg_values): Removed member function safe_aggval_at.
1619 Use ipa_argagg_values for m_known_aggs.
1620 (ipa_get_indirect_edge_target): Removed declaration.
1621 (ipa_find_agg_cst_for_param): Likewise.
1622 (ipa_find_agg_cst_from_init): New declaration.
1623 (ipa_agg_value_from_jfunc): Likewise.
1624 (ipa_agg_value_set_from_jfunc): Removed declaration.
1625 (ipa_push_agg_values_from_jfunc): New declaration.
1626 * ipa-cp.cc (ipa_agg_value_from_node): Renamed to
1627 ipa_agg_value_from_jfunc, made public.
1628 (ipa_agg_value_set_from_jfunc): Removed.
1629 (ipa_push_agg_values_from_jfunc): New function.
1630 (ipa_get_indirect_edge_target_1): Removed known_aggs parameter, use
1631 avs for this purpose too.
1632 (ipa_get_indirect_edge_target): Removed the overload working on
1633 ipa_auto_call_arg_values, use ipa_argagg_value_list in the remaining
1635 (devirtualization_time_bonus): Use ipa_argagg_value_list and
1636 ipa_get_indirect_edge_target_1 instead of
1637 ipa_get_indirect_edge_target.
1638 (context_independent_aggregate_values): Removed function.
1639 (gather_context_independent_values): Work on ipa_argagg_value_list.
1640 (estimate_local_effects): Likewise, define some iterator variables
1641 only in the construct where necessary.
1642 (ipcp_discover_new_direct_edges): Adjust the call to
1643 ipa_get_indirect_edge_target_1.
1644 (push_agg_values_for_index_from_edge): Adjust the call
1645 ipa_agg_value_from_node which has been renamed to
1646 ipa_agg_value_from_jfunc.
1647 * ipa-fnsummary.cc (evaluate_conditions_for_known_args): Work on
1648 ipa_argagg_value_list.
1649 (evaluate_properties_for_edge): Replace manual filling in aggregate
1650 values with call to ipa_push_agg_values_from_jfunc.
1651 (estimate_calls_size_and_time): Work on ipa_argagg_value_list.
1652 (ipa_cached_call_context::duplicate_from): Likewise.
1653 (ipa_cached_call_context::release): Likewise.
1654 (ipa_call_context::equal_to): Likewise.
1655 * ipa-prop.cc (ipa_find_agg_cst_from_init): Make public.
1656 (ipa_find_agg_cst_for_param): Removed function.
1657 (ipa_find_agg_cst_from_jfunc_items): New function.
1658 (try_make_edge_direct_simple_call): Replace calls to
1659 ipa_agg_value_set_from_jfunc and ipa_find_agg_cst_for_param with
1660 ipa_find_agg_cst_from_init and ipa_find_agg_cst_from_jfunc_items.
1661 (try_make_edge_direct_virtual_call): Replace calls to
1662 ipa_agg_value_set_from_jfunc and ipa_find_agg_cst_for_param with
1663 simple query of constant jump function and a call to
1664 ipa_find_agg_cst_from_jfunc_items.
1665 (ipa_auto_call_arg_values::~ipa_auto_call_arg_values): Removed.
1667 2022-10-18 Martin Jambor <mjambor@suse.cz>
1669 * ipa-prop.h (IPA_PROP_ARG_INDEX_LIMIT_BITS): New.
1670 (ipcp_transformation): Added forward declaration.
1671 (ipa_argagg_value): New type.
1672 (ipa_argagg_value_list): New type.
1673 (ipa_agg_replacement_value): Removed type.
1674 (ipcp_transformation): Switch from using ipa_agg_replacement_value
1675 to ipa_argagg_value_list.
1676 (ipa_get_agg_replacements_for_node): Removed.
1677 (ipa_dump_agg_replacement_values): Removed declaration.
1678 * ipa-cp.cc: Define INCLUDE_ALGORITHM.
1679 (values_equal_for_ipcp_p): Moved up in the file.
1680 (ipa_argagg_value_list::dump): New function.
1681 (ipa_argagg_value_list::debug): Likewise.
1682 (ipa_argagg_value_list::get_elt): Likewise.
1683 (ipa_argagg_value_list::get_elt_for_index): Likewise.
1684 (ipa_argagg_value_list::get_value): New overloaded functions.
1685 (ipa_argagg_value_list::superset_of_p): New function.
1686 (new ipa_argagg_value_list::push_adjusted_values): Likewise.
1687 (push_agg_values_from_plats): Likewise.
1688 (intersect_argaggs_with): Likewise.
1689 (get_clone_agg_value): Removed.
1690 (ipa_agg_value_from_node): Make last parameter const, use
1691 ipa_argagg_value_list to search values coming from clones.
1692 (ipa_get_indirect_edge_target_1): Use ipa_argagg_value_list to search
1693 values coming from clones.
1694 (ipcp_discover_new_direct_edges): Pass around a vector of
1695 ipa_argagg_values rather than a link list of replacement values.
1696 (cgraph_edge_brings_value_p): Use ipa_argagg_value_list to search
1697 values coming from clones.
1698 (create_specialized_node): Work with a vector of ipa_argagg_values
1699 rather than a link list of replacement values.
1700 (self_recursive_agg_pass_through_p): Make the pointer parameters
1702 (copy_plats_to_inter): Removed.
1703 (intersect_with_plats): Likewise.
1704 (agg_replacements_to_vector): Likewise.
1705 (intersect_with_agg_replacements): Likewise.
1706 (intersect_aggregates_with_edge): Likewise.
1707 (push_agg_values_for_index_from_edge): Likewise.
1708 (push_agg_values_from_edge): Likewise.
1709 (find_aggregate_values_for_callers_subset): Rewrite.
1710 (cgraph_edge_brings_all_agg_vals_for_node): Likewise.
1711 (ipcp_val_agg_replacement_ok_p): Use ipa_argagg_value_list to search
1713 (decide_about_value): Work with a vector of ipa_argagg_values rather
1714 than a link list of replacement values.
1715 (decide_whether_version_node): Likewise.
1716 (ipa_analyze_node): Check number of parameters, assert that there
1717 are no descriptors when bailing out.
1718 * ipa-prop.cc (ipa_set_node_agg_value_chain): Switch to a vector of
1720 (ipa_node_params_t::duplicate): Removed superfluous handling of
1721 ipa_agg_replacement_values. Name of src parameter removed because
1722 it is no longer used.
1723 (ipcp_transformation_t::duplicate): Replaced duplication of
1724 ipa_agg_replacement_values with copying vector m_agg_values.
1725 (ipa_dump_agg_replacement_values): Removed.
1726 (write_ipcp_transformation_info): Stream the new data-structure
1728 (read_ipcp_transformation_info): Likewise.
1729 (adjust_agg_replacement_values): Work with ipa_argagg_values instead
1730 of linked lists of ipa_agg_replacement_values, copy the items and
1731 truncate the vector as necessary to keep it sorted instead of marking
1732 items as invalid. Return one bool if CFG should be updated.
1733 (ipcp_modif_dom_walker): Store ipcp_transformation instead of
1734 linked list of ipa_agg_replacement_values.
1735 (ipcp_modif_dom_walker::before_dom_children): Use
1736 ipa_argagg_value_list instead of walking a list of
1737 ipa_agg_replacement_values.
1738 (ipcp_transform_function): Switch to the new data structure, adjust
1741 2022-10-18 Richard Biener <rguenther@suse.de>
1743 PR tree-optimization/107302
1744 * tree-vect-loop.cc (vectorizable_recurrence): Fix vec_perm
1745 placement for a PHI latch def.
1747 2022-10-18 Andre Vieira <andre.simoesdiasvieira@arm.com>
1749 PR tree-optimization/107275
1750 * tree-if-conv.cc (if_convertible_loop_p_1): Move
1751 find_data_references_in_loop call from here...
1752 (if_convertible_loop_p): And move data-reference vector initialization
1754 (tree_if_conversion):... to here.
1756 2022-10-18 Jakub Jelinek <jakub@redhat.com>
1759 * gimple.def (GIMPLE_ASSUME): New statement kind.
1760 * gimple.h (struct gimple_statement_assume): New type.
1761 (is_a_helper <gimple_statement_assume *>::test,
1762 is_a_helper <const gimple_statement_assume *>::test): New.
1763 (gimple_build_assume): Declare.
1764 (gimple_has_substatements): Return true for GIMPLE_ASSUME.
1765 (gimple_assume_guard, gimple_assume_set_guard,
1766 gimple_assume_guard_ptr, gimple_assume_body_ptr, gimple_assume_body):
1767 New inline functions.
1768 * gsstruct.def (GSS_ASSUME): New.
1769 * gimple.cc (gimple_build_assume): New function.
1770 (gimple_copy): Handle GIMPLE_ASSUME.
1771 * gimple-pretty-print.cc (dump_gimple_assume): New function.
1772 (pp_gimple_stmt_1): Handle GIMPLE_ASSUME.
1773 * gimple-walk.cc (walk_gimple_op): Handle GIMPLE_ASSUME.
1774 * omp-low.cc (WALK_SUBSTMTS): Likewise.
1775 (lower_omp_1): Likewise.
1776 * omp-oacc-kernels-decompose.cc (adjust_region_code_walk_stmt_fn):
1778 * tree-cfg.cc (verify_gimple_stmt, verify_gimple_in_seq_2): Likewise.
1779 * function.h (struct function): Add assume_function bitfield.
1780 * gimplify.cc (gimplify_call_expr): If the assumption isn't
1781 simple enough, expand it into GIMPLE_ASSUME wrapped block or
1783 * gimple-low.cc: Include attribs.h.
1784 (create_assumption_fn): New function.
1785 (struct lower_assumption_data): New type.
1786 (find_assumption_locals_r, assumption_copy_decl,
1787 adjust_assumption_stmt_r, adjust_assumption_stmt_op,
1788 lower_assumption): New functions.
1789 (lower_stmt): Handle GIMPLE_ASSUME.
1790 * tree-ssa-ccp.cc (pass_fold_builtins::execute): Remove
1792 * lto-streamer-out.cc (output_struct_function_base): Pack
1793 assume_function bit.
1794 * lto-streamer-in.cc (input_struct_function_base): And unpack it.
1795 * cgraphunit.cc (cgraph_node::expand): Don't verify assume_function
1796 has TREE_ASM_WRITTEN set and don't release its body.
1797 (symbol_table::compile): Allow assume functions not to have released
1799 * internal-fn.cc (expand_ASSUME): Remove gcc_unreachable.
1800 * passes.cc (execute_one_pass): For TODO_discard_function don't
1801 release body of assume functions.
1802 * cgraph.cc (cgraph_node::verify_node): Don't verify cgraph nodes
1803 of PROP_assumptions_done functions.
1804 * tree-pass.h (PROP_assumptions_done): Define.
1805 (TODO_discard_function): Adjust comment.
1806 (make_pass_assumptions): Declare.
1807 * passes.def (pass_assumptions): Add.
1808 * timevar.def (TV_TREE_ASSUMPTIONS): New.
1809 * tree-inline.cc (remap_gimple_stmt): Handle GIMPLE_ASSUME.
1810 * tree-vrp.cc (pass_data_assumptions): New variable.
1811 (pass_assumptions): New class.
1812 (make_pass_assumptions): New function.
1814 2022-10-18 Richard Biener <rguenther@suse.de>
1816 PR tree-optimization/107301
1817 * gimple-ssa-isolate-paths.cc (handle_return_addr_local_phi_arg):
1818 Check whether we can duplicate the block.
1819 (find_implicit_erroneous_behavior): Likewise.
1821 2022-10-18 Andrew MacLeod <amacleod@redhat.com>
1823 PR tree-optimization/107273
1824 * value-relation.cc (equiv_oracle::add_partial_equiv): Merge
1825 instead of copying precison of each member.
1827 2022-10-17 Jeff Law <jeffreyalaw@gmail.com>
1830 * config/h8300/combiner.md: Replace '<' preincment constraint with
1831 ZA/Z1..ZH/Z7 combinations.
1832 * config/h8300/movepush.md: Similarly
1834 2022-10-17 Jeff Law <jeffreyalaw@gmail.com>
1836 * config/h8300/constraints.md (Za..Zh): New constraints for
1837 autoinc addresses using a specific register.
1838 * config/h8300/h8300.cc (pre_incdec_with_reg): New function.
1839 * config/h8300/h8300-protos.h (pre_incdec_with_reg): Add prototype.
1841 2022-10-17 Jeff Law <jlaw@ventanamicro.com>
1843 * config/i386/cet.c: Remove accidental commit.
1844 * config/i386/driver-mingw32.c: Likewise.
1845 * config/i386/i386-builtins.c: Likewise.
1846 * config/i386/i386-d.c: Likewise.
1847 * config/i386/i386-expand.c: Likewise.
1848 * config/i386/i386-features.c: Likewise.
1849 * config/i386/i386-options.c: Likewise.
1850 * config/i386/t-cet: Likewise.
1851 * config/i386/x86-tune-sched-atom.c: Likewise.
1852 * config/i386/x86-tune-sched-bd.c: Likewise.
1853 * config/i386/x86-tune-sched-core.c: Likewise.
1854 * config/i386/x86-tune-sched.c: Likewise.
1856 2022-10-17 Jeff Law <jeffreyalaw@gmail.com>
1858 * common/config/h8300/h8300-common.cc (h8300_option_optimization_table):
1859 Enable redundant extension elimination at -O2 and above.
1860 * config/i386/cet.c: New file.
1861 * config/i386/driver-mingw32.c: New file.
1862 * config/i386/i386-builtins.c: New file.
1863 * config/i386/i386-d.c: New file.
1864 * config/i386/i386-expand.c: New file.
1865 * config/i386/i386-features.c: New file.
1866 * config/i386/i386-options.c: New file.
1867 * config/i386/t-cet: New file.
1868 * config/i386/x86-tune-sched-atom.c: New file.
1869 * config/i386/x86-tune-sched-bd.c: New file.
1870 * config/i386/x86-tune-sched-core.c: New file.
1871 * config/i386/x86-tune-sched.c: New file.
1873 2022-10-17 Jeff Law <jeffreyalaw@gmail.com>
1875 * config/h8300/extensions.md (CCZN setting zero extended load): Add
1878 2022-10-17 Aldy Hernandez <aldyh@redhat.com>
1880 PR tree-optimization/107293
1882 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges):
1883 Check that condition post-dominates the definition point.
1885 2022-10-17 Aldy Hernandez <aldyh@redhat.com>
1887 PR tree-optimization/107286
1888 * value-range.cc (range_tests_floats): Do not test for -Inf when
1889 flag_finite_math_only.
1891 2022-10-17 Andrew MacLeod <amacleod@redhat.com>
1893 * gimple-range-fold.cc (fold_using_range::range_of_range_op):
1894 Provide relation_trio class.
1895 * gimple-range-gori.cc (gori_compute::refine_using_relation):
1896 Provide relation_trio class.
1897 (gori_compute::refine_using_relation): Ditto.
1898 (gori_compute::compute_operand1_range): Provide lhs_op2 and
1899 op1_op2 relations via relation_trio class.
1900 (gori_compute::compute_operand2_range): Ditto.
1901 * gimple-range-op.cc (gimple_range_op_handler::calc_op1): Use
1902 relation_trio instead of relation_kind.
1903 (gimple_range_op_handler::calc_op2): Ditto.
1904 (*::fold_range): Ditto.
1905 * gimple-range-op.h (gimple_range_op::calc_op1): Adjust prototypes.
1906 (gimple_range_op::calc_op2): Adjust prototypes.
1907 * range-op-float.cc (*::fold_range): Use relation_trio instead of
1909 (*::op1_range): Ditto.
1910 (*::op2_range): Ditto.
1911 * range-op.cc (*::fold_range): Use relation_trio instead of
1913 (*::op1_range): Ditto.
1914 (*::op2_range): Ditto.
1915 * range-op.h (class range_operator): Adjust prototypes.
1916 (class range_operator_float): Ditto.
1917 (class range_op_handler): Adjust prototypes.
1918 (relop_early_resolve): Pickup op1_op2 relation from relation_trio.
1919 * value-relation.cc (VREL_LAST): Adjust use to be one past the end of
1921 (relation_oracle::validate_relation): Use relation_trio in call
1923 * value-relation.h (enum relation_kind_t): Add VREL_LAST as
1925 (class relation_trio): New.
1926 (TRIO_VARYING, TRIO_SHIFT, TRIO_MASK): New.
1928 2022-10-17 Andrew MacLeod <amacleod@redhat.com>
1930 * range-op-float.cc (foperator_not_equal::op1_range): Check for
1931 VREL_EQ after singleton.
1932 (foperator_unordered::op1_range): Set VARYING before calling
1934 (foperator_ordered::op1_range): Set rather than clear NAN if both
1935 operands are the same.
1937 2022-10-17 Andrew MacLeod <amacleod@redhat.com>
1939 * value-relation.cc (value_relation::dump): Change message.
1940 * value-relation.h (value_relation::set_relation): If op1 is the
1941 same as op2 do not create a relation.
1943 2022-10-17 Thomas Schwinge <thomas@codesourcery.com>
1945 * config/gcn/gcn.cc (VnMODE): Use 'case E_QImode:' instead of
1946 'case QImode:', etc.
1948 2022-10-17 Thomas Schwinge <thomas@codesourcery.com>
1950 * gimple-expr.cc (mark_addressable_2): Tag as 'static'.
1952 2022-10-17 Richard Biener <rguenther@suse.de>
1953 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1955 PR tree-optimization/99409
1956 PR tree-optimization/99394
1957 * tree-vectorizer.h (vect_def_type::vect_first_order_recurrence): Add.
1958 (stmt_vec_info_type::recurr_info_type): Likewise.
1959 (vectorizable_recurr): New function.
1960 * tree-vect-loop.cc (vect_phi_first_order_recurrence_p): New
1962 (vect_analyze_scalar_cycles_1): Look for first order
1964 (vect_analyze_loop_operations): Handle them.
1965 (vect_transform_loop): Likewise.
1966 (vectorizable_recurr): New function.
1967 (maybe_set_vectorized_backedge_value): Handle the backedge value
1968 setting in the first order recurrence PHI and the permutes.
1969 * tree-vect-stmts.cc (vect_analyze_stmt): Handle first order
1971 (vect_transform_stmt): Likewise.
1972 (vect_is_simple_use): Likewise.
1973 (vect_is_simple_use): Likewise.
1974 * tree-vect-slp.cc (vect_get_and_check_slp_defs): Likewise.
1975 (vect_build_slp_tree_2): Likewise.
1976 (vect_schedule_scc): Handle the backedge value setting in the
1977 first order recurrence PHI and the permutes.
1979 2022-10-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1981 * config/riscv/t-riscv: Change Tab into 2 space.
1983 2022-10-17 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
1985 * config/riscv/riscv-vector-builtins.cc (mangle_builtin_type): Move down the function.
1987 2022-10-17 Arsen Arsenovic <arsen@aarsen.me>
1989 * config.gcc: --target=*-elf --without-{newlib,headers} should
1992 2022-10-17 Hu, Lin1 <lin1.hu@intel.com>
1994 * common/config/i386/cpuinfo.h:
1995 (get_intel_cpu): Handle Meteorlake.
1996 * common/config/i386/i386-common.cc:
1997 (processor_alias_table): Add Meteorlake.
1999 2022-10-17 Haochen Jiang <haochen.jiang@intel.com>
2001 * common/config/i386/cpuinfo.h:
2002 (get_intel_cpu): Handle Raptorlake.
2003 * common/config/i386/i386-common.cc:
2004 (processor_alias_table): Add Raptorlake.
2006 2022-10-16 Jeff Law <jeffreyalaw@gmail.com>
2008 * config/h8300/constraints.md (Z0..Z7): New register
2010 * config/h8300/h8300.h (reg_class): Add new classes.
2011 (REG_CLASS_NAMES): Similarly.
2012 (REG_CLASS_CONTENTS): Similarly.
2014 2022-10-16 Jeff Law <jeffreyalaw@gmail.com>
2016 * config/h8300/constraints.md (Zz constraint): Renamed
2018 * config/h8300/movepush.md (movqi_h8sx, movhi_h8sx): Adjust
2019 constraint to use Zz instead of Z.
2021 2022-10-16 Jeff Law <jeffreyalaw@gmail.com>
2023 * config/h8300/h8300.cc (h8300_register_move_cost): Fix typo.
2025 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2027 * value-range.cc (frange::set): Implement distinction between
2028 HONOR_SIGNED_ZEROS and MODE_HAS_SIGNED_ZEROS.
2030 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2032 * gimple-range-op.cc (class cfn_copysign): New.
2033 (gimple_range_op_handler::maybe_builtin_call): Add
2034 CFN_BUILT_IN_COPYSIGN*.
2036 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2038 * real.h (real_isdenormal): Check rvc_normal.
2039 * value-range.cc (range_tests_floats): New test.
2041 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2043 * gimple-range-op.cc
2044 (gimple_range_op_handler::maybe_builtin_call): Replace
2045 CFN_BUILTIN_SIGNBIT* cases with CASE_FLT_FN.
2047 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2049 * value-range.cc (frange::set): Normalize ranges for both bounds.
2051 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2053 * value-range.cc (frange::set): Drop -0.0 for !HONOR_SIGNED_ZEROS.
2055 2022-10-14 Aldy Hernandez <aldyh@redhat.com>
2057 * gimple-range-op.cc
2058 (gimple_range_op_handler::maybe_builtin_call): Add
2059 CFN_BUILT_IN_SIGNBIT[FL]* entries.
2061 2022-10-14 Richard Biener <rguenther@suse.de>
2063 PR tree-optimization/107254
2064 * tree-vect-slp.cc (vect_slp_analyze_node_operations_1):
2065 For permutes also analyze live lanes.
2066 (vect_schedule_slp_node): For permutes also code generate
2069 2022-10-14 Eric Botcazou <ebotcazou@adacore.com>
2072 * config/sparc/sparc.cc (sparc_expand_prologue): Emit a frame
2073 blockage for leaf functions.
2074 (sparc_flat_expand_prologue): Emit frame instead of full blockage.
2075 (sparc_expand_epilogue): Emit a frame blockage for leaf functions.
2076 (sparc_flat_expand_epilogue): Emit frame instead of full blockage.
2078 2022-10-14 Jakub Jelinek <jakub@redhat.com>
2080 * tree-core.h (enum tree_index): Add TI_BFLOAT16_TYPE.
2081 * tree.h (bfloat16_type_node): Define.
2082 * tree.cc (excess_precision_type): Promote bfloat16_type_mode
2083 like float16_type_mode.
2084 (build_common_tree_nodes): Initialize bfloat16_type_node if
2085 BFmode is supported.
2086 * expmed.h (maybe_expand_shift): Declare.
2087 * expmed.cc (maybe_expand_shift): No longer static.
2088 * expr.cc (convert_mode_scalar): Don't ICE on BF -> HF or HF -> BF
2089 conversions. If there is no optab, handle BF -> {DF,XF,TF,HF}
2090 conversions as separate BF -> SF -> {DF,XF,TF,HF} conversions, add
2091 -ffast-math generic implementation for BF -> SF and SF -> BF
2093 * builtin-types.def (BT_BFLOAT16, BT_FN_BFLOAT16_CONST_STRING): New.
2094 * builtins.def (BUILT_IN_NANSF16B): New builtin.
2095 * fold-const-call.cc (fold_const_call): Handle CFN_BUILT_IN_NANSF16B.
2096 * config/i386/i386.cc (classify_argument): Handle E_BCmode.
2097 (ix86_libgcc_floating_mode_supported_p): Also return true for BFmode
2099 (ix86_mangle_type): Mangle BFmode as DF16b.
2100 (ix86_invalid_conversion, ix86_invalid_unary_op,
2101 ix86_invalid_binary_op): Remove.
2102 (TARGET_INVALID_CONVERSION, TARGET_INVALID_UNARY_OP,
2103 TARGET_INVALID_BINARY_OP): Don't redefine.
2104 * config/i386/i386-builtins.cc (ix86_bf16_type_node): Remove.
2105 (ix86_register_bf16_builtin_type): Use bfloat16_type_node rather than
2106 ix86_bf16_type_node, only create it if still NULL.
2107 * config/i386/i386-builtin-types.def (BFLOAT16): Likewise.
2108 * config/i386/i386.md (cbranchbf4, cstorebf4): New expanders.
2110 2022-10-14 Jakub Jelinek <jakub@redhat.com>
2114 * doc/invoke.texi (-fexcess-precision=standard): Mention that the
2115 option now also works in C++.
2117 2022-10-13 Eric Botcazou <ebotcazou@adacore.com>
2119 * gimple-ssa-warn-access.cc (pass_waccess::check_call): Return
2120 early for calls made from thunks.
2122 2022-10-13 Eric Botcazou <ebotcazou@adacore.com>
2124 * expr.cc (emit_group_stote): Fix handling of modes of different
2125 sizes for big-endian targets in latest change and add commentary.
2127 2022-10-13 Martin Liska <mliska@suse.cz>
2129 * output.h (assemble_vtv_preinit_initializer): Remove.
2130 * varasm.cc (assemble_vtv_preinit_initializer): Remove.
2132 2022-10-13 Andrew MacLeod <amacleod@redhat.com>
2134 PR tree-optimization/102540
2135 PR tree-optimization/102872
2136 * gimple-range-cache.cc (ranger_cache::fill_block_cache):
2137 Handle partial equivs.
2138 (ranger_cache::range_from_dom): Cleanup dump output.
2140 2022-10-13 Andrew MacLeod <amacleod@redhat.com>
2142 * range-op.cc (operator_cast::lhs_op1_relation): New.
2143 (operator_bitwise_and::lhs_op1_relation): New.
2145 2022-10-13 Andrew MacLeod <amacleod@redhat.com>
2147 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Use
2150 (equiv_relation_iterator::equiv_relation_iterator): New.
2151 (equiv_relation_iterator::next): New.
2152 (equiv_relation_iterator::get_name): New.
2153 * value-relation.h (class relation_oracle): Privatize some methods.
2154 (class equiv_relation_iterator): New.
2155 (FOR_EACH_EQUIVALENCE): New.
2156 (FOR_EACH_PARTIAL_EQUIV): New.
2157 (FOR_EACH_PARTIAL_AND_FULL_EQUIV): New.
2159 2022-10-13 Andrew MacLeod <amacleod@redhat.com>
2161 * value-relation.cc (equiv_chain::dump): Don't print empty
2163 (equiv_oracle::equiv_oracle): Allocate a partial equiv table.
2164 (equiv_oracle::~equiv_oracle): Release the partial equiv table.
2165 (equiv_oracle::add_partial_equiv): New.
2166 (equiv_oracle::partial_equiv_set): New.
2167 (equiv_oracle::partial_equiv): New.
2168 (equiv_oracle::query_relation): Check for partial equivs too.
2169 (equiv_oracle::dump): Also dump partial equivs.
2170 (dom_oracle::register_relation): Handle partial equivs.
2171 (dom_oracle::query_relation): Check for partial equivs.
2172 * value-relation.h (enum relation_kind_t): Add partial equivs.
2173 (relation_partial_equiv_p): New.
2174 (relation_equiv_p): New.
2175 (class pe_slice): New.
2176 (class equiv_oracle): Add prototypes.
2181 2022-10-13 Richard Biener <rguenther@suse.de>
2183 PR tree-optimization/107247
2184 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
2185 Reduce multi vector SLP reduction accumulators. Check
2186 the adjusted number of accumulator vectors against
2187 one for the re-use in the epilogue.
2189 2022-10-13 Jakub Jelinek <jakub@redhat.com>
2191 * genmodes.cc (emit_mode_wider): Emit previous content of
2192 mode_wider array into mode_next array and for mode_wider
2193 emit always VOIDmode for !CLASS_HAS_WIDER_MODES_P classes,
2194 otherwise skip through modes with the same precision.
2195 * machmode.h (mode_next): Declare.
2196 (GET_MODE_NEXT_MODE): New inline function.
2197 (mode_iterator::get_next, mode_iterator::get_known_next): New
2199 (FOR_EACH_MODE_IN_CLASS): Use get_next instead of get_wider.
2200 (FOR_EACH_MODE): Use get_known_next instead of get_known_wider.
2201 (FOR_EACH_MODE_FROM): Use get_next instead of get_wider.
2202 (FOR_EACH_WIDER_MODE_FROM): Define.
2203 (FOR_EACH_NEXT_MODE): Define.
2204 * expmed.cc (emit_store_flag_1): Use FOR_EACH_WIDER_MODE_FROM
2205 instead of FOR_EACH_MODE_FROM.
2206 * optabs.cc (prepare_cmp_insn): Likewise. Remove redundant
2207 !CLASS_HAS_WIDER_MODES_P check.
2208 (prepare_float_lib_cmp): Use FOR_EACH_WIDER_MODE_FROM instead of
2210 * config/i386/i386-expand.cc (get_mode_wider_vector): Use
2211 GET_MODE_NEXT_MODE instead of GET_MODE_WIDER_MODE.
2213 2022-10-13 Wilco Dijkstra <wdijkstr@arm.com>
2216 * config/aarch64/aarch64.cc (aarch64_select_cc_mode): Allow
2217 GT/LE for merging compare with zero into AND.
2218 (aarch64_get_condition_code_1): Add CC_NZVmode support.
2219 * config/aarch64/aarch64-modes.def: Add CC_NZV.
2220 * config/aarch64/aarch64.md: Use CC_NZV in cmp+and patterns.
2222 2022-10-13 Richard Biener <rguenther@suse.de>
2224 PR tree-optimization/107160
2225 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
2226 Do not register accumulator if we failed to reduce it
2229 2022-10-13 Aldy Hernandez <aldyh@redhat.com>
2231 * range-op-float.cc (range_operator_float::op1_op2_relation): New.
2232 (class foperator_equal): Add using.
2233 (class foperator_not_equal): Same.
2234 (class foperator_lt): Same.
2235 (class foperator_le): Same.
2236 (class foperator_gt): Same.
2237 (class foperator_ge): Same.
2238 * range-op.cc (range_op_handler::op1_op2_relation): New.
2239 * range-op.h (range_operator_float::op1_op2_relation): New.
2241 2022-10-13 Richard Biener <rguenther@suse.de>
2243 * genmatch.cc (parser::parse_c_expr): Diagnose 'return'.
2244 * match.pd: Replace 'return' statements in with expressions
2245 with appropriate variants.
2247 2022-10-13 Andre Vieira <andre.simoesdiasvieira@arm.com>
2249 PR tree-optimization/107229
2250 * tree-if-conv.cc (get_bitfield_rep): Fix bitposition calculation.
2252 2022-10-13 Andre Vieira <andre.simoesdiasvieira@arm.com>
2254 PR tree-optimization/107226
2255 * tree-vect-patterns.cc (vect_recog_bitfield_ref_pattern): Reject
2256 BITFIELD_REF's with non integral typed first operands.
2258 2022-10-13 Joseph Myers <joseph@codesourcery.com>
2260 * ginclude/float.h (FLT_IS_IEC_60559, DBL_IS_IEC_60559)
2261 (LDBL_IS_IEC_60559): Update comment.
2263 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2265 * range-op-float.cc (class foperator_negate): New.
2266 (floating_op_table::floating_op_table): Add NEGATE_EXPR
2267 (range_op_float_tests): Add negate tests.
2269 2022-10-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2271 * config/riscv/riscv-vector-builtins.h: Remove unused macro.
2273 2022-10-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2275 * config/riscv/riscv-vector-builtins.cc (DEF_RVV_TYPE): Apply
2277 (add_vector_type_attribute): Ditto.
2278 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Apply
2280 * config/riscv/riscv-vector-builtins.h (DEF_RVV_TYPE): Apply
2283 2022-10-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2285 * config/riscv/riscv-vector-builtins.cc (builtin_types): Redefine
2287 (build_const_pointer): New function.
2288 (register_builtin_type): Ditto.
2289 (DEF_RVV_TYPE): Simplify macro.
2290 (register_vector_type): Refine implementation.
2291 * config/riscv/riscv-vector-builtins.h (rvv_builtin_types_t): New.
2293 2022-10-12 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2295 * config/riscv/riscv-vector-builtins.h (class rvv_switcher): Move to
2297 * config/riscv/riscv-vector-builtins.cc (class rvv_switcher):
2300 2022-10-12 Cui,Lili <lili.cui@intel.com>
2302 * config/i386/driver-i386.cc (host_detect_local_cpu):
2303 Move sapphirerapids out of AVX512_VP2INTERSECT.
2304 * config/i386/i386.h: Remove AVX512_VP2INTERSECT from PTA_SAPPHIRERAPIDS
2305 * doc/invoke.texi: Remove AVX512_VP2INTERSECT from SAPPHIRERAPIDS
2307 2022-10-12 Martin Liska <mliska@suse.cz>
2309 * gcov-io.cc (gcov_write_summary): Rename to ...
2310 (gcov_write_object_summary): ... this.
2311 * gcov-io.h (GCOV_TAG_OBJECT_SUMMARY_LENGTH): Rename from ...
2312 (GCOV_TAG_SUMMARY_LENGTH): ... this.
2314 2022-10-12 Martin Liska <mliska@suse.cz>
2316 * configure: Regenerate.
2318 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2320 * range-op-float.cc (frange_float): New.
2321 (range_op_float_tests): New.
2322 * range-op.cc (range_op_tests): Call range_op_float_tests.
2324 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2326 * value-range.h (frange::nan_signbit_p): New.
2328 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2330 * value-range.cc (frange::set_nonnegative): Pass bool to
2332 * value-range.h: Disallow conversion to bool in update_nan().
2334 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2336 * value-range.h (frange::frange): Add constructor taking type.
2338 2022-10-12 Aldy Hernandez <aldyh@redhat.com>
2340 * range-op-float.cc: Add relation_kind = VREL_VARYING to all
2343 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2345 * gimple-range-gori.cc (gori_compute::logical_combine): Avoid
2346 calling tracer.trailer().
2348 2022-10-11 Jakub Jelinek <jakub@redhat.com>
2351 * config/i386/i386.md (*notxor<mode>_1): Use MASK_REG_P (x) instead of
2352 MASK_REGNO_P (REGNO (x)).
2354 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2356 * range-op-float.cc (class foperator_abs): New.
2357 (floating_op_table::floating_op_table): Add ABS_EXPR entry.
2359 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2361 * range-op-float.cc (foperator_unordered_le::op1_range): New.
2362 (foperator_unordered_le::op2_range): New.
2363 (foperator_unordered_gt::op1_range): New.
2364 (foperator_unordered_gt::op2_range): New.
2365 (foperator_unordered_ge::op1_range): New.
2366 (foperator_unordered_ge::op2_range): New.
2367 (foperator_unordered_equal::op1_range): New.
2369 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2371 * range-op-float.cc (class foperator_unordered_lt): New.
2372 (class foperator_relop_unknown): Remove
2373 (class foperator_unordered_le): New.
2374 (class foperator_unordered_gt): New.
2375 (class foperator_unordered_ge): New.
2376 (class foperator_unordered_equal): New.
2377 (floating_op_table::floating_op_table): Replace all UN_EXPR
2378 entries with their appropriate fop_unordered_* counterpart.
2380 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2382 * range-op.cc (operator_equal::op1_range): Move BRS_TRUE case up.
2383 (operator_lt::op2_range): Same.
2384 (operator_le::op2_range): Same.
2385 (operator_gt::op2_range): Same.
2386 (operator_ge::op2_range): Same.
2388 2022-10-11 Richard Biener <rguenther@suse.de>
2390 PR tree-optimization/107212
2391 * tree-vect-loop.cc (vectorizable_reduction): Make sure to
2392 set STMT_VINFO_REDUC_DEF for all live lanes in a SLP
2394 (vectorizable_live_operation): Do not pun to the SLP
2395 node representative for reduction epilogue generation.
2397 2022-10-11 Andrew Stubbs <ams@codesourcery.com>
2399 * config/gcn/gcn-valu.md (neg<mode>2): New define_expand.
2401 2022-10-11 Andrew Stubbs <ams@codesourcery.com>
2403 * config/gcn/gcn-valu.md (vec_init<V_ALL:mode><V_ALL_ALT:mode>): New.
2404 * config/gcn/gcn.cc (GEN_VN): Add andvNsi3, subvNsi3.
2405 (GEN_VNM): Add gathervNm_expr.
2406 (GEN_VN_NOEXEC): Add vec_seriesvNsi.
2407 (gcn_expand_vector_init): Add initialization of vectors from smaller
2410 2022-10-11 Andrew Stubbs <ams@codesourcery.com>
2412 * config/gcn/gcn-protos.h (get_exec): Add prototypes for two variants.
2413 * config/gcn/gcn-valu.md
2414 (vec_extract<V_ALL:mode><V_ALL_ALT:mode>): New define_expand.
2415 * config/gcn/gcn.cc (get_exec): Export the existing function. Add a
2416 new overload variant.
2418 2022-10-11 Andrew Stubbs <ams@codesourcery.com>
2420 * config/gcn/gcn-valu.md
2421 (<cvt_name><VCVT_MODE:mode><VCVT_FMODE:mode>2<exec>): Use MODE_VF.
2422 (<cvt_name><VCVT_FMODE:mode><VCVT_IMODE:mode>2<exec>): Likewise.
2423 * config/gcn/gcn.h (MODE_VF): New macro.
2425 2022-10-11 Andrew Stubbs <ams@codesourcery.com>
2427 * config/gcn/gcn-modes.def (VECTOR_MODE): Add new modes
2428 V32QI, V32HI, V32SI, V32DI, V32TI, V32HF, V32SF, V32DF,
2429 V16QI, V16HI, V16SI, V16DI, V16TI, V16HF, V16SF, V16DF,
2430 V8QI, V8HI, V8SI, V8DI, V8TI, V8HF, V8SF, V8DF,
2431 V4QI, V4HI, V4SI, V4DI, V4TI, V4HF, V4SF, V4DF,
2432 V2QI, V2HI, V2SI, V2DI, V2TI, V2HF, V2SF, V2DF.
2433 (ADJUST_ALIGNMENT): Likewise.
2434 * config/gcn/gcn-protos.h (gcn_full_exec): Delete.
2435 (gcn_full_exec_reg): Delete.
2436 (gcn_scalar_exec): Delete.
2437 (gcn_scalar_exec_reg): Delete.
2438 (vgpr_1reg_mode_p): Use inner mode to identify vector registers.
2439 (vgpr_2reg_mode_p): Likewise.
2440 (vgpr_vector_mode_p): Use VECTOR_MODE_P.
2441 * config/gcn/gcn-valu.md (V_QI, V_HI, V_HF, V_SI, V_SF, V_DI, V_DF,
2442 V_QIHI, V_1REG, V_INT_1REG, V_INT_1REG_ALT, V_FP_1REG, V_2REG, V_noQI,
2443 V_noHI, V_INT_noQI, V_INT_noHI, V_ALL, V_ALL_ALT, V_INT, V_FP):
2444 Add additional vector modes.
2445 (V64_SI, V64_DI, V64_ALL, V64_FP): New iterators.
2446 (scalar_mode, SCALAR_MODE, vnsi, VnSI, vndi, VnDI, sdwa):
2447 Add additional vector mode mappings.
2448 (mov<mode>): Implement vector length conversions.
2449 (ldexp<mode>3<exec>): Use VnSI.
2450 (frexp<mode>_exp2<exec>): Likewise.
2451 (VCVT_MODE, VCVT_FMODE, VCVT_IMODE): Add additional vector modes.
2452 (reduc_<reduc_op>_scal_<mode>): Use V64_ALL.
2453 (fold_left_plus_<mode>): Use V64_FP.
2454 (*<reduc_op>_dpp_shr_<mode>): Use V64_1REG.
2455 (*<reduc_op>_dpp_shr_<mode>): Use V64_DI.
2456 (*plus_carry_dpp_shr_<mode>): Use V64_INT_1REG.
2457 (*plus_carry_in_dpp_shr_<mode>): Use V64_SI.
2458 (*plus_carry_dpp_shr_<mode>): Use V64_DI.
2459 (mov_from_lane63_<mode>): Use V64_2REG.
2460 * config/gcn/gcn.cc (VnMODE): New function.
2461 (gcn_can_change_mode_class): Support multiple vector sizes.
2462 (gcn_modes_tieable_p): Likewise.
2463 (gcn_operand_part): Likewise.
2464 (gcn_scalar_exec): Delete function.
2465 (gcn_scalar_exec_reg): Delete function.
2466 (gcn_full_exec): Delete function.
2467 (gcn_full_exec_reg): Delete function.
2468 (gcn_inline_fp_constant_p): Support multiple vector sizes.
2469 (gcn_fp_constant_p): Likewise.
2471 (GEN_VN_NOEXEC): New macro.
2472 (GEN_VNM_NOEXEC): New macro.
2473 (GEN_VN): New macro.
2474 (GEN_VNM): New macro.
2475 (GET_VN_FN): New macro.
2476 (CODE_FOR): New macro.
2477 (CODE_FOR_OP): New macro.
2478 (gen_mov_with_exec): Delete function.
2479 (gen_duplicate_load): Delete function.
2480 (gcn_expand_vector_init): Support multiple vector sizes.
2481 (strided_constant): Likewise.
2482 (gcn_addr_space_legitimize_address): Likewise.
2483 (gcn_expand_scalar_to_vector_address): Likewise.
2484 (gcn_expand_scaled_offsets): Likewise.
2485 (gcn_secondary_reload): Likewise.
2486 (gcn_valid_cvt_p): Likewise.
2487 (gcn_expand_builtin_1): Likewise.
2488 (gcn_make_vec_perm_address): Likewise.
2489 (gcn_vectorize_vec_perm_const): Likewise.
2490 (gcn_vector_mode_supported_p): Likewise.
2491 (gcn_autovectorize_vector_modes): New hook.
2492 (gcn_related_vector_mode): Support multiple vector sizes.
2493 (gcn_expand_dpp_shr_insn): Add FIXME comment.
2494 (gcn_md_reorg): Support multiple vector sizes.
2495 (print_reg): Likewise.
2496 (print_operand): Likewise.
2497 (TARGET_VECTORIZE_AUTOVECTORIZE_VECTOR_MODES): New hook.
2499 2022-10-11 Andre Vieira <andre.simoesdiasvieira@arm.com>
2501 * tree-if-conv.cc (if_convertible_loop_p_1): Move ordering of loop bb's from
2503 (tree_if_conversion): ... to here. Also call bitfield lowering when
2505 (version_loop_for_if_conversion): Adapt to enable loop versioning when we only
2506 need to lower bitfields.
2507 (ifcvt_split_critical_edges): Relax condition of expected loop form as this is
2509 (get_bitfield_rep): New function.
2510 (lower_bitfield): Likewise.
2511 (bitfields_to_lower_p): Likewise.
2512 (need_to_lower_bitfields): New global boolean.
2513 (need_to_ifcvt): Likewise.
2514 * tree-vect-data-refs.cc (vect_find_stmt_data_reference): Improve diagnostic
2516 * tree-vect-patterns.cc (vect_recog_temp_ssa_var): Add default value for last
2518 (vect_recog_bitfield_ref_pattern): New.
2519 (vect_recog_bit_insert_pattern): New.
2521 2022-10-11 liuhongt <hongtao.liu@intel.com>
2524 * config/i386/i386.md (*notxor<mode>_1): New post_reload
2525 define_insn_and_split.
2526 (*notxorqi_1): Ditto.
2528 2022-10-11 Aldy Hernandez <aldyh@redhat.com>
2530 PR tree-optimization/107195
2531 * value-range.cc (irange::set_range_from_nonzero_bits): Set range
2532 to [0,0] when nonzero mask is 0.
2534 2022-10-11 Olivier Hainque <hainque@adacore.com>
2535 Olivier Hainque <hainque@adacore.com>
2537 * configure: Regenerate.
2539 2022-10-11 Olivier Hainque <hainque@adacore.com>
2541 * config.gcc (*vxworks*): Add t-slibgcc fragment
2544 2022-10-11 Olivier Hainque <hainque@adacore.com>
2546 * config/vxworks.h (VX_LGCC_EH_SO0, VX_LGCC_EH_SO1): New
2548 (VXWORKS_LIBGCC_SPEC): Use them and document.
2550 2022-10-11 Martin Liska <mliska@suse.cz>
2552 * gimple-range-op.cc: Add override keyword.
2554 2022-10-11 Eugene Rozenfeld <erozen@microsoft.com>
2557 * tree-cfg.cc (assign_discriminators): Move declaration of cur_locus_e
2560 2022-10-11 Liwei Xu <liwei.xu@intel.com>
2561 liuhongt <hongtao.liu@intel.com>
2563 PR tree-optimization/54346
2564 * match.pd: Merge the index of VCST then generates the new vec_perm.
2566 2022-10-11 Jeff Law <jeffreyalaw@gmail.com>
2568 PR rtl-optimization/107182
2569 * cfgrtl.cc (fixup_reorder_chain): When optimizing a jump to a
2570 return, clear EDGE_CROSSING on the appropriate edge.
2572 2022-10-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2574 * config/riscv/riscv-vector-builtins.cc (struct vector_type_info): Move
2575 from config/riscv/riscv-vector-builtins.h.
2576 (DEF_RVV_TYPE): Change USER_NAME to NAME.
2577 (register_vector_type): Change user_name to name.
2578 * config/riscv/riscv-vector-builtins.def (DEF_RVV_TYPE): Change
2580 * config/riscv/riscv-vector-builtins.h (struct vector_type_info): Move
2581 to riscv-vector-builtins.cc.
2582 (DEF_RVV_TYPE): Change USER_NAME to NAME.
2584 2022-10-11 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
2586 * config/riscv/riscv.md: Add vsetvl instruction type.
2588 2022-10-10 Nathan Sidwell <nathan@acm.org>
2590 * common.opt (-fabi-version=): Document 18.
2591 * doc/invoke.texi (-fabi-version): Document 18.
2593 2022-10-10 Andrea Corallo <andrea.corallo@arm.com>
2596 * toplev.cc (toplev::main): Don't run self tests in case of
2599 2022-10-10 Kito Cheng <kito.cheng@sifive.com>
2601 * config/riscv/riscv-c.cc: Add newline to the end of file.
2603 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2605 * range-op-float.cc (class foperator_identity): Make members public.
2606 (class foperator_equal): Same.
2607 (class foperator_not_equal): Same.
2608 (class foperator_lt): Same.
2609 (class foperator_le): Same.
2610 (class foperator_gt): Same.
2611 (class foperator_ge): Same.
2612 (class foperator_unordered): Same.
2613 (class foperator_ordered): Same.
2615 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2617 * value-range.h (frange::maybe_isnan): New.
2619 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2621 * range-op-float.cc (foperator_not_equal::op1_range): Set NAN on
2622 TRUE side for x != x.
2624 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2626 * range-op-float.cc (foperator_unordered::op1_range): Set NAN when
2627 operands are equal and result is TRUE.
2629 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2631 * range.h (range_true): Return int_range<2>.
2632 (range_false): Same.
2633 (range_true_and_false): Same.
2635 2022-10-10 Aldy Hernandez <aldyh@redhat.com>
2637 * gimple-range-op.cc: Add op1_range entry for __builtin_signbit.
2639 2022-10-10 liuhongt <hongtao.liu@intel.com>
2642 * config/i386/i386.md (lrint<MODEF:mode><SWI48:mode>2): Swap
2643 predicate of operands[0] and operands[1].
2645 2022-10-10 Claudiu Zissulescu <claziss@gmail.com>
2647 * common/config/arc/arc-common.cc (arc_option_optimization_table):
2648 Remove Rcq and Rcw options.
2649 * config/arc/arc.opt (mRcq): Ignore option, preserve it for
2650 backwards compatibility.
2652 * doc/invoke.texi (mRcw, mRcq): Update document.
2654 2022-10-10 Claudiu Zissulescu <claziss@gmail.com>
2656 * config/arc/arc.cc (arc_check_short_reg_p): New function.
2657 (arc_address_cost): Replace satisfies_constraint_Rcq with the
2659 (arc_output_addsi): Likewise.
2660 (split_addsi): Likewise.
2661 (split_subsi): Likewise.
2662 * config/arc/arc.md (movqi_insn): Remove Rcq constraint.
2663 (movhi_insn): Likewise.
2664 (movsi_insn): Likewise.
2665 (tst_movb): Likewise.
2667 (tst_bitfield): Likewise.
2669 (addsi3_mixed): Likewise.
2670 (mulhisi3_reg): Likewise.
2671 (umulhisi3_reg): Likewise.
2672 (mulsi_600): Likewise.
2674 (subsi3_insn): Likewise.
2675 (bicsi3_insn): Likewise.
2678 (one_cmplsi2): Likewise.
2679 (lshrsi3_insn): Likewise.
2680 (cmpsi_cc_insn_mixed): Likewise.
2681 (cmpsi_cc_zn_insn): Likewise.
2683 (cmpsi_cc_z_insn): Likewise.
2684 (cmpsi_cc_c_insn): Likewise.
2685 (indirect_jump): Likewise.
2686 (casesi_jump): Likewise.
2688 (call_value_i): Likewise.
2691 (ashlsi2_cnt1): Likewise.
2692 (lshrsi3_cnt1): Likewise.
2693 (ashrsi3_cnt1): Likewise.
2694 * config/arc/constraints.md (Rcq): Remove.
2696 2022-10-10 Claudiu Zissulescu <claziss@gmail.com>
2698 * config/arc/arc.md (smaxsi3): Remove Rcw.
2699 (sminsi3): Likewise.
2700 (addsi3_mixed): Likewise.
2701 (add_f_2): Likewise.
2702 (subsi3_insn): Likewise.
2708 (bset_insn): Likewise.
2709 (bxor_insn): Likewise.
2710 (bclr_insn): Likewise.
2711 (bmsk_insn): Likewise.
2712 (bicsi3_insn): Likewise.
2715 (lshrsi3_insn): Likewise.
2718 * config/arc/constraints.md(Rcw): Remove it.
2720 2022-10-10 Claudiu Zissulescu <claziss@gmail.com>
2722 * config/arc/arc.md(mulsi3_700): Remove Rcr.
2723 (mulsi3_highpart): Likewise.
2724 (umulsi3_highpart_i): Likewise.
2725 (umulsi3_highpart_int): Likewise.
2728 * config/arc/constraints.md (Rcr): Remove it.
2730 2022-10-10 Claudiu Zissulescu <claziss@gmail.com>
2732 * config/arc/arc.cc (arc_save_callee_enter): Use negative offsets.
2734 2022-10-09 Dimitar Dimitrov <dimitar@dinux.eu>
2737 * config/pru/pru-protos.h (pru_noteq_condition): New
2738 function declaration.
2739 * config/pru/pru.cc (pru_noteq_condition): New function.
2740 * config/pru/pru.md (cbranchdi4): Define new pattern.
2742 2022-10-09 Dimitar Dimitrov <dimitar@dinux.eu>
2744 * config/pru/pru.md (lshrdi3): New expand pattern.
2747 2022-10-09 YunQiang Su <yunqiang.su@cipunited.com>
2749 * config.gcc: set with_arch to default_mips_arch if no defined.
2750 * config/mips/driver-native.cc (host_detect_local_cpu):
2751 try getauxval(AT_BASE_PLATFORM) and _MIPS_ARCH, too.
2752 pass -mnan=2008 if __mips_nan2008__ is defined.
2753 * config.in: define HAVE_SYS_AUXV_H and HAVE_GETAUXVAL.
2754 * configure.ac: detect sys/auxv.h and getauxval.
2755 * configure: regenerated.
2757 2022-10-07 Eugene Rozenfeld <erozen@microsoft.com>
2759 * tree-cfg.cc (assign_discriminators): Set discriminators for call stmts
2760 on the same line within the same basic block.
2762 2022-10-07 Qing Zhao <qing.zhao@oracle.com>
2764 PR tree-optimization/101836
2765 * tree-object-size.cc (addr_object_size): Use array_at_struct_end_p
2766 to determine a flexible array member reference.
2768 2022-10-07 Qing Zhao <qing.zhao@oracle.com>
2770 * doc/extend.texi: Document strict_flex_array attribute.
2771 * doc/invoke.texi: Document -fstrict-flex-arrays[=n] option.
2772 * print-tree.cc (print_node): Print new bit decl_not_flexarray.
2773 * tree-core.h (struct tree_decl_common): New bit field
2775 * tree-streamer-in.cc (unpack_ts_decl_common_value_fields): Stream
2776 in new bit decl_not_flexarray.
2777 * tree-streamer-out.cc (pack_ts_decl_common_value_fields): Stream
2778 out new bit decl_not_flexarray.
2779 * tree.cc (array_at_struct_end_p): Update it with the new bit field
2781 * tree.h (DECL_NOT_FLEXARRAY): New flag.
2783 2022-10-07 Olivier Hainque <hainque@adacore.com>
2785 * config/vxworks/_vxworks-versions.h: Use OS specific
2786 paths in #include of version.h.
2788 2022-10-07 Martin Liska <mliska@suse.cz>
2790 * opts.cc (finish_options): Print sorry message only
2791 for -flive-patching=inline-only-static.
2793 2022-10-07 Jason Merrill <jason@redhat.com>
2795 * gimplify.cc (gimplify_modify_expr_rhs): Don't optimize
2796 x = *(A*)&<expr> to x = <expr> for a TREE_ADDRESSABLE type.
2798 2022-10-07 Richard Biener <rguenther@suse.de>
2800 PR tree-optimization/107153
2801 * tree-cfg.cc (gimple_duplicate_sese_tail): Do not update
2803 * tree-parloops.cc (gen_parallel_loop): Update SSA form
2804 after to-exit-first transform, no PHI insertion is necessary.
2805 (pass_parallelize_loops::execute): Force re-write of the
2806 virtual operand SSA web.
2808 2022-10-07 Jonathan Wakely <jwakely@redhat.com>
2810 * value-range.cc (irange::irange_contains_p): Fix comment typo.
2812 2022-10-07 Eric Botcazou <ebotcazou@adacore.com>
2814 * function.cc (thread_prologue_and_epilogue_insns): Update only
2815 entry and exit blocks when not optimizing. Remove dead statement.
2817 2022-10-07 Aldy Hernandez <aldyh@redhat.com>
2819 * value-range.cc (irange::irange_set): Convert nonzero mask to
2821 (irange::irange_set_anti_range): Same.
2822 (irange::set): Same.
2823 (irange::verify_range): Same.
2824 (irange::contains_p): Same.
2825 (irange::invert): Same.
2826 (irange::set_range_from_nonzero_bits): Same.
2827 (irange::set_nonzero_bits): Same.
2829 (irange::intersect_nonzero_bits): Same.
2830 (irange::union_nonzero_bits): Same.
2831 * value-range.h (irange::varying_compatible_p): Same.
2834 (irange::set_undefined): Same.
2835 (irange::set_varying): Same.
2837 2022-10-07 Martin Liska <mliska@suse.cz>
2839 * config/i386/i386-protos.h (ix86_binary_operator_ok): Add array
2840 size to function parameter.
2841 (ix86_unary_operator_ok): Likewise.
2843 2022-10-07 Martin Liska <mliska@suse.cz>
2845 * auto-profile.cc (get_inline_stack): Remove unused variable.
2847 2022-10-07 Jakub Jelinek <jakub@redhat.com>
2849 * ipa-prop.h (ipa_constant_data): Fix comment typo.
2850 * value-range.cc (irange::irange_contains_p): Likewise.
2851 * value-relation.cc (dom_oracle::set_one_relation): Likewise.
2852 * gimple-predicate-analysis.cc (predicate::simplify_4): Likewise.
2853 * tree-inline.cc (remap_ssa_name): Likewise.
2855 2022-10-07 Jakub Jelinek <jakub@redhat.com>
2857 * attribs.h (is_attribute_namespace_p): New inline function.
2858 (lookup_attribute): Document meaning of ATTR_NS equal to "".
2859 * attribs.cc (remove_attribute): Use is_attribute_namespace_p.
2860 (private_lookup_attribute): For ATTR_NS "" match either standard
2861 attribute or "gnu" namespace one.
2863 2022-10-07 Olivier Hainque <hainque@adacore.com>
2865 * config/vxworks.h (DWARF_VERSION_DEFAULT): Adjust from
2866 4 to 3 for VxWorks >= 7.
2868 2022-10-07 Olivier Hainque <hainque@adacore.com>
2870 * defaults.h (DWARF_DEFAULT_VERSION): Define if not
2872 * common.opt (gdwarf-): Use it.
2873 * doc/tm.texi.in (DWARF_DEFAULT_VERSION): Document.
2874 * doc/tm.texi: Update accordingly.
2875 * config/vxworks.h (DWARF_DEFAULT_VERSION): Redefine.
2876 * config/vxworks.cc: Remove code setting dwarf_version, now
2877 handled by the DWARF_DEFAULT_VERSION redefinition.
2879 2022-10-07 Olivier Hainque <hainque@adacore.com>
2881 * ginclude/stddef.h: #undef offsetof before #define.
2883 2022-10-06 Aldy Hernandez <aldyh@redhat.com>
2885 PR tree-optimization/107170
2886 * vr-values.cc (vr_values::range_of_expr): Do not die on
2889 2022-10-06 Joseph Myers <joseph@codesourcery.com>
2891 * doc/invoke.texi (-fno-asm): Update description of effects on
2894 2022-10-06 Eric Botcazou <ebotcazou@adacore.com>
2896 * gimple-loop-jam.cc (tree_loop_unroll_and_jam): Bail out for a self
2897 dependency that is a write-after-write if the access function is not
2900 2022-10-06 Eric Botcazou <ebotcazou@adacore.com>
2902 * df-scan.cc (df_ref_create_structure): Minor cleanup.
2904 2022-10-06 Richard Biener <rguenther@suse.de>
2906 PR middle-end/107115
2907 * expr.cc (store_expr): Check mems_same_for_tbaa_p before
2908 eliding a seemingly redundant store.
2910 2022-10-06 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
2912 * config/aarch64/atomics.md (*atomic_load<ALLX:mode>_zext<SD_HSDI:mode>):
2915 2022-10-06 Philipp Tomsich <philipp.tomsich@vrull.eu>
2917 * config/aarch64/aarch64-cores.def (AARCH64_CORE): Update
2918 Ampere-1 core entry.
2920 2022-10-06 Philipp Tomsich <philipp.tomsich@vrull.eu>
2922 * config/aarch64/driver-aarch64.cc (readline): Fix off-by-one.
2924 2022-10-06 Richard Biener <rguenther@suse.de>
2926 PR tree-optimization/107107
2927 * tree-ssa-sccvn.cc (visit_reference_op_store): Do not
2928 affect value-numbering when doing the tail merging
2931 2022-10-06 Claudiu Zissulescu <claziss@synopsys.com>
2933 * config/arc/linux.h (LINK_SPEC): Remove max-page-size and
2936 2022-10-06 Jakub Jelinek <jakub@redhat.com>
2939 * internal-fn.def (ASSUME): New internal function.
2940 * internal-fn.h (expand_ASSUME): Declare.
2941 * internal-fn.cc (expand_ASSUME): Define.
2942 * gimplify.cc (gimplify_call_expr): Gimplify IFN_ASSUME.
2943 * fold-const.h (simple_condition_p): Declare.
2944 * fold-const.cc (simple_operand_p_2): Rename to ...
2945 (simple_condition_p): ... this. Remove forward declaration.
2946 No longer static. Adjust function comment and fix a typo in it.
2947 Adjust recursive call.
2948 (simple_operand_p): Adjust function comment.
2949 (fold_truth_andor): Adjust simple_operand_p_2 callers to call
2951 * doc/extend.texi: Document assume attribute. Move fallthrough
2952 attribute example to its section.
2954 2022-10-06 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
2956 PR rtl-optimization/107088
2957 * cselib.cc (new_cselib_val): Skip BImode while keeping track of
2960 2022-10-06 Aldy Hernandez <aldyh@redhat.com>
2962 * value-range.cc (frange::set): Call set_nan unconditionally.
2963 (range_tests_nan): Adjust tests.
2964 (range_tests_signed_zeros): Same.
2965 (range_tests_floats): Same.
2966 * value-range.h (frange::update_nan): Guard with HONOR_NANS.
2967 (frange::set_nan): Set undefined if !HONOR_NANS.
2969 2022-10-06 Aldy Hernandez <aldyh@redhat.com>
2971 * range-op-float.cc (foperator_lt::fold_range): Remove extra check
2972 to finite_operands_p.
2973 (foperator_le::fold_range): Same.
2974 (foperator_gt::fold_range): Same.
2975 (foperator_ge::fold_range): Same.
2977 2022-10-06 Aldy Hernandez <aldyh@redhat.com>
2979 * value-range-pretty-print.cc (vrange_printer::print_real_value):
2980 Avoid printing INF and NAN twice.
2982 2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
2984 * config/rs6000/constraints.md (wD): Delete.
2985 * doc/md.texi (Machine Constraints): Adjust.
2987 2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
2989 * config/rs6000/vsx.md (vsx_extract_<mode>): Replace define_insn by a
2990 define_expand. Split the contents to...
2991 (*vsx_extract_<mode>_0): ... this. Rewrite.
2992 (*vsx_extract_<mode>_1): ... and this. Rewrite.
2994 2022-10-05 Segher Boessenkool <segher@kernel.crashing.org>
2996 * config/rs6000/vsx.md (*vsx_extract_<mode>_store): Use "n" instead of
2999 2022-10-05 David Malcolm <dmalcolm@redhat.com>
3002 * doc/analyzer.texi (__analyzer_get_unknown_ptr): Document.
3004 2022-10-05 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
3006 * config.gcc: Add riscv_vector.h.
3007 * config/riscv/riscv-builtins.cc: Add RVV builtin types support.
3008 * config/riscv/riscv-c.cc (riscv_pragma_intrinsic): New function.
3009 (riscv_register_pragmas): Ditto.
3010 * config/riscv/riscv-protos.h (riscv_register_pragmas): Ditto.
3011 (init_builtins): Move declaration from riscv-vector-builtins.h to riscv-protos.h.
3012 (mangle_builtin_type): Ditto.
3013 (verify_type_context): Ditto.
3014 (handle_pragma_vector): New function.
3015 * config/riscv/riscv-vector-builtins.cc (GTY): New variable.
3016 (register_vector_type): New function.
3017 (init_builtins): Add RVV builtin types support.
3018 (handle_pragma_vector): New function.
3019 * config/riscv/riscv-vector-builtins.h (GCC_RISCV_V_BUILTINS_H): Change
3020 name according to file name.
3021 (GCC_RISCV_VECTOR_BUILTINS_H): Ditto.
3022 (init_builtins): Remove declaration in riscv-vector-builtins.h.
3023 (mangle_builtin_type): Ditto.
3024 (verify_type_context): Ditto.
3025 * config/riscv/riscv.cc: Adjust for RVV builtin types support.
3026 * config/riscv/riscv.h (REGISTER_TARGET_PRAGMAS): New macro.
3027 * config/riscv/t-riscv: Remove redundant file including.
3028 * config/riscv/riscv_vector.h: New file.
3030 2022-10-05 Aldy Hernandez <aldyh@redhat.com>
3032 * range-op.cc (operator_cast::fold_range): Handle truncating casts
3035 2022-10-05 Martin Liska <mliska@suse.cz>
3038 * attribs.h (lookup_attribute_by_prefix): Support the attribute
3039 starting with underscore (_Noreturn).
3041 2022-10-05 Aldy Hernandez <aldyh@redhat.com>
3043 PR tree-optimization/107052
3044 * gimple-range-op.cc (cfn_popcount::fold_range): Take into account
3047 2022-10-05 Aldy Hernandez <aldyh@redhat.com>
3049 PR tree-optimization/107052
3050 * range-op.cc (operator_cast::fold_range): Set nonzero mask.
3052 2022-10-05 Eric Botcazou <ebotcazou@adacore.com>
3054 PR tree-optimization/106698
3055 * pointer-query.cc (handle_array_ref): Fix handling of low bound.
3057 2022-10-05 Vineet Gupta <vineetg@rivosinc.com>
3059 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins):
3060 Remove __riscv_cmodel_pic, that deprecated in last version.
3062 2022-10-05 David Malcolm <dmalcolm@redhat.com>
3065 * Makefile.in (ANALYZER_OBJS): Add analyzer/call-summary.o.
3067 2022-10-04 Jakub Jelinek <jakub@redhat.com>
3069 * attribs.h (remove_attribute): Declare overload with additional
3071 (private_lookup_attribute): Declare overload with additional
3072 attr_ns and attr_ns_len arguments.
3073 (lookup_attribute): New overload with additional attr_ns argument.
3074 * attribs.cc (remove_attribute): New overload with additional
3076 (private_lookup_attribute): New overload with additional
3077 attr_ns and attr_ns_len arguments.
3079 2022-10-04 Jakub Jelinek <jakub@redhat.com>
3081 * attribs.cc (handle_ignored_attributes_option, decl_attributes,
3082 common_function_versions): Use auto_diagnostic_group.
3084 2022-10-04 Aldy Hernandez <aldyh@redhat.com>
3086 * value-range.cc (irange::set_nonzero_bits): Remove assert.
3088 2022-10-04 Richard Sandiford <richard.sandiford@arm.com>
3090 * config/aarch64/aarch64.h (AARCH64_ISA_RCPC): New macro.
3091 * config/aarch64/aarch64-arches.def (armv8.3-a): Include RCPC.
3092 * config/aarch64/aarch64-cores.def (thunderx3t110, zeus, neoverse-v1)
3093 (neoverse-512tvb, saphira): Remove RCPC from these Armv8.3-A+ cores.
3094 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Define
3095 __ARM_FEATURE_RCPC when appropriate.
3097 2022-10-04 Tobias Burnus <tobias@codesourcery.com>
3099 * doc/invoke.texi (-fopenmp): Mention C++ attribut syntax.
3100 (-fopenmp-simd): Likewise; update permitted directives.
3102 2022-10-04 Tobias Burnus <tobias@codesourcery.com>
3104 * doc/install.texi (Specific): Add missing items to bullet list.
3105 (amdgcn): Update LLVM requirements, use version not date for newlib.
3106 (nvptx): Use version not git hash for newlib.
3108 2022-10-04 Aldy Hernandez <aldyh@redhat.com>
3110 * value-range-storage.cc (irange_storage_slot::set_irange): Remove
3112 * value-range.cc (irange::irange_set): Adjust for nonzero mask
3114 (irange::irange_set_anti_range): Same.
3115 (irange::set): Same.
3116 (irange::verify_range): Same.
3117 (irange::legacy_equal_p): Same.
3118 (irange::operator==): Same.
3119 (irange::contains_p): Same.
3120 (irange::legacy_intersect): Same.
3121 (irange::legacy_union): Same.
3122 (irange::irange_single_pair_union): Call union_nonzero_bits.
3123 (irange::irange_union): Same.
3124 (irange::irange_intersect): Call intersect_nonzero_bits.
3125 (irange::intersect): Adjust for nonzero mask being a wide int.
3126 (irange::invert): Same.
3127 (irange::set_nonzero_bits): Same.
3128 (irange::get_nonzero_bits_from_range): New.
3129 (irange::set_range_from_nonzero_bits): New.
3130 (irange::get_nonzero_bits): Adjust for nonzero mask being a wide
3132 (irange::intersect_nonzero_bits): Same.
3133 (irange::union_nonzero_bits): Same.
3134 (range_tests_nonzero_bits): Remove test.
3135 * value-range.h (irange::varying_compatible_p): Adjust for nonzero
3136 mask being a wide int.
3139 (irange::set_undefined): Same.
3140 (irange::set_varying): Same.
3141 (irange::normalize_kind): Same.
3143 2022-10-04 Aldy Hernandez <aldyh@redhat.com>
3145 PR tree-optimization/107130
3146 * gimple-range-op.cc (class cfn_popcount): Call op_cfn_ffs.
3147 (class cfn_ffs): New.
3148 (gimple_range_op_handler::maybe_builtin_call): Separate out
3149 CASE_CFN_FFS into its own case.
3151 2022-10-03 Sergei Trofimovich <siarheit@google.com>
3153 * config/i386/t-i386: Add build-time dependencies against
3154 i386-builtin-types.inc to i386-builtins.o, i386-expand.o,
3157 2022-10-03 Andrew Stubbs <ams@codesourcery.com>
3159 * config/gcn/gcn-valu.md (while_ultsidi): Limit mask length using
3161 * doc/md.texi (while_ult): Document new operand 3 usage.
3162 * internal-fn.cc (expand_while_optab_fn): Set operand 3 when lhs_type
3163 maps to a non-vector mode.
3165 2022-10-03 Andrew MacLeod <amacleod@redhat.com>
3167 PR tree-optimization/107109
3168 * range-op.cc (adjust_op1_for_overflow): Don't process undefined.
3170 2022-10-03 Christophe Lyon <christophe.lyon@arm.com>
3172 * config/arm/mve.md (mve_vrev64q_m_<supf><mode>): Add early
3174 (mve_vrev64q_m_f<mode>): Likewise.
3176 2022-10-03 Aldy Hernandez <aldyh@redhat.com>
3178 * value-range.cc (irange::set_nonzero_bits): Do not pessimize range.
3179 (range_tests_nonzero_bits): New test.
3181 2022-10-03 Aldy Hernandez <aldyh@redhat.com>
3183 * value-range.cc (irange::operator==): Early bail on m_num_ranges
3186 2022-10-03 Aldy Hernandez <aldyh@redhat.com>
3188 * value-range.cc (irange::legacy_equal_p): Remove nonozero mask
3189 check when comparing VR_VARYING ranges.
3191 2022-10-03 Aldy Hernandez <aldyh@redhat.com>
3193 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Do not compare
3194 incompatible ranges in ipa-prop.
3196 2022-10-02 Jakub Jelinek <jakub@redhat.com>
3198 PR tree-optimization/107121
3199 * tree-cfg.cc (verify_gimple_call): Fix a typo in diagnostics,
3200 DEFFERED_INIT -> DEFERRED_INIT.
3202 2022-10-02 Marc Poulhiès <poulhies@adacore.com>
3204 * config/vxworks-driver.cc: New.
3205 * config.gcc (*vxworks*): Add vxworks-driver.o in extra_gcc_objs.
3206 * config/t-vxworks: Add vxworks-driver.o.
3207 * config/vxworks.h (GCC_DRIVER_HOST_INITIALIZATION): New.
3209 2022-10-02 Olivier Hainque <hainque@adacore.com>
3211 * config/vxworks.h (VX_CRTBEGIN_SPEC, VX_CRTEND_SPEC): If
3212 HAVE_INITFINI_ARRAY_SUPPORT, pick crtstuff objects regardless
3213 of ENABLE_SHARED_LIBGCC.
3215 2022-10-01 Stafford Horne <shorne@gmail.com>
3217 * config/or1k/or1k.cc (TARGET_HAVE_TLS): Only define if
3218 HAVE_AS_TLS is defined.
3220 2022-10-01 Julian Brown <julian@codesourcery.com>
3222 * gimplify.cc (omp_group_base): Fix IF_PRESENT (no_create)
3225 2022-10-01 Jeff Law <jeffreyalaw@gmail.com>
3227 * config/h8300/h8300.md (HSI2): New iterator.
3228 (eqne_invert): Similarly.
3229 * config/h8300/testcompare.md (testhi_upper_z): New pattern.
3230 (cmpqi_z, cmphi_z, cmpsi_z): Likewise.
3231 (store_z_qi, store_z_i_qi, store_z_hi, store_z_hi_sb): New
3232 define_insn_and_splits and/or define_insns.
3233 (store_z_hi_neg, store_z_hi_and, store_z_<mode>): Likewise.
3234 (store_z_<mode>_neg, store_z_<mode>_and, store_z): Likewise.
3236 2022-10-01 Jeff Law <jeffreyalaw@gmail.com>
3238 * tree-ssa-dom.cc (record_edge_info): Install correct version of
3241 2022-10-01 Eugene Rozenfeld <erozen@microsoft.com>
3243 * dwarf2out.cc (add_call_src_coords_attributes): Emit discriminators for inlined call sites.
3245 2022-09-30 Jeff Law <jeffreyalaw@gmail.com>
3247 * tree-ssa-dom.cc (single_block_loop_p): New function.
3248 (record_edge_info): Also record equivalences for the outgoing
3249 edge of a single block loop where the condition is an invariant.
3251 2022-09-30 Jeff Law <jeffreyalaw@gmail.com>
3253 * tree-ssa-dom.cc (free_dom_edge_info): Clear e->aux too.
3254 (free_all_edge_infos): Do not clear e->aux here.
3256 2022-09-30 H.J. Lu <hjl.tools@gmail.com>
3258 * target.def (TARGET_C_EXCESS_PRECISION): Document
3259 -fexcess-precision=16.
3261 2022-09-30 Palmer Dabbelt <palmer@rivosinc.com>
3263 * doc/tm.texi (TARGET_C_EXCESS_PRECISION): Add 16.
3265 2022-09-30 Palmer Dabbelt <palmer@rivosinc.com>
3268 * config/riscv/riscv.cc (riscv_excess_precision): Add support
3269 for EXCESS_PRECISION_TYPE_FLOAT16.
3271 2022-09-30 Jakub Jelinek <jakub@redhat.com>
3274 * config/arm/arm.cc (arm_mangle_type): Mangle just __fp16 as Dh
3275 and _Float16 as DF16_.
3276 * config/csky/csky.cc (csky_init_builtins): Fix a comment typo.
3277 (csky_mangle_type): Mangle __fp16 as Dh and _Float16 as DF16_
3278 rather than mangling __fp16 as __fp16.
3280 2022-09-30 Lewis Hyatt <lhyatt@gmail.com>
3282 PR preprocessor/69543
3283 * tree-ssa-uninit.cc (warn_uninit): Stop stripping macro tracking
3284 information away from the diagnostic location.
3285 (maybe_warn_read_write_only): Likewise.
3286 (maybe_warn_operand): Likewise.
3288 2022-09-30 Jakub Jelinek <jakub@redhat.com>
3291 * config/aarch64/aarch64.cc (aarch64_mangle_type): Mangle just __fp16
3292 as Dh and _Float16 as DF16_.
3294 2022-09-30 Jakub Jelinek <jakub@redhat.com>
3297 * config/i386/i386.cc (ix86_mangle_type): Always return NULL
3298 for float128_type_node or float64x_type_node, don't check
3299 float128t_type_node later on.
3300 * config/ia64/ia64.cc (ia64_mangle_type): Always return NULL
3301 for float128_type_node or float64x_type_node.
3302 * config/rs6000/rs6000.cc (rs6000_mangle_type): Likewise.
3303 Don't check float128_type_node later on.
3304 * config/s390/s390.cc (s390_mangle_type): Don't use
3305 TYPE_MAIN_VARIANT on type which was set to TYPE_MAIN_VARIANT
3306 a few lines earlier.
3308 2022-09-30 Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
3309 Yvan ROUX <yvan.roux@foss.st.com>
3311 * doc/sourcebuild.texi: Document new vma_equals_lma effective
3314 2022-09-30 Jiawei <jiawei@iscas.ac.cn>
3316 * config/riscv/riscv.cc (riscv_file_start): New .option.
3317 * config/riscv/riscv.opt: New options.
3318 * doc/invoke.texi: New definations.
3320 2022-09-30 Kewen Lin <linkw@linux.ibm.com>
3324 * doc/invoke.texi (option -fpatchable-function-entry): Adjust the
3325 documentation for PowerPC ELFv2 ABI dual entry points.
3326 * config/rs6000/rs6000-internal.h
3327 (rs6000_print_patchable_function_entry): New function declaration.
3328 * config/rs6000/rs6000-logue.cc (rs6000_output_function_prologue):
3329 Support patchable-function-entry by emitting nops before and after
3330 local entry for the function that needs global entry.
3331 * config/rs6000/rs6000.cc (rs6000_print_patchable_function_entry): Skip
3332 the function that needs global entry till global entry has been
3334 * config/rs6000/rs6000.h (struct machine_function): New bool member
3335 global_entry_emitted.
3337 2022-09-30 Richard Biener <rguenther@suse.de>
3339 PR tree-optimization/107095
3340 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Use data arg
3341 for .MASK_STORE size.
3343 2022-09-29 Andrew MacLeod <amacleod@redhat.com>
3345 * range-op.cc (plus_minus_ranges): New.
3346 (adjust_op1_for_overflow): New.
3347 (operator_plus::op1_range): Use new adjustment.
3348 (operator_plus::op2_range): Ditto.
3349 (operator_minus::op1_range): Ditto.
3350 * value-relation.h (relation_lt_le_gt_ge_p): New.
3352 2022-09-29 Andrew MacLeod <amacleod@redhat.com>
3354 * gimple-range-gori.cc (gori_compute::refine_using_relation): New.
3355 (gori_compute::compute_operand1_range): Invoke
3356 refine_using_relation when applicable.
3357 (gori_compute::compute_operand2_range): Ditto.
3358 * gimple-range-gori.h (class gori_compute): Adjust prototypes.
3360 2022-09-29 Andrew MacLeod <amacleod@redhat.com>
3362 * gimple-range-gori.cc (ori_compute::compute_operand_range):
3363 Create a relation record and pass it along when possible.
3364 (gori_compute::compute_operand1_range): Pass relation along.
3365 (gori_compute::compute_operand2_range): Ditto.
3366 (gori_compute::compute_operand1_and_operand2_range): Ditto.
3367 * gimple-range-gori.h (class gori_compute): Adjust prototypes.
3368 * gimple-range-op.cc (gimple_range_op_handler::calc_op1): Pass
3369 relation to op1_range call.
3370 (gimple_range_op_handler::calc_op2): Pass relation to op2_range call.
3371 * gimple-range-op.h (class gimple_range_op_handler): Adjust
3374 2022-09-29 Andrew MacLeod <amacleod@redhat.com>
3376 * value-relation.cc (class value_relation): Move to .h file.
3377 (value_relation::set_relation): Ditto.
3378 (value_relation::value_relation): ditto.
3379 * value-relation.h (class value_relation): Move from .cc file.
3380 (value_relation::set_relation): Ditto
3381 (value_relation::value_relation): Ditto.
3383 2022-09-29 Andrew MacLeod <amacleod@redhat.com>
3385 * range-op.cc (operator_minus::op2_range): Check for undefined.
3386 (operator_mult::op1_range): Ditto.
3387 (operator_exact_divide::op1_range): Ditto.
3388 (operator_lshift::op1_range): Ditto.
3389 (operator_rshift::op1_range): Ditto.
3390 (operator_cast::op1_range): Ditto.
3391 (operator_bitwise_and::op1_range): Ditto.
3392 (operator_bitwise_or::op1_range): Ditto.
3393 (operator_trunc_mod::op1_range): Ditto.
3394 (operator_trunc_mod::op2_range): Ditto.
3395 (operator_bitwise_not::op1_range): Ditto.
3396 (pointer_or_operator::op1_range): Ditto.
3397 (range_op_handler::op1_range): Ditto.
3398 (range_op_handler::op2_range): Ditto.
3400 2022-09-29 Andrew Stubbs <ams@codesourcery.com>
3402 * config/gcn/gcn.cc (gcn_simd_clone_compute_vecsize_and_simdlen):
3403 Remove unused elt_bits variable.
3405 2022-09-29 Olivier Hainque <hainque@adacore.com>
3407 * config/vxworks.h: Add comment on our use of
3408 HAVE_INITFINI_ARRAY_SUPPORT.
3410 2022-09-29 Olivier Hainque <hainque@adacore.com>
3412 * config/aarch64/t-aarch64-vxworks: Request multilib
3413 variants for mcmodel=large.
3415 2022-09-29 Olivier Hainque <hainque@adacore.com>
3417 * config/rs6000/vxworks.h (TARGET_FLOAT128_ENABLE_TYPE): Remove
3420 2022-09-29 Olivier Hainque <hainque@adacore.com>
3422 * config/vx-common.h (DWARF2_UNWIND_INFO): #define to 0
3423 when ARM_UNWIND_INFO is set.
3425 2022-09-29 Julian Brown <julian@codesourcery.com>
3427 PR middle-end/107028
3428 * gimplify.cc (omp_check_mapping_compatibility,
3429 oacc_resolve_clause_dependencies): New functions.
3430 (omp_accumulate_sibling_list): Remove redundant duplicate clause
3431 detection for OpenACC.
3432 (build_struct_sibling_lists): Skip deleted groups. Don't build sibling
3433 list for struct variables that are fully mapped on the same directive
3435 (gimplify_scan_omp_clauses): Call oacc_resolve_clause_dependencies.
3437 2022-09-29 Jose E. Marchesi <jose.marchesi@oracle.com>
3440 * varasm.cc (categorize_decl_for_section): Place `const volatile'
3441 objects in read-only sections.
3442 (default_select_section): Likewise.
3444 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3446 * tree-data-ref.cc (dr_may_alias_p): Use to_poly_widest instead
3449 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3451 * config/aarch64/aarch64.h (TARGET_CRYPTO, TARGET_SHA3, TARGET_SM4)
3452 (TARGET_DOTPROD): Don't depend on TARGET_SIMD.
3453 (TARGET_AES, TARGET_SHA2): Likewise. Remove TARGET_CRYPTO test.
3454 (TARGET_FP_F16INST): Don't depend on TARGET_FLOAT.
3455 (TARGET_SVE2, TARGET_SVE_F32MM, TARGET_SVE_F64MM): Don't depend
3457 (TARGET_SVE2_AES, TARGET_SVE2_BITPERM, TARGET_SVE2_SHA3)
3458 (TARGET_SVE2_SM4): Don't depend on TARGET_SVE2.
3459 (TARGET_F32MM, TARGET_F64MM): Delete.
3460 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Guard
3461 float macros with just TARGET_FLOAT rather than TARGET_FLOAT
3463 * config/aarch64/aarch64-simd.md (copysign<mode>3): Depend
3464 only on TARGET_SIMD, rather than TARGET_FLOAT && TARGET_SIMD.
3465 (aarch64_crypto_aes<aes_op>v16qi): Depend only on TARGET_AES,
3466 rather than TARGET_SIMD && TARGET_AES.
3467 (aarch64_crypto_aes<aesmc_op>v16qi): Likewise.
3468 (*aarch64_crypto_aese_fused): Likewise.
3469 (*aarch64_crypto_aesd_fused): Likewise.
3470 (aarch64_crypto_pmulldi): Likewise.
3471 (aarch64_crypto_pmullv2di): Likewise.
3472 (aarch64_crypto_sha1hsi): Likewise TARGET_SHA2.
3473 (aarch64_crypto_sha1hv4si): Likewise.
3474 (aarch64_be_crypto_sha1hv4si): Likewise.
3475 (aarch64_crypto_sha1su1v4si): Likewise.
3476 (aarch64_crypto_sha1<sha1_op>v4si): Likewise.
3477 (aarch64_crypto_sha1su0v4si): Likewise.
3478 (aarch64_crypto_sha256h<sha256_op>v4si): Likewise.
3479 (aarch64_crypto_sha256su0v4si): Likewise.
3480 (aarch64_crypto_sha256su1v4si): Likewise.
3481 (aarch64_crypto_sha512h<sha512_op>qv2di): Likewise TARGET_SHA3.
3482 (aarch64_crypto_sha512su0qv2di): Likewise.
3483 (aarch64_crypto_sha512su1qv2di, eor3q<mode>4): Likewise.
3484 (aarch64_rax1qv2di, aarch64_xarqv2di, bcaxq<mode>4): Likewise.
3485 (aarch64_sm3ss1qv4si): Likewise TARGET_SM4.
3486 (aarch64_sm3tt<sm3tt_op>qv4si): Likewise.
3487 (aarch64_sm3partw<sm3part_op>qv4si): Likewise.
3488 (aarch64_sm4eqv4si, aarch64_sm4ekeyqv4si): Likewise.
3489 * config/aarch64/aarch64.md (<FLOATUORS:optab>dihf2)
3490 (copysign<GPF:mode>3, copysign<GPF:mode>3_insn)
3491 (xorsign<mode>3): Remove redundant TARGET_FLOAT condition.
3493 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3495 * config/aarch64/aarch64.opt (aarch64_asm_isa_flags): New variable.
3496 * config/aarch64/aarch64.h (aarch64_asm_isa_flags)
3497 (aarch64_isa_flags): Redefine as read-only macros.
3498 (TARGET_SIMD, TARGET_FLOAT, TARGET_SVE): Don't depend on
3499 !TARGET_GENERAL_REGS_ONLY.
3500 * common/config/aarch64/aarch64-common.cc
3501 (aarch64_set_asm_isa_flags): New function.
3502 (aarch64_handle_option): Call it when updating -mgeneral-regs.
3503 * config/aarch64/aarch64-protos.h (aarch64_simd_switcher): Replace
3504 m_old_isa_flags with m_old_asm_isa_flags.
3505 (aarch64_set_asm_isa_flags): Declare.
3506 * config/aarch64/aarch64-builtins.cc
3507 (aarch64_simd_switcher::aarch64_simd_switcher)
3508 (aarch64_simd_switcher::~aarch64_simd_switcher): Save and restore
3509 aarch64_asm_isa_flags instead of aarch64_isa_flags.
3510 * config/aarch64/aarch64-sve-builtins.cc
3511 (check_required_extensions): Use aarch64_asm_isa_flags instead
3512 of aarch64_isa_flags.
3513 * config/aarch64/aarch64.cc (aarch64_set_asm_isa_flags): New function.
3514 (aarch64_override_options, aarch64_handle_attr_arch)
3515 (aarch64_handle_attr_cpu, aarch64_handle_attr_isa_flags): Use
3516 aarch64_set_asm_isa_flags to set the ISA flags.
3517 (aarch64_option_print, aarch64_declare_function_name)
3518 (aarch64_start_file): Use aarch64_asm_isa_flags instead
3519 of aarch64_isa_flags.
3520 (aarch64_can_inline_p): Check aarch64_asm_isa_flags as well as
3523 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3525 * common/config/aarch64/aarch64-common.cc (all_extensions):
3526 Include the feature flag in flags_on and flags_off.
3527 (aarch64_parse_extension): Update accordingly.
3528 (aarch64_get_extension_string_for_isa_flags): Likewise.
3530 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3532 * common/config/aarch64/aarch64-common.cc: Use aarch64_feature_flags
3533 for feature flags throughout.
3534 * config/aarch64/aarch64-protos.h: Likewise.
3535 * config/aarch64/aarch64-sve-builtins.h: Likewise.
3536 * config/aarch64/aarch64-sve-builtins.cc: Likewise.
3537 * config/aarch64/aarch64.cc: Likewise.
3538 * config/aarch64/aarch64.opt: Likewise.
3539 * config/aarch64/driver-aarch64.cc: Likewise.
3541 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3543 * common/config/aarch64/aarch64-common.cc (aarch64_option_extension)
3544 (processor_name_to_arch, arch_to_arch_name): Remove const from
3546 (all_extensions, all_cores, all_architectures): Make a constexpr.
3547 * config/aarch64/aarch64.cc (processor): Remove const from
3549 (all_architectures): Make a constexpr.
3550 * config/aarch64/driver-aarch64.cc (aarch64_core_data)
3551 (aarch64_arch_driver_info): Remove const from member variables.
3552 (aarch64_cpu_data, aarch64_arches): Make a constexpr.
3553 (get_arch_from_id): Return a pointer to const.
3554 (host_detect_local_cpu): Update accordingly.
3556 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3558 * common/config/aarch64/aarch64-common.cc (processor_name_to_arch)
3559 (arch_to_arch_name): Use const char * instead of std::string.
3561 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3563 * common/config/aarch64/aarch64-common.cc
3564 (TARGET_OPTION_INIT_STRUCT): Delete.
3565 (aarch64_option_extension): Remove is_synthetic_flag.
3566 (all_extensions): Update accordingly.
3567 (all_extensions_by_on, opt_ext, opt_ext_cmp): Delete.
3568 (aarch64_option_init_struct, aarch64_contains_opt): Delete.
3569 (aarch64_get_extension_string_for_isa_flags): Rewrite to use
3570 all_extensions instead of all_extensions_on.
3572 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3574 * config/aarch64/aarch64-option-extensions.def: Switch to a new format.
3575 * config/aarch64/aarch64-cores.def: Use the same format to specify
3577 * config/aarch64/aarch64-arches.def: Likewise, moving that information
3579 * config/aarch64/aarch64-opts.h (aarch64_feature_flags): New typedef.
3580 * config/aarch64/aarch64.h (aarch64_feature): New class enum.
3581 Turn AARCH64_FL_* macros into constexprs, getting the definitions
3582 from aarch64-option-extensions.def. Remove AARCH64_FL_FOR_* macros.
3583 * common/config/aarch64/aarch64-common.cc: Include
3584 aarch64-feature-deps.h.
3585 (all_extensions): Update for new .def format.
3586 (all_extensions_by_on, all_cores, all_architectures): Likewise.
3587 * config/aarch64/driver-aarch64.cc: Include aarch64-feature-deps.h.
3588 (aarch64_extensions): Update for new .def format.
3589 (aarch64_cpu_data, aarch64_arches): Likewise.
3590 * config/aarch64/aarch64.cc: Include aarch64-feature-deps.h.
3591 (all_architectures, all_cores): Update for new .def format.
3592 * config/aarch64/aarch64-sve-builtins.cc
3593 (check_required_extensions): Likewise.
3594 * config/aarch64/aarch64-feature-deps.h: New file.
3596 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3598 * config/aarch64/aarch64-option-extensions.def: Move crypto
3601 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3603 * config/aarch64/aarch64-option-extensions.def (dotprod): Depend
3604 on fp as well as simd.
3606 (aes): Likewise. Make +noaes disable crypto.
3607 (sha2): Likewise +nosha2. Also make +nosha2 disable sha3 and
3609 (sve2-sha3): Depend on sha2 as well as sha3.
3611 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3614 * config/aarch64/aarch64.h (oAARCH64_FL_RCPC8_4): Delete.
3615 (AARCH64_FL_FOR_V8_4A): Update accordingly.
3616 (AARCH64_ISA_RCPC8_4): Use AARCH64_FL_V8_4A directly.
3617 * config/aarch64/aarch64-cores.def (thunderx3t110): Remove
3620 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3622 * config/aarch64/aarch64-cores.def: Remove AARCH64_FL_FOR_<ARCH>
3623 from the flags field.
3624 * common/config/aarch64/aarch64-common.cc (all_cores): Add it
3626 * config/aarch64/aarch64.cc (all_cores): Likewise.
3627 * config/aarch64/driver-aarch64.cc (all_cores): Likewise.
3629 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3631 * config.gcc: Remove dead aarch64-option-extensions.def code.
3632 * config/aarch64/aarch64-arches.def: Update comment.
3634 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3636 * config/aarch64/aarch64-arches.def: Add a leading "V" to the
3638 * config/aarch64/aarch64-cores.def: Update accordingly.
3639 * common/config/aarch64/aarch64-common.cc (all_cores): Likewise.
3640 * config/aarch64/aarch64.cc (all_cores): Likewise.
3641 * config/aarch64/driver-aarch64.cc (aarch64_arches): Skip the
3644 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3646 * config/aarch64/aarch64.h (AARCH64_FL_FOR_ARCH8): Rename to...
3647 (AARCH64_FL_FOR_V8A): ...this.
3648 (AARCH64_FL_FOR_ARCH8_1): Rename to...
3649 (AARCH64_FL_FOR_V8_1A): ...this.
3650 (AARCH64_FL_FOR_ARCH8_2): Rename to...
3651 (AARCH64_FL_FOR_V8_2A): ...this.
3652 (AARCH64_FL_FOR_ARCH8_3): Rename to...
3653 (AARCH64_FL_FOR_V8_3A): ...this.
3654 (AARCH64_FL_FOR_ARCH8_4): Rename to...
3655 (AARCH64_FL_FOR_V8_4A): ...this.
3656 (AARCH64_FL_FOR_ARCH8_5): Rename to...
3657 (AARCH64_FL_FOR_V8_5A): ...this.
3658 (AARCH64_FL_FOR_ARCH8_6): Rename to...
3659 (AARCH64_FL_FOR_V8_6A): ...this.
3660 (AARCH64_FL_FOR_ARCH8_7): Rename to...
3661 (AARCH64_FL_FOR_V8_7A): ...this.
3662 (AARCH64_FL_FOR_ARCH8_8): Rename to...
3663 (AARCH64_FL_FOR_V8_8A): ...this.
3664 (AARCH64_FL_FOR_ARCH8_R): Rename to...
3665 (AARCH64_FL_FOR_V8R): ...this.
3666 (AARCH64_FL_FOR_ARCH9): Rename to...
3667 (AARCH64_FL_FOR_V9A): ...this.
3668 (AARCH64_FL_FOR_ARCH9_1): Rename to...
3669 (AARCH64_FL_FOR_V9_1A): ...this.
3670 (AARCH64_FL_FOR_ARCH9_2): Rename to...
3671 (AARCH64_FL_FOR_V9_2A): ...this.
3672 (AARCH64_FL_FOR_ARCH9_3): Rename to...
3673 (AARCH64_FL_FOR_V9_3A): ...this.
3674 * common/config/aarch64/aarch64-common.cc (all_cores): Update
3676 * config/aarch64/aarch64-arches.def: Likewise.
3677 * config/aarch64/aarch64-cores.def: Likewise.
3678 * config/aarch64/aarch64.cc (all_cores): Likewise.
3680 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3682 * config/aarch64/aarch64.h (AARCH64_FL_V8_1, AARCH64_FL_V8_2)
3683 (AARCH64_FL_V8_3, AARCH64_FL_V8_4, AARCH64_FL_V8_5, AARCH64_FL_V8_6)
3684 (AARCH64_FL_V9, AARCH64_FL_V8_7, AARCH64_FL_V8_8, AARCH64_FL_V9_1)
3685 (AARCH64_FL_V9_2, AARCH64_FL_V9_3): Add "A" to the end of the name.
3686 (AARCH64_FL_V8_R): Rename to AARCH64_FL_V8R.
3687 (AARCH64_FL_FOR_ARCH8_1, AARCH64_FL_FOR_ARCH8_2): Update accordingly.
3688 (AARCH64_FL_FOR_ARCH8_3, AARCH64_FL_FOR_ARCH8_4): Likewise.
3689 (AARCH64_FL_FOR_ARCH8_5, AARCH64_FL_FOR_ARCH8_6): Likewise.
3690 (AARCH64_FL_FOR_ARCH8_7, AARCH64_FL_FOR_ARCH8_8): Likewise.
3691 (AARCH64_FL_FOR_ARCH8_R, AARCH64_FL_FOR_ARCH9): Likewise.
3692 (AARCH64_FL_FOR_ARCH9_1, AARCH64_FL_FOR_ARCH9_2): Likewise.
3693 (AARCH64_FL_FOR_ARCH9_3, AARCH64_ISA_V8_2A, AARCH64_ISA_V8_3A)
3694 (AARCH64_ISA_V8_4A, AARCH64_ISA_V8_5A, AARCH64_ISA_V8_6A): Likewise.
3695 (AARCH64_ISA_V8R, AARCH64_ISA_V9A, AARCH64_ISA_V9_1A): Likewise.
3696 (AARCH64_ISA_V9_2A, AARCH64_ISA_V9_3A): Likewise.
3698 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3700 * config/aarch64/aarch64.h (AARCH64_ISA_V8_2, AARCH64_ISA_V8_3)
3701 (AARCH64_ISA_V8_4, AARCH64_ISA_V8_5, AARCH64_ISA_V8_6)
3702 (AARCH64_ISA_V9, AARCH64_ISA_V9_1, AARCH64_ISA_V9_2)
3703 (AARCH64_ISA_V9_3): Add "A" to the end of the name.
3704 (AARCH64_ISA_V8_R): Rename to AARCH64_ISA_V8R.
3705 (TARGET_ARMV8_3, TARGET_JSCVT, TARGET_FRINT, TARGET_MEMTAG): Update
3707 * common/config/aarch64/aarch64-common.cc
3708 (aarch64_get_extension_string_for_isa_flags): Likewise.
3709 * config/aarch64/aarch64-c.cc
3710 (aarch64_define_unconditional_macros): Likewise.
3712 2022-09-29 Richard Sandiford <richard.sandiford@arm.com>
3714 * Makefile.in (GTFILES): Add OPTIONS_H_EXTRA.
3716 2022-09-29 Jakub Jelinek <jakub@redhat.com>
3719 * cppdefault.cc (cpp_include_defaults): If SYSROOT_HEADERS_SUFFIX_SPEC
3720 isn't defined, add FIXED_INCLUDE_DIR entry with multilib flag 2
3721 before FIXED_INCLUDE_DIR entry with multilib flag 0.
3722 * gcc.cc (do_spec_1): If multiarch_dir, add
3723 include-fixed/multiarch_dir paths before include-fixed paths.
3725 2022-09-29 Martin Liska <mliska@suse.cz>
3728 * common.opt: Add -gz=zstd value.
3729 * configure.ac: Detect --compress-debug-sections=zstd
3730 for both linker and assembler.
3731 * configure: Regenerate.
3732 * gcc.cc (LINK_COMPRESS_DEBUG_SPEC): Handle -gz=zstd.
3733 (ASM_COMPRESS_DEBUG_SPEC): Likewise.
3735 2022-09-29 Richard Biener <rguenther@suse.de>
3737 PR tree-optimization/105646
3738 * tree-ssa-uninit.cc (warn_uninitialized_vars): Pre-compute
3739 the set of fallthru reachable blocks from function entry
3740 and use that to determine wlims.always_executed.
3742 2022-09-29 liuhongt <hongtao.liu@intel.com>
3744 PR tree-optimization/107055
3745 * tree-vect-loop-manip.cc (vect_can_advance_ivs_p): Check for
3746 nonlinear induction variables.
3747 * tree-vect-loop.cc (vect_can_peel_nonlinear_iv_p): New
3749 (vectorizable_nonlinear_induction): Put part codes into
3750 vect_can_peel_nonlinear_iv_p.
3751 * tree-vectorizer.h (vect_can_peel_nonlinear_iv_p): Declare.
3753 2022-09-28 Eugene Rozenfeld <erozen@microsoft.com>
3755 * basic-block.h: Remove discriminator from basic blocks.
3756 * cfghooks.cc (split_block_1): Remove discriminator from basic blocks.
3757 * final.cc (final_start_function_1): Switch from per-bb to per statement
3759 (final_scan_insn_1): Don't keep track of basic block discriminators.
3760 (compute_discriminator): Switch from basic block discriminators to
3761 instruction discriminators.
3762 (insn_discriminator): New function to return instruction discriminator.
3763 (notice_source_line): Use insn_discriminator.
3764 * gimple-pretty-print.cc (dump_gimple_bb_header): Remove dumping of
3765 basic block discriminators.
3766 * gimple-streamer-in.cc (input_bb): Remove reading of basic block
3768 * gimple-streamer-out.cc (output_bb): Remove writing of basic block
3770 * input.cc (make_location): Pass 0 discriminator to COMBINE_LOCATION_DATA.
3771 (location_with_discriminator): New function to combine locus with
3773 (has_discriminator): New function to check if a location has a discriminator.
3774 (get_discriminator_from_loc): New function to get the discriminator
3776 * input.h: Declarations of new functions.
3777 * lto-streamer-in.cc (cmp_loc): Use discriminators in location comparison.
3778 (apply_location_cache): Keep track of current discriminator.
3779 (input_location_and_block): Read discriminator from stream.
3780 * lto-streamer-out.cc (clear_line_info): Set current discriminator to
3782 (lto_output_location_1): Write discriminator to stream.
3783 * lto-streamer.h: Add discriminator to cached_location.
3784 Add current_discr to lto_location_cache.
3785 Add current_discr to output_block.
3786 * print-rtl.cc (print_rtx_operand_code_i): Print discriminator.
3787 * rtl.h: Add extern declaration of insn_discriminator.
3788 * tree-cfg.cc (assign_discriminator): New function to assign a unique
3789 discriminator value to all statements in a basic block that have the given
3791 (assign_discriminators): Assign discriminators to statement locations.
3792 * tree-pretty-print.cc (dump_location): Dump discriminators.
3793 * tree.cc (set_block): Preserve discriminator when setting block.
3794 (set_source_range): Preserve discriminator when setting source range.
3796 2022-09-28 H.J. Lu <hjl.tools@gmail.com>
3799 * config/i386/predicates.md (encodekey128_operation): Check
3800 XMM4-XMM6 as clobbered.
3801 (encodekey256_operation): Likewise.
3802 * config/i386/sse.md (encodekey128u32): Clobber XMM4-XMM6.
3803 (encodekey256u32): Likewise.
3805 2022-09-28 Ju-Zhe Zhong <juzhe.zhong@rivai.ai>
3807 * config.gcc: Add riscv-vector-builtins.o.
3808 * config/riscv/riscv-builtins.cc (riscv_init_builtins): Add RVV builtin function.
3809 * config/riscv/riscv-protos.h (riscv_v_ext_enabled_vector_mode_p): New function.
3810 * config/riscv/riscv.cc (ENTRY): New macro.
3811 (riscv_v_ext_enabled_vector_mode_p): New function.
3812 (riscv_mangle_type): Add RVV mangle.
3813 (riscv_vector_mode_supported_p): Adjust RVV machine mode.
3814 (riscv_verify_type_context): Add context check for RVV.
3815 (riscv_vector_alignment): Add RVV alignment target hook support.
3816 (TARGET_VECTOR_MODE_SUPPORTED_P): New target hook support.
3817 (TARGET_VERIFY_TYPE_CONTEXT): Ditto.
3818 (TARGET_VECTOR_ALIGNMENT): Ditto.
3819 * config/riscv/t-riscv: Add riscv-vector-builtins.o
3820 * config/riscv/riscv-vector-builtins.cc: New file.
3821 * config/riscv/riscv-vector-builtins.def: New file.
3822 * config/riscv/riscv-vector-builtins.h: New file.
3823 * config/riscv/riscv-vector-switch.def: New file.
3825 2022-09-28 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
3827 * var-tracking.cc (vt_add_function_parameter): Add entry values
3828 up to maximal register mode.
3830 2022-09-28 Stefan Schulze Frielinghaus <stefansf@linux.ibm.com>
3832 * cselib.cc (new_cselib_val): Keep track of further subvalue
3835 2022-09-28 Andrea Corallo <andrea.corallo@arm.com>
3837 * config/arm/arm-c.cc (arm_cpu_builtins): Define
3838 __ARM_FEATURE_AES and __ARM_FEATURE_SHA2.
3840 2022-09-28 Xi Ruoyao <xry111@xry111.site>
3842 PR tree-optimization/105414
3843 * config/loongarch/loongarch.md (UNSPEC_FMAX): New unspec.
3844 (UNSPEC_FMIN): Likewise.
3845 (fmax<mode>3): Use UNSPEC_FMAX instead of smax.
3846 (fmin<mode>3): Use UNSPEC_FMIN instead of smin.
3848 2022-09-28 Lulu Cheng <chenglulu@loongson.cn>
3850 * config/loongarch/loongarch.cc (loongarch_asan_shadow_offset):
3851 Fixed typo in "asan_mapping.h".
3853 2022-09-28 H.J. Lu <hjl.tools@gmail.com>
3856 * calls.cc: Include "tree-eh.h".
3857 (expand_call): Check stack canary before throwing exception.
3859 2022-09-27 Eugene Rozenfeld <erozen@microsoft.com>
3861 * ipa-cp.cc (good_cloning_opportunity_p): Fix profile count comparison.
3863 2022-09-27 Kim Kuparinen <kim.kuparinen@rightware.com>
3865 * doc/invoke.texi: Update ABI version info.
3867 2022-09-27 Aldy Hernandez <aldyh@redhat.com>
3869 * gimple-range-op.cc (cfn_popcount): Calculate the popcount of a
3872 2022-09-27 Aldy Hernandez <aldyh@redhat.com>
3874 * value-range.cc (irange::set_nonzero_bits): Set range when known.
3876 2022-09-27 Aldy Hernandez <aldyh@redhat.com>
3878 * value-range.h (irange::set): New version taking wide_int_ref.
3880 2022-09-27 Jakub Jelinek <jakub@redhat.com>
3882 PR tree-optimization/107029
3883 * tree-ssa-reassoc.cc (optimize_range_tests_cmp_bitwise): Treat
3884 OFFSET_TYPE like POINTER_TYPE, except that OFFSET_TYPE may be
3885 signed and so can trigger even the (b % 4) == 3 case.
3887 2022-09-27 Jeff Law <jeffreyalaw@gmail.com>
3889 * cfgrtl.cc (fixup_reorder_chain): Verify that simple_return
3890 and return are available before trying to use them.
3892 2022-09-27 Jakub Jelinek <jakub@redhat.com>
3896 * tree-core.h (enum tree_index): Add TI_FLOAT128T_TYPE
3898 * tree.h (float128t_type_node): Define.
3899 * tree.cc (build_common_tree_nodes): Initialize float128t_type_node.
3900 * builtins.def (DEF_FLOATN_BUILTIN): Adjust comment now that
3901 _Float<N> is supported in C++ too.
3902 * config/i386/i386.cc (ix86_mangle_type): Only mangle as "g"
3903 float128t_type_node.
3904 * config/i386/i386-builtins.cc (ix86_init_builtin_types): Use
3905 float128t_type_node for __float128 instead of float128_type_node
3906 and create it if NULL.
3907 * config/i386/avx512fp16intrin.h (_mm_setzero_ph, _mm256_setzero_ph,
3908 _mm512_setzero_ph, _mm_set_sh, _mm_load_sh): Use 0.0f16 instead of
3910 * config/ia64/ia64.cc (ia64_init_builtins): Use
3911 float128t_type_node for __float128 instead of float128_type_node
3912 and create it if NULL.
3913 * config/rs6000/rs6000-c.cc (is_float128_p): Also return true
3914 for float128t_type_node if non-NULL.
3915 * config/rs6000/rs6000.cc (rs6000_mangle_type): Don't mangle
3916 float128_type_node as "u9__ieee128".
3917 * config/rs6000/rs6000-builtin.cc (rs6000_init_builtins): Use
3918 float128t_type_node for __float128 instead of float128_type_node
3919 and create it if NULL.
3921 2022-09-26 Martin Liska <mliska@suse.cz>
3923 * doc/invoke.texi: Add missing dash for
3924 Wanalyzer-exposure-through-uninit-copy.
3926 2022-09-26 Aldy Hernandez <aldyh@redhat.com>
3928 PR tree-optimization/107009
3929 * range-op.cc (operator_bitwise_and::op1_range): Optimize 0 = x & MASK.
3930 (range_op_bitwise_and_tests): New test.
3932 2022-09-26 Aldy Hernandez <aldyh@redhat.com>
3934 PR tree-optimization/107009
3936 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges):
3937 Iterate over exports.
3939 2022-09-26 Thomas Schwinge <thomas@codesourcery.com>
3941 * config.gcc (with_arch) [nvptx]: Allow '--with-arch' to override
3943 * config/nvptx/gen-multilib-matches.sh: New.
3944 * config/nvptx/t-nvptx (MULTILIB_OPTIONS, MULTILIB_MATCHES)
3945 (MULTILIB_EXCEPTIONS): Handle this.
3946 * doc/install.texi (Specific) <nvptx-*-none>: Document this.
3947 * doc/invoke.texi (Nvidia PTX Options): Likewise.
3949 2022-09-26 Thomas Schwinge <thomas@codesourcery.com>
3951 * config.gcc (TM_MULTILIB_CONFIG) [nvptx]: Set to '$with_arch'.
3952 * config/nvptx/t-nvptx (MULTILIB_OPTIONS, MULTILIB_MATCHES)
3953 (MULTILIB_EXCEPTIONS): Handle it.
3955 2022-09-26 Thomas Schwinge <thomas@codesourcery.com>
3957 * config.gcc (with_arch) [nvptx]: Set to 'sm_30'.
3958 * config/nvptx/nvptx.cc (nvptx_option_override): Assert that
3960 * config/nvptx/nvptx.h (OPTION_DEFAULT_SPECS): Define.
3961 * config/nvptx/nvptx.opt (misa=): Remove 'Init'.
3963 2022-09-26 Thomas Schwinge <thomas@codesourcery.com>
3965 * config/nvptx/nvptx.h (ASM_SPEC): Define.
3967 2022-09-26 Jeff Law <jeffreyalaw@gmail.com>
3969 * cfgcleanup.cc (bb_is_just_return): No longer static.
3970 * cfgcleanup.h (bb_is_just_return): Add prototype.
3971 * cfgrtl.cc (fixup_reorder_chain): Do not create an
3972 unconditional jump to a return block. Conditionally
3973 remove unreachable blocks.
3975 2022-09-26 Tobias Burnus <tobias@codesourcery.com>
3977 PR middle-end/106982
3978 * omp-low.cc (lower_oacc_reductions): Add some unshare_expr.
3980 2022-09-26 Martin Liska <mliska@suse.cz>
3982 * config/s390/s390.cc (s390_rtx_costs): Remove dest variable
3985 2022-09-26 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
3987 * config/aarch64/aarch64-arches.def (armv9.1-a): Define.
3988 (armv9.2-a): Likewise.
3989 (armv9.3-a): Likewise.
3990 * config/aarch64/aarch64.h (AARCH64_FL_V9_1): Likewise.
3991 (AARCH64_FL_V9_2): Likewise.
3992 (AARCH64_FL_V9_3): Likewise.
3993 (AARCH64_FL_FOR_ARCH9_1): Likewise.
3994 (AARCH64_FL_FOR_ARCH9_2): Likewise.
3995 (AARCH64_FL_FOR_ARCH9_3): Likewise.
3996 (AARCH64_ISA_V9_1): Likewise.
3997 (AARCH64_ISA_V9_2): Likewise.
3998 (AARCH64_ISA_V9_3): Likewise.
3999 * doc/invoke.texi (AArch64 Options): Document armv9.1-a, armv9.2-a,
4000 armv9.3-a values to -march.
4002 2022-09-26 Martin Liska <mliska@suse.cz>
4004 * value-range.cc (tree_compare): Remove unused function.
4006 2022-09-26 Kewen Lin <linkw@linux.ibm.com>
4009 * config/rs6000/rs6000-logue.cc (rs6000_emit_epilogue): Update the
4010 condition for adding REG_CFA_DEF_CFA reg note with
4011 frame_pointer_needed_indeed.
4013 2022-09-26 Kewen Lin <linkw@linux.ibm.com>
4016 * config/rs6000/vector.md (vec_shr_<mode>): Replace condition
4017 TARGET_ALTIVEC with VECTOR_UNIT_ALTIVEC_OR_VSX_P.
4019 2022-09-26 Hongtao Liu <hongtao.liu@intel.com>
4020 Liwei Xu <liwei.xu@intel.com>
4023 * config/i386/i386-expand.cc (expand_vec_perm_shufps_shufps):
4025 (ix86_expand_vec_perm_const_1): Insert
4026 expand_vec_perm_shufps_shufps at the end of 2-instruction
4029 2022-09-25 Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
4031 * doc/sourcebuild.texi: Fix chapter level.
4033 2022-09-24 Jakub Jelinek <jakub@redhat.com>
4036 * omp-low.cc (lower_omp_taskgroup): Don't add GOMP_RETURN statement
4038 * omp-expand.cc (build_omp_regions_1): Clarify GF_OMP_TARGET_KIND_DATA
4039 is not stand-alone directive. For GIMPLE_OMP_TASKGROUP, also don't
4041 (omp_make_gimple_edges) <case GIMPLE_OMP_TASKGROUP>: Reset
4042 cur_region back after new_omp_region.
4044 2022-09-23 Vineet Gupta <vineetg@rivosinc.com>
4046 * config/riscv/riscv.h (LOCAL_SYM_P): New.
4047 (USE_LOAD_ADDRESS_MACRO): Simplify by calling LOCAL_SYM_P.
4049 2022-09-23 zhongjuzhe <juzhe.zhong@rivai.ai>
4051 * config/riscv/riscv-modes.def (VECTOR_BOOL_MODE): Add RVV mask modes.
4052 (ADJUST_NUNITS): Adjust nunits using riscv_vector_chunks.
4053 (ADJUST_ALIGNMENT): Adjust alignment.
4054 (ADJUST_BYTESIZE): Adjust bytesize using riscv_vector_chunks.
4055 (RVV_MODES): New macro.
4056 (VECTOR_MODE_WITH_PREFIX): Add RVV vector modes.
4057 (VECTOR_MODES_WITH_PREFIX): Add RVV vector modes.
4059 2022-09-23 zhongjuzhe <juzhe.zhong@rivai.ai>
4061 * common/config/riscv/riscv-common.cc: Change "static void" to "void".
4062 * config.gcc: Add riscv-selftests.o
4063 * config/riscv/predicates.md: Allow const_poly_int.
4064 * config/riscv/riscv-protos.h (riscv_reinit): New function.
4065 (riscv_parse_arch_string): change as exten function.
4066 (riscv_run_selftests): New function.
4067 * config/riscv/riscv.cc (riscv_cannot_force_const_mem): Don't allow poly
4069 (riscv_report_v_required): New function.
4070 (riscv_expand_op): New function.
4071 (riscv_expand_mult_with_const_int): New function.
4072 (riscv_legitimize_poly_move): Ditto.
4073 (riscv_legitimize_move): New function.
4074 (riscv_hard_regno_mode_ok): Add VL/VTYPE register allocation and fix
4076 (riscv_convert_vector_bits): Fix riscv_vector_chunks configuration for
4078 (riscv_reinit): New function.
4079 (TARGET_RUN_TARGET_SELFTESTS): New target hook support.
4080 * config/riscv/t-riscv: Add riscv-selftests.o.
4081 * config/riscv/riscv-selftests.cc: New file.
4083 2022-09-23 Richard Biener <rguenther@suse.de>
4085 PR tree-optimization/106922
4086 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Allow
4087 an arbitrary number of same valued skipped stores.
4089 2022-09-23 Aldy Hernandez <aldyh@redhat.com>
4091 * value-range.cc (frange::set): Swap setters such that the one
4092 accepting REAL_VALUE_TYPE does all the work.
4094 2022-09-23 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4096 * config/aarch64/aarch64-cores.def (neoverse-v2): New entry.
4097 (demeter): Update tunings to neoversev2.
4098 * config/aarch64/aarch64-tune.md: Regenerate.
4099 * config/aarch64/aarch64.cc (demeter_addrcost_table): Rename to
4100 neoversev2_addrcost_table.
4101 (demeter_regmove_cost): Rename to neoversev2_addrcost_table.
4102 (demeter_advsimd_vector_cost): Rename to neoversev2_advsimd_vector_cost.
4103 (demeter_sve_vector_cost): Rename to neoversev2_sve_vector_cost.
4104 (demeter_scalar_issue_info): Rename to neoversev2_scalar_issue_info.
4105 (demeter_advsimd_issue_info): Rename to neoversev2_advsimd_issue_info.
4106 (demeter_sve_issue_info): Rename to neoversev2_sve_issue_info.
4107 (demeter_vec_issue_info): Rename to neoversev2_vec_issue_info.
4108 Update references to above.
4109 (demeter_vector_cost): Rename to neoversev2_vector_cost.
4110 (demeter_tunings): Rename to neoversev2_tunings.
4111 (aarch64_vec_op_count::rename_cycles_per_iter): Use
4112 neoversev2_sve_issue_info instead of demeter_sve_issue_info.
4113 * doc/invoke.texi (AArch64 Options): Document neoverse-v2.
4115 2022-09-23 Aldy Hernandez <aldyh@redhat.com>
4117 * range-op-float.cc (build_le): Use vrp_val_*.
4121 * value-range.cc (frange::set): Chop ranges outside of the
4122 representable numbers for -ffinite-math-only.
4123 (frange::normalize_kind): Use vrp_val*.
4124 (frange::verify_range): Same.
4125 (frange::set_nonnegative): Same.
4126 (range_tests_floats): Remove tests that depend on -INF and +INF.
4127 * value-range.h (real_max_representable): Add prototype.
4128 (real_min_representable): Same.
4129 (vrp_val_max): Set max representable number for
4131 (vrp_val_min): Same but for min.
4132 (frange::set_varying): Use vrp_val*.
4134 2022-09-23 Aldy Hernandez <aldyh@redhat.com>
4136 * real.cc (debug): New.
4138 2022-09-23 Aldy Hernandez <aldyh@redhat.com>
4140 * value-range-pretty-print.cc (vrange_printer::print_real_value): New.
4141 (vrange_printer::visit): Call print_real_value.
4142 * value-range-pretty-print.h: New print_real_value.
4144 2022-09-23 Martin Liska <mliska@suse.cz>
4146 * common.opt: Update -flto-compression-level documentation.
4147 * opts.cc (print_filtered_help): Do not append range to an
4148 option that uses \t syntax.
4150 2022-09-23 Jakub Jelinek <jakub@redhat.com>
4152 * attribs.cc (decl_attributes): Improve diagnostics, instead of
4153 saying expected between 1 and 1, found 2 just say expected 1, found 2.
4155 2022-09-23 Hu, Lin1 <lin1.hu@intel.com>
4158 * config/i386/constraints.md (BH): New define_constraint.
4159 * config/i386/i386.cc (standard_sse_constant_p): Add return
4160 3/4 when operand matches new predicate.
4161 (standard_sse_constant_opcode): Add new alternative branch to
4163 * config/i386/predicates.md
4164 (vector_all_ones_zero_extend_half_operand): New define_predicate.
4165 (vector_all_ones_zero_extend_quarter_operand): Ditto.
4166 * config/i386/sse.md: Add constraint to insn "mov<mode>_internal".
4168 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4170 * gimple-range-fold.cc (range_of_range_op): Handle no operands.
4171 (range_of_call): Do not check for builtins.
4172 (fold_using_range::range_of_builtin_call): Delete.
4173 (fold_using_range::range_of_builtin_int_call): Delete.
4174 * gimple-range-fold.h: Adjust prototypes.
4175 * gimple-range-op.cc (class cfn_parity): New.
4176 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4178 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4180 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4181 for CFN_GOACC_DIM_*.
4182 * gimple-range-op.cc (class cfn_goacc_dim): New.
4183 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4185 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4187 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4188 for CFN_BUILT_IN_STRLEN.
4189 * gimple-range-op.cc (class cfn_strlen): New.
4190 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4192 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4194 * gimple-range-fold.cc (range_of_builtin_ubsan_call): Delete.
4195 (range_of_builtin_int_call): Remove cases for
4196 CFN_BUILT_IN_UBSAN_CHECK.
4197 * gimple-range-op.cc (class cfn_ubsan): New.
4198 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4200 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4202 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4203 for CFN_BUILT_IN_CLRSB.
4204 * gimple-range-op.cc (class cfn_clrsb): New.
4205 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4207 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4209 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4211 * gimple-range-op.cc (class cfn_ctz): New.
4212 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4214 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4216 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4218 * gimple-range-op.cc (class cfn_clz): New.
4219 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4221 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4223 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4224 for CFN_FFS and CFN_POPCOUNT.
4225 * gimple-range-op.cc (class cfn_pocount): New.
4226 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4228 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4230 * gimple-range-fold.cc (get_letter_range): Move to new class.
4231 (range_of_builtin_int_call): Remove case for CFN_BUILT_IN_TOUPPER
4232 and CFN_BUILT_IN_TOLOWER.
4233 * gimple-range-op.cc (class cfn_toupper_tolower): New.
4234 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4236 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4238 * gimple-range-fold.cc (range_of_builtin_int_call): Remove case
4239 for CFN_BUILT_IN_SIGNBIT.
4240 * gimple-range-op.cc (class cfn_signbit): New.
4241 (gimple_range_op_handler::maybe_builtin_call): Set arguments.
4243 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4245 * gimple-range-fold.cc
4246 (fold_using_range::range_of_builtin_int_call): Remove case for
4247 CFN_BUILT_IN_CONSTANT_P.
4248 * gimple-range-op.cc (gimple_range_op_handler::supported_p):
4249 Check if a call also creates a range-op object.
4250 (gimple_range_op_handler): Also check builtin calls.
4251 (class cfn_constant_float_p): New. Float CFN_BUILT_IN_CONSTANT_P.
4252 (class cfn_constant_p): New. Integral CFN_BUILT_IN_CONSTANT_P.
4253 (gimple_range_op_handler::maybe_builtin_call): Set arguments and
4254 handler for supported built-in calls.
4255 * gimple-range-op.h (maybe_builtin_call): New prototype.
4257 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4259 * gimple-range-fold.cc (range_of_range_op): Set result to
4260 VARYING if the call to fold_range fails.
4261 * tree-data-ref.cc (compute_distributive_range): Ditto.
4262 * tree-vrp.cc (range_fold_binary_expr): Ditto.
4263 (range_fold_unary_expr): Ditto.
4264 * value-query.cc (range_query::get_tree_range): Ditto.
4266 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4268 * range-op-float.cc (range_operator_float::fold_range): New base
4269 method for "int = float op int".
4270 * range-op.cc (range_op_handler::fold_range): New case.
4271 * range-op.h: Update prototypes.
4273 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4275 * gimple-range-op.cc (gimple_range_op_handler::calc_op1): Use
4276 operand 1 for second range if there is no operand 2.
4278 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4280 * Makefile.in (OBJS): Add gimple-range-op.o.
4281 * gimple-range-edge.cc (gimple_outgoing_range_stmt_p): Use
4282 gimple_range_op_handler.
4283 * gimple-range-fold.cc (gimple_range_base_of_assignment): Move
4284 to a method in gimple_range_op_handler.
4285 (gimple_range_operand1): Ditto.
4286 (gimple_range_operand2): Ditto.
4287 (fold_using_range::fold_stmt): Use gimple_range_op_handler.
4288 (fold_using_range::range_of_range_op): Ditto.
4289 (fold_using_range::relation_fold_and_or): Ditto.
4290 (fur_source::register_outgoing_edges): Ditto.
4291 (gimple_range_ssa_names): Relocate to gimple-range-op.cc.
4292 * gimple-range-fold.h: Adjust prototypes.
4293 * gimple-range-gori.cc (gimple_range_calc_op1): Move
4294 to a method in gimple_range_op_handler.
4295 (gimple_range_calc_op2): Ditto.
4296 (gori_compute::compute_operand_range): Use
4297 gimple_range_op_handler.
4298 (gori_compute::compute_logical_operands): Ditto.
4299 (compute_operand1_range): Ditto.
4300 (gori_compute::compute_operand2_range): Ditto.
4301 (gori_compute::compute_operand1_and_operand2_range): Ditto.
4302 * gimple-range-gori.h: Adjust protoypes.
4303 * gimple-range-op.cc: New. Supply gimple_range_op_handler methods.
4304 * gimple-range-op.h: New. Supply gimple_range_op_handler class.
4305 * gimple-range.cc (gimple_ranger::prefill_name): Use
4306 gimple_range_op_handler.
4307 (gimple_ranger::prefill_stmt_dependencies): Ditto.
4308 * gimple-range.h: Include gimple-range-op.h.
4309 * range-op.cc (range_op_handler::range_op_handler): Adjust and
4310 remove gimple * parameter option.
4311 * range-op.h: Adjust prototypes.
4313 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4315 * range-op.cc (range_op_handler::set_op_handler): Set new fields.
4316 (ange_op_handler::range_op_handler): Likewise.
4317 (range_op_handler::operator bool): Remove.
4318 (range_op_handler::fold_range): Use appropriate handler.
4319 (range_op_handler::op1_range): Likewise.
4320 (range_op_handler::op2_range): Likewise.
4321 (range_op_handler::lhs_op1_relation): Likewise.
4322 (range_op_handler::lhs_op2_relation): Likewise.
4323 (range_op_handler::op1_op2_relation): Likewise.
4324 * range-op.h (class range_op_handler): Store handler pointers.
4325 (range_op_handler:: operator bool): Inline.
4327 2022-09-22 Andrew MacLeod <amacleod@redhat.com>
4329 * tree-ssa-threadbackward.cc
4330 (back_threader::find_paths_to_names): Replace sequence with
4331 a call to gimple_range_ssa_names.
4333 2022-09-22 Martin Liska <mliska@suse.cz>
4334 Fangrui Song <i@maskray.me>
4336 * configure: Regenerate.
4337 * configure.ac: Simplify to gcc_cv_ld_compress_debug={0,1}
4338 and gcc_cv_as_compress_debug={0,1}.
4339 * doc/invoke.texi: Document the removal.
4340 * gcc.cc (LINK_COMPRESS_DEBUG_SPEC): Simplify and ignore
4342 (ASM_COMPRESS_DEBUG_SPEC): Likewise.
4344 2022-09-22 Richard Biener <rguenther@suse.de>
4346 PR tree-optimization/106922
4347 * tree-ssa-sccvn.cc (vn_walk_cb_data::same_val): New member.
4348 (vn_walk_cb_data::finish): Perform delayed verification of
4349 a skipped may-alias.
4350 (vn_reference_lookup_pieces): Likewise.
4351 (vn_reference_lookup): Likewise.
4352 (vn_reference_lookup_3): When skipping stores of the same
4353 value also handle constant stores that are more than a
4354 single VDEF away by delaying the verification.
4356 2022-09-22 Max Filippov <jcmvbkbc@gmail.com>
4358 * config/xtensa/xtensa.cc (TARGET_MAX_ANCHOR_OFFSET): New
4361 2022-09-22 Max Filippov <jcmvbkbc@gmail.com>
4363 * config/xtensa/xtensa.cc (xtensa_can_output_mi_thunk)
4364 (xtensa_output_mi_thunk): New functions.
4365 (TARGET_ASM_CAN_OUTPUT_MI_THUNK)
4366 (TARGET_ASM_OUTPUT_MI_THUNK): New macro definitions.
4367 (xtensa_prepare_expand_call): Use fixed register a8 as temporary
4368 when called with reload_completed set to 1.
4370 2022-09-22 Richard Biener <rguenther@suse.de>
4372 PR tree-optimization/99407
4373 * tree-ssa-dse.cc (dse_stmt_to_dr_map): New global.
4374 (dse_classify_store): Use data-ref analysis to disambiguate more uses.
4375 (pass_dse::use_dr_analysis_p): New pass parameter.
4376 (pass_dse::set_pass_param): Implement.
4377 (pass_dse::execute): Allocate and deallocate dse_stmt_to_dr_map.
4378 * passes.def: Allow DR analysis for the DSE pass before loop.
4380 2022-09-22 Richard Biener <rguenther@suse.de>
4382 * tree-ssa-sccvn.cc (can_track_predicate_on_edge): New
4383 function split out from ...
4384 (vn_nary_op_insert_pieces_predicated): ... here.
4386 2022-09-22 liuhongt <hongtao.liu@intel.com>
4389 * config/i386/mmx.md (floorv2sf2): Fix typo, use
4390 register_operand instead of vector_operand for operands[1].
4392 2022-09-21 Aldy Hernandez <aldyh@redhat.com>
4394 PR tree-optimization/106967
4395 * value-range.cc (frange::set): Set known NANs to undefined for
4396 flag_finite_math_only.
4398 2022-09-21 Aldy Hernandez <aldyh@redhat.com>
4400 * value-range.cc (frange::combine_zeros): Call set_undefined.
4401 (frange::intersect_nans): Same.
4402 (frange::intersect): Same.
4403 (frange::verify_range): Undefined ranges do not have a type.
4404 * value-range.h (frange::set_undefined): Clear NAN flags and type.
4406 2022-09-21 Richard Biener <rguenther@suse.de>
4408 PR tree-optimization/106984
4409 * tsan.cc (instrument_builtin_call): Build the COND_EXPR condition in
4410 a separate statement.
4412 2022-09-21 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
4414 * config/aarch64/aarch64.h (HAVE_LOCAL_CPU_DETECT,
4415 EXTRA_SPEC_FUNCTIONS, MCPU_MTUNE_NATIVE_SPECS): Move definitions up before
4416 OPTION_DEFAULT_SPECS.
4417 (MCPU_MTUNE_NATIVE_SPECS): Pass "cpu" to
4418 local_cpu_detect when rewriting -march=native and no -mcpu or -mtune
4420 (CONFIG_TUNE_SPEC): Define.
4421 (OPTION_DEFAULT_SPECS): Use CONFIG_TUNE_SPEC for "tune".
4423 2022-09-21 Aldy Hernandez <aldyh@redhat.com>
4425 PR tree-optimization/106967
4426 * range-op-float.cc (foperator_equal::fold_range): Adjust for NAN.
4427 (foperator_equal::op1_range): Same.
4428 (foperator_not_equal::fold_range): Same.
4429 (foperator_not_equal::op1_range): Same.
4430 (foperator_lt::fold_range): Same.
4431 (foperator_lt::op1_range): Same.
4432 (foperator_lt::op2_range): Same.
4433 (foperator_le::fold_range): Same.
4434 (foperator_le::op1_range): Same.
4435 (foperator_le::op2_range): Same.
4436 (foperator_gt::fold_range): Same.
4437 (foperator_gt::op1_range): Same.
4438 (foperator_gt::op2_range): Same.
4439 (foperator_ge::fold_range): Same.
4440 (foperator_ge::op1_range): Same.
4441 (foperator_ge::op2_range): Same.
4442 (foperator_unordered::op1_range): Same.
4443 (foperator_ordered::fold_range): Same.
4444 (foperator_ordered::op1_range): Same.
4445 (build_le): Assert that we don't have a NAN.
4450 2022-09-21 liuhongt <hongtao.liu@intel.com>
4452 PR tree-optimization/106963
4453 * tree-vect-loop.cc (vect_create_nonlinear_iv_init): Use
4454 vec_gen_perm_mask_any instead of vec_gen_perm_mask_check.
4456 2022-09-20 Aldy Hernandez <aldyh@redhat.com>
4458 * value-range.h (frange::maybe_isnan): Return false for
4461 2022-09-20 Aldy Hernandez <aldyh@redhat.com>
4463 * value-range.cc (frange::set_nonnegative): Set +NAN.
4464 (range_tests_signed_zeros): New test.
4465 * value-range.h (frange::update_nan): New overload to set NAN sign.
4467 2022-09-20 Richard Sandiford <richard.sandiford@arm.com>
4470 * config/aarch64/aarch64-sve-builtins.cc (scalar_types)
4471 (acle_vector_types, acle_svpattern, acle_svprfop): Add GTY
4472 markup to (new) extern declarations instead of to the main
4475 2022-09-20 Richard Sandiford <richard.sandiford@arm.com>
4477 PR tree-optimization/106794
4478 PR tree-optimization/106914
4479 * tree-vect-slp.cc (vect_optimize_slp_pass::internal_node_cost):
4480 Only consider loads that already have a permutation.
4481 (vect_optimize_slp_pass::start_choosing_layouts): Assert that
4482 loads with permutations are leaf nodes. Prevent any kind of grouped
4483 access from changing layout if it doesn't have a load permutation.
4485 2022-09-20 Richard Sandiford <richard.sandiford@arm.com>
4487 * tree-vect-data-refs.cc (vect_check_gather_scatter): Restrict
4488 early-out optimisation to SSA_NAMEs.
4490 2022-09-20 Martin Liska <mliska@suse.cz>
4492 * ctfc.cc (ctf_add_string): Replace "the the" with "the".
4493 * doc/md.texi: Likewise.
4494 * gimple-range-infer.cc (non_null_loadstore): Likewise.
4496 2022-09-20 liuhongt <hongtao.liu@intel.com>
4499 * config/i386/mmx.md (nearbyintv2sf2): New expander.
4502 (lceilv2sfv2si2): Ditto.
4503 (floorv2sf2): Ditto.
4504 (lfloorv2sfv2si2): Ditto.
4505 (btruncv2sf2): Ditto.
4506 (lrintv2sfv2si2): Ditto.
4507 (roundv2sf2): Ditto.
4508 (lroundv2sfv2si2): Ditto.
4509 (*mmx_roundv2sf2): New define_insn.
4511 2022-09-20 konglin1 <lingling.kong@intel.com>
4513 PR middle-end/105735
4514 * tree-scalar-evolution.cc
4515 (analyze_and_compute_bitop_with_inv_effect): New function.
4516 (final_value_replacement_loop): Enhanced to handle bitop
4519 2022-09-20 Xi Ruoyao <xry111@xry111.site>
4521 * config/loongarch/gnu-user.h (GNU_USER_TARGET_LINK_SPEC): For
4522 -static-pie, pass -static -pie --no-dynamic-linker -z text to
4523 the linker, and do not pass --dynamic-linker.
4525 2022-09-20 Aldy Hernandez <aldyh@redhat.com>
4527 * value-range.cc (frange::flush_denormals_to_zero): New.
4528 (frange::set): Call flush_denormals_to_zero.
4529 * value-range.h (class frange): Add flush_denormals_to_zero.
4531 2022-09-20 liuhongt <hongtao.liu@intel.com>
4533 * config/i386/x86-tune-sched.cc (ix86_issue_rate): Adjust for
4534 latest Intel processors.
4536 2022-09-20 konglin1 <lingling.kong@intel.com>
4539 * config/i386/i386-expand.cc (ix86_expand_vector_init_duplicate):
4540 Fixed V16BF mode case.
4542 2022-09-19 Torbjörn SVENSSON <torbjorn.svensson@foss.st.com>
4544 * targhooks.cc (default_zero_call_used_regs): Improve sorry
4547 2022-09-18 Julian Brown <julian@codesourcery.com>
4549 * gimplify.cc (omp_segregate_mapping_groups): Update comment.
4550 (gimplify_adjust_omp_clauses): Move ATTACH and
4551 ATTACH_ZERO_LENGTH_ARRAY_SECTION nodes to the end of the clause list
4552 for offloaded OpenMP regions.
4554 2022-09-18 Jakub Jelinek <jakub@redhat.com>
4556 PR middle-end/106831
4557 * value-range.cc (frange::singleton_p): Avoid propagating long
4558 doubles that may have multiple representations.
4560 2022-09-18 Aldy Hernandez <aldyh@redhat.com>
4562 * range-op-float.cc (frange_add_zeros): Replace set_signbit with
4564 * value-query.cc (range_query::get_tree_range): Remove set_signbit
4566 * value-range-pretty-print.cc (vrange_printer::print_frange_prop):
4568 (vrange_printer::print_frange_nan): New.
4569 * value-range-pretty-print.h (print_frange_prop): Remove.
4570 (print_frange_nan): New.
4571 * value-range-storage.cc (frange_storage_slot::set_frange): Set
4572 kind and NAN fields.
4573 (frange_storage_slot::get_frange): Restore kind and NAN fields.
4574 * value-range-storage.h (class frange_storage_slot): Add kind and
4576 * value-range.cc (frange::update_nan): Remove.
4577 (frange::set_signbit): Remove.
4578 (frange::set): Adjust for NAN fields.
4579 (frange::normalize_kind): Remove m_props.
4580 (frange::combine_zeros): New.
4581 (frange::union_nans): New.
4582 (frange::union_): Handle new NAN fields.
4583 (frange::intersect_nans): New.
4584 (frange::intersect): Handle new NAN fields.
4585 (frange::operator=): Same.
4586 (frange::operator==): Same.
4587 (frange::contains_p): Same.
4588 (frange::singleton_p): Remove special case for signed zeros.
4589 (frange::verify_range): Adjust for new NAN fields.
4590 (frange::set_zero): Handle signed zeros.
4591 (frange::set_nonnegative): Same.
4592 (range_tests_nan): Adjust tests.
4593 (range_tests_signed_zeros): Same.
4594 (range_tests_signbit): Same.
4595 (range_tests_floats): Same.
4596 * value-range.h (class fp_prop): Remove.
4597 (FP_PROP_ACCESSOR): Remove.
4598 (class frange_props): Remove
4599 (frange::lower_bound): NANs don't have endpoints.
4600 (frange::upper_bound): Same.
4601 (frange_props::operator==): Remove.
4602 (frange_props::union_): Remove.
4603 (frange_props::intersect): Remove.
4604 (frange::update_nan): New.
4605 (frange::clear_nan): New.
4606 (frange::undefined_p): New.
4607 (frange::set_nan): New.
4608 (frange::known_finite): Adjust for new NAN representation.
4609 (frange::maybe_isnan): Same.
4610 (frange::known_isnan): Same.
4611 (frange::signbit_p): Same.
4612 * gimple-range-fold.cc (range_of_builtin_int_call): Rename
4613 known_signbit_p into signbit_p.
4615 2022-09-17 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4617 * config/csky/csky.h (FUNCTION_ARG_REGNO_P): Cast REGNO to (int)
4620 2022-09-17 Jakub Jelinek <jakub@redhat.com>
4622 PR tree-optimization/106958
4623 * tree-ssa-reassoc.cc (optimize_range_tests_cmp_bitwise): If
4624 id >= l, cast op to type1, otherwise to pointer_sized_int_node.
4625 If type has pointer type, cast exp to pointer_sized_int_node
4628 2022-09-16 liuhongt <hongtao.liu@intel.com>
4631 * config/i386/i386-builtins.cc
4632 (ix86_builtin_vectorized_function): Modernized with
4633 corresponding expanders.
4634 * config/i386/sse.md (lrint<mode><sseintvecmodelower>2): New
4636 (floor<mode>2): Ditto.
4637 (lfloor<mode><sseintvecmodelower>2): Ditto.
4638 (ceil<mode>2): Ditto.
4639 (lceil<mode><sseintvecmodelower>2): Ditto.
4640 (btrunc<mode>2): Ditto.
4641 (lround<mode><sseintvecmodelower>2): Ditto.
4642 (exp2<mode>2): Ditto.
4644 2022-09-15 Joseph Myers <joseph@codesourcery.com>
4646 * ginclude/float.h (INFINITY): Define only if
4647 [__FLT_HAS_INFINITY__].
4649 2022-09-15 Richard Biener <rguenther@suse.de>
4651 PR tree-optimization/106922
4652 * tree-ssa-pre.cc (translate_vuse_through_block): Only
4653 keep the VUSE if its def dominates PHIBLOCK.
4654 (prune_clobbered_mems): Rewrite logic so we check whether
4655 a value dies in a block when the VUSE def doesn't dominate it.
4657 2022-09-15 Richard Biener <rguenther@suse.de>
4659 * tree.cc (build_common_tree_nodes): Initialize void_list_node
4662 2022-09-15 Jiufu Guo <guojiufu@linux.ibm.com>
4665 * config/rs6000/rs6000.cc (rs6000_emit_set_long_const): Use pli.
4667 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4669 * range-op-float.cc (frange_add_zeros): New.
4670 (build_le): Call frange_add_zeros.
4672 (foperator_equal::op1_range): Same.
4673 (foperator_not_equal::op1_range): Same.
4675 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4677 * range-op-float.cc (build_le): Accept frange instead of number.
4681 (foperator_lt::op1_range): Pass full range to build_*.
4682 (foperator_lt::op2_range): Same.
4683 (foperator_le::op1_range): Same.
4684 (foperator_le::op2_range): Same.
4685 (foperator_gt::op1_range): Same.
4686 (foperator_gt::op2_range): Same.
4687 (foperator_ge::op1_range): Same.
4688 (foperator_ge::op2_range): Same.
4690 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4692 * value-range.cc (frange::set): Use set_nan.
4693 * value-range.h (frange::set_nan): Inline code originally in
4696 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4698 * range-op-float.cc (frange_set_nan): Remove.
4699 (build_lt): Use set_nan, update_nan, clear_nan.
4701 (foperator_equal::op1_range): Same.
4702 (foperator_not_equal::op1_range): Same.
4703 (foperator_lt::op1_range): Same.
4704 (foperator_lt::op2_range): Same.
4705 (foperator_le::op1_range): Same.
4706 (foperator_le::op2_range): Same.
4707 (foperator_gt::op1_range): Same.
4708 (foperator_gt::op2_range): Same.
4709 (foperator_ge::op1_range): Same.
4710 (foperator_ge::op2_range): Same.
4711 (foperator_unordered::op1_range): Same.
4712 (foperator_ordered::op1_range): Same.
4713 * value-query.cc (range_query::get_tree_range): Same.
4714 * value-range.cc (frange::set_nan): Same.
4715 (frange::update_nan): Same.
4716 (frange::union_): Same.
4717 (frange::intersect): Same.
4718 (range_tests_nan): Same.
4719 (range_tests_signed_zeros): Same.
4720 (range_tests_signbit): Same.
4721 (range_tests_floats): Same.
4722 * value-range.h (class frange): Add update_nan and clear_nan.
4723 (frange::set_nan): New.
4725 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4727 * value-query.cc (range_query::get_tree_range): Remove check for overflow.
4728 * value-range-pretty-print.cc (vrange_printer::visit): Move read
4729 of type until after undefined_p is checked.
4730 * value-range.cc (frange::set): Remove asserts for REAL_CST.
4731 (frange::contains_p): Tidy up.
4732 (range_tests_nan): Add comment.
4733 * value-range.h (frange::type): Check for undefined_p.
4734 (frange::set_undefined): Remove set of endpoints.
4736 2022-09-14 Julian Brown <julian@codesourcery.com>
4738 * gimplify.cc (omp_group_last): Allow GOMP_MAP_ATTACH_DETACH after
4739 GOMP_MAP_STRUCT (for reindexing).
4740 (omp_gather_mapping_groups): Reimplement using...
4741 (omp_gather_mapping_groups_1): This new function. Stop processing at
4743 (omp_group_base): Allow GOMP_MAP_TO_PSET without any following node.
4744 (omp_index_mapping_groups): Reimplement using...
4745 (omp_index_mapping_groups_1): This new function. Handle
4747 (omp_reindex_mapping_groups, omp_mapped_by_containing_struct): New
4749 (omp_tsort_mapping_groups_1): Adjust handling of base group being the
4750 same as current group. Use omp_mapped_by_containing_struct.
4751 (omp_build_struct_sibling_lists): Use omp_mapped_by_containing_struct
4752 and omp_reindex_mapping_groups. Robustify group deletion for reordered
4754 (gimplify_scan_omp_clauses): Update calls to
4755 omp_build_struct_sibling_lists.
4757 2022-09-14 Julian Brown <julian@codesourcery.com>
4759 * gimplify.cc (gimplify_omp_var_data): Remove GOVD_MAP_HAS_ATTACHMENTS.
4760 (GOMP_FIRSTPRIVATE_IMPLICIT): Renumber.
4761 (insert_struct_comp_map): Refactor function into...
4762 (build_omp_struct_comp_nodes): This new function. Remove list handling
4763 and improve self-documentation.
4764 (extract_base_bit_offset): Remove BASE_REF, OFFSETP parameters. Move
4765 code to strip outer parts of address out of function, but strip no-op
4767 (omp_mapping_group): Add DELETED field for use during reindexing.
4768 (omp_strip_components_and_deref, omp_strip_indirections): New functions.
4769 (omp_group_last, omp_group_base): Add GOMP_MAP_STRUCT handling.
4770 (omp_gather_mapping_groups): Initialise DELETED field for new groups.
4771 (omp_index_mapping_groups): Notice DELETED groups when (re)indexing.
4772 (omp_siblist_insert_node_after, omp_siblist_move_node_after,
4773 omp_siblist_move_nodes_after, omp_siblist_move_concat_nodes_after): New
4775 (omp_accumulate_sibling_list): New function to build up GOMP_MAP_STRUCT
4776 node groups for sibling lists. Outlined from gimplify_scan_omp_clauses.
4777 (omp_build_struct_sibling_lists): New function.
4778 (gimplify_scan_omp_clauses): Remove struct_map_to_clause,
4779 struct_seen_clause, struct_deref_set. Call
4780 omp_build_struct_sibling_lists as pre-pass instead of handling sibling
4781 lists in the function's main processing loop.
4782 (gimplify_adjust_omp_clauses_1): Remove GOVD_MAP_HAS_ATTACHMENTS
4783 handling, unused now.
4784 * omp-low.cc (scan_sharing_clauses): Handle pointer-type indirect
4785 struct references, and references to pointers to structs also.
4787 2022-09-14 Richard Biener <rguenther@suse.de>
4789 PR tree-optimization/106938
4790 * tree-cfg.cc (execute_fixup_cfg): Purge dead abnormal
4791 edges for all last stmts in a block. Do EH cleanup
4792 only on the last stmt in a block.
4794 2022-09-14 Aldy Hernandez <aldyh@redhat.com>
4796 PR tree-optimization/106936
4797 * value-query.cc (range_query::get_value_range): Remove assert.
4799 2022-09-14 Jan-Benedict Glaw <jbglaw@lug-owl.de>
4801 * config/mips/mips.cc (mips_option_override): Drop unused variable.
4803 2022-09-14 Julian Brown <julian@codesourcery.com>
4805 * gimplify.cc (is_or_contains_p, omp_target_reorder_clauses): Delete
4807 (omp_tsort_mark): Add enum.
4808 (omp_mapping_group): Add struct.
4809 (debug_mapping_group, omp_get_base_pointer, omp_get_attachment,
4810 omp_group_last, omp_gather_mapping_groups, omp_group_base,
4811 omp_index_mapping_groups, omp_containing_struct,
4812 omp_tsort_mapping_groups_1, omp_tsort_mapping_groups,
4813 omp_segregate_mapping_groups, omp_reorder_mapping_groups): New
4815 (gimplify_scan_omp_clauses): Call above functions instead of
4816 omp_target_reorder_clauses, unless we've seen an error.
4817 * omp-low.cc (scan_sharing_clauses): Avoid strict test if we haven't
4818 sorted mapping groups.
4820 2022-09-14 Jakub Jelinek <jakub@redhat.com>
4822 PR tree-optimization/106878
4823 * tree-cfg.cc (verify_gimple_assign_binary): Disallow pointer,
4824 reference or OFFSET_TYPE BIT_IOR_EXPR, BIT_XOR_EXPR or, unless
4825 the second argument is INTEGER_CST, BIT_AND_EXPR.
4826 * match.pd ((type) X op CST -> (type) (X op ((type-x) CST)),
4827 (type) (((type2) X) op Y) -> (X op (type) Y)): Punt for
4828 POINTER_TYPE_P or OFFSET_TYPE.
4829 * tree-ssa-reassoc.cc (optimize_range_tests_cmp_bitwise): For
4830 pointers cast them to pointer sized integers first.
4832 2022-09-14 Richard Biener <rguenther@suse.de>
4834 PR tree-optimization/106934
4835 * tree-ssa.cc (non_rewritable_mem_ref_base): Avoid BIT_FIELD_REFs
4837 (maybe_rewrite_mem_ref_base): Likewise.
4839 2022-09-14 liuhongt <hongtao.liu@intel.com>
4841 PR tree-optimization/106905
4842 * tree-vect-loop.cc (vectorizable_nonlinear_induction): Return
4843 false when !vect_use_loop_mask_for_alignment_p (loop_vinfo) &&
4844 LOOP_VINFO_PEELING_FOR_ALIGNMENT (loop_vinfo) < 0.
4846 2022-09-13 Roger Sayle <roger@nextmovesoftware.com>
4849 * reg-stack.cc (move_for_stack_reg): Check for any_malformed_asm
4852 2022-09-13 Max Filippov <jcmvbkbc@gmail.com>
4855 2022-09-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
4857 * config/xtensa/xtensa.cc (TARGET_CONSTANT_OK_FOR_CPROP_P):
4858 New macro definition.
4859 (xtensa_constant_ok_for_cprop_p):
4860 Implement the hook as mentioned above.
4862 2022-09-13 Kewen Lin <linkw@linux.ibm.com>
4865 * config/rs6000/rs6000-c.cc (altivec_resolve_overloaded_builtin): Fix
4866 the equality check for argument number, and move this hunk ahead.
4868 2022-09-13 Kewen.Lin <linkw@gcc.gnu.org>
4871 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_builtin): Add
4872 the handling for unresolved overloaded builtin function.
4873 (rs6000_expand_builtin): Likewise.
4875 2022-09-13 Kewen Lin <linkw@linux.ibm.com>
4877 * config/rs6000/rs6000.cc (class rs6000_cost_data): Add new members
4878 m_nstores, m_reduc_factor, m_gather_load and member function
4879 determine_suggested_unroll_factor.
4880 (rs6000_cost_data::update_target_cost_per_stmt): Update for m_nstores,
4881 m_reduc_factor and m_gather_load.
4882 (rs6000_cost_data::determine_suggested_unroll_factor): New function.
4883 (rs6000_cost_data::finish_cost): Use determine_suggested_unroll_factor.
4884 * config/rs6000/rs6000.opt (rs6000-vect-unroll-limit): New parameter.
4885 (rs6000-vect-unroll-issue): Likewise.
4886 (rs6000-vect-unroll-reduc-threshold): Likewise.
4887 * doc/invoke.texi (rs6000-vect-unroll-limit): Document new parameter.
4889 2022-09-13 Richard Biener <rguenther@suse.de>
4891 PR middle-end/106909
4892 * gimple-fold.cc (gimple_fold_call): Clear the ctrl-altering
4893 flag of a unreachable call.
4895 2022-09-13 Richard Biener <rguenther@suse.de>
4897 PR tree-optimization/106913
4898 * tree-ssa-uninit.cc (warn_uninitialized_vars): Do not set
4899 ft_reachable on EXIT_BLOCK.
4901 2022-09-13 Richard Sandiford <richard.sandiford@arm.com>
4903 * config/aarch64/aarch64.cc (aarch64_classify_vector_mode): Use
4904 TARGET_FLOAT instead of TARGET_SIMD.
4905 (aarch64_vectorize_related_mode): Restrict ADVSIMD handling to
4907 (aarch64_hard_regno_mode_ok): Don't allow tuples of 2 64-bit vectors
4909 (aarch64_classify_address): Treat little-endian structure moves
4910 like big-endian for TARGET_FLOAT && !TARGET_SIMD.
4911 (aarch64_secondary_memory_needed): New function.
4912 (aarch64_secondary_reload): Handle 128-bit Advanced SIMD vectors
4913 in the same way as TF, TI and TD.
4914 (aarch64_rtx_mult_cost): Restrict ADVSIMD handling to TARGET_SIMD.
4915 (aarch64_rtx_costs): Likewise.
4916 (aarch64_register_move_cost): Treat a pair of 64-bit vectors
4917 separately from a single 128-bit vector. Handle the cost implied
4918 by aarch64_secondary_memory_needed.
4919 (aarch64_simd_valid_immediate): Restrict ADVSIMD handling to
4921 (aarch64_expand_vec_perm_const_1): Likewise.
4922 (TARGET_SECONDARY_MEMORY_NEEDED): New macro.
4923 * config/aarch64/iterators.md (VTX): New iterator.
4924 * config/aarch64/aarch64.md (arches): Add fp_q as a synonym of simd.
4925 (arch_enabled): Adjust accordingly.
4926 (@aarch64_reload_mov<TX:mode>): Extend to...
4927 (@aarch64_reload_mov<VTX:mode>): ...this.
4928 * config/aarch64/aarch64-simd.md (mov<mode>): Require TARGET_FLOAT
4929 rather than TARGET_SIMD.
4930 (movmisalign<mode>): Likewise.
4931 (load_pair<DREG:mode><DREG2:mode>): Likewise.
4932 (vec_store_pair<DREG:mode><DREG2:mode>): Likewise.
4933 (load_pair<VQ:mode><VQ2:mode>): Likewise.
4934 (vec_store_pair<VQ:mode><VQ2:mode>): Likewise.
4935 (@aarch64_split_simd_mov<mode>): Likewise.
4936 (aarch64_get_low<mode>): Likewise.
4937 (aarch64_get_high<mode>): Likewise.
4938 (aarch64_get_half<mode>): Likewise. Canonicalize to a move for
4940 (*aarch64_simd_mov<VDMOV:mode>): Require TARGET_FLOAT rather than
4941 TARGET_SIMD. Use different w<-w and r<-w instructions for
4942 !TARGET_SIMD. Disable immediate moves for !TARGET_SIMD but
4943 add an alternative specifically for w<-Z.
4944 (*aarch64_simd_mov<VQMOV:mode>): Require TARGET_FLOAT rather than
4945 TARGET_SIMD. Likewise for the associated define_splits. Disable
4946 FPR moves and immediate moves for !TARGET_SIMD but add an alternative
4947 specifically for w<-Z.
4948 (aarch64_simd_mov_from_<mode>high): Require TARGET_FLOAT rather than
4949 TARGET_SIMD. Restrict the existing alternatives to TARGET_SIMD
4950 but add a new r<-w one for !TARGET_SIMD.
4951 (*aarch64_get_high<mode>): New pattern.
4952 (load_pair_lanes<mode>): Require TARGET_FLOAT rather than TARGET_SIMD.
4953 (store_pair_lanes<mode>): Likewise.
4954 (*aarch64_combine_internal<mode>): Likewise. Restrict existing
4955 w<-w, w<-r and w<-m alternatives to TARGET_SIMD but add a new w<-r
4956 alternative for !TARGET_SIMD.
4957 (*aarch64_combine_internal_be<mode>): Likewise.
4958 (aarch64_combinez<mode>): Require TARGET_FLOAT rather than TARGET_SIMD.
4959 Remove bogus arch attribute.
4960 (*aarch64_combinez_be<mode>): Likewise.
4961 (@aarch64_vec_concat<mode>): Require TARGET_FLOAT rather than
4963 (aarch64_combine<mode>): Likewise.
4964 (aarch64_rev_reglist<mode>): Likewise.
4965 (mov<mode>): Likewise.
4966 (*aarch64_be_mov<VSTRUCT_2D:mode>): Extend to TARGET_FLOAT &&
4967 !TARGET_SIMD, regardless of endianness. Extend associated
4968 define_splits in the same way, both for this pattern and the
4970 (*aarch64_be_mov<VSTRUCT_2Qmode>): Likewise. Restrict w<-w
4971 alternative to TARGET_SIMD.
4972 (*aarch64_be_movoi): Likewise.
4973 (*aarch64_be_movci): Likewise.
4974 (*aarch64_be_movxi): Likewise.
4975 (*aarch64_be_mov<VSTRUCT_4QD:mode>): Extend to TARGET_FLOAT
4976 && !TARGET_SIMD, regardless of endianness. Restrict w<-w alternative
4977 to TARGET_SIMD for tuples of 128-bit vectors.
4978 (*aarch64_be_mov<VSTRUCT_4QD:mode>): Likewise.
4979 * config/aarch64/aarch64-ldpstp.md: Remove TARGET_SIMD condition
4980 from DREG STP peephole. Change TARGET_SIMD to TARGET_FLOAT in
4981 the VQ and VP_2E LDP and STP peepholes.
4983 2022-09-13 Richard Sandiford <richard.sandiford@arm.com>
4985 * config/aarch64/aarch64-simd.md (movv8di): Remove TARGET_SIMD
4986 condition. Likewise for the related define_split. Tweak formatting.
4988 2022-09-12 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
4990 * config/xtensa/xtensa.cc (TARGET_CONSTANT_OK_FOR_CPROP_P):
4991 New macro definition.
4992 (xtensa_constant_ok_for_cprop_p):
4993 Implement the hook as mentioned above.
4995 2022-09-12 Joseph Myers <joseph@codesourcery.com>
4997 * ginclude/stdatomic.h [defined __STDC_VERSION__ &&
4998 __STDC_VERSION__ > 201710L] (ATOMIC_VAR_INIT): Do not define.
5000 2022-09-12 Tobias Burnus <tobias@codesourcery.com>
5002 * config/nvptx/mkoffload.cc (process): Replace a fatal_error by
5003 a warning + not enabling offloading if -misa=sm_30 prevents
5005 (main): Use tool_name as progname for diagnostic.
5006 * config/gcn/mkoffload.cc (main): Likewise.
5008 2022-09-12 Aldy Hernandez <aldyh@redhat.com>
5010 * value-range.cc (frange::set_signbit): Avoid changing sign when
5011 already in the correct sign.
5013 2022-09-12 Max Filippov <jcmvbkbc@gmail.com>
5015 * config/xtensa/xtensa.cc (xtensa_function_value_regno_p):
5016 Recognize all 4 return registers.
5017 * config/xtensa/xtensa.h (GP_RETURN_REG_COUNT): New definition.
5018 * config/xtensa/xtensa.md (untyped_call): New pattern.
5020 2022-09-12 Jonathan Wakely <jwakely@redhat.com>
5022 * doc/extend.texi (Floating Types): Fix "_float128" typo.
5024 2022-09-10 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
5026 * config/xtensa/xtensa.cc (xtensa_constantsynth):
5027 Add new pattern for the abovementioned case.
5029 2022-09-10 Akari Takahashi <akaritakahashioss@gmail.com>
5030 Segher Boessenkool <segher@kernel.crashing.org>
5032 * config/rs6000/rs6000.cc (get_memref_parts): Regularize some code.
5034 2022-09-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
5036 * config/xtensa/xtensa.md: Rewrite the split pattern that performs
5037 the abovementioned process so that insns that overwrite clobbered
5038 register no longer need to be contiguous.
5039 (DSC): Remove as no longer needed.
5041 2022-09-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
5043 * config/xtensa/xtensa.cc (machine_function): New boolean member as
5044 a flag that controls whether to emit the insns for stack pointer
5045 adjustment inside of the pro/epilogue.
5046 (xtensa_emit_adjust_stack_ptr): New function to share the common
5047 codes and to emit insns if not inhibited.
5048 (xtensa_expand_epilogue): Change to use the function mentioned
5049 above when using the CALL0 ABI.
5050 (xtensa_expand_prologue): Ditto.
5051 And also change to set the inhibit flag used by
5052 xtensa_emit_adjust_stack_ptr() to true if the stack pointer is only
5053 used for its own adjustment.
5055 2022-09-09 David Malcolm <dmalcolm@redhat.com>
5057 * doc/invoke.texi (Static Analyzer Options): Add
5058 -Wanalyzer-exposure-through-uninit-copy.
5060 2022-09-09 David Malcolm <dmalcolm@redhat.com>
5062 * Makefile.in (ANALYZER_OBJS): Add
5063 analyzer/known-function-manager.o.
5065 2022-09-09 Tobias Burnus <tobias@codesourcery.com>
5067 * config/nvptx/mkoffload.cc (struct id_map): Add 'dim' member.
5068 (record_id): Store func name without quotes, store dim separately.
5069 (process): For GOMP_REQUIRES_REVERSE_OFFLOAD, check that -march is
5070 at least sm_35, create '$offload_func_table' global array and init
5071 with reverse-offload function addresses.
5072 * config/nvptx/nvptx.cc (write_fn_proto_1, write_fn_proto): New
5073 force_public attribute to force .visible.
5074 (nvptx_declare_function_name): For "omp target
5075 device_ancestor_nohost" attribut, force .visible/TREE_PUBLIC.
5077 2022-09-09 Tobias Burnus <tobias@codesourcery.com>
5079 * config/gcn/mkoffload.cc (process_asm): Create .offload_func_table,
5080 similar to pre-existing .offload_var_table.
5082 2022-09-09 Joseph Myers <joseph@codesourcery.com>
5084 * ginclude/stddef.h [__STDC_VERSION__ > 201710L] (unreachable):
5087 2022-09-09 Kewen Lin <linkw@linux.ibm.com>
5089 PR middle-end/106833
5090 * tree.cc (verify_opaque_type): New function.
5091 (verify_type): Call verify_opaque_type for OPAQUE_TYPE.
5093 2022-09-09 Kwok Cheung Yeung <kcy@codesourcery.com>
5095 * config/gcn/gcn-builtins.def (FABSVF, LDEXPVF, LDEXPV, FREXPVF_EXP,
5096 FREXPVF_MANT, FREXPV_EXP, FREXPV_MANT): Add new builtins.
5097 * config/gcn/gcn-protos.h (gcn_dconst1over2pi): New prototype.
5098 * config/gcn/gcn-valu.md (MATH_UNOP_1OR2REG, MATH_UNOP_1REG,
5099 MATH_UNOP_TRIG): New iterators.
5100 (math_unop): New attributes.
5101 (<math_unop><mode>2, <math_unop><mode>2<exec>,
5102 <math_unop><mode>2, <math_unop><mode>2<exec>,
5103 *<math_unop><mode>2_insn, *<math_unop><mode>2<exec>_insn,
5104 ldexp<mode>3, ldexp<mode>3<exec>,
5105 frexp<mode>_exp2, frexp<mode>_mant2,
5106 frexp<mode>_exp2<exec>, frexp<mode>_mant2<exec>): New instructions.
5107 (<math_unop><mode>2, <math_unop><mode>2<exec>): New expanders.
5108 * config/gcn/gcn.cc (init_ext_gcn_constants): Update definition of
5110 (gcn_dconst1over2pi): New.
5111 (gcn_builtin_type_index): Add entry for v64df type.
5112 (v64df_type_node): New.
5113 (gcn_init_builtin_types): Initialize v64df_type_node.
5114 (gcn_expand_builtin_1): Expand new builtins to instructions.
5115 (print_operand): Fix assembler output for 1/(2*PI) constant.
5116 * config/gcn/gcn.md (unspec): Add new entries.
5118 2022-09-09 Richard Biener <rguenther@suse.de>
5120 PR tree-optimization/106722
5121 * tree-predcom.cc (ref_at_iteration): Do not associate the
5122 constant part of the offset into the MEM_REF offset
5123 operand, across a non-zero offset.
5125 2022-09-09 Kito Cheng <kito.cheng@sifive.com>
5127 * common/config/riscv/riscv-common.cc (RISCV_USE_CUSTOMISED_MULTI_LIB):
5128 Move forward for cover all all necessary functions for suppress
5129 unused function warnings.
5130 (riscv_multi_lib_check): Move forward, and tweak message to suppress
5131 -Werror=format-diag warning.
5133 2022-09-09 Richard Biener <rguenther@suse.de>
5135 PR tree-optimization/106881
5136 * gimple-predicate-analysis.cc (simple_control_dep_chain):
5137 Add only non-fallthru edges and avoid the same set of edges
5138 as compute_control_dep_chain_pdom does.
5140 2022-09-09 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5142 * doc/tm.texi.in (TARGET_OVERRIDES_FORMAT_ATTRIBUTES): Document requirement
5143 of TARGET_OVERRIDES_FORMAT_ATTRIBUTES_COUNT being defined as well.
5144 * doc/tm.texi: Regenerate.
5146 2022-09-09 Martin Liska <mliska@suse.cz>
5148 * optc-save-gen.awk: Always compare array option values with memcmp.
5150 2022-09-08 Jonathan Wakely <jwakely@redhat.com>
5153 * doc/extend.texi (Type Traits): Fix requirements. Document
5154 __is_aggregate and __is_final.
5156 2022-09-08 Tim Lange <mail@tim-lange.me>
5160 State that the checker also reasons about symbolic values.
5162 2022-09-08 Richard Sandiford <richard.sandiford@arm.com>
5164 PR tree-optimization/106886
5165 * tree-vect-slp.cc (vect_optimize_slp_pass::get_result_with_layout):
5166 Fix copying of scalar stmts.
5168 2022-09-08 Chung-Lin Tang <cltang@codesourcery.com>
5170 * config/nios2/linux.h (MUSL_DYNAMIC_LINKER): Add #undef before #define.
5172 2022-09-08 Richard Biener <rguenther@suse.de>
5174 PR middle-end/106870
5175 * gimple-harden-conditionals.cc (insert_check_and_trap):
5176 Set the control-altering flag on the built IFN_TRAP.
5177 * gimple.cc (gimple_build_builtin_unreachable): Likewise.
5178 * tree-cfg.cc (handle_abnormal_edges): Set the control-altering
5179 flag on the .ABNORMAL_DISPATCHER call.
5180 * tree-cfgcleanup.cc (cleanup_call_ctrl_altering_flag): Avoid
5181 resetting the control altering flag for ECF_NORETURN calls.
5182 (cleanup_control_flow_bb): Set the control altering flag on
5183 discovered noreturn calls.
5184 * symtab-thunks.cc (expand_thunk): Set the control altering
5185 flag for the noreturn tailcall case.
5186 * tree-eh.cc (lower_resx): Likewisw for trap and unwind_resume
5189 2022-09-08 Jakub Jelinek <jakub@redhat.com>
5191 * omp-expand.cc (expand_omp_ordered_sink): Add CONT_BB argument.
5192 Add doacross(sink:omp_cur_iteration-1) support.
5193 (expand_omp_ordered_source_sink): Clear counts[fd->ordered + 1].
5194 Adjust expand_omp_ordered_sink caller.
5195 (expand_omp_for_ordered_loops): If counts[fd->ordered + 1] is
5196 non-NULL, set that variable to true at the start of outermost
5197 non-collapsed loop and set it to false at the end of innermost
5199 (expand_omp_for_generic): If fd->ordered, allocate
5200 1 + (fd->ordered - fd->collapse) further elements in counts array.
5201 Copy to counts + 2 + fd->ordered the counts of fd->collapse ..
5202 fd->ordered - 1 loop if any.
5204 2022-09-08 Richard Biener <rguenther@suse.de>
5206 PR tree-optimization/106881
5207 * gimple-predicate-analysis.cc (compute_control_dep_chain_pdom):
5208 Add only non-fallthru edges and avoid the same set of edges
5211 2022-09-08 Aldy Hernandez <aldyh@redhat.com>
5213 * gimple-range-fold.cc
5214 (fold_using_range::range_of_builtin_int_call): Use fpclassify like API.
5215 * range-op-float.cc (finite_operand_p): Same.
5216 (finite_operands_p): Same.
5217 (foperator_lt::fold_range): Same.
5218 (foperator_le::fold_range): Same.
5219 (foperator_gt::fold_range): Same.
5220 (foperator_ge::fold_range): Same.
5221 (foperator_unordered::fold_range): Same.
5222 (foperator_unordered::op1_range): Same.
5223 (foperator_ordered::fold_range): Same.
5224 * value-range.cc (frange::set_nan): Same.
5225 (frange::set_signbit): Same.
5226 (frange::union_): Same.
5227 (frange::intersect): Same.
5228 (frange::operator==): Same.
5229 (frange::singleton_p): Same.
5230 (frange::verify_range): Same.
5231 (range_tests_nan): Same.
5232 (range_tests_floats): Same.
5233 * value-range.h(frange::known_finite): New.
5234 (frange::maybe_inf): New.
5235 (frange::known_inf): New.
5236 (frange::maybe_nan): New.
5237 (frange::known_nan): New.
5238 (frange::known_signbit): New.
5240 2022-09-08 Iain Buclaw <ibuclaw@gdcproject.org>
5242 * config/darwin-d.cc: Include tm.h.
5243 * config/dragonfly-d.cc: Likewise.
5244 * config/freebsd-d.cc: Remove memmodel.h.
5245 * config/glibc-d.cc: Likewise.
5246 * config/netbsd-d.cc: Include tm.h.
5247 * config/openbsd-d.cc: Likewise.
5248 * config/sol2-d.cc: Likewise.
5250 2022-09-08 Christophe Lyon <christophe.lyon@arm.com>
5252 * config/arm/mve.md (mve_vqshluq_n_s<mode>): Use
5253 MVE_pred/MVE_constraint instead of mve_imm_7/Ra.
5254 (mve_vqshluq_m_n_s<mode>): Likewise.
5255 (mve_vqrshrnbq_n_<supf><mode>): Use MVE_pred3/MVE_constraint3
5256 instead of mve_imm_8/Rb.
5257 (mve_vqrshrunbq_n_s<mode>): Likewise.
5258 (mve_vqrshrntq_n_<supf><mode>): Likewise.
5259 (mve_vqrshruntq_n_s<mode>): Likewise.
5260 (mve_vrshrnbq_n_<supf><mode>): Likewise.
5261 (mve_vrshrntq_n_<supf><mode>): Likewise.
5262 (mve_vqrshrnbq_m_n_<supf><mode>): Likewise.
5263 (mve_vqrshrntq_m_n_<supf><mode>): Likewise.
5264 (mve_vrshrnbq_m_n_<supf><mode>): Likewise.
5265 (mve_vrshrntq_m_n_<supf><mode>): Likewise.
5266 (mve_vqrshrunbq_m_n_s<mode>): Likewise.
5267 (mve_vsriq_n_<supf><mode): Use MVE_pred2/MVE_constraint2 instead
5268 of mve_imm_selective_upto_8/Rg.
5269 (mve_vsriq_m_n_<supf><mode>): Likewise.
5271 2022-09-08 Jiufu Guo <guojiufu@linux.ibm.com>
5273 * config/rs6000/rs6000.md (splitter for set to and_mask constants):
5274 Use int_reg_operand (instead of int_reg_operand_not_pseudo).
5275 (splitter for multi-insn constant loads): Ditto.
5277 2022-09-08 Sebastian Huber <sebastian.huber@embedded-brains.de>
5279 * config/rs6000/rtems.h (CPP_OS_DEFAULT_SPEC): Define __PPC_VRSAVE__ if
5280 -mvrsave is present.
5281 * config/rs6000/t-rtems: Add -mvrsave multilib variants for
5284 2022-09-07 Martin Liska <mliska@suse.cz>
5286 * configure.ac: Restore detection of HAVE_XCOFF_DWARF_EXTRAS.
5287 * config/rs6000/rs6000.cc (HAVE_XCOFF_DWARF_EXTRAS): Reset it.
5288 * configure: Regenerate.
5289 * config.in: Regenerate.
5291 2022-09-07 Surya Kumari Jangala <jskumari@linux.ibm.com>
5293 PR rtl-optimization/105586
5294 * sched-rgn.cc (save_state_for_fallthru_edge): New function.
5295 (schedule_region): Use it for all blocks.
5297 2022-09-07 Joseph Myers <joseph@codesourcery.com>
5299 * ginclude/stdalign.h [defined __STDC_VERSION__ &&
5300 __STDC_VERSION__ > 201710L]: Disable all content.
5301 * ginclude/stdbool.h [defined __STDC_VERSION__ && __STDC_VERSION__
5302 > 201710L] (bool, true, false): Do not define.
5304 2022-09-07 Martin Liska <mliska@suse.cz>
5307 * collect2.cc (scan_prog_file): Restore if XCOFF_DEBUGGING_INFO.
5308 * config/rs6000/rs6000.cc (rs6000_option_override_internal):
5309 Restore usage of XCOFF_DEBUGGING_INFO.
5310 * config/rs6000/xcoff.h (XCOFF_DEBUGGING_INFO): Restore.
5311 * dwarf2asm.cc (XCOFF_DEBUGGING_INFO): Restore support for
5312 XCOFF_DEBUGGING_INFO.
5313 (dw2_asm_output_nstring): Likewise.
5314 (USE_LINKONCE_INDIRECT): Likewise.
5315 * dwarf2out.cc (XCOFF_DEBUGGING_INFO): Likewise.
5316 (HAVE_XCOFF_DWARF_EXTRAS): Likewise.
5317 (output_fde): Likewise.
5318 (output_call_frame_info): Likewise.
5319 (have_macinfo): Likewise.
5320 (add_AT_loc_list): Likewise.
5321 (add_AT_view_list): Likewise.
5322 (output_compilation_unit_header): Likewise.
5323 (output_pubnames): Likewise.
5324 (output_aranges): Likewise.
5325 (output_line_info): Likewise.
5326 (output_macinfo): Likewise.
5327 (dwarf2out_finish): Likewise.
5328 (dwarf2out_early_finish): Likewise.
5330 2022-09-07 Richard Biener <rguenther@suse.de>
5332 PR tree-optimization/106866
5333 * tree-ssa-dce.cc (eliminate_unnecessary_stmts): When
5334 we changed cfun->calls_setjmp make sure to purge all
5335 abnormal call edges.
5337 2022-09-07 Xianmiao Qu <cooper.qu@linux.alibaba.com>
5339 * config/csky/csky.cc (csky_emit_compare_float): Fix the expanding of
5340 float LE comparing with zero for fpuv3.
5341 * config/csky/csky.h (TARGET_SUPPORT_FPV2): New, true if any fpuv2
5342 features are enabled.
5344 2022-09-07 Richard Biener <rguenther@suse.de>
5346 PR tree-optimization/106860
5347 * tree-ssa-loop-split.cc (split_loop): Find the exit to
5348 latch edge from the loop exit edge instead of from the
5349 latch. Verify we're going to find it.
5351 2022-09-07 Richard Biener <rguenther@suse.de>
5353 * gimple-predicate-analysis.cc (dfs_mark_dominating_region):
5354 Adjust to take the region exit source as argument.
5355 (uninit_analysis::init_from_phi_def): Adjust.
5356 (uninit_analysis::init_use_preds): Mark the dominating region
5357 before computing control dependences.
5359 2022-09-07 Richard Sandiford <richard.sandiford@arm.com>
5361 * config/aarch64/aarch64.md (*mov<SHORT:mode>_aarch64): Extend
5362 w<-w, r<-w and w<-r alternatives to !simd, using 32-bit moves
5363 in that case. Extend w<-r to w<-Z.
5364 (*mov<HFBF:mode>_aarch64): Likewise, but with Y instead of Z.
5365 (*movti_aarch64): Use an FMOV from XZR for w<-Z if MOVI is not
5367 (define_split): Do not apply the floating-point immediate-to-register
5368 split to zeros, even if MOVI is not available.
5370 2022-09-07 Richard Sandiford <richard.sandiford@arm.com>
5372 * config/aarch64/aarch64.cc (aarch64_conditional_register_usage):
5373 Disallow use of FPRs in register asms for !TARGET_FLOAT.
5375 2022-09-07 Lulu Cheng <chenglulu@loongson.cn>
5378 * config/loongarch/loongarch.cc (loongarch_asan_shadow_offset): New.
5379 (TARGET_ASAN_SHADOW_OFFSET): New.
5381 2022-09-07 Jakub Jelinek <jakub@redhat.com>
5383 * doc/invoke.texi (-Wno-unicode): Document.
5385 2022-09-07 Jiufu Guo <guojiufu@linux.ibm.com>
5387 * config/rs6000/rs6000.md (const_scalar_int splitter): Remove.
5389 2022-09-07 Jiufu Guo <guojiufu@linux.ibm.com>
5391 * config/rs6000/rs6000.md: (constant splitters): Use "(pc)" as the
5394 2022-09-07 liuhongt <hongtao.liu@intel.com>
5396 PR tree-optimization/103144
5397 * tree-vect-loop.cc (vect_is_nonlinear_iv_evolution): New function.
5398 (vect_analyze_scalar_cycles_1): Detect nonlinear iv by upper function.
5399 (vect_create_nonlinear_iv_init): New function.
5400 (vect_peel_nonlinear_iv_init): Ditto.
5401 (vect_create_nonlinear_iv_step): Ditto
5402 (vect_create_nonlinear_iv_vec_step): Ditto
5403 (vect_update_nonlinear_iv): Ditto
5404 (vectorizable_nonlinear_induction): Ditto.
5405 (vectorizable_induction): Call
5406 vectorizable_nonlinear_induction when induction_type is not
5408 * tree-vect-loop-manip.cc (vect_update_ivs_after_vectorizer):
5409 Update nonlinear iv for epilogue loop.
5410 * tree-vectorizer.h (enum vect_induction_op_type): New enum.
5411 (STMT_VINFO_LOOP_PHI_EVOLUTION_TYPE): New Macro.
5413 2022-09-06 Richard Biener <rguenther@suse.de>
5415 PR tree-optimization/106754
5416 * gimple-predicate-analysis.cc (compute_control_dep_chain_pdom):
5417 New function, split out from compute_control_dep_chain. Handle
5418 loop-exit like conditions here by pushing to the control vector.
5419 (compute_control_dep_chain): Adjust and streamline dumping.
5420 In the wrapper perform a post-dominator walk as well.
5421 (uninit_analysis::init_use_preds): Remove premature early exit.
5423 2022-09-06 Max Filippov <jcmvbkbc@gmail.com>
5425 * config/xtensa/linux.h (LINK_SPEC): Add static-pie.
5427 2022-09-06 Aldy Hernandez <aldyh@redhat.com>
5429 * range-op-float.cc (build_le): Handle NANs and going past infinity.
5433 (foperator_lt::op1_range): Avoid adjustments to range if build_*
5435 (foperator_lt::op2_range): Same.
5436 (foperator_le::op1_range): Same.
5437 (foperator_le::op2_range): Same.
5438 (foperator_gt::op1_range): Same.
5439 (foperator_gt::op2_range): Same.
5441 2022-09-06 Richard Biener <rguenther@suse.de>
5443 * gimple-predicate-analysis.cc (compute_control_dep_chain):
5444 Add output flag to indicate whether we possibly have dropped
5445 any chains. Return whether the info is complete from the
5447 (uninit_analysis::init_use_preds): Adjust accordingly, with
5448 a workaround for PR106754.
5449 (uninit_analysis::init_from_phi_def): Properly guard the
5450 case where we complete an empty chain.
5452 2022-09-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5454 * config/msp430/msp430.cc (msp430_single_op_cost): Document unused argument.
5456 2022-09-06 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
5458 * tree-ssa-forwprop.cc (simplify_permutation): Set res_type to a vector
5459 type with same element type as arg0, and length as op2.
5461 2022-09-06 Richard Biener <rguenther@suse.de>
5463 PR tree-optimization/106844
5464 * gimple-predicate-analysis.cc (compute_control_dep_chain):
5465 Return whether we found a chain.
5467 2022-09-06 Richard Biener <rguenther@suse.de>
5469 PR tree-optimization/106841
5470 * tree-vect-slp.cc (vect_detect_hybrid_slp): Also process
5471 scatter/gather offset.
5473 2022-09-06 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5475 * config/bpf/bpf.h (REGNO_REG_CLASS): Reference arguments as (void).
5477 2022-09-06 Jakub Jelinek <jakub@redhat.com>
5479 * gimple.h (enum gf_mask): Add GF_OMP_ORDERED_STANDALONE enumerator.
5480 (gimple_omp_subcode): Use GIMPLE_OMP_ORDERED instead of
5481 GIMPLE_OMP_TEAMS as upper bound.
5482 (gimple_omp_ordered_standalone_p, gimple_omp_ordered_standalone): New
5484 * gimplify.cc (find_standalone_omp_ordered): Look for OMP_ORDERED with
5485 NULL OMP_ORDERED_BODY rather than with OMP_DOACROSS clause.
5486 (gimplify_expr): Call gimple_omp_ordered_standalone for OMP_ORDERED
5487 with NULL OMP_ORDERED_BODY.
5488 * omp-low.cc (check_omp_nesting_restrictions): Use
5489 gimple_omp_ordered_standalone_p test instead of
5490 omp_find_clause (..., OMP_CLAUSE_DOACROSS).
5491 (lower_omp_ordered): Likewise.
5492 * omp-expand.cc (expand_omp, build_omp_regions_1,
5493 omp_make_gimple_edges): Likewise.
5495 2022-09-06 Xianmiao Qu <cooper.qu@linux.alibaba.com>
5497 * config/csky/csky.md (cskyv2_addcc): Fix missing operand.
5498 (cskyv2_addcc_invert): Likewise.
5500 2022-09-06 Jose E. Marchesi <jose.marchesi@oracle.com>
5502 * config/bpf/bpf.cc (bpf_expand_prologue): Remove unused automatic
5504 (bpf_expand_epilogue): Likewise.
5506 2022-09-06 liuhongt <hongtao.liu@intel.com>
5508 * config/i386/i386-builtin.def (IX86_BUILTIN_CVTPS2PH512):
5509 Map to CODE_FOR_avx512f_vcvtps2ph512_mask_sae.
5510 * config/i386/sse.md (<mask_codefor>avx512f_vcvtps2ph512<mask_name>): Extend to ..
5511 (<mask_codefor>avx512f_vcvtps2ph512<mask_name><round_saeonly_name>): .. this.
5512 (avx512f_vcvtps2ph512_mask_sae): New expander
5514 2022-09-05 Aldy Hernandez <aldyh@redhat.com>
5516 * gimple-range-fold.cc
5517 (fold_using_range::range_of_builtin_int_call): Ignore sign bit
5518 when there's the possibility of a NAN.
5520 2022-09-05 Aldy Hernandez <aldyh@redhat.com>
5522 * gimple-range-fold.cc
5523 (fold_using_range::range_of_builtin_int_call): Fold a set signbit
5524 in __builtin_signbit to nonzero.
5526 2022-09-05 Alexander Monakov <amonakov@ispras.ru>
5529 * config/i386/i386.md (sse4_2_crc32di): Model that only low 32
5530 bits of operand 0 are consumed, and the result is zero-extended
5533 2022-09-05 Aldy Hernandez <aldyh@redhat.com>
5535 PR middle-end/106831
5536 * value-range.h (frange::supports_p): Disable decimal floats.
5537 * range-op-float.cc (frange_drop_inf): Remove DECIMAL_FLOAT_MODE_P
5539 (frange_drop_ninf): Same.
5541 2022-09-05 Aldy Hernandez <aldyh@redhat.com>
5543 PR middle-end/106824
5544 * value-range.cc (frange::set_nan): Set undefined when updating a
5547 2022-09-05 Kito Cheng <kito.cheng@sifive.com>
5549 * config/riscv/riscv.cc (riscv_option_override): Fix wrong
5550 condition for MASK_DIV and simplify incompatible checking.
5551 * config/riscv/riscv.md (muldi3): Adding parentheses.
5553 2022-09-05 Kyrylo Tkachov <kyrylo.tkachov@arm.com>
5555 * config/aarch64/aarch64.cc (aarch64_validate_march): Check if invalid arch
5556 string is a valid -mcpu string and emit hint.
5558 2022-09-05 LiaoShihua <shihua@iscas.ac.cn>
5560 * common/config/riscv/riscv-common.cc: Add Zmmul.
5561 * config/riscv/riscv-opts.h (MASK_ZMMUL): New.
5562 (TARGET_ZMMUL): Ditto.
5563 * config/riscv/riscv.cc (riscv_option_override):Ditto.
5564 * config/riscv/riscv.md: Add Zmmul
5565 * config/riscv/riscv.opt: Ditto.
5567 2022-09-05 Richard Biener <rguenther@suse.de>
5569 * tree-cfg.h (get_cases_for_edge): Declare.
5570 * tree-cfg.cc (get_cases_for_edge): Export.
5571 * tree-ssa-uninit.cc (execute_late_warn_uninitialized):
5572 Start and end recording case labels.
5573 * gimple-predicate-analysis.cc (MAX_SWITCH_CASES): Remove.
5574 (predicate::init_from_control_deps): Use get_cases_for_edge.
5576 2022-09-05 Richard Biener <rguenther@suse.de>
5578 * gimple-predicate-analysis.cc (MAX_POSTDOM_CHECK): Remove.
5579 (compute_control_dep_chain): Move uninit-control-dep-attempts
5580 checking where it also counts the post-dominator check
5583 2022-09-05 Richard Biener <rguenther@suse.de>
5585 * gimple-predicate-analysis.h (predicate::debug): New.
5586 (predicate::dump): Add FILE * argument, add base overload.
5587 * gimple-predicate-analysis.cc (debug): New.
5588 (dump_pred_info): Add FILE * argument.
5589 (dump_pred_chain): Likewise.
5590 (predicate::dump): Split out preamble into overload. Add
5592 (predicate::debug): New.
5593 (predicate::simplify): Adjust.
5594 (predicate::normalize): Likewise.
5595 (predicate::init_from_control_deps): Likewise.
5597 2022-09-05 Richard Sandiford <richard.sandiford@arm.com>
5599 * config/aarch64/aarch64-protos.h
5600 (aarch64_init_simd_builtins): Remove prototype.
5601 * config/aarch64/aarch64-builtins.cc
5602 (aarch64_simd_builtins_initialized_p): Delete.
5603 (aarch64_init_simd_builtins): Make static. Remove protection
5604 against multiple calls.
5605 * config/aarch64/aarch64-c.cc (aarch64_pragma_target_parse): Remove
5606 lazy SIMD builtin initialization.
5607 * config/aarch64/aarch64.cc
5608 (aarch64_option_valid_attribute_p): Likewise.
5610 2022-09-05 Xi Ruoyao <xry111@xry111.site>
5612 * config/loongarch/genopts/loongarch.opt.in: Add
5613 -mdirect-extern-access option.
5614 * config/loongarch/loongarch.opt: Regenerate.
5615 * config/loongarch/loongarch.cc
5616 (loongarch_symbol_binds_local_p): Return true if
5617 TARGET_DIRECT_EXTERN_ACCESS.
5618 (loongarch_option_override_internal): Complain if
5619 -mdirect-extern-access is used with -fPIC or -fpic.
5620 * doc/invoke.texi: Document -mdirect-extern-access for
5623 2022-09-05 Martin Liska <mliska@suse.cz>
5625 * doc/tm.texi.in: Rename DEBUGGER_REGISTER_NUMBER to
5627 * doc/tm.texi: Regenerate.
5629 2022-09-05 Martin Liska <mliska@suse.cz>
5631 * config/aarch64/aarch64-protos.h (aarch64_dbx_regno):
5632 Rename DBX_REGISTER_NUMBER to DEBUGGER_REGNO.
5633 (aarch64_debugger_regno): Likewise.
5634 * config/aarch64/aarch64.cc (aarch64_dbx_regno): Likewise.
5635 (aarch64_debugger_regno): Likewise.
5636 * config/aarch64/aarch64.h (DBX_REGISTER_NUMBER): Likewise.
5637 (DEBUGGER_REGNO): Likewise.
5638 (DWARF_FRAME_REGNUM): Likewise.
5639 * config/alpha/alpha.h (DWARF_FRAME_REGNUM): Likewise.
5640 * config/arc/arc.cc (arc_init_reg_tables): Likewise.
5641 * config/arc/arc.h (DBX_REGISTER_NUMBER): Likewise.
5642 (DEBUGGER_REGNO): Likewise.
5643 * config/arm/arm-protos.h (arm_dbx_regno): Likewise.
5644 (arm_debugger_regno): Likewise.
5645 * config/arm/arm.cc (arm_dbx_regno): Likewise.
5646 (arm_debugger_regno): Likewise.
5647 * config/arm/arm.h (DBX_REGISTER_NUMBER): Likewise.
5648 (DEBUGGER_REGNO): Likewise.
5649 * config/bfin/bfin.h (DBX_REGISTER_NUMBER): Likewise.
5650 (DEBUGGER_REGNO): Likewise.
5651 * config/c6x/c6x.cc: Likewise.
5652 * config/c6x/c6x.h (DBX_REGISTER_NUMBER): Likewise.
5653 (DEBUGGER_REGNO): Likewise.
5654 * config/cris/cris.h (enum reg_class): Likewise.
5655 (DBX_REGISTER_NUMBER): Likewise.
5656 (DEBUGGER_REGNO): Likewise.
5657 * config/csky/csky.cc (enum reg_class): Likewise.
5658 * config/csky/csky.h (DWARF_FRAME_REGNUM): Likewise.
5659 (DBX_REGISTER_NUMBER): Likewise.
5660 (DEBUGGER_REGNO): Likewise.
5661 * config/frv/frv.h (DBX_REGISTER_NUMBER): Likewise.
5662 (DEBUGGER_REGNO): Likewise.
5663 * config/gcn/gcn-hsa.h (DBX_REGISTER_NUMBER): Likewise.
5664 (DEBUGGER_REGNO): Likewise.
5665 * config/gcn/gcn.cc (print_operand): Likewise.
5666 * config/i386/bsd.h (ASM_QUAD): Likewise.
5667 * config/i386/cygming.h (DBX_REGISTER_NUMBER): Likewise.
5668 (DEBUGGER_REGNO): Likewise.
5669 (DWARF_FRAME_REGNUM): Likewise.
5670 * config/i386/darwin.h (DBX_REGISTER_NUMBER): Likewise.
5671 (DEBUGGER_REGNO): Likewise.
5672 * config/i386/djgpp.h (DBX_REGISTER_NUMBER): Likewise.
5673 (DEBUGGER_REGNO): Likewise.
5674 * config/i386/dragonfly.h (DBX_REGISTER_NUMBER): Likewise.
5675 (DEBUGGER_REGNO): Likewise.
5676 * config/i386/freebsd.h (DBX_REGISTER_NUMBER): Likewise.
5677 (DEBUGGER_REGNO): Likewise.
5678 * config/i386/gas.h: Likewise.
5679 * config/i386/gnu-user.h (DBX_REGISTER_NUMBER): Likewise.
5680 (DEBUGGER_REGNO): Likewise.
5681 * config/i386/i386.cc (enum reg_class): Likewise.
5682 * config/i386/i386.h (DBX_REGISTER_NUMBER): Likewise.
5683 (DEBUGGER_REGNO): Likewise.
5684 * config/i386/i386elf.h (DBX_REGISTER_NUMBER): Likewise.
5685 (DEBUGGER_REGNO): Likewise.
5686 * config/i386/iamcu.h (DBX_REGISTER_NUMBER): Likewise.
5687 (DEBUGGER_REGNO): Likewise.
5688 * config/i386/lynx.h (DBX_REGISTER_NUMBER): Likewise.
5689 (DEBUGGER_REGNO): Likewise.
5690 * config/i386/netbsd-elf.h (DBX_REGISTER_NUMBER): Likewise.
5691 (DEBUGGER_REGNO): Likewise.
5692 * config/i386/nto.h (DBX_REGISTER_NUMBER): Likewise.
5693 (DEBUGGER_REGNO): Likewise.
5694 * config/i386/openbsdelf.h (DBX_REGISTER_NUMBER): Likewise.
5695 (DEBUGGER_REGNO): Likewise.
5696 * config/i386/sysv4.h (DBX_REGISTER_NUMBER): Likewise.
5697 (DEBUGGER_REGNO): Likewise.
5698 * config/i386/vxworks.h (DBX_REGISTER_NUMBER): Likewise.
5699 (DEBUGGER_REGNO): Likewise.
5700 * config/i386/x86-64.h (DBX_REGISTER_NUMBER): Likewise.
5701 (DEBUGGER_REGNO): Likewise.
5702 * config/ia64/ia64-protos.h (ia64_dbx_regno): Likewise.
5703 (ia64_debugger_regno): Likewise.
5704 * config/ia64/ia64.cc (ia64_output_function_prologue): Likewise.
5705 (ia64_dbx_regno): Likewise.
5706 (ia64_debugger_regno): Likewise.
5707 (process_cfa_adjust_cfa): Likewise.
5708 (process_cfa_register): Likewise.
5709 (ia64_asm_unwind_emit): Likewise.
5710 * config/ia64/ia64.h: Likewise.
5711 * config/ia64/sysv4.h (DBX_REGISTER_NUMBER): Likewise.
5712 (DEBUGGER_REGNO): Likewise.
5713 * config/lm32/lm32.h (DBX_REGISTER_NUMBER): Likewise.
5714 (DEBUGGER_REGNO): Likewise.
5715 * config/m32c/m32c.cc (m32c_eh_return_stackadj_rtx): Likewise.
5716 * config/m32c/m32c.h (DBX_REGISTER_NUMBER): Likewise.
5717 (DEBUGGER_REGNO): Likewise.
5718 * config/m68k/linux.h (DBX_REGISTER_NUMBER): Likewise.
5719 (DEBUGGER_REGNO): Likewise.
5720 * config/m68k/m68k.h (DBX_REGISTER_NUMBER): Likewise.
5721 (DEBUGGER_REGNO): Likewise.
5722 (__transfer_from_trampoline): Likewise.
5723 * config/m68k/m68kelf.h (DBX_REGISTER_NUMBER): Likewise.
5724 (DEBUGGER_REGNO): Likewise.
5725 * config/m68k/netbsd-elf.h (M68K_STATIC_CHAIN_REG_NAME): Likewise.
5726 (DBX_REGISTER_NUMBER): Likewise.
5727 (DEBUGGER_REGNO): Likewise.
5728 * config/microblaze/microblaze.cc (microblaze_option_override): Likewise.
5729 * config/microblaze/microblaze.h (DBX_REGISTER_NUMBER): Likewise.
5730 (DEBUGGER_REGNO): Likewise.
5731 (GP_DBX_FIRST): Likewise.
5732 (GP_DEBUGGER_FIRST): Likewise.
5733 * config/mips/vxworks.h (DBX_REGISTER_NUMBER): Likewise.
5734 (DEBUGGER_REGNO): Likewise.
5735 * config/mmix/mmix-protos.h (mmix_dbx_regno): Likewise.
5736 (mmix_debugger_regno): Likewise.
5737 * config/mmix/mmix.cc (mmix_dbx_regno): Likewise.
5738 (mmix_debugger_regno): Likewise.
5739 * config/mmix/mmix.h (DBX_REGISTER_NUMBER): Likewise.
5740 (DEBUGGER_REGNO): Likewise.
5741 * config/nds32/nds32-protos.h (nds32_dbx_regno): Likewise.
5742 (nds32_debugger_regno): Likewise.
5743 * config/nds32/nds32.cc (nds32_dbx_regno): Likewise.
5744 (nds32_debugger_regno): Likewise.
5745 (nds32_use_blocks_for_constant_p): Likewise.
5746 * config/nds32/nds32.h (DBX_REGISTER_NUMBER): Likewise.
5747 (DEBUGGER_REGNO): Likewise.
5748 * config/nvptx/nvptx.h (DBX_REGISTER_NUMBER): Likewise.
5749 (DEBUGGER_REGNO): Likewise.
5750 * config/or1k/or1k.h (DBX_REGISTER_NUMBER): Likewise.
5751 (DEBUGGER_REGNO): Likewise.
5752 * config/pa/pa32-regs.h (DBX_REGISTER_NUMBER): Likewise.
5753 (DEBUGGER_REGNO): Likewise.
5754 * config/pa/pa64-regs.h (DBX_REGISTER_NUMBER): Likewise.
5755 (DEBUGGER_REGNO): Likewise.
5756 * config/rs6000/rs6000-protos.h (rs6000_dbx_regno): Likewise.
5757 (rs6000_debugger_regno): Likewise.
5758 * config/rs6000/rs6000.cc (rs6000_dbx_regno): Likewise.
5759 (rs6000_debugger_regno): Likewise.
5760 * config/rs6000/rs6000.h (DBX_REGISTER_NUMBER): Likewise.
5761 (DEBUGGER_REGNO): Likewise.
5762 (DWARF2_FRAME_REG_OUT): Likewise.
5763 * config/s390/s390.h (DBX_REGISTER_NUMBER): Likewise.
5764 (DEBUGGER_REGNO): Likewise.
5765 * config/sh/elf.h (DBX_REGISTER_NUMBER): Likewise.
5766 (DEBUGGER_REGNO): Likewise.
5767 * config/sh/linux.h (DBX_REGISTER_NUMBER): Likewise.
5768 (DEBUGGER_REGNO): Likewise.
5769 * config/sh/sh.h (DBX_REGISTER_NUMBER): Likewise.
5770 (DEBUGGER_REGNO): Likewise.
5771 (SH_DBX_REGISTER_NUMBER): Likewise.
5772 (SH_DEBUGGER_REGNO): Likewise.
5773 * config/visium/visium.h (DBX_REGISTER_NUMBER): Likewise.
5774 (DEBUGGER_REGNO): Likewise.
5775 * config/xtensa/elf.h (DBX_REGISTER_NUMBER): Likewise.
5776 (DEBUGGER_REGNO): Likewise.
5777 * config/xtensa/linux.h (DBX_REGISTER_NUMBER): Likewise.
5778 (DEBUGGER_REGNO): Likewise.
5779 * config/xtensa/uclinux.h (DBX_REGISTER_NUMBER): Likewise.
5780 (DEBUGGER_REGNO): Likewise.
5781 * config/xtensa/xtensa-protos.h (xtensa_dbx_regno): Likewise.
5782 (xtensa_debugger_regno): Likewise.
5783 * config/xtensa/xtensa.cc (xtensa_dbx_regno): Likewise.
5784 (xtensa_debugger_regno): Likewise.
5785 * config/xtensa/xtensa.h (DBX_REGISTER_NUMBER): Likewise.
5786 (DEBUGGER_REGNO): Likewise.
5787 * defaults.h (DBX_REGISTER_NUMBER): Likewise.
5788 (DEBUGGER_REGNO): Likewise.
5789 (DWARF_FRAME_REGNUM): Likewise.
5790 * doc/tm.texi: Likewise.
5791 * doc/tm.texi.in: Likewise.
5792 * dwarf2out.cc (dbx_reg_number): Likewise.
5793 (debugger_reg_number): Likewise.
5794 (reg_loc_descriptor): Likewise.
5795 (multiple_reg_loc_descriptor): Likewise.
5796 (mem_loc_descriptor): Likewise.
5797 * except.cc: Likewise.
5799 2022-09-05 konglin1 <lingling.kong@intel.com>
5802 * config/i386/i386-expand.cc (ix86_expand_vector_init_duplicate):
5804 (expand_vec_perm_broadcast_1): Ditto.
5805 * config/i386/sse.md (avx512fmaskhalfmode): Add BF vector mode.
5806 (vec_set<mode>_0): Add @ to it.
5807 (@vec_set<mode>_0): Ditto.
5808 (vec_interleave_high<mode><mask_name>): Ditto.
5809 (@vec_interleave_high<mode><mask_name>): Ditto.
5810 (vec_interleave_low<mode><mask_name>): Ditto.
5811 (@vec_interleave_low<mode><mask_name>): Ditto.
5812 * config/i386/subst.md (SUBST_V): Add BF vector mode.
5814 2022-09-05 Aldy Hernandez <aldyh@redhat.com>
5816 * value-range.cc (early_nan_resolve): Remove.
5817 (frange::intersect): Handle NANs.
5819 2022-09-04 Aldy Hernandez <aldyh@redhat.com>
5821 * value-range.cc (frange::union_): Do not drop properties when
5822 unioning a NAN with something else.
5823 (range_tests_signed_zeros): Add tests.
5825 2022-09-04 Aldy Hernandez <aldyh@redhat.com>
5827 * real.cc (encode_ieee_single): Use real_isdenormal.
5828 (encode_ieee_double): Same.
5829 (encode_ieee_extended): Same.
5830 (encode_ieee_quad): Same.
5831 (encode_ieee_half): Same.
5832 (encode_arm_bfloat_half): Same.
5833 * real.h (real_isdenormal): New.
5835 2022-09-03 Aldy Hernandez <aldyh@redhat.com>
5837 * value-range.cc (frange::singleton_p): Move NAN check to the top.
5839 2022-09-03 Jan-Benedict Glaw <jbglaw@lug-owl.de>
5841 * config/nvptx/nvptx.h (ASM_OUTPUT_DEF): Reference macro arguments.
5843 2022-09-03 Jakub Jelinek <jakub@redhat.com>
5845 * tree-core.h (enum omp_clause_code): Add OMP_CLAUSE_DOACROSS.
5846 (enum omp_clause_depend_kind): Remove OMP_CLAUSE_DEPEND_SOURCE
5847 and OMP_CLAUSE_DEPEND_SINK, add OMP_CLAUSE_DEPEND_INVALID.
5848 (enum omp_clause_doacross_kind): New type.
5849 (struct tree_omp_clause): Add subcode.doacross_kind member.
5850 * tree.h (OMP_CLAUSE_DEPEND_SINK_NEGATIVE): Remove.
5851 (OMP_CLAUSE_DOACROSS_KIND): Define.
5852 (OMP_CLAUSE_DOACROSS_SINK_NEGATIVE): Define.
5853 (OMP_CLAUSE_DOACROSS_DEPEND): Define.
5854 (OMP_CLAUSE_ORDERED_DOACROSS): Define.
5855 * tree.cc (omp_clause_num_ops, omp_clause_code_name): Add
5856 OMP_CLAUSE_DOACROSS entries.
5857 * tree-nested.cc (convert_nonlocal_omp_clauses,
5858 convert_local_omp_clauses): Handle OMP_CLAUSE_DOACROSS.
5859 * tree-pretty-print.cc (dump_omp_clause): Don't handle
5860 OMP_CLAUSE_DEPEND_SOURCE and OMP_CLAUSE_DEPEND_SINK. Handle
5861 OMP_CLAUSE_DOACROSS.
5862 * gimplify.cc (gimplify_omp_depend): Don't handle
5863 OMP_CLAUSE_DEPEND_SOURCE and OMP_CLAUSE_DEPEND_SINK.
5864 (gimplify_scan_omp_clauses): Likewise. Handle OMP_CLAUSE_DOACROSS.
5865 (gimplify_adjust_omp_clauses): Handle OMP_CLAUSE_DOACROSS.
5866 (find_standalone_omp_ordered): New function.
5867 (gimplify_omp_for): When OMP_CLAUSE_ORDERED is present, search
5868 body for OMP_ORDERED with OMP_CLAUSE_DOACROSS and if found,
5869 set OMP_CLAUSE_ORDERED_DOACROSS.
5870 (gimplify_omp_ordered): Don't handle OMP_CLAUSE_DEPEND_SINK or
5871 OMP_CLAUSE_DEPEND_SOURCE, instead check OMP_CLAUSE_DOACROSS, adjust
5872 diagnostics that presence or absence of ordered clause parameter
5873 is irrelevant. Handle doacross(sink:omp_cur_iteration-1). Use
5874 actual user name of the clause - doacross or depend - in diagnostics.
5875 * omp-general.cc (omp_extract_for_data): Don't set fd->ordered
5876 if !OMP_CLAUSE_ORDERED_DOACROSS (t). If
5877 OMP_CLAUSE_ORDERED_DOACROSS (t) but !OMP_CLAUSE_ORDERED_EXPR (t),
5878 set fd->ordered to -1 and set it after the loop in that case to
5880 * omp-low.cc (check_omp_nesting_restrictions): Don't handle
5881 OMP_CLAUSE_DEPEND_SOURCE nor OMP_CLAUSE_DEPEND_SINK, instead check
5882 OMP_CLAUSE_DOACROSS. Use actual user name of the clause - doacross
5883 or depend - in diagnostics. Diagnose mixing of stand-alone and
5884 block associated ordered constructs binding to the same loop.
5885 (lower_omp_ordered_clauses): Don't handle OMP_CLAUSE_DEPEND_SINK,
5886 instead handle OMP_CLAUSE_DOACROSS.
5887 (lower_omp_ordered): Look for OMP_CLAUSE_DOACROSS instead of
5889 (lower_depend_clauses): Don't handle OMP_CLAUSE_DEPEND_SOURCE and
5890 OMP_CLAUSE_DEPEND_SINK.
5891 * omp-expand.cc (expand_omp_ordered_sink): Emit a sorry for
5892 doacross(sink:omp_cur_iteration-1).
5893 (expand_omp_ordered_source_sink): Use
5894 OMP_CLAUSE_DOACROSS_SINK_NEGATIVE instead of
5895 OMP_CLAUSE_DEPEND_SINK_NEGATIVE. Use actual user name of the clause
5896 - doacross or depend - in diagnostics.
5897 (expand_omp): Look for OMP_CLAUSE_DOACROSS clause instead of
5899 (build_omp_regions_1): Likewise.
5900 (omp_make_gimple_edges): Likewise.
5901 * lto-streamer-out.cc (hash_tree): Handle OMP_CLAUSE_DOACROSS.
5902 * tree-streamer-in.cc (unpack_ts_omp_clause_value_fields): Likewise.
5903 * tree-streamer-out.cc (pack_ts_omp_clause_value_fields): Likewise.
5905 2022-09-02 David Malcolm <dmalcolm@redhat.com>
5908 * doc/invoke.texi (Warning Options): Add -Wxor-used-as-pow.
5910 2022-09-02 Iain Buclaw <ibuclaw@gdcproject.org>
5913 * config.gcc: Set tm_d_file to ${cpu_type}/${cpu_type}-d.h.
5914 * config/aarch64/aarch64-d.cc: Include tm_d.h.
5915 * config/aarch64/aarch64-protos.h (aarch64_d_target_versions): Move to
5916 config/aarch64/aarch64-d.h.
5917 (aarch64_d_register_target_info): Likewise.
5918 * config/aarch64/aarch64.h (TARGET_D_CPU_VERSIONS): Likewise.
5919 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5920 * config/arm/arm-d.cc: Include tm_d.h and arm-protos.h instead of
5922 * config/arm/arm-protos.h (arm_d_target_versions): Move to
5924 (arm_d_register_target_info): Likewise.
5925 * config/arm/arm.h (TARGET_D_CPU_VERSIONS): Likewise.
5926 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5927 * config/default-d.cc: Remove memmodel.h include.
5928 * config/freebsd-d.cc: Include tm_d.h instead of tm_p.h.
5929 * config/glibc-d.cc: Likewise.
5930 * config/i386/i386-d.cc: Include tm_d.h.
5931 * config/i386/i386-protos.h (ix86_d_target_versions): Move to
5932 config/i386/i386-d.h.
5933 (ix86_d_register_target_info): Likewise.
5934 (ix86_d_has_stdcall_convention): Likewise.
5935 * config/i386/i386.h (TARGET_D_CPU_VERSIONS): Likewise.
5936 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5937 (TARGET_D_HAS_STDCALL_CONVENTION): Likewise.
5938 * config/i386/winnt-d.cc: Include tm_d.h instead of tm_p.h.
5939 * config/mips/mips-d.cc: Include tm_d.h.
5940 * config/mips/mips-protos.h (mips_d_target_versions): Move to
5941 config/mips/mips-d.h.
5942 (mips_d_register_target_info): Likewise.
5943 * config/mips/mips.h (TARGET_D_CPU_VERSIONS): Likewise.
5944 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5945 * config/netbsd-d.cc: Include tm_d.h instead of tm.h and memmodel.h.
5946 * config/openbsd-d.cc: Likewise.
5947 * config/pa/pa-d.cc: Include tm_d.h.
5948 * config/pa/pa-protos.h (pa_d_target_versions): Move to
5950 (pa_d_register_target_info): Likewise.
5951 * config/pa/pa.h (TARGET_D_CPU_VERSIONS): Likewise.
5952 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5953 * config/riscv/riscv-d.cc: Include tm_d.h.
5954 * config/riscv/riscv-protos.h (riscv_d_target_versions): Move to
5955 config/riscv/riscv-d.h.
5956 (riscv_d_register_target_info): Likewise.
5957 * config/riscv/riscv.h (TARGET_D_CPU_VERSIONS): Likewise.
5958 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5959 * config/rs6000/rs6000-d.cc: Include tm_d.h.
5960 * config/rs6000/rs6000-protos.h (rs6000_d_target_versions): Move to
5961 config/rs6000/rs6000-d.h.
5962 (rs6000_d_register_target_info): Likewise.
5963 * config/rs6000/rs6000.h (TARGET_D_CPU_VERSIONS) Likewise.:
5964 (TARGET_D_REGISTER_CPU_TARGET_INFO) Likewise.:
5965 * config/s390/s390-d.cc: Include tm_d.h.
5966 * config/s390/s390-protos.h (s390_d_target_versions): Move to
5967 config/s390/s390-d.h.
5968 (s390_d_register_target_info): Likewise.
5969 * config/s390/s390.h (TARGET_D_CPU_VERSIONS): Likewise.
5970 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5971 * config/sol2-d.cc: Include tm_d.h instead of tm.h and memmodel.h.
5972 * config/sparc/sparc-d.cc: Include tm_d.h.
5973 * config/sparc/sparc-protos.h (sparc_d_target_versions): Move to
5974 config/sparc/sparc-d.h.
5975 (sparc_d_register_target_info): Likewise.
5976 * config/sparc/sparc.h (TARGET_D_CPU_VERSIONS): Likewise.
5977 (TARGET_D_REGISTER_CPU_TARGET_INFO): Likewise.
5978 * configure: Regenerate.
5979 * configure.ac (tm_d_file): Remove defaults.h.
5980 (tm_d_include_list): Remove options.h and insn-constants.h.
5981 * config/aarch64/aarch64-d.h: New file.
5982 * config/arm/arm-d.h: New file.
5983 * config/i386/i386-d.h: New file.
5984 * config/mips/mips-d.h: New file.
5985 * config/pa/pa-d.h: New file.
5986 * config/riscv/riscv-d.h: New file.
5987 * config/rs6000/rs6000-d.h: New file.
5988 * config/s390/s390-d.h: New file.
5989 * config/sparc/sparc-d.h: New file.
5991 2022-09-02 Aldy Hernandez <aldyh@redhat.com>
5993 * builtins.cc (fold_builtin_inf): Convert use of real_info to dconstinf.
5994 (fold_builtin_fpclassify): Same.
5995 * fold-const-call.cc (fold_const_call_cc): Same.
5997 * omp-low.cc (omp_reduction_init_op): Same.
5998 * realmpfr.cc (real_from_mpfr): Same.
5999 * tree.cc (build_complex_inf): Same.
6001 2022-09-02 Peter Bergner <bergner@linux.ibm.com>
6003 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_mma_builtin): Use
6004 NOP_EXPR for MMA pointer casting.
6006 2022-09-02 Richard Sandiford <richard.sandiford@arm.com>
6008 * tree-vect-slp.cc (vect_optimize_slp_pass::internal_node_cost):
6009 Reduce the fallback cost to 1. Only use it if the number of
6010 input lanes is equal to the number of output lanes.
6012 2022-09-02 Richard Sandiford <richard.sandiford@arm.com>
6014 PR tree-optimization/106787
6015 * tree-vect-slp.cc (vect_map_to_instance): New function, split out
6017 (vect_bb_partition_graph_r): ...here. Replace the visited set
6018 with a map from nodes to instances. Ensure that a node only
6019 appears in one partition.
6020 (vect_bb_partition_graph): Update accordingly.
6022 2022-09-02 Richard Biener <rguenther@suse.de>
6024 * timevar.def (TV_TREE_RPO_VN): New.
6025 * tree-ssa-sccvn.h (do_rpo_vn): Remove one overload.
6026 * tree-ssa-sccvn.cc (do_rpo_vn_1): Rename the worker.
6027 (do_rpo_vn): Unify the public API, track with TV_TREE_RPO_VN.
6028 (pass_fre::execute): Adjust.
6029 * tree-ssa-uninit.cc (execute_early_warn_uninitialized): Adjust.
6031 2022-09-02 Richard Biener <rguenther@suse.de>
6033 PR tree-optimization/106809
6034 * tree-ssa-sccvn.cc (dominaged_by_p_w_unex): Check we have
6035 more than one successor before doing extra work.
6037 2022-09-02 Kito Cheng <kito.cheng@sifive.com>
6039 * common/config/riscv/riscv-common.cc: Include <vector>.
6040 (struct riscv_multi_lib_info_t): New.
6041 (riscv_subset_list::match_score): Ditto.
6042 (find_last_appear_switch): Ditto.
6043 (prefixed_with): Ditto.
6044 (struct multi_lib_info_t): Ditto.
6045 (riscv_current_arch_str): Ditto.
6046 (riscv_current_abi_str): Ditto.
6047 (riscv_multi_lib_info_t::parse): Ditto.
6048 (riscv_check_cond): Ditto.
6049 (riscv_check_conds): Ditto.
6050 (riscv_compute_multilib): Ditto.
6051 (TARGET_COMPUTE_MULTILIB): Defined.
6052 * config/riscv/elf.h (LIB_SPEC): Call riscv_multi_lib_check if
6054 (RISCV_USE_CUSTOMISED_MULTI_LIB): New.
6055 * config/riscv/riscv.h (riscv_multi_lib_check): New.
6056 (EXTRA_SPEC_FUNCTIONS): Add riscv_multi_lib_check.
6057 * config/riscv/riscv-subset.h (riscv_subset_list::match_score): New.
6059 2022-09-02 Kito Cheng <kito.cheng@sifive.com>
6061 * common/common-target.def (compute_multilib): New.
6062 * common/common-targhooks.h (default_compute_multilib): New.
6063 * common/common-targhooks.cc (default_compute_multilib): New.
6064 * doc/tm.texi.in (TARGET_COMPUTE_MULTILIB): New.
6065 * doc/tm.texi: Regen.
6066 * gcc.cc: Include common/common-target.h.
6067 (set_multilib_dir) Call targetm_common.compute_multilib.
6068 (SWITCH_LIVE): Move to opts.h.
6069 (SWITCH_FALSE): Ditto.
6070 (SWITCH_IGNORE): Ditto.
6071 (SWITCH_IGNORE_PERMANENTLY): Ditto.
6072 (SWITCH_KEEP_FOR_GCC): Ditto.
6073 (struct switchstr): Ditto.
6074 * opts.h (SWITCH_LIVE): Move from gcc.c.
6075 (SWITCH_FALSE): Ditto.
6076 (SWITCH_IGNORE): Ditto.
6077 (SWITCH_IGNORE_PERMANENTLY): Ditto.
6078 (SWITCH_KEEP_FOR_GCC): Ditto.
6079 (struct switchstr): Ditto.
6081 2022-09-02 Martin Liska <mliska@suse.cz>
6083 * config/pdp11/pdp11.h (PREFERRED_DEBUGGING_TYPE): Disable
6086 2022-09-02 Martin Liska <mliska@suse.cz>
6088 * Makefile.in: Remove -gstabs option support, DBX-related
6089 macros and DBX debugging info support.
6090 * collect2.cc (scan_prog_file): Likewise.
6091 * common.opt: Likewise.
6092 * config.gcc: Likewise.
6093 * config.in: Likewise.
6094 * config/aarch64/aarch64-elf.h (DBX_DEBUGGING_INFO): Likewise.
6095 * config/alpha/alpha.cc: Likewise.
6096 * config/alpha/elf.h (ASM_SPEC): Likewise.
6097 * config/arc/arc.h (DBX_DEBUGGING_INFO): Likewise.
6098 (DBX_CONTIN_LENGTH): Likewise.
6099 * config/arm/aout.h (DBX_DEBUGGING_INFO): Likewise.
6100 (DBX_CONTIN_LENGTH): Likewise.
6101 * config/arm/netbsd-elf.h (DBX_CONTIN_LENGTH): Likewise.
6102 * config/darwin.h (DSYMUTIL_SPEC): Likewise.
6103 (ASM_DEBUG_SPEC): Likewise.
6104 (DBX_DEBUGGING_INFO): Likewise.
6105 (DBX_USE_BINCL): Likewise.
6106 (DBX_CONTIN_LENGTH): Likewise.
6107 (DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END): Likewise.
6108 * config/epiphany/epiphany.h (DBX_DEBUGGING_INFO): Likewise.
6109 (DBX_CONTIN_LENGTH): Likewise.
6110 * config/i386/bsd.h (DBX_NO_XREFS): Likewise.
6111 (DBX_CONTIN_LENGTH): Likewise.
6112 * config/i386/gas.h (DBX_NO_XREFS): Likewise.
6113 (DBX_CONTIN_LENGTH): Likewise.
6114 * config/ia64/ia64.h: Likewise.
6115 * config/ia64/sysv4.h (DBX_DEBUGGING_INFO): Likewise.
6116 * config/m68k/linux.h (DBX_CONTIN_LENGTH): Likewise.
6117 * config/m68k/openbsd.h (DBX_DEBUGGING_INFO): Likewise.
6118 (DBX_CONTIN_LENGTH): Likewise.
6119 (DBX_CONTIN_CHAR): Likewise.
6120 * config/mips/mips.cc (mips_output_filename): Likewise.
6121 (mips_option_override): Likewise.
6122 * config/mips/mips.h (SUBTARGET_ASM_DEBUGGING_SPEC): Likewise.
6123 (DBX_DEBUGGING_INFO): Likewise.
6124 (DBX_CONTIN_LENGTH): Likewise.
6125 (DBX_REGISTER_NUMBER): Likewise.
6126 (GP_DBX_FIRST): Likewise.
6127 (FP_DBX_FIRST): Likewise.
6128 (MD_DBX_FIRST): Likewise.
6129 * config/nvptx/nvptx.cc: Likewise.
6130 * config/openbsd.h (DBX_NO_XREFS): Likewise.
6131 * config/pa/pa-64.h (DBX_DEBUGGING_INFO): Likewise.
6132 * config/pa/pa.h (ASSEMBLER_DIALECT): Likewise.
6133 (DBX_CONTIN_LENGTH): Likewise.
6134 * config/pa/som.h (PREFERRED_DEBUGGING_TYPE): Likewise.
6135 (DBX_USE_BINCL): Likewise.
6136 (DBX_LINES_FUNCTION_RELATIVE): Likewise.
6137 (DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END): Likewise.
6138 * config/pdp11/pdp11.cc: Likewise.
6139 * config/pdp11/pdp11.h (DBX_DEBUGGING_INFO): Likewise.
6140 (PREFERRED_DEBUGGING_TYPE): Likewise.
6141 (DBX_CONTIN_LENGTH): Likewise.
6142 * config/rs6000/rs6000-builtin.cc: Likewise.
6143 * config/rs6000/rs6000-call.cc: Likewise.
6144 * config/rs6000/rs6000-logue.cc (defined): Likewise.
6145 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Likewise.
6146 (HAVE_XCOFF_DWARF_EXTRAS): Likewise.
6147 (rs6000_xcoff_declare_function_name): Likewise.
6148 * config/rs6000/sysv4.h (DBX_DEBUGGING_INFO): Likewise.
6149 (DBX_FUNCTION_FIRST): Likewise.
6150 * config/rs6000/xcoff.h (XCOFF_DEBUGGING_INFO): Likewise.
6151 * config/rx/rx.h (DBX_DEBUGGING_INFO): Likewise.
6152 * config/sh/elf.h (DBX_LINES_FUNCTION_RELATIVE): Likewise.
6153 (DBX_OUTPUT_NULL_N_SO_AT_MAIN_SOURCE_FILE_END): Likewise.
6154 * config/sol2.h (NO_DBX_BNSYM_ENSYM): Likewise.
6155 * config/sparc/freebsd.h (DBX_CONTIN_CHAR): Likewise.
6156 * config/sparc/netbsd-elf.h (DBX_CONTIN_CHAR): Likewise.
6157 * config/sparc/sparc.h (DBX_CONTIN_LENGTH): Likewise.
6158 * config/vax/vax.cc (vax_file_start): Likewise.
6159 * config/vax/vax.h (DBX_DEBUGGING_INFO): Likewise.
6160 (DBX_CONTIN_LENGTH): Likewise.
6161 (DBX_CONTIN_CHAR): Likewise.
6162 (DBX_NO_XREFS): Likewise.
6163 (DBX_STATIC_STAB_DATA_SECTION): Likewise.
6164 * config/vx-common.h (DBX_DEBUGGING_INFO): Likewise.
6165 (XCOFF_DEBUGGING_INFO): Likewise.
6166 * configure: Regenerate. Likewise.
6167 * configure.ac: Likewise.
6168 * debug.h: Likewise.
6169 * doc/install.texi: Likewise.
6170 * doc/invoke.texi: Likewise.
6171 * doc/passes.texi: Likewise.
6172 * doc/tm.texi: Likewise.
6173 * doc/tm.texi.in: Likewise.
6174 * dwarf2asm.cc (XCOFF_DEBUGGING_INFO): Likewise.
6175 (dw2_asm_output_nstring): Likewise.
6176 (USE_LINKONCE_INDIRECT): Likewise.
6177 * dwarf2out.cc (XCOFF_DEBUGGING_INFO): Likewise.
6178 (HAVE_XCOFF_DWARF_EXTRAS): Likewise.
6179 (output_fde): Likewise.
6180 (output_call_frame_info): Likewise.
6181 (have_macinfo): Likewise.
6182 (add_AT_loc_list): Likewise.
6183 (add_AT_view_list): Likewise.
6184 (output_compilation_unit_header): Likewise.
6185 (output_pubnames): Likewise.
6186 (output_aranges): Likewise.
6187 (output_line_info): Likewise.
6188 (output_macinfo): Likewise.
6189 (dwarf2out_finish): Likewise.
6190 (dwarf2out_early_finish): Likewise.
6191 * final.cc (final_scan_insn_1): Likewise.
6192 (rest_of_handle_final): Likewise.
6193 * flag-types.h (enum debug_info_type): Likewise.
6194 (DBX_DEBUG): Likewise.
6195 (XCOFF_DEBUG): Likewise.
6196 * function.cc (defined): Likewise.
6197 * gcc.cc (defined): Likewise.
6198 (ASM_DEBUG_SPEC): Likewise.
6199 (ASM_DEBUG_OPTION_SPEC): Likewise.
6200 * opts.cc (common_handle_option): Likewise.
6201 (set_debug_level): Likewise.
6202 * system.h (fancy_abort): Likewise.
6203 * target-def.h (TARGET_ASM_CONSTRUCTOR): Likewise.
6204 (TARGET_ASM_DESTRUCTOR): Likewise.
6205 * toplev.cc (defined): Likewise.
6206 * varasm.cc: Likewise.
6207 * config/dbxcoff.h: Removed.
6208 * config/dbxelf.h: Removed.
6209 * dbxout.cc: Removed.
6210 * dbxout.h: Removed.
6212 * stab.def: Removed.
6213 * xcoffout.cc: Removed.
6214 * xcoffout.h: Removed.
6216 2022-09-02 Simon Rainer <gcc.gnu@vvalter.com>
6219 * config/i386/i386-features.cc (ix86_get_function_versions_dispatcher):
6220 Set TREE_NOTHROW correctly for dispatcher declaration.
6221 * config/rs6000/rs6000.cc (rs6000_get_function_versions_dispatcher):
6224 2022-09-01 Tamar Christina <tamar.christina@arm.com>
6227 * config/aarch64/aarch64.cc
6228 (aarch64_vector_costs::prefer_unrolled_loop): Replace %u with
6229 HOST_WIDE_INT_PRINT_UNSIGNED.
6231 2022-09-01 Richard Sandiford <richard.sandiford@arm.com>
6233 * tree-vect-slp.cc (vect_build_slp_tree_2): When building a
6234 VEC_PERM_EXPR of an existing vector, set the SLP_TREE_LANES
6235 to the number of vector elements, if that's a known constant.
6236 (vect_optimize_slp_pass::is_compatible_layout): Remove associated
6237 comment about zero SLP_TREE_LANES.
6238 (vect_optimize_slp_pass::start_choosing_layouts): Iterate over
6239 all partition members when looking for potential layouts.
6240 Handle existing permutes of fixed-length vectors.
6242 2022-09-01 Uroš Bizjak <ubizjak@gmail.com>
6245 * config/i386/i386.md (moves to/from AX_REG into xchg peephole2):
6246 Do not convert a move pattern where both operands are AX_REG.
6248 2022-09-01 Aldy Hernandez <aldyh@redhat.com>
6250 * range-op-float.cc (build_le): Convert to dconst*inf.
6252 * value-range.cc (frange::set_signbit): Same.
6253 (frange::normalize_kind): Same.
6254 (range_tests_floats): Same.
6255 * value-range.h (vrp_val_max): Same.
6256 (vrp_val_min): Same.
6257 (frange::set_varying): Same.
6259 2022-09-01 Aldy Hernandez <aldyh@redhat.com>
6261 * emit-rtl.cc (init_emit_once): Initialize dconstinf and
6263 * real.h: Add dconstinf and dconstninf.
6265 2022-09-01 Richard Biener <rguenther@suse.de>
6267 * gimple-predicate-analysis.cc (compute_control_dep_chain):
6268 Remove cycle detection, instead avoid walking backedges.
6270 2022-09-01 Richard Biener <rguenther@suse.de>
6272 * gimple-predicate-analysis.cc (compute_control_dep_chain):
6273 New wrapping overload.
6274 (uninit_analysis::init_use_preds): Simplify.
6275 (uninit_analysis::init_from_phi_def): Likewise.
6277 2022-09-01 Aldy Hernandez <aldyh@redhat.com>
6279 * gimple-range-fold.cc
6280 (fold_using_range::range_of_builtin_int_call): Add case for
6281 CFN_BUILT_IN_SIGNBIT.
6283 2022-09-01 Aldy Hernandez <aldyh@redhat.com>
6285 * range-op-float.cc (foperator_equal::op1_range): Do not copy sign
6287 (foperator_not_equal::op1_range): Same.
6288 * value-query.cc (range_query::get_tree_range): Set sign bit.
6289 * value-range-pretty-print.cc (vrange_printer::visit): Dump sign bit.
6290 * value-range.cc (frange::set_signbit): New.
6291 (frange::set): Adjust for sign bit.
6292 (frange::normalize_kind): Same.
6293 (frange::union_): Remove useless comment.
6294 (frange::intersect): Same.
6295 (frange::contains_p): Adjust for sign bit.
6296 (frange::singleton_p): Same.
6297 (frange::verify_range): Same.
6298 (range_tests_signbit): New tests.
6299 (range_tests_floats): Call range_tests_signbit.
6300 * value-range.h (class frange_props): Add signbit
6301 (class frange): Same.
6303 2022-09-01 Jakub Jelinek <jakub@redhat.com>
6306 * dumpfile.h (dump_printf_loc): Use ATTRIBUTE_GCC_DUMP_PRINTF (3, 4)
6307 instead of ATTRIBUTE_GCC_DUMP_PRINTF (3, 0).
6308 * tree-parloops.cc (parloops_is_slp_reduction): Cast pointers to
6309 derived types of gimple to gimple * to avoid -Wformat warnings.
6310 * tree-vect-loop-manip.cc (vect_set_loop_condition,
6311 vect_update_ivs_after_vectorizer): Likewise.
6312 * tree-vect-stmts.cc (vectorizable_load): Likewise.
6313 * tree-vect-patterns.cc (vect_split_statement,
6314 vect_recog_mulhs_pattern, vect_recog_average_pattern,
6315 vect_determine_precisions_from_range,
6316 vect_determine_precisions_from_users): Likewise.
6317 * gimple-loop-versioning.cc
6318 (loop_versioning::analyze_term_using_scevs): Likewise.
6319 * tree-vect-slp.cc (vect_build_slp_tree_1): Likewise.
6320 (vect_build_slp_tree): Cast slp_tree to void * to avoid
6322 (optimize_load_redistribution_1, vect_match_slp_patterns,
6323 vect_build_slp_instance, vect_optimize_slp_pass::materialize,
6324 vect_optimize_slp_pass::dump, vect_slp_convert_to_external,
6325 vect_slp_analyze_node_operations, vect_bb_partition_graph): Likewise.
6326 (vect_print_slp_tree): Likewise. Also use
6327 HOST_WIDE_INT_PRINT_UNSIGNED instead of %u.
6328 * tree-vect-loop.cc (vect_determine_vectorization_factor,
6329 vect_analyze_scalar_cycles_1, vect_analyze_loop_operations,
6330 vectorizable_induction, vect_transform_loop): Cast pointers to derived
6331 types of gimple to gimple * to avoid -Wformat warnings.
6332 (vect_analyze_loop_2): Cast slp_tree to void * to avoid
6334 (vect_estimate_min_profitable_iters): Use HOST_WIDE_INT_PRINT_UNSIGNED
6336 * tree-vect-slp-patterns.cc (vect_pattern_validate_optab): Use %G
6337 instead of %T and STMT_VINFO_STMT (SLP_TREE_REPRESENTATIVE (node))
6338 instead of SLP_TREE_DEF_TYPE (node).
6340 2022-09-01 Jakub Jelinek <jakub@redhat.com>
6343 * doc/invoke.texi (-Winvalid-utf8): Document it.
6345 2022-09-01 Aldy Hernandez <aldyh@redhat.com>
6347 PR tree-optimization/106785
6348 * value-range.cc (range_tests_nan): Adjust tests for !HONOR_NANS.
6349 (range_tests_floats): Same.
6351 2022-09-01 konglin1 <lingling.kong@intel.com>
6353 * tree-if-conv.cc (is_cond_scalar_reduction): Add MULT_EXPR
6356 2022-09-01 Peter Bergner <bergner@linux.ibm.com>
6359 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_mma_builtin):
6360 Enforce the use of a valid MMA pointer type.
6362 2022-09-01 zhongjuzhe <juzhe.zhong@rivai.ai>
6364 * config/riscv/riscv.cc (riscv_conditional_register_usage): Add vector
6367 2022-09-01 zhongjuzhe <juzhe.zhong@rivai.ai>
6369 * config/riscv/riscv.cc (riscv_const_insns): Add cost of poly_int.
6370 (riscv_output_move): Add csrr vlenb assembly.
6371 * config/riscv/riscv.md (move_type): Add csrr vlenb type.
6372 (ext): New attribute.
6373 (ext_enabled): Ditto.
6376 2022-09-01 zhongjuzhe <juzhe.zhong@rivai.ai>
6378 * config/riscv/constraints.md (TARGET_VECTOR ? V_REGS : NO_REGS): Add
6380 (TARGET_VECTOR ? VD_REGS : NO_REGS): Add "vd" constraint.
6381 (TARGET_VECTOR ? VM_REGS : NO_REGS): Add "vm" constraint.
6382 (vp): Add poly constraint.
6384 2022-09-01 zhongjuzhe <juzhe.zhong@rivai.ai>
6386 * config/riscv/riscv.h (enum reg_class): Change vype to vtype.
6388 2022-09-01 zhongjuzhe <juzhe.zhong@rivai.ai>
6390 * config/riscv/riscv.cc (riscv_convert_vector_bits): Change
6391 configuration according to TARGET_MIN_VLEN.
6392 * config/riscv/riscv.h (UNITS_PER_FP_REG): Fix comment.
6394 2022-08-31 Martin Liska <mliska@suse.cz>
6396 * config.build: Remove deprecated ports.
6397 * config.gcc: Likewise.
6398 * config.host: Likewise.
6399 * configure.ac: Likewise.
6400 * configure: Regenerate.
6401 * config/pa/pa-hpux10.h: Removed.
6402 * config/pa/pa-hpux10.opt: Removed.
6403 * config/pa/t-dce-thr: Removed.
6405 2022-08-31 Martin Liska <mliska@suse.cz>
6407 * Makefile.in: Always install limits.h and syslimits.h to
6409 * configure.ac: Assign STMP_FIXINC blank if
6410 --disable-fixincludes is used.
6411 * configure: Regenerate.
6413 2022-08-31 Richard Biener <rguenther@suse.de>
6415 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
6416 Assert the guard_bb isn't empty and has more than one successor.
6417 Drop appropriate parts of the predicate when an edge fails to
6418 register a predicate.
6419 (predicate::dump): Dump empty predicate as TRUE.
6421 2022-08-31 Richard Biener <rguenther@suse.de>
6423 PR tree-optimization/90994
6424 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
6425 Ignore exceptional control flow and skip the edge for the purpose of
6426 predicate generation also for non-calls.
6428 2022-08-31 Aldy Hernandez <aldyh@redhat.com>
6430 * value-range-storage.cc (frange_storage_slot::set_frange): Save
6432 (frange_storage_slot::get_frange): Restore endpoints.
6433 * value-range-storage.h (class frange_storage_slot): Add endpoint
6436 2022-08-31 Martin Liska <mliska@suse.cz>
6438 PR tree-optimization/106789
6439 * range-op-float.cc (default_frelop_fold_range): Remove the
6442 2022-08-31 Martin Liska <mliska@suse.cz>
6444 * value-range.h: Add more override keywords.
6446 2022-08-31 Martin Liska <mliska@suse.cz>
6448 * value-range.h: Add override.
6450 2022-08-31 Richard Biener <rguenther@suse.de>
6452 PR tree-optimization/65244
6453 * gimple-predicate-analysis.h (predicate::init_from_control_deps):
6454 Add argument to specify whether the predicate is for the USE.
6455 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
6456 Also include predicates effective fallthru control edges when
6457 the predicate is for the USE.
6459 2022-08-31 Richard Biener <rguenther@suse.de>
6461 PR tree-optimization/73550
6462 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
6463 Sanitize debug dumping. Handle case labels with a CASE_HIGH.
6464 (predicate::dump): Adjust for better readability.
6466 2022-08-31 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
6467 Jakub Jelinek <jakub@redhat.com>
6469 * omp-simd-clone.cc (simd_clone_adjust_return_type,
6470 simd_clone_adjust_argument_types): Use known_eq (veclen, 0U)
6471 instead of known_eq (veclen, 0) to avoid -Wsign-compare warnings.
6473 2022-08-31 Richard Sandiford <richard.sandiford@arm.com>
6475 * tree-vect-slp.cc (vect_optimize_slp_pass::dump): Remove bogus
6478 2022-08-31 zhongjuzhe <juzhe.zhong@rivai.ai>
6480 * tree-vect-loop-manip.cc (vect_gen_vector_loop_niters): Simply
6481 initialize const_vf to 0.
6483 2022-08-31 Martin Liska <mliska@suse.cz>
6485 * config.gcc: Remove cr16.
6487 2022-08-30 Martin Jambor <mjambor@suse.cz>
6489 * vec.h (array_slice): Add constructors for non-const reference to
6490 heap vector and pointers to heap vectors.
6492 2022-08-30 Aldy Hernandez <aldyh@redhat.com>
6494 * value-range.cc (early_nan_resolve): Change comment.
6495 (frange::union_): Handle union when one side is a NAN.
6496 (range_tests_nan): Add tests for NAN union.
6498 2022-08-30 Andrew Stubbs <ams@codesourcery.com>
6500 * config/gcn/gcn.cc (gcn_simd_clone_compute_vecsize_and_simdlen): New.
6501 (gcn_simd_clone_adjust): New.
6502 (gcn_simd_clone_usable): New.
6503 (TARGET_SIMD_CLONE_ADJUST): New.
6504 (TARGET_SIMD_CLONE_COMPUTE_VECSIZE_AND_SIMDLEN): New.
6505 (TARGET_SIMD_CLONE_USABLE): New.
6507 2022-08-30 Andrew Stubbs <ams@codesourcery.com>
6509 * doc/tm.texi: Regenerate.
6510 * omp-simd-clone.cc (simd_clone_adjust_return_type): Allow zero
6512 (simd_clone_adjust_argument_types): Likewise.
6513 * target.def (compute_vecsize_and_simdlen): Document the new
6514 vecsize_int and vecsize_float semantics.
6516 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6518 * expmed.cc (store_bit_field_1): Fix byte offset calculation
6519 for undefined structures.
6521 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6523 * params.opt (-param=vect-max-layout-candidates=): New parameter.
6524 * doc/invoke.texi (vect-max-layout-candidates): Document it.
6525 * tree-vectorizer.h (auto_lane_permutation_t): New typedef.
6526 (auto_load_permutation_t): Likewise.
6527 * tree-vect-slp.cc (vect_slp_node_weight): New function.
6528 (slpg_layout_cost): New class.
6529 (slpg_vertex): Replace perm_in and perm_out with partition,
6530 out_degree, weight and out_weight.
6531 (slpg_partition_info, slpg_partition_layout_costs): New classes.
6532 (vect_optimize_slp_pass): Likewise, cannibalizing some part of
6533 the previous vect_optimize_slp.
6534 (vect_optimize_slp): Use it.
6536 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6538 * hash-traits.h (vec_hash_base): New class.
6539 (vec_free_hash_base): Likewise.
6541 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6543 * hash-traits.h (int_hash_base): New struct, split out from...
6544 (int_hash): ...this class, which now inherits from int_hash_base.
6545 * hash-map-traits.h (unbounded_hashmap_traits): Take a template
6546 parameter for the key that provides hash and equality functions.
6547 (unbounded_int_hashmap_traits): Turn into a type alias of
6548 unbounded_hashmap_traits.
6550 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6552 * graphds.cc (graphds_scc): Add a pass-back parameter for the
6554 * graphds.h (graphds_scc): Update prototype accordingly.
6556 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6558 * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Split out from...
6559 (vect_transform_slp_perm_load): ...here. Use SLP_TREE_VECTYPE instead
6560 of STMT_VINFO_VECTYPE.
6562 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6564 * tree-vect-slp.cc (vectorizable_slp_permutation_1): Split out from...
6565 (vectorizable_slp_permutation): ...here.
6567 2022-08-30 Richard Sandiford <richard.sandiford@arm.com>
6569 * tree-vect-stmts.cc (get_related_vectype_for_scalar_type): Check
6570 that the requested number of units is interoperable with the requested
6573 2022-08-30 Martin Liska <mliska@suse.cz>
6575 * config.gcc: Remove the port.
6576 * config/m32c/rtems.h: Removed.
6578 2022-08-30 Richard Biener <rguenther@suse.de>
6580 PR tree-optimization/73550
6581 * gimple-predicate-analysis.cc (predicate::init_from_control_deps):
6582 Do not apply MAX_NUM_CHAINS again.
6584 2022-08-30 Richard Biener <rguenther@suse.de>
6586 * gimple-predicate-analysis.cc (format_edge_vec): Dump
6587 both source and destination.
6588 (dump_dep_chains): Remove.
6589 (uninit_analysis::init_use_preds): Remove redundant
6592 2022-08-30 Aldy Hernandez <aldyh@redhat.com>
6594 * value-range-storage.cc (frange_storage_slot::get_frange): Use
6596 * value-range.cc (frange::set_nan): New.
6597 (frange_nan): Move to header file.
6598 (range_tests_nan): Adjust frange_nan callers to pass type.
6600 * value-range.h (FRANGE_PROP_ACCESSOR): Remove.
6603 2022-08-30 Richard Biener <rguenther@suse.de>
6605 PR tree-optimization/67196
6606 * gimple-predicate-analysis.cc (uninit_analysis::is_use_guarded):
6607 Simplify and normalize use prediates before first use.
6609 2022-08-30 Richard Biener <rguenther@suse.de>
6611 * gimple-predicate-analysis.cc (dump_pred_chain): Fix
6612 parentizing and AND prepending.
6613 (predicate::dump): Do not dump the GENERIC expanded
6614 predicate, properly parentize and prepend ORs to the
6615 piecewise predicate dump.
6616 (build_pred_expr): Remove.
6618 2022-08-30 Aldy Hernandez <aldyh@redhat.com>
6620 * range-op-float.cc (finite_operand_p): New.
6625 (foperator_equal::fold_range): New implementation with endpoints.
6626 (foperator_equal::op1_range): Same.
6627 (foperator_not_equal::fold_range): Same.
6628 (foperator_not_equal::op1_range): Same.
6629 (foperator_lt::fold_range): Same.
6630 (foperator_lt::op1_range): Same.
6631 (foperator_lt::op2_range): Same.
6632 (foperator_le::fold_range): Same.
6633 (foperator_le::op1_range): Same.
6634 (foperator_le::op2_range): Same.
6635 (foperator_gt::fold_range): Same.
6636 (foperator_gt::op1_range): Same.
6637 (foperator_gt::op2_range): Same.
6638 (foperator_ge::fold_range): Same.
6639 (foperator_ge::op1_range): Same.
6640 (foperator_ge::op2_range): Same.
6642 2022-08-30 Aldy Hernandez <aldyh@redhat.com>
6644 * range-op-float.cc (frange_set_nan): New.
6645 (frange_drop_inf): New.
6646 (frange_drop_ninf): New.
6647 (foperator_equal::op1_range): Adjust for endpoints.
6648 (foperator_lt::op1_range): Same.
6649 (foperator_lt::op2_range): Same.
6650 (foperator_gt::op1_range): Same.
6651 (foperator_gt::op2_range): Same.
6652 (foperator_unordered::op1_range): Same.
6653 * value-query.cc (range_query::get_tree_range): Same.
6654 * value-range-pretty-print.cc (vrange_printer::visit): Same.
6655 * value-range-storage.cc (frange_storage_slot::get_frange): Same.
6656 * value-range.cc (frange::set): Same.
6657 (frange::normalize_kind): Same.
6658 (frange::union_): Same.
6659 (frange::intersect): Same.
6660 (frange::operator=): Same.
6661 (early_nan_resolve): New.
6662 (frange::contains_p): New.
6663 (frange::singleton_p): New.
6664 (frange::set_nonzero): New.
6665 (frange::nonzero_p): New.
6666 (frange::set_zero): New.
6667 (frange::zero_p): New.
6668 (frange::set_nonnegative): New.
6669 (frange_float): New.
6671 (range_tests_nan): New.
6672 (range_tests_signed_zeros): New.
6673 (range_tests_floats): New.
6675 * value-range.h (frange::lower_bound): New.
6676 (frange::upper_bound): New.
6677 (vrp_val_min): Use real_inf with a sign instead of negating inf.
6678 (frange::frange): New.
6679 (frange::set_varying): Adjust for endpoints.
6680 (real_max_representable): New.
6681 (real_min_representable): New.
6683 2022-08-30 Aldy Hernandez <aldyh@redhat.com>
6685 * match.pd ((cmp @0 zerop) real_zerop (negate@1 @0)): Add variant
6688 2022-08-30 Martin Liska <mliska@suse.cz>
6690 * config/s390/s390.cc (s390_rtx_costs): Use proper type as
6693 2022-08-30 Richard Biener <rguenther@suse.de>
6695 * tree-ssa-uninit.cc (warn_uninitialized_vars): Pre-compute
6696 the set of fallthru reachable blocks from function entry
6697 and use that to determine wlims.always_executed.
6699 2022-08-30 Richard Biener <rguenther@suse.de>
6701 PR tree-optimization/56654
6702 * tree-ssa-uninit.cc (cand_cmp): New.
6703 (find_uninit_use): First process all PHIs and collect candidate
6704 stmts, then sort those after RPO.
6705 (warn_uninitialized_phi): Pass on bb_to_rpo.
6706 (execute_late_warn_uninitialized): Compute and pass on
6707 reverse lookup of RPO number from basic block index.
6709 2022-08-30 Richard Biener <rguenther@suse.de>
6711 * gimple-predicate-analysis.h (uninit_analysis::operator()):
6713 * gimple-predicate-analysis.cc
6714 (uninit_analysis::collect_phi_def_edges): Use phi_arg_set,
6716 * tree-ssa-uninit.cc (defined_args): New global.
6717 (compute_uninit_opnds_pos): Mask with the recorded set
6718 of guarded maybe-uninitialized uses.
6719 (uninit_undef_val_t::operator()): Remove.
6720 (find_uninit_use): Process all PHI uses, recording the
6721 guarded ones and marking the PHI result as uninitialized
6723 (warn_uninitialized_phi): Adjust.
6724 (execute_late_warn_uninitialized): Get rid of the PHI worklist
6725 and instead walk the function in RPO order.
6726 * spellcheck.h (best_match::m_best_candidate_len): Initialize.
6728 2022-08-30 Tamar Christina <tamar.christina@arm.com>
6730 PR tree-optimization/106744
6731 * tree-ssa-phiopt.cc (minmax_replacement): Correct arguments.
6733 2022-08-30 Tamar Christina <tamar.christina@arm.com>
6735 * expmed.cc (store_bit_field_1): Initialize regnum to 0.
6737 2022-08-29 David Faust <david.faust@oracle.com>
6740 * config/bpf/coreout.cc (bpf_core_get_sou_member_index): Fix
6741 computation of index for anonymous members.
6743 2022-08-29 Jose E. Marchesi <jose.marchesi@oracle.com>
6745 * config/bpf/bpf.cc (bpf_target_macros): Define __bpf__ as a
6748 2022-08-29 H.J. Lu <hjl.tools@gmail.com>
6751 * config/i386/i386-expand.cc
6752 (ix86_avx256_split_vector_move_misalign): Handle E_V16BFmode.
6753 * config/i386/sse.md (V_256H): Add V16BF.
6755 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6757 * config/s390/s390.cc (s390_address_cost): Declare.
6758 (s390_hard_regno_nregs): Declare.
6759 (s390_rtx_costs): Add handling for REG and MEM in SET.
6761 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6763 * config/s390/s390.cc (expand_perm_with_vpdi): Recognize swap pattern.
6764 (is_reverse_perm_mask): New function.
6765 (expand_perm_with_rot): Recognize reverse pattern.
6766 (expand_perm_with_vstbrq): New function.
6767 (expand_perm_with_vster): Use vler/vster for element reversal on z15.
6768 (vectorize_vec_perm_const_1): Use.
6769 (s390_vectorize_vec_perm_const): Add expand functions.
6770 * config/s390/vx-builtins.md: Prefer vster over vler.
6772 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6774 * config/s390/s390.md: Remove UNSPEC_VEC_EXTRACT.
6775 * config/s390/vector.md: Rewrite patterns to use vec_select.
6776 * config/s390/vx-builtins.md (vec_scatter_element<V_HW_2:mode>_SI):
6779 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6782 * config/s390/vector.md (@vpdi4_2<mode>): New pattern.
6783 (rotl<mode>3_di): New pattern.
6784 * config/s390/vx-builtins.md: Use vpdi and verll for reversing
6787 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6789 * config/s390/s390.cc (s390_issue_rate): Add z15.
6791 2022-08-29 Robin Dapp <rdapp@linux.ibm.com>
6793 * common/config/s390/s390-common.cc: Enable -funroll-loops and
6794 -munroll-only-small-loops for OPT_LEVELS_2_PLUS_SPEED_ONLY.
6795 * config/s390/s390.cc (s390_loop_unroll_adjust): Do not unroll
6796 loops larger than 12 instructions.
6797 (s390_override_options_after_change): Set unroll options.
6798 (s390_option_override_internal): Likewise.
6799 * config/s390/s390.opt: Document munroll-only-small-loops.
6801 2022-08-29 Richard Biener <rguenther@suse.de>
6803 * gimple-predicate-analysis.cc (is_loop_exit,
6804 find_control_equiv_block): Inline into single caller ...
6805 (uninit_analysis::init_use_preds): ... here and refactor.
6807 2022-08-29 Richard Biener <rguenther@suse.de>
6809 * gimple-predicate-analysis.cc (compute_control_dep_chain):
6810 Inline is_loop_exit and refactor, add comment about
6813 2022-08-29 Kito Cheng <kito.cheng@sifive.com>
6815 * config/riscv/riscv.cc (riscv_frame_info): Introduce `reset(void)`;
6816 (riscv_frame_info::reset(void)): New.
6817 (riscv_compute_frame_info): Use riscv_frame_info::reset instead
6818 of memset when clean frame.
6820 2022-08-29 zhongjuzhe <juzhe.zhong@rivai.ai>
6822 * config/riscv/riscv.cc (riscv_v_ext_vector_mode_p): New function.
6823 (riscv_classify_address): Disallow PLUS/LO_SUM/CONST_INT address types for RVV.
6824 (riscv_address_insns): Add RVV modes condition.
6825 (riscv_binary_cost): Ditto.
6826 (riscv_rtx_costs): Adjust cost for RVV.
6827 (riscv_secondary_memory_needed): Add RVV modes condition.
6828 (riscv_hard_regno_nregs): Add RVV register allocation.
6829 (riscv_hard_regno_mode_ok): Add RVV register allocation.
6830 (riscv_class_max_nregs): Add RVV register allocation.
6831 * config/riscv/riscv.h (DWARF_FRAME_REGNUM): Add VL/VTYPE and vector registers in Dwarf.
6832 (UNITS_PER_V_REG): New macro.
6833 (FIRST_PSEUDO_REGISTER): Adjust first pseudo num for RVV.
6834 (V_REG_FIRST): New macro.
6835 (V_REG_LAST): Ditto.
6839 (VTYPE_REG_P): Ditto.
6840 (RISCV_DWARF_VL): Ditto.
6841 (RISCV_DWARF_VTYPE): Ditto.
6842 (enum reg_class): Add RVV register types.
6843 (REG_CLASS_CONTENTS): Add RVV register types.
6844 * config/riscv/riscv.md: Add VL/VTYPE register number constants.
6846 2022-08-29 zhongjuzhe <juzhe.zhong@rivai.ai>
6848 * config/riscv/riscv.md: Add new type for vector instructions.
6850 2022-08-28 Peter Bergner <bergner@linux.ibm.com>
6853 * config/rs6000/rs6000.cc (rs6000_invalid_conversion): Remove handling
6854 of MMA pointer conversions.
6856 2022-08-27 Xi Ruoyao <xry111@xry111.site>
6858 * config/i386/gcc-auto-profile: Regenerate.
6860 2022-08-26 Aldy Hernandez <aldyh@redhat.com>
6862 * real.cc (real_iszero): New.
6863 * real.h (real_iszero): New.
6865 2022-08-26 Aldy Hernandez <aldyh@redhat.com>
6867 * real.cc (real_isinf): New overload.
6868 (real_inf): Add sign argument.
6869 * real.h (real_isinf): New overload.
6870 (real_inf): Add sign argument.
6872 2022-08-26 Marek Polacek <polacek@redhat.com>
6875 * doc/invoke.texi: Document -Wself-move.
6877 2022-08-26 Aldy Hernandez <aldyh@redhat.com>
6879 * value-range.cc (vrange::set): Set varying.
6880 (vrange::set_nonzero): Same.
6881 (vrange::set_zero): Same.
6882 (vrange::set_nonnegative): Same.
6884 2022-08-26 Aldy Hernandez <aldyh@redhat.com>
6886 * range-op-float.cc (foperator_equal::op1_range): Do not blindly
6887 copy op2 range when honoring signed zeros.
6889 2022-08-26 Aldy Hernandez <aldyh@redhat.com>
6891 * tree-ssa-threadbackward.cc (possibly_profitable_path_p): Always
6893 (profitable_path_p): Same.
6895 2022-08-26 Richard Biener <rguenther@suse.de>
6897 * gimple-predicate-analysis.h
6898 (uninit_analysis::use_cannot_happen): Remove.
6899 * gimple-predicate-analysis.cc (can_be_invalidated_p): Remove.
6900 (uninit_analysis::use_cannot_happen): Likewise.
6901 (uninit_analysis::is_use_guarded): Do not call
6903 (dump_predicates): Remove.
6904 (simple_control_dep_chain): Remove edge overload.
6906 2022-08-26 Tobias Burnus <tobias@codesourcery.com>
6908 * internal-fn.cc (expand_GOMP_TARGET_REV): New.
6909 * internal-fn.def (GOMP_TARGET_REV): New.
6910 * lto-cgraph.cc (lto_output_node, verify_node_partition): Mark
6911 'omp target device_ancestor_host' as in_other_partition and don't
6913 * omp-low.cc (create_omp_child_function): Mark as 'noclone'.
6914 * omp-expand.cc (expand_omp_target): For reverse offload, remove
6915 sorry, use device = GOMP_DEVICE_HOST_FALLBACK and create
6916 empty-body nohost function.
6917 * omp-offload.cc (execute_omp_device_lower): Handle
6918 IFN_GOMP_TARGET_REV.
6919 (pass_omp_target_link::execute): For ACCEL_COMPILER, don't
6920 nullify fn argument for reverse offload
6922 2022-08-26 Jakub Jelinek <jakub@redhat.com>
6924 * builtins.def (BUILT_IN_ISSIGNALING): New built-in.
6925 * builtins.cc (expand_builtin_issignaling): New function.
6926 (expand_builtin_signbit): Don't overwrite target.
6927 (expand_builtin): Handle BUILT_IN_ISSIGNALING.
6928 (fold_builtin_classify): Likewise.
6929 (fold_builtin_1): Likewise.
6930 * optabs.def (issignaling_optab): New.
6931 * fold-const-call.cc (fold_const_call_ss): Handle
6932 BUILT_IN_ISSIGNALING.
6933 * config/i386/i386.md (issignalingxf2): New expander.
6934 * doc/extend.texi (__builtin_issignaling): Document.
6935 (__builtin_isinf, __builtin_isnan): Clarify behavior with
6937 * doc/md.texi (issignaling<mode>2): Likewise.
6939 2022-08-26 Jakub Jelinek <jakub@redhat.com>
6941 PR tree-optimization/106099
6942 * internal-fn.def (TRAP): Add ECF_LOOPING_CONST_OR_PURE flag.
6943 * tree-cfg.cc (execute_fixup_cfg): Add IFN_TRAP instead of
6944 __builtin_trap to avoid the need of vops.
6946 2022-08-26 Richard Biener <rguenther@suse.de>
6948 * gimple-predicate-analysis.cc (dfs_mark_dominating_region):
6950 (compute_control_dep_chain): Adjust to honor marked region
6952 (uninit_analysis::init_from_phi_def): Pre-mark the dominating
6953 region to improve compute_control_dep_chain walking.
6954 * vec.h (vec<T, va_heap, vl_ptr>::allocated): Add forwarder.
6956 2022-08-26 Richard Biener <rguenther@suse.de>
6958 * gimple-predicate-analysis.cc
6959 (uninit_analysis::collect_phi_def_edges): Only expand a
6960 PHI def edge when it is possibly undefined.
6962 2022-08-26 Martin Liska <mliska@suse.cz>
6964 * doc/extend.texi: Remove cr16 related stuff.
6965 * doc/install.texi: Likewise.
6966 * doc/invoke.texi: Likewise.
6967 * doc/md.texi: Likewise.
6968 * function-tests.cc (test_expansion_to_rtl): Likewise.
6969 * common/config/cr16/cr16-common.cc: Removed.
6970 * config/cr16/constraints.md: Removed.
6971 * config/cr16/cr16-protos.h: Removed.
6972 * config/cr16/cr16.cc: Removed.
6973 * config/cr16/cr16.h: Removed.
6974 * config/cr16/cr16.md: Removed.
6975 * config/cr16/cr16.opt: Removed.
6976 * config/cr16/predicates.md: Removed.
6977 * config/cr16/t-cr16: Removed.
6979 2022-08-26 liuhongt <hongtao.liu@intel.com>
6982 * config/i386/i386-builtin.def (BDESC): Add
6983 CODE_FOR_avx_blendvpd256/CODE_FOR_avx_blendvps256 to
6984 corresponding builtins.
6985 * config/i386/i386.cc (ix86_gimple_fold_builtin):
6986 Don't fold IX86_BUILTIN_PBLENDVB256, IX86_BUILTIN_BLENDVPS256,
6987 IX86_BUILTIN_BLENDVPD256 w/o TARGET_AVX2.
6989 2022-08-25 Marek Polacek <polacek@redhat.com>
6991 * ginclude/stddef.h: Define nullptr_t.
6993 2022-08-25 Joseph Myers <joseph@codesourcery.com>
6995 * gimplify.cc (gimplify_modify_expr): Convert initialization from
6996 a variable-size CONSTRUCTOR to memset before call to
6997 gimplify_modify_expr_rhs.
6999 2022-08-25 Jason Merrill <jason@redhat.com>
7001 * dwarf2out.cc (base_type_die): Also use DW_ATE_UTF for char8_t.
7003 2022-08-25 Andreas Krebbel <krebbel@linux.ibm.com>
7006 * config/s390/predicates.md (subreg_register_operand): New
7008 * config/s390/s390-protos.h (s390_gen_lowpart_subreg): New
7010 * config/s390/s390.cc (s390_gen_lowpart_subreg): New function.
7011 (s390_expand_insv): Use s390_gen_lowpart_subreg instead of
7013 * config/s390/s390.md ("*get_tp_64", "*zero_extendhisi2_31")
7014 ("*zero_extendqisi2_31", "*zero_extendqihi2_31"): Likewise.
7015 ("movstrictqi", "movstricthi", "movstrictsi"): Use the
7016 subreg_register_operand predicate instead of register_operand.
7018 2022-08-25 Xi Ruoyao <xry111@xry111.site>
7020 * config/loongarch/loongarch-protos.h (loongarch_symbol_type):
7021 Add SYMBOL_PCREL64 and change the description for SYMBOL_PCREL.
7022 * config/loongarch/loongarch.cc (loongarch_attribute_table):
7023 New attribute table.
7024 (TARGET_ATTRIBUTE_TABLE): Define the target hook.
7025 (loongarch_handle_model_attribute): New static function.
7026 (loongarch_classify_symbol): Take TARGET_CMODEL_EXTREME and the
7027 model attribute of SYMBOL_REF_DECL into account returning
7028 SYMBOL_PCREL or SYMBOL_PCREL64.
7029 (loongarch_use_anchors_for_symbol_p): New static function.
7030 (TARGET_USE_ANCHORS_FOR_SYMBOL_P): Define the target hook.
7031 (loongarch_symbol_extreme_p): New static function.
7032 (loongarch_symbolic_constant_p): Handle SYMBOL_PCREL64.
7033 (loongarch_symbol_insns): Likewise.
7034 (loongarch_split_symbol_type): Likewise.
7035 (loongarch_split_symbol): Check SYMBOL_PCREL64 instead of
7036 TARGET_CMODEL_EXTREME for PC-relative addressing.
7037 (loongarch_print_operand_reloc): Likewise.
7038 * doc/extend.texi (Variable Attributes): Document new
7039 LoongArch specific attribute.
7041 2022-08-25 Xi Ruoyao <xry111@xry111.site>
7043 * config/loongarch/loongarch.cc (loongarch_classify_symbol):
7044 Return early if the rtx is not SYMBOL_REF.
7046 2022-08-25 Richard Biener <rguenther@suse.de>
7048 PR tree-optimization/106737
7049 * tree-parloops.cc (transform_to_exit_first_loop_alt): Do not
7052 2022-08-25 Chenghua Xu <xuchenghua@loongson.cn>
7055 * config/loongarch/loongarch.cc (loongarch_build_integer):
7057 * config/loongarch/loongarch.h (IMM_REACH): Likewise.
7058 (HWIT_1U): New Defined.
7059 (LU12I_OPERAND): Use HOST_WIDE_INT.
7060 (LU32I_OPERAND): Likewise.
7061 (LU52I_OPERAND): Likewise.
7062 (HWIT_UC_0xFFF): Likwise.
7064 2022-08-24 Andrew Pinski <apinski@marvell.com>
7068 * config/riscv/bitmanip.md (*shNadduw): Use n constraint
7070 (*slliuw): Likewise.
7071 (*bexti): Likewise. Also add a check for operands[2] to be less
7072 than the mode bitsize.
7074 2022-08-24 Andrew Pinski <apinski@marvell.com>
7076 * config/riscv/constraints.md (DbS): New constraint.
7077 (DnS): New constraint.
7078 * config/riscv/bitmanip.md (*bset<mode>_1_mask): Use new constraint.
7079 (*bclr<mode>): Likewise.
7080 (*binvi<mode>): Likewise.
7082 2022-08-24 Andrew Pinski <apinski@marvell.com>
7085 * config/riscv/predicates.md (single_bit_mask_operand):
7086 Use SINGLE_BIT_MASK_OPERAND instead of directly calling pow2p_hwi.
7087 (not_single_bit_mask_operand): Likewise.
7088 * config/riscv/riscv.cc (riscv_build_integer_1): Don't special case
7089 1<<31 for 32bits as it is already handled.
7090 Call trunc_int_for_mode on the upper part after the subtraction.
7091 (riscv_move_integer): Call trunc_int_for_mode before generating
7092 the integer just make sure the constant has been sign extended
7094 (riscv_emit_int_compare): Call trunc_int_for_mode after doing the
7095 addition for the new rhs.
7096 * config/riscv/riscv.h (SINGLE_BIT_MASK_OPERAND): If !TARGET64BIT,
7097 then mask off the upper 32bits of the HWI as it will be sign extended.
7099 2022-08-24 Andrew Pinski <apinski@marvell.com>
7101 * config/riscv/constraints.md (DsS): New constraint.
7102 (DsD): New constraint.
7103 * config/riscv/iterators.md (shiftm1c): New iterator.
7104 * config/riscv/bitmanip.md (*bset<mode>_mask):
7106 (*bset<mode>_1_mask): Likewise.
7108 2022-08-24 Andrew Pinski <apinski@marvell.com>
7110 * config/riscv/constraints.md (Ds3): New constraint.
7111 * config/riscv/predicates.md (imm123_operand): New predicate.
7112 * config/riscv/bitmanip.md (*shNadd): Use Ds3 and imm123_operand.
7113 (*shNadduw): Likewise.
7115 2022-08-24 Andrew Pinski <apinski@marvell.com>
7117 * config/riscv/riscv.cc (riscv_print_operand):
7119 (riscv_print_operand_punct_valid_p): New function
7120 (TARGET_PRINT_OPERAND_PUNCT_VALID_P): Define.
7121 * config/riscv/bitmanip.md (<bitmanip_optab>si2/clz_ctz_pcnt):
7122 Use %~ instead of conditional the pattern on TARGET_64BIT.
7123 (rotrsi3): Likewise.
7124 (rotlsi3): Likewise.
7125 * config/riscv/riscv.md: Add ~ to the list of modifiers.
7126 (addsi3): Use %~ instead of conditional the pattern on TARGET_64BIT.
7130 (optab>si3/any_div): Likewise.
7131 (*add<mode>hi3): Likewise.
7132 (<optab>si3/any_shift): Likewise.
7134 2022-08-24 Andrew Pinski <apinski@marvell.com>
7136 * config/riscv/riscv.cc (riscv_print_operand): Make a mention to
7137 keep the list in riscv.md in sync with this list.
7138 * config/riscv/riscv.md: Add list of modifiers as comments.
7140 2022-08-24 Andrew Pinski <apinski@marvell.com>
7142 * config/riscv/sync.md (any_atomic, atomic_optab): Move to ...
7143 * config/riscv/iterators.md: Here.
7145 2022-08-24 Andrew Pinski <apinski@marvell.com>
7147 * config/riscv/bitmanip.md
7148 (bitmanip_bitwise, bitmanip_minmax, clz_ctz_pcna,
7149 tbitmanip_optab, bitmanip_insn, shiftm1): Move to ...
7150 * config/riscv/iterators.md: Here.
7152 2022-08-24 Andrew Pinski <apinski@marvell.com>
7154 * config/riscv/riscv.md (GPR): Move to new file.
7155 (P, X, BR): Likewise.
7156 (MOVE32, MOVE64, SHORT): Likewise.
7157 (HISI, SUPERQI, SUBX): Likewise.
7158 (ANYI, ANYF, SOFTF): Likewise.
7159 (size, load, default_load): Likewise.
7160 (softload, store, softstore): Likewise.
7161 (reg, fmt, ifmt, amo): Likewise.
7162 (UNITMODE, HALFMODE): Likewise.
7163 (RINT, rint_pattern, rint_rm): Likewise.
7164 (QUIET_COMPARISON, quiet_pattern, QUIET_PATTERN): Likewise.
7165 (any_extend, any_shiftrt, any_shift): Likewise.
7166 (any_bitwise): Likewise.
7167 (any_div, any_mod): Likewise.
7168 (any_gt, any_ge, any_lt, any_le): Likewise.
7170 (optab, insn): Likewise.
7171 * config/riscv/iterators.md: New file.
7173 2022-08-24 Andrew Pinski <apinski@marvell.com>
7176 * config/riscv/bitmanip.md (bswaphi2): New pattern.
7178 2022-08-24 Andrew Pinski <apinski@marvell.com>
7181 * config/riscv/bitmanip.md (bswap<mode>2): Remove
7182 condition on TARGET_64BIT as X is already conditional there.
7184 2022-08-24 Joseph Myers <joseph@codesourcery.com>
7186 * tree.cc (build_real): Give DFP dconst0 the minimum quantum
7187 exponent for the type.
7189 2022-08-24 Jose E. Marchesi <jose.marchesi@oracle.com>
7192 * config/bpf/bpf.cc (bpf_legitimate_address_p): Recognize integer
7193 constants as legitimate addresses for functions.
7194 (bpf_small_register_classes_for_mode_p): Define target hook.
7196 2022-08-24 Richard Biener <rguenther@suse.de>
7198 * gimple-predicate-analysis.cc: Move predicate normalization
7199 after the comment documenting it.
7201 2022-08-24 Richard Biener <rguenther@suse.de>
7203 * gimple-predicate-analysis.h (predicate): Split out
7204 non-predicate related functionality into ..
7205 (uninit_analysis): .. this new class.
7206 * gimple-predicate-analysis.cc: Refactor into two classes.
7207 * tree-ssa-uninit.cc (find_uninit_use): Use uninit_analysis.
7209 2022-08-24 Richard Biener <rguenther@suse.de>
7211 * gimple-predicate-analysis.cc (predicate::use_cannot_happen):
7212 Do simple_control_dep_chain only up to cd_root, add the PHI
7213 operand edge to the chains like init_from_phi_def does.
7214 (predicate::is_use_guarded): Speedup early out, avoid half-way
7215 initializing the PHI def predicate.
7217 2022-08-24 Jakub Jelinek <jakub@redhat.com>
7220 * config/i386/sse.md (shuffletype): Add V32BF, V16BF and V8BF entries.
7221 Change V32HF, V16HF and V8HF entries from "f" to "i".
7222 (iptr): Add V32BF, V16BF, V8BF and BF entries.
7223 (i128vldq): Add V16HF and V16BF entries.
7224 (avx512er_vmrcp28<mode><mask_name><round_saeonly_name>): Fix typo,
7225 mask_opernad3 -> mask_operand3.
7227 2022-08-24 Martin Liska <mliska@suse.cz>
7228 Jørgen Kvalsvik <j@lambda.is>
7230 * gcov.cc (add_line_counts): Add group functions to coverage
7232 (accumulate_line_counts): Similarly for files.
7234 2022-08-24 Lulu Cheng <chenglulu@loongson.cn>
7236 * config/loongarch/genopts/loongarch-strings: Support code model medium.
7237 * config/loongarch/genopts/loongarch.opt.in: Likewise.
7238 * config/loongarch/loongarch-def.c: Likewise.
7239 * config/loongarch/loongarch-def.h (CMODEL_LARGE): Likewise.
7240 (CMODEL_EXTREME): Likewise.
7241 (N_CMODEL_TYPES): Likewise.
7242 (CMODEL_MEDIUM): Likewise.
7243 * config/loongarch/loongarch-opts.cc: Likewise.
7244 * config/loongarch/loongarch-opts.h (TARGET_CMODEL_MEDIUM): Likewise.
7245 * config/loongarch/loongarch-str.h (STR_CMODEL_MEDIUM): Likewise.
7246 * config/loongarch/loongarch.cc (loongarch_call_tls_get_addr):
7247 Tls symbol Loading support medium mode.
7248 (loongarch_legitimize_call_address): When medium mode, make a symbolic
7249 jump with two instructions.
7250 (loongarch_option_override_internal): Support medium.
7251 * config/loongarch/loongarch.md (@pcalau12i<mode>): New template.
7252 (@sibcall_internal_1<mode>): New function call templates added to support
7254 (@sibcall_value_internal_1<mode>): Likewise.
7255 (@sibcall_value_multiple_internal_1<mode>): Likewise.
7256 (@call_internal_1<mode>): Likewise.
7257 (@call_value_internal_1<mode>): Likewise.
7258 (@call_value_multiple_internal_1<mode>): Likewise.
7259 * config/loongarch/loongarch.opt: Support medium.
7260 * config/loongarch/predicates.md: Add processing about medium mode.
7261 * doc/invoke.texi: Document for '-mcmodel=medium'.
7263 2022-08-24 Richard Biener <rguenther@suse.de>
7265 * gimple-predicate-analysis.cc (predicate::use_cannot_happen):
7266 Start the compute_control_dep_chain walk from the immediate
7267 dominator of the PHI.
7269 2022-08-23 H.J. Lu <hjl.tools@gmail.com>
7272 * config/i386/amxtileintrin.h (_tile_loadd_internal): Cast to
7274 (_tile_stream_loadd_internal): Likewise.
7275 (_tile_stored_internal): Likewise.
7277 2022-08-23 Richard Biener <rguenther@suse.de>
7279 PR tree-optimization/106722
7280 * gimple-predicate-analysis.h (MAX_NUM_CHAINS, MAX_CHAIN_LEN,
7281 MAX_POSTDOM_CHECK, MAX_SWITCH_CASES): Move ...
7282 * gimple-predicate-analysis.cc: ... here and document.
7283 (simple_control_dep_chain): New function, factored from
7284 predicate::use_cannot_happen.
7285 (predicate::use_cannot_happen): Adjust.
7286 (predicate::predicate): Use simple_control_dep_chain as fallback.
7288 2022-08-23 Aldy Hernandez <aldyh@redhat.com>
7290 * range-op-float.cc (foperator_equal::op1_range): Set range to
7293 2022-08-23 Richard Biener <rguenther@suse.de>
7295 * gimple-predicate-analysis.cc (is_loop_exit): Split out
7297 (is_non_loop_exit_postdominating): ... here. Remove after
7299 (find_control_equiv_block): ... here.
7300 (compute_control_dep_chain): ... and here.
7301 (predicate::is_use_guarded): Do not excempt loop exits
7302 from short-cutting the case of the use post-dominating the
7305 2022-08-23 Andrew MacLeod <amacleod@redhat.com>
7307 PR tree-optimization/106687
7308 * range-op.cc (operator_minus::lhs_op1_relation): Return VREL_LE
7309 for the VREL_GT case as well.
7311 2022-08-22 Dimitar Dimitrov <dimitar@dinux.eu>
7313 * config/pru/pru.md (pru_<code>di3): New alternative for
7314 two operands but without earlyclobber.
7316 2022-08-22 Dimitar Dimitrov <dimitar@dinux.eu>
7318 * config/pru/pru.md (prumov<mode>, mov<mode>): Add
7319 variants for loading -1 consts.
7321 2022-08-22 Dimitar Dimitrov <dimitar@dinux.eu>
7324 * config/pru/constraints.md (Um): New constraint for -1.
7325 (Uf): New constraint for IOR fill-bytes constants.
7326 (Uz): New constraint for AND zero-bytes constants.
7327 * config/pru/predicates.md (const_fillbytes_operand): New
7328 predicate for IOR fill-bytes constants.
7329 (const_zerobytes_operand): New predicate for AND zero-bytes
7331 * config/pru/pru-protos.h (pru_output_sign_extend): Remove.
7332 (struct pru_byterange): New struct to describe a byte range.
7333 (pru_calc_byterange): New declaration.
7334 * config/pru/pru.cc (pru_rtx_costs): Add penalty for
7336 (pru_output_sign_extend): Remove.
7337 (pru_calc_byterange): New helper function to extract byte
7338 range info from a constant.
7339 (pru_print_operand): Remove 'y' and 'z' print modifiers.
7340 * config/pru/pru.md (zero_extendqidi2): New pattern.
7341 (zero_extendhidi2): New pattern.
7342 (zero_extendsidi2): New pattern.
7343 (extend<EQS0:mode><EQD:mode>2): Rewrite as an expand.
7344 (@pru_ior_fillbytes<mode>): New pattern.
7345 (@pru_and_zerobytes<mode>): New pattern.
7346 (<code>di3): Rewrite as an expand and handle ZERO and FILL
7348 (pru_<code>di3): New name for <code>di3.
7349 (@cbranch_qbbx_const_<BIT_TEST:code><HIDI:mode>): New pattern to
7350 handle bit-test for 64-bit registers.
7352 2022-08-22 Richard Biener <rguenther@suse.de>
7354 * gimple-predicate-analysis.h (predicate::m_use_expr): Remove.
7355 (predicate::def_expr): Likewise.
7356 (predicate::use_expr): Likewise.
7357 (predicate::expr): Likewise.
7358 * gimple-predicate-analysis.cc (predicate::def_expr): Remove.
7359 (predicate::use_expr): Likewise.
7360 (predicate::expr): Likewise.
7361 (predicate::is_use_guarded): Do not build m_use_expr.
7363 2022-08-22 Martin Liska <mliska@suse.cz>
7366 * configure.ac: Detect O_NONBLOCK flag for open.
7367 * config.in: Regenerate.
7368 * configure: Regenerate.
7369 * opts-common.cc (jobserver_info::connect): Set is_connected
7370 properly based on O_NONBLOCK.
7371 * opts-jobserver.h (struct jobserver_info): Add is_connected
7374 2022-08-22 zhongjuzhe <juzhe.zhong@rivai.ai>
7376 * simplify-rtx.cc (test_vector_subregs_fore_back): Make first value
7377 and repeat value different.
7379 2022-08-22 Tobias Burnus <tobias@codesourcery.com>
7382 * lto-wrapper.cc (free_array_of_ptrs): Move before tool_cleanup.
7383 (tool_cleanup): Unlink offload_names.
7384 (compile_offload_image): Take filename argument to set it early.
7385 (compile_images_for_offload_targets): Update call; set
7386 offload_names to NULL after freeing the array.
7388 2022-08-22 Richard Biener <rguenther@suse.de>
7390 PR tree-optimization/105937
7391 * tree-ssa-uninit.cc (find_uninit_use): Do not queue PHIs
7393 (execute_late_warn_uninitialized): Mark backedges.
7395 2022-08-22 Richard Biener <rguenther@suse.de>
7397 * gimple-predicate-analysis.cc (predicate::use_cannot_happen):
7398 If the use is guarded with multiple predicate paths compute
7399 the predicates intersection before going forward. When
7400 compute_control_dep_chain wasn't able to come up with at
7401 least one path from function entry to the PHI edge compute
7402 a conservative sparse path instead.
7404 2022-08-20 Lulu Cheng <chenglulu@loongson.cn>
7406 * config/loongarch/loongarch-opts.cc: Allow cmodel to be extreme.
7407 * config/loongarch/loongarch.cc (loongarch_call_tls_get_addr):
7408 Add extreme support for TLS GD and LD types.
7409 (loongarch_legitimize_tls_address): Add extreme support for TLS LE
7411 (loongarch_split_symbol): When compiling with -mcmodel=extreme,
7412 the symbol address will be obtained through five instructions.
7413 (loongarch_print_operand_reloc): Add support.
7414 (loongarch_print_operand): Add support.
7415 (loongarch_print_operand_address): Add support.
7416 (loongarch_option_override_internal): Set '-mcmodel=extreme' option
7417 incompatible with '-mno-explicit-relocs'.
7418 * config/loongarch/loongarch.md (@lui_l_hi20<mode>):
7419 Loads bits 12-31 of data into registers.
7420 (lui_h_lo20): Load bits 32-51 of the data and spell bits 0-31 of
7421 the source register.
7422 (lui_h_hi12): Load bits 52-63 of the data and spell bits 0-51 of
7423 the source register.
7424 * config/loongarch/predicates.md: Symbols need to be decomposed
7425 when defining the macro TARGET_CMODEL_EXTREME
7426 * doc/invoke.texi: Modify the description information of cmodel in the document.
7427 Document -W[no-]extreme-plt.
7429 2022-08-19 Tobias Burnus <tobias@codesourcery.com>
7431 * config/gcn/mkoffload.cc (main): Add omp_requires_file and dbgobj to
7433 * config/i386/intelmic-mkoffload.cc (prepare_target_image): Add
7434 omp_requires_file to temp_files.
7435 * config/nvptx/mkoffload.cc (omp_requires_file): New global static var.
7436 (main): Remove local omp_requires_file var.
7437 (tool_cleanup): Handle omp_requires_file.
7439 2022-08-19 Aldy Hernandez <aldyh@redhat.com>
7441 * gimple-range-path.cc (path_range_query::path_range_query):
7442 Remove constructor that takes edge.
7443 * gimple-range-path.h (class path_range_query): Same.
7444 * tree-ssa-loop-ch.cc (edge_range_query): New.
7445 (entry_loop_condition_is_static): Call edge_range_query.
7447 2022-08-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7449 * config/xtensa/xtensa.h
7450 (enum reg_class, REG_CLASS_NAMES, REG_CLASS_CONTENTS):
7451 Add new register class "ISC_REGS".
7452 * config/xtensa/constraints.md (c): Add new register constraint.
7453 * config/xtensa/xtensa.md (define_constants): Remove "A11_REG".
7454 (sibcall_internal, sibcall_value_internal):
7455 Change to use the new register constraint, and remove two split
7456 patterns for fixups that are no longer needed.
7458 2022-08-18 Maciej W. Rozycki <macro@embecosm.com>
7460 * config/riscv/riscv.md (*mov<GPR:mode><X:mode>cc): Fix output
7463 2022-08-18 Tim Lange <mail@tim-lange.me>
7466 * doc/invoke.texi: Add Wanalyzer-imprecise-fp-arithmetic.
7468 2022-08-18 Aldy Hernandez <aldyh@redhat.com>
7470 * gimple-range-path.cc (path_range_query::path_range_query): Add
7471 various constructors to take a path.
7472 (path_range_query::~path_range_query): Remove m_alloced_ranger.
7473 (path_range_query::range_on_path_entry): Adjust for m_ranger being
7475 (path_range_query::set_path): Rename to...
7476 (path_range_query::reset_path): ...this and call compute_ranges.
7477 (path_range_query::ssa_range_in_phi): Adjust for m_ranger
7479 (path_range_query::range_defined_in_block): Same.
7480 (path_range_query::compute_ranges_in_block): Same.
7481 (path_range_query::adjust_for_non_null_uses): Same.
7482 (path_range_query::compute_exit_dependencies): Use m_path instead
7484 (path_range_query::compute_ranges): Remove path argument.
7485 (path_range_query::range_of_stmt): Adjust for m_ranger reference.
7486 (path_range_query::compute_outgoing_relations): Same.
7487 * gimple-range-path.h (class path_range_query): Add various
7489 Make compute_ranges and compute_exit_dependencies private.
7490 Rename set_path to reset_path.
7491 Make m_ranger a reference.
7492 Remove m_alloced_ranger.
7493 * tree-ssa-dom.cc (pass_dominator::execute): Adjust constructor to
7495 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Take a
7496 ranger and instantiate a new path_range_query every time.
7497 (ch_base::copy_headers): Pass ranger instead of path_range_query.
7498 * tree-ssa-threadbackward.cc (class back_threader): Remove m_solver.
7499 (back_threader::~back_threader): Remove m_solver.
7500 (back_threader::find_taken_edge_switch): Adjust for m_ranger
7502 (back_threader::find_taken_edge_cond): Same.
7503 (back_threader::dump): Remove m_solver.
7504 (back_threader::back_threader): Move verify_marked_backedges
7505 here from the path_range_query constructor.
7506 * tree-ssa-threadedge.cc (hybrid_jt_simplifier::simplify): Move
7507 some code from compute_ranges_from_state here.
7508 (hybrid_jt_simplifier::compute_ranges_from_state): Rename...
7509 (hybrid_jt_simplifier::compute_exit_dependencies): ...to this.
7510 * tree-ssa-threadedge.h (class hybrid_jt_simplifier): Rename
7511 compute_ranges_from_state to compute_exit_dependencies.
7514 2022-08-18 Richard Biener <rguenther@suse.de>
7516 PR middle-end/106617
7517 * match.pd ((a ? b : c) > d -> a ? (b > d) : (c > d)): Fix
7518 guard, disable on GENERIC to not cause quadratic behavior
7519 with the fold-const.cc implementation and the use of !
7521 2022-08-18 Andrew Pinski <apinski@marvell.com>
7523 PR gcov-profile/106659
7524 * gcov-dump.cc (INCLUDE_VECTOR): Include vector.h with
7527 2022-08-18 konglin1 <lingling.kong@intel.com>
7529 * config/i386/i386-expand.cc (ix86_expand_sse_movcc): Handle vector
7531 (ix86_expand_vector_init_duplicate): Support vector BFmode.
7532 (ix86_expand_vector_init_one_nonzero): Ditto.
7533 (ix86_expand_vector_init_one_var): Ditto.
7534 (ix86_expand_vector_init_concat): Ditto.
7535 (ix86_expand_vector_init_interleave): Ditto.
7536 (ix86_expand_vector_init_general): Ditto.
7537 (ix86_expand_vector_init): Ditto.
7538 (ix86_expand_vector_set_var): Ditto.
7539 (ix86_expand_vector_set): Ditto.
7540 (ix86_expand_vector_extract): Ditto.
7541 * config/i386/i386.cc (classify_argument): Add BF vector modes.
7542 (function_arg_64): Ditto.
7543 (ix86_gimplify_va_arg): Ditto.
7544 (ix86_get_ssemov): Ditto.
7545 * config/i386/i386.h (VALID_AVX256_REG_MODE): Add BF vector modes.
7546 (VALID_AVX512F_REG_MODE): Ditto.
7547 (host_detect_local_cpu): Ditto.
7548 (VALID_SSE2_REG_MODE): Ditto.
7549 * config/i386/i386.md: Add BF vector modes.
7551 (ssemodesuffix): Add bf suffix for BF vector modes.
7552 (ssevecmode): Ditto.
7553 * config/i386/sse.md (VMOVE): Adjust for BF vector modes.
7554 (VI12HFBF_AVX512VL): Ditto.
7556 (VF_AVX512HFBF16): Ditto.
7557 (VF_AVX512BWHFBF16): Ditto.
7560 (VIHFBF_256): Ditto.
7561 (VIHFBF_AVX512BW): Ditto.
7562 (VI2F_256_512):Ditto.
7566 (sseinsnmode): Ditto.
7567 (sseconstm1): Ditto.
7568 (sseintmodesuffix): New mode_attr.
7569 (avx512fmaskmode): Ditto.
7570 (avx512fmaskmodelower): Ditto.
7571 (ssedoublevecmode): Ditto.
7572 (ssehalfvecmode): Ditto.
7573 (ssehalfvecmodelower): Ditto.
7574 (ssescalarmode): Add vector BFmode mapping.
7575 (ssescalarmodelower): Ditto.
7576 (ssexmmmode): Ditto.
7577 (ternlogsuffix): Ditto.
7578 (ssescalarsize): Ditto.
7579 (sseintprefix): Ditto.
7582 (bcstscalarsuff): Ditto.
7583 (<avx512>_blendm<mode>): New define_insn for BFmode.
7584 (<avx512>_store<mode>_mask): Ditto.
7585 (vcond_mask_<mode><avx512fmaskmodelower>): Ditto.
7586 (vec_set<mode>_0): New define_insn for BF vector set.
7587 (V8BFH_128): New mode_iterator for BFmode.
7588 (avx512fp16_mov<mode>): Ditto.
7589 (vec_set<mode>): New define_insn for BF vector set.
7590 (@vec_extract_hi_<mode>): Ditto.
7591 (@vec_extract_lo_<mode>): Ditto.
7592 (vec_set_hi_<mode>): Ditto.
7593 (vec_set_lo_<mode>): Ditto.
7594 (*vec_extract<mode>_0): New define_insn_and_split for BF
7596 (*vec_extract<mode>): New define_insn.
7597 (VEC_EXTRACT_MODE): Add BF vector modes.
7598 (PINSR_MODE): Add V8BF.
7600 (pinsr_evex_isa): Ditto.
7601 (<sse2p4_1>_pinsr<ssemodesuffix>): Adjust to support
7602 insert for V8BFmode.
7603 (pbroadcast_evex_isa): Add BF vector modes.
7604 (AVX2_VEC_DUP_MODE): Ditto.
7605 (VEC_INIT_MODE): Ditto.
7606 (VEC_INIT_HALF_MODE): Ditto.
7607 (avx2_pbroadcast<mode>): Adjust to support BF vector mode
7609 (avx2_pbroadcast<mode>_1): Ditto.
7610 (<avx512>_vec_dup<mode>_1): Ditto.
7611 (<mask_codefor><avx512>_vec_dup_gpr<mode><mask_name>):
7614 2022-08-18 Martin Liska <mliska@suse.cz>
7616 * configure: Regenerate.
7618 2022-08-18 Haochen Gui <guihaoc@gcc.gnu.org>
7621 * config/rs6000/rs6000.md (<u>maddditi4): New pattern for multiply-add.
7622 (<u>madddi4_highpart): New.
7623 (<u>madddi4_highpart_le): New.
7625 2022-08-18 Aldy Hernandez <aldyh@redhat.com>
7627 * gimple-range-path.cc
7628 (path_range_query::compute_exit_dependencies): Use
7629 gimple_range_ssa_names.
7631 2022-08-18 zhongjuzhe <juzhe.zhong@rivai.ai>
7633 * config/riscv/predicates.md: Adjust runtime invariant.
7634 * config/riscv/riscv-modes.def (MAX_BITSIZE_MODE_ANY_MODE): New.
7635 (NUM_POLY_INT_COEFFS): New.
7636 * config/riscv/riscv-protos.h (riscv_initial_elimination_offset):Adjust
7638 * config/riscv/riscv-sr.cc (riscv_remove_unneeded_save_restore_calls):
7639 Adjust runtime invariant.
7640 * config/riscv/riscv.cc (struct riscv_frame_info): Adjust runtime
7642 (enum riscv_microarchitecture_type): Ditto.
7643 (riscv_valid_offset_p): Ditto.
7644 (riscv_valid_lo_sum_p): Ditto.
7645 (riscv_address_insns): Ditto.
7646 (riscv_load_store_insns): Ditto.
7647 (riscv_legitimize_move): Ditto.
7648 (riscv_binary_cost): Ditto.
7649 (riscv_rtx_costs): Ditto.
7650 (riscv_output_move): Ditto.
7651 (riscv_extend_comparands): Ditto.
7652 (riscv_flatten_aggregate_field): Ditto.
7653 (riscv_get_arg_info): Ditto.
7654 (riscv_pass_by_reference): Ditto.
7655 (riscv_elf_select_rtx_section): Ditto.
7656 (riscv_stack_align): Ditto.
7657 (riscv_compute_frame_info): Ditto.
7658 (riscv_initial_elimination_offset): Ditto.
7659 (riscv_set_return_address): Ditto.
7660 (riscv_for_each_saved_reg): Ditto.
7661 (riscv_first_stack_step): Ditto.
7662 (riscv_expand_prologue): Ditto.
7663 (riscv_expand_epilogue): Ditto.
7664 (riscv_can_use_return_insn): Ditto.
7665 (riscv_secondary_memory_needed): Ditto.
7666 (riscv_hard_regno_nregs): Ditto.
7667 (riscv_convert_vector_bits): New.
7668 (riscv_option_override): Adjust runtime invariant.
7669 (riscv_promote_function_mode): Ditto.
7670 * config/riscv/riscv.h (POLY_SMALL_OPERAND_P): New.
7671 (BITS_PER_RISCV_VECTOR): New.
7672 (BYTES_PER_RISCV_VECTOR): New.
7673 * config/riscv/riscv.md: Adjust runtime invariant.
7675 2022-08-18 Lulu Cheng <chenglulu@loongson.cn>
7677 * config/loongarch/loongarch.cc (loongarch_call_tls_get_addr):
7678 Get __tls_get_addr address through got table when disable plt.
7680 2022-08-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7682 * config/xtensa/xtensa.cc (xtensa_expand_prologue):
7683 Use an "addmi" machine instruction for updating the stack pointer
7684 rather than addition/subtraction via hard register A9, if the amount
7685 of change satisfies the literal value conditions of that instruction
7686 when the CALL0 ABI is used.
7687 (xtensa_expand_epilogue): Ditto.
7688 And also inhibit the stack pointer addition of constant zero.
7690 2022-08-17 Roger Sayle <roger@nextmovesoftware.com>
7693 * config/i386/i386-features.cc
7694 (timde_scalar_chain::compute_convert_gain): Replace incorrect use
7695 of XINT with INTVAL (XEXP (src, 1)).
7697 2022-08-17 Aldy Hernandez <aldyh@redhat.com>
7699 * gimple-range-path.cc
7700 (path_range_query::compute_ranges_in_block): Remove
7701 set_root_oracle call.
7702 (path_range_query::compute_ranges): Pass ranger oracle to
7704 * value-relation.cc (path_oracle::reset_path): Set root oracle.
7705 * value-relation.h (path_oracle::reset_path): Add root oracle
7708 2022-08-17 Marek Polacek <polacek@redhat.com>
7711 * diagnostic-spec.cc (nowarn_spec_t::nowarn_spec_t): Handle
7712 OPT_Wpessimizing_move and OPT_Wredundant_move.
7713 * diagnostic-spec.h (nowarn_spec_t): Add NW_REDUNDANT enumerator.
7715 2022-08-17 Francois-Xavier Coudert <fxcoudert@gcc.gnu.org>
7716 Jakub Jelinek <jakub@redhat.com>
7719 * common.opt (static-libquadmath): New option.
7720 * gcc.cc (driver_handle_option): Always accept -static-libquadmath.
7721 * config/darwin.h (LINK_SPEC): Handle -static-libquadmath.
7723 2022-08-17 Tobias Burnus <tobias@codesourcery.com>
7725 * lto-cgraph.cc (input_offload_tables): Improve requires diagnostic
7726 when filenames come out identically.
7728 2022-08-17 Tobias Burnus <tobias@codesourcery.com>
7730 PR middle-end/106548
7731 * omp-low.cc (lower_rec_input_clauses): Use build_outer_var_ref
7732 for 'simd' linear-step values that are variable.
7734 2022-08-17 Tobias Burnus <tobias@codesourcery.com>
7735 Chung-Lin Tang <cltang@codesourcery.com>
7738 * gimplify.cc (omp_notice_variable): Call omp_mappable_type
7739 instead of removed langhook.
7740 * omp-general.h (omp_mappable_type): New prototype.
7741 * omp-general.cc (omp_mappable_type): New; moved from ...
7742 * langhooks.cc (lhd_omp_mappable_type): ... here.
7743 * langhooks-def.h (lhd_omp_mappable_type,
7744 LANG_HOOKS_OMP_MAPPABLE_TYPE): Remove.
7745 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Remote the latter.
7746 * langhooks.h (struct lang_hooks_for_types): Remove
7749 2022-08-17 Christophe Lyon <christophe.lyon@arm.com>
7751 * config.gcc (arm): Define with_float to hard if target name ends
7754 2022-08-17 Richard Biener <rguenther@suse.de>
7756 * tree-ssa-threadbackward.cc
7757 (back_threader_profitability): Split profitable_path_p
7758 into possibly_profitable_path_p and itself, keep state
7760 (back_threader::m_profit): Remove.
7761 (back_threader::find_paths): Likewise.
7762 (back_threader::maybe_register_path): Take profitability
7763 instance as parameter.
7764 (back_threader::find_paths_to_names): Likewise. Use
7765 possibly_profitable_path_p and avoid the path range query
7766 when the path is currently too large.
7767 (back_threader::find_paths): Fold into ...
7768 (back_threader::maybe_thread_block): ... this.
7769 (get_gimple_control_stmt): Remove.
7770 (back_threader_profitability::possibly_profitable_path_p):
7771 Split out from profitable_path_p, do early profitability
7773 (back_threader_profitability::profitable_path_p): Do final
7774 profitability path after the taken edge has been determined.
7776 2022-08-17 Xi Ruoyao <xry111@xry111.site>
7778 * config/loongarch/loongarch.md (fmax<mode>3): New RTL pattern.
7779 (fmin<mode>3): Likewise.
7781 2022-08-17 Andrew MacLeod <amacleod@redhat.com>
7783 * gimple-range-fold.cc (gimple_range_ssa_names): New.
7784 * gimple-range-fold.h (gimple_range_ssa_names): New prototype.
7785 * gimple-range-gori.cc (range_def_chain::get_def_chain): Move
7786 code to new routine.
7788 2022-08-16 Martin Liska <mliska@suse.cz>
7791 * doc/install.texi: Remove link to www.bullfreeware.com
7793 2022-08-16 Kito Cheng <kito.cheng@sifive.com>
7795 * common/config/riscv/riscv-common.cc (riscv_implied_info): Add
7797 (riscv_ext_version_table): Ditto.
7798 (riscv_ext_flag_table): Ditto.
7799 * config/riscv/riscv-opts.h (MASK_ZFHMIN): New.
7801 (TARGET_ZFHMIN): Ditto.
7802 (TARGET_ZFH): Ditto.
7803 * config/riscv/riscv.cc (riscv_output_move): Handle HFmode move
7805 (riscv_emit_float_compare): Handle HFmode.
7806 * config/riscv/riscv.md (ANYF): Add HF.
7811 (truncdfhf2): Ditto.
7812 (extendhfsf2): Ditto.
7813 (extendhfdf2): Ditto.
7814 (*movhf_hardfloat): Ditto.
7815 (*movhf_softfloat): Make sure not ZFHMIN.
7816 * config/riscv/riscv.opt (riscv_zf_subext): New.
7818 2022-08-16 Kito Cheng <kito.cheng@sifive.com>
7820 * config/riscv/riscv-builtins.cc: include stringpool.h
7821 (riscv_float16_type_node): New.
7822 (riscv_init_builtin_types): Ditto.
7823 (riscv_init_builtins): Call riscv_init_builtin_types.
7824 * config/riscv/riscv-modes.def (HF): New.
7825 * config/riscv/riscv.cc (riscv_output_move): Handle HFmode.
7826 (riscv_mangle_type): New.
7827 (riscv_scalar_mode_supported_p): Ditto.
7828 (riscv_libgcc_floating_mode_supported_p): Ditto.
7829 (riscv_excess_precision): Ditto.
7830 (riscv_floatn_mode): Ditto.
7831 (riscv_init_libfuncs): Ditto.
7832 (TARGET_MANGLE_TYPE): Ditto.
7833 (TARGET_SCALAR_MODE_SUPPORTED_P): Ditto.
7834 (TARGET_LIBGCC_FLOATING_MODE_SUPPORTED_P): Ditto.
7835 (TARGET_INIT_LIBFUNCS): Ditto.
7836 (TARGET_C_EXCESS_PRECISION): Ditto.
7837 (TARGET_FLOATN_MODE): Ditto.
7838 * config/riscv/riscv.md (mode): Add HF.
7844 (*movhf_softfloat): New.
7846 2022-08-16 Richard Biener <rguenther@suse.de>
7848 * tree-ssa-threadbackward.cc (back_threader::find_paths_to_names):
7849 Do not walk further if we are leaving the current loop.
7851 2022-08-16 Sergei Trofimovich <siarheit@google.com>
7854 * gcc.cc (driver::detect_jobserver): Allocate storage xputenv()
7855 argument using xstrdup().
7857 2022-08-16 Aldy Hernandez <aldyh@redhat.com>
7859 * gimple-range-path.cc (path_range_query::import_p): Rename to...
7860 (path_range_query::exit_dependency_p): ...this.
7861 (path_range_query::dump): Rename imports to exit dependencies.
7862 (path_range_query::compute_ranges_in_phis): Same.
7863 (path_range_query::compute_ranges_in_block): Same.
7864 (path_range_query::adjust_for_non_null_uses): Same.
7865 (path_range_query::compute_ranges): Same.
7866 (path_range_query::compute_phi_relations): Same.
7867 (path_range_query::add_to_imports): Rename to...
7868 (path_range_query::add_to_exit_dependencies): ...this.
7869 (path_range_query::compute_imports): Rename to...
7870 (path_range_query::compute_exit_dependencies): ...this.
7871 * gimple-range-path.h (class path_range_query): Rename imports to
7874 2022-08-16 Martin Liska <mliska@suse.cz>
7876 * value-range-storage.h (class obstack_vrange_allocator): Mark
7878 (class ggc_vrange_allocator): Likewise.
7880 2022-08-16 Martin Liska <mliska@suse.cz>
7882 * value-range-equiv.h (class value_range_equiv): Add virtual
7884 * value-range.h: Likewise.
7886 2022-08-16 Richard Biener <rguenther@suse.de>
7888 PR middle-end/106630
7889 * match.pd ((T)(x * CST) -> (T)x * CST): Restrict to
7890 narrowing conversions.
7892 2022-08-16 Martin Liska <mliska@suse.cz>
7894 * value-range-equiv.h (class value_range_equiv):
7896 2022-08-16 Martin Liska <mliska@suse.cz>
7898 * config/i386/i386-features.h (class general_scalar_chain): Add
7899 final override for a method.
7900 (class timode_scalar_chain): Likewise.
7902 2022-08-16 Richard Biener <rguenther@suse.de>
7904 * doc/invoke.texi (max-jump-thread-paths): Adjust.
7906 2022-08-16 Martin Liska <mliska@suse.cz>
7908 * opts-common.cc (jobserver_info::connect): Open fifo
7909 in non-blocking mode.
7911 2022-08-16 Kewen.Lin <linkw@gcc.gnu.org>
7914 * config/rs6000/mma.md (define_expand movoo): Move TARGET_MMA condition
7915 check to preparation statements and add handlings for !TARGET_MMA.
7916 (define_expand movxo): Likewise.
7918 2022-08-16 Kewen Lin <linkw@linux.ibm.com>
7920 PR tree-optimization/106322
7921 * tree-vect-stmts.cc (vectorizable_call): Don't allow
7922 vect_emulated_vector_p type for both vectype_in and vectype_out.
7924 2022-08-16 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
7926 * common/config/xtensa/xtensa-common.cc
7927 (xtensa_option_optimization_table): Add OPT_fsplit_wide_types_early
7928 for OPT_LEVELS_ALL in order to restore pre-GCC10 behavior.
7930 2022-08-15 Andrew MacLeod <amacleod@redhat.com>
7932 PR tree-optimization/106621
7933 * value-range.cc (irange::set): Check for POLY_INT_CST early.
7935 2022-08-15 Roger Sayle <roger@nextmovesoftware.com>
7937 * config/i386/i386-features.cc
7938 (timode_scalar_chain::compute_convert_gain): Provide costs for
7940 (timode_scalar_chain::convert_insn): Handle ASHIFTRT, ROTATERT
7941 and ROTATE just like existing ASHIFT and LSHIFTRT cases.
7942 (timode_scalar_to_vector_candidate_p): Handle all shifts and
7943 rotates by integer constants between 0 and 127.
7945 2022-08-15 Roger Sayle <roger@nextmovesoftware.com>
7947 * config/i386/i386-features.cc
7948 (timode_scalar_chain::compute_convert_gain): Provide gains for
7949 comparisons against 0/-1, including "*testti" patterns.
7951 2022-08-15 Roger Sayle <roger@nextmovesoftware.com>
7953 PR tree-optimization/64992
7954 PR tree-optimization/98956
7955 * match.pd (ne (lshift @0 @1) 0): Simplify (X << C) != 0 to X
7956 when X is zero_one_valued_p and the shift constant C is valid.
7957 (eq (lshift @0 @1) 0): Likewise, simplify (X << C) == 0 to !X
7958 when X is zero_one_valued_p and the shift constant C is valid.
7960 2022-08-15 Roger Sayle <roger@nextmovesoftware.com>
7961 Richard Biener <rguenther@suse.de>
7963 PR tree-optimization/71343
7964 * match.pd (op (lshift @0 @1) (lshift @2 @1)): Optimize the
7965 expression (X<<C) + (Y<<C) to (X+Y)<<C for multiple operators.
7966 (op (rshift @0 @1) (rshift @2 @1)): Likewise, simplify (X>>C)^(Y>>C)
7967 to (X^Y)>>C for binary logical operators, AND, IOR and XOR.
7969 2022-08-15 Richard Biener <rguenther@suse.de>
7971 * gimple-range-path.cc (range_on_path_entry): Just
7972 call range_on_entry.
7974 2022-08-15 Jakub Jelinek <jakub@redhat.com>
7976 PR rtl-optimization/106590
7977 * ifcvt.cc (check_for_cc_cmp_clobbers): New function.
7978 (noce_convert_multiple_sets_1): If SEQ sets or clobbers any regs
7979 mentioned in cc_cmp or rev_cc_cmp, don't consider seq2 for any
7980 further conditional moves.
7982 2022-08-15 konglin1 <lingling.kong@intel.com>
7984 * config/i386/i386-builtin-types.def (BFLOAT16): New primitive type.
7985 * config/i386/i386-builtins.cc : Support __bf16 type for i386 backend.
7986 (ix86_register_bf16_builtin_type): New function.
7987 (ix86_bf16_type_node): New.
7988 (ix86_bf16_ptr_type_node): Ditto.
7989 (ix86_init_builtin_types): Add ix86_register_bf16_builtin_type function call.
7990 * config/i386/i386-modes.def (FLOAT_MODE): Add BFmode.
7991 (ADJUST_FLOAT_FORMAT): Ditto.
7992 * config/i386/i386.cc (classify_argument): Handle BFmode.
7993 (construct_container): Ditto.
7994 (function_value_32): Return __bf16 by %xmm0.
7995 (function_value_64): Return __bf16 by SSE register.
7996 (ix86_output_ssemov): Handle BFmode.
7997 (ix86_legitimate_constant_p): Disable BFmode constant double.
7998 (ix86_secondary_reload): Require gpr as intermediate register
7999 to store __bf16 from sse register when sse4 is not available.
8000 (ix86_scalar_mode_supported_p): Enable __bf16 under sse2.
8001 (ix86_mangle_type): Add manlging for __bf16 type.
8002 (ix86_invalid_conversion): New function for target hook.
8003 (ix86_invalid_unary_op): Ditto.
8004 (ix86_invalid_binary_op): Ditto.
8005 (TARGET_INVALID_CONVERSION): New define for target hook.
8006 (TARGET_INVALID_UNARY_OP): Ditto.
8007 (TARGET_INVALID_BINARY_OP): Ditto.
8008 * config/i386/i386.h (host_detect_local_cpu): Add BFmode.
8009 * config/i386/i386.md ("mode"): Add BFmode.
8012 (HFBF): Add new define_mode_iterator.
8013 (*pushhf_rex64): Change for BFmode.
8014 (*push<mode>_rex64): Ditto.
8016 (*push<mode>): Ditto.
8018 (hfbfconstf): Add new define_mode_attr.
8019 (*mov<mode>_internal): Add BFmode.
8021 2022-08-13 Roger Sayle <roger@nextmovesoftware.com>
8022 Uroš Bizjak <ubizjak@gmail.com>
8024 * config/i386/predicates.md (const_0_to_255_not_mul_8_operand):
8025 New predicate for values between 0/1 and 255, not multiples of 8.
8026 * config/i386/sse.md (ashlv1ti3): Delay lowering of logical left
8027 shifts by constant bit counts.
8028 (*ashlvti3_internal): New define_insn_and_split that lowers
8029 logical left shifts by constant bit counts, that aren't multiples
8030 of 8, before reload.
8031 (lshrv1ti3): Delay lowering of logical right shifts by constant.
8032 (*lshrv1ti3_internal): New define_insn_and_split that lowers
8033 logical right shifts by constant bit counts, that aren't multiples
8034 of 8, before reload.
8035 (ashrv1ti3):: Delay lowering of arithmetic right shifts by
8036 constant bit counts.
8037 (*ashrv1ti3_internal): New define_insn_and_split that lowers
8038 arithmetic right shifts by constant bit counts before reload.
8039 (rotlv1ti3): Delay lowering of rotate left by constant.
8040 (*rotlv1ti3_internal): New define_insn_and_split that lowers
8041 rotate left by constant bits counts before reload.
8042 (rotrv1ti3): Delay lowering of rotate right by constant.
8043 (*rotrv1ti3_internal): New define_insn_and_split that lowers
8044 rotate right by constant bits counts before reload.
8046 2022-08-12 Srinath Parvathaneni <srinath.parvathaneni@arm.com>
8048 * doc/invoke.texi (Arm Options): Document -mcpu=cortex-m55 options.
8050 2022-08-12 Jan Hubicka <hubicka@ucw.cz>
8052 PR middle-end/106057
8053 * ipa-devirt.cc (type_or_derived_type_possibly_instantiated_p): New
8055 (possible_polymorphic_call_targets): Use it.
8057 2022-08-12 Andrew Carlotti <andrew.carlotti@arm.com>
8059 * tree-ssa-loop.h: Improve comment
8061 2022-08-12 Jakub Jelinek <jakub@redhat.com>
8063 PR tree-optimization/106506
8064 * tree-ssa-phiopt.cc (spaceship_replacement): Don't punt for
8065 is_cast or orig_use_lhs cases if phi_bb has 3 predecessors.
8067 2022-08-12 Richard Biener <rguenther@suse.de>
8069 PR tree-optimization/106593
8070 * tree-ssa-threadbackward.cc (back_threader::find_paths):
8071 If the imports from the conditional do not satisfy
8072 gimple_range_ssa_p don't try to thread anything.
8074 2022-08-12 Tamar Christina <tamar.christina@arm.com>
8077 * config/aarch64/aarch64-sve.md (*fcmuo<mode>_nor_combine,
8078 *fcmuo<mode>_bic_combine): Don't accept comparisons against zero.
8080 2022-08-12 Tim Lange <mail@tim-lange.me>
8083 * doc/invoke.texi: Add Wanalyzer-out-of-bounds.
8085 2022-08-12 Andrew Pinski <apinski@marvell.com>
8087 * config/aarch64/aarch64.md: Remove comment
8088 about MD_INCLUDES as it is out of date and not needed.
8090 2022-08-11 Richard Biener <rguenther@suse.de>
8092 * gimple-range-path.cc (path_range_query::compute_imports):
8093 Restrict walking SSA defs to blocks inside the path. Track
8094 the same operands as range_def_chain::get_def_chain does.
8096 2022-08-11 Richard Biener <rguenther@suse.de>
8098 PR tree-optimization/106514
8099 * tree-ssa-threadbackward.cc (back_threader::find_paths_to_names):
8100 Compute and unwind both m_imports and interesting on the fly during
8102 (back_threader::find_paths): Compute the original m_imports
8103 from just the SSA uses of the exit conditional. Drop
8104 handling single_succ_to_potentially_threadable_block.
8105 * gimple-range-path.cc (path_range_query::ssa_range_in_phi): Handle
8106 constant PHI arguments without crashing. Use PHI_ARG_DEF_FROM_EDGE.
8108 2022-08-11 Richard Biener <rguenther@suse.de>
8110 * gimple-range-path.h (path_range_query::compute_imports):
8111 Take path as argument, not the exit block.
8112 * gimple-range-path.cc (path_range_query::compute_imports):
8113 Likewise, and adjust, avoiding possibly stale m_path.
8114 (path_range_query::compute_outgoing_relations): Register
8115 relations for all conditionals.
8116 * tree-ssa-threadbackward.cc (back_threader::find_paths):
8119 2022-08-11 Kewen Lin <linkw@linux.ibm.com>
8121 * config/rs6000/rs6000-builtin.cc (rs6000_init_builtins): Fix the
8122 oversight on ENB_CELL by simplifying with rs6000_builtin_is_supported.
8123 (rs6000_expand_builtin): Simplify with rs6000_builtin_is_supported.
8125 2022-08-11 Kewen Lin <linkw@linux.ibm.com>
8127 * config/rs6000/rs6000-internal.h (rs6000_global_entry_point_needed_p):
8128 Remove function declaration.
8130 2022-08-10 Richard Biener <rguenther@suse.de>
8132 PR tree-optimization/106513
8133 * gimple-ssa-store-merging.cc (do_shift_rotate): Use uint64_t
8136 2022-08-10 Martin Liska <mliska@suse.cz>
8139 * opts-jobserver.h (struct jobserver_info): Add pipefd.
8140 (jobserver_info::connect): New.
8141 (jobserver_info::disconnect): Likewise.
8142 (jobserver_info::get_token): Likewise.
8143 (jobserver_info::return_token): Likewise.
8144 * opts-common.cc: Implement the new functions.
8146 2022-08-10 Martin Liska <mliska@suse.cz>
8148 * opts-jobserver.h: Add one member.
8149 * opts-common.cc (jobserver_info::jobserver_info): Parse FIFO
8150 format of --jobserver-auth.
8152 2022-08-10 Martin Liska <mliska@suse.cz>
8154 * gcc.cc (driver::detect_jobserver): Remove and move to
8156 * lto-wrapper.cc (jobserver_active_p): Likewise.
8157 (run_gcc): Likewise.
8158 * opts-jobserver.h: New file.
8159 * opts-common.cc (jobserver_info::jobserver_info): New function.
8161 2022-08-09 Roger Sayle <roger@nextmovesoftware.com>
8163 * config/i386/i386-features.cc (scalar_chain::convert_compare):
8164 Create new pseudos only when/if needed. Add support for TEST,
8165 i.e. (COMPARE (AND x y) (const_int 0)), using UNSPEC_PTEST.
8166 When broadcasting V2DImode and V4SImode use new pseudo register.
8167 (timode_scalar_chain::convert_op): Do nothing if operand is
8168 already V1TImode. Avoid generating useless SUBREG conversions,
8169 i.e. (SUBREG:V1TImode (REG:V1TImode) 0). Handle CONST_WIDE_INT
8170 in addition to CONST_INT by using CONST_SCALAR_INT_P.
8171 (convertible_comparison_p): Use CONST_SCALAR_INT_P to match both
8172 CONST_WIDE_INT and CONST_INT. Recognize new *testti_doubleword
8173 pattern as an STV candidate.
8174 (timode_scalar_to_vector_candidate_p): Allow CONST_SCALAR_INT_P
8175 operands in binary logic operations.
8176 * config/i386/i386.cc (ix86_rtx_costs) <case UNSPEC>: Add costs
8177 for UNSPEC_PTEST; a PTEST that performs an AND has the same cost
8178 as regular PTEST, i.e. cost->sse_op.
8179 * config/i386/i386.md (*testti_doubleword): New pre-reload
8180 define_insn_and_split that recognizes comparison of TI mode AND
8182 * config/i386/sse.md (*ptest<mode>_and): New pre-reload
8183 define_insn_and_split that recognizes UNSPEC_PTEST of identical
8186 2022-08-09 Roger Sayle <roger@nextmovesoftware.com>
8187 Richard Biener <rguenther@suse.de>
8190 PR tree-optimization/98954
8191 * fold-const.cc (fold_binary_loc): Remove optimizations to
8192 optimize ((X >> C1) & C2) ==/!= 0.
8193 * match.pd (cmp (bit_and (lshift @0 @1) @2) @3): Remove wi::ctz
8194 check, and handle all values of INTEGER_CSTs @2 and @3.
8195 (cmp (bit_and (rshift @0 @1) @2) @3): Likewise, remove wi::clz
8196 checks, and handle all values of INTEGER_CSTs @2 and @3.
8198 2022-08-09 David Malcolm <dmalcolm@redhat.com>
8200 * doc/invoke.texi (Static Analyzer Options): Add notes on which
8201 functions the analyzer has hardcoded knowledge of.
8203 2022-08-09 Andrew Stubbs <ams@codesourcery.com>
8205 * config/gcn/gcn.cc (gcn_function_value): Allow vector return values.
8206 (num_arg_regs): Allow vector arguments.
8207 (gcn_function_arg): Likewise.
8208 (gcn_function_arg_advance): Likewise.
8209 (gcn_arg_partial_bytes): Likewise.
8210 (gcn_return_in_memory): Likewise.
8211 (gcn_expand_epilogue): Get return value from v8.
8212 * config/gcn/gcn.h (RETURN_VALUE_REG): Set to v8.
8213 (FIRST_PARM_REG): USE FIRST_SGPR_REG for clarity.
8214 (FIRST_VPARM_REG): New.
8215 (FUNCTION_ARG_REGNO_P): Allow vector parameters.
8216 (struct gcn_args): Add vnum field.
8217 (LIBCALL_VALUE): All vector return values.
8218 * config/gcn/gcn.md (gcn_call_value): Add vector constraints.
8219 (gcn_call_value_indirect): Likewise.
8221 2022-08-09 Richard Biener <rguenther@suse.de>
8223 * omp-expand.cc (expand_omp_atomic_load): Emit GIMPLE
8224 directly. Avoid update_ssa when in SSA form.
8225 (expand_omp_atomic_store): Likewise.
8226 (expand_omp_atomic_fetch_op): Avoid update_ssa when in SSA
8228 (expand_omp_atomic_pipeline): Likewise.
8229 (expand_omp_atomic_mutex): Likewise.
8230 * tree-parloops.cc (gen_parallel_loop): Use
8231 TODO_update_ssa_no_phi after loop_version.
8233 2022-08-09 Richard Biener <rguenther@suse.de>
8235 * doc/invoke.texi (max-fsm-thread-length): Remove.
8236 * params.opt (max-fsm-thread-length): Likewise.
8237 * tree-ssa-threadbackward.cc
8238 (back_threader_profitability::profitable_path_p): Do not
8239 check max-fsm-thread-length.
8241 2022-08-09 Richard Biener <rguenther@suse.de>
8243 PR tree-optimization/106514
8244 * params.opt (max-jump-thread-paths): New.
8245 * doc/invoke.texi (max-jump-thread-paths): Document.
8246 * tree-ssa-threadbackward.cc (back_threader::find_paths_to_names):
8247 Honor max-jump-thread-paths, take overall_path argument.
8248 (back_threader::find_paths): Pass 1 as initial overall_path.
8250 2022-08-09 Tobias Burnus <tobias@codesourcery.com>
8252 PR middle-end/106492
8253 * omp-low.cc (lower_rec_input_clauses): Add missing folding
8254 to data type of linear-clause list item.
8256 2022-08-08 Andrew MacLeod <amacleod@redhat.com>
8258 PR tree-optimization/106556
8259 * gimple-range-gori.cc (gori_compute::condexpr_adjust): Use the
8260 type of the cond_expr operands being evaluted.
8262 2022-08-08 Tom Honermann <tom@honermann.net>
8264 * ginclude/stdatomic.h (atomic_char8_t,
8265 ATOMIC_CHAR8_T_LOCK_FREE): New typedef and macro.
8267 2022-08-08 Andrew Pinski <apinski@marvell.com>
8269 PR middle-end/103645
8270 * gimplify.cc (gimplify_init_constructor): Don't build/add
8271 gimple assignment of an empty type.
8273 2022-08-08 Richard Biener <rguenther@suse.de>
8277 * dwarf2out.cc (dwarf2out_register_external_die): Restore
8279 * lto-streamer-in.cc (lto_read_tree_1): Use lto_input_tree_1
8280 to input DECL_INITIAL, avoiding to commit drefs.
8282 2022-08-07 Roger Sayle <roger@nextmovesoftware.com>
8284 * config/i386/i386.md (*cmp<dwi>_doubleword): Change predicate
8285 for x86_64_hilo_general_operand to general operand. Call
8286 force_reg on parts that are not x86_64_immediate_operand.
8288 2022-08-05 David Malcolm <dmalcolm@redhat.com>
8291 * doc/invoke.texi: Add -Wanalyzer-jump-through-null.
8293 2022-08-05 Roger Sayle <roger@nextmovesoftware.com>
8295 * expmed.cc (emit_store_flag_1): Move code to expand double word
8296 equality and inequality against zero or -1, using word operations,
8297 to after trying to use the backend's cstore<mode>4 optab/expander.
8299 2022-08-05 Tamar Christina <tamar.christina@arm.com>
8301 PR middle-end/106534
8302 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Guard the
8303 value_replacement and store_elim from diamonds.
8305 2022-08-05 Richard Biener <rguenther@suse.de>
8307 * tree-ssa-threadbackward.cc (back_threader::maybe_register_path):
8308 Check whether the registry register_path rejected the path.
8309 (back_threader_registry::register_path): Return whether
8310 register_jump_thread succeeded.
8312 2022-08-05 Aldy Hernandez <aldyh@redhat.com>
8314 PR tree-optimization/106514
8315 * value-range.cc (unsupported_range::unsupported_range): Move...
8316 * value-range.h (unsupported_range::unsupported_range): ...here.
8317 (unsupported_range::set_undefined): New.
8319 2022-08-05 Richard Biener <rguenther@suse.de>
8321 PR tree-optimization/106533
8322 * tree-loop-distribution.cc (loop_distribution::execute): Continue
8323 analyzing the inner loops when find_seed_stmts_for_distribution
8326 2022-08-05 Andrew Pinski <apinski@marvell.com>
8328 * config/riscv/predicates.md (splittable_const_int_operand):
8329 Remove the check for TARGET_64BIT for single bit const values.
8331 2022-08-04 Andrew MacLeod <amacleod@redhat.com>
8333 PR tree-optimization/106514
8334 * gimple-range-path.cc (path_range_query::compute_ranges_in_block):
8335 Use EXECUTE_IF_AND_IN_BITMAP to loop over 2 bitmaps.
8337 2022-08-04 Tamar Christina <tamar.christina@arm.com>
8339 * match.pd: New bit_not rule.
8341 2022-08-04 Tamar Christina <tamar.christina@arm.com>
8343 PR middle-end/106519
8344 * tree-ssa-phiopt.cc (tree_ssa_phiopt_worker): Check final phi edge for
8347 2022-08-04 Sam Feifer <sfeifer@redhat.com>
8349 PR tree-optimization/106243
8350 * match.pd (-x & 1): New simplification.
8352 2022-08-04 Richard Biener <rguenther@suse.de>
8354 PR tree-optimization/106521
8355 * gimple-loop-jam.cc (tree_loop_unroll_and_jam): Perform
8356 CFG cleanup manually before rewriting into LC SSA.
8358 2022-08-04 Richard Biener <rguenther@suse.de>
8360 * tree-ssa-threadbackward.cc (populate_worklist): Remove.
8361 (back_threader::resolve_phi): Likewise.
8362 (back_threader::find_paths_to_names): Rewrite greedy search.
8364 2022-08-04 Ilya Leoshkevich <iii@linux.ibm.com>
8366 * config/s390/vector.md (V_HW_FT): New iterator.
8367 * config/s390/vx-builtins.md (vsel<mode>): Use V_HW_FT instead
8370 2022-08-03 Michael Meissner <meissner@linux.ibm.com>
8372 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Remove code
8373 setting -mblock-ops-vector-pair.
8375 2022-08-03 Andrew MacLeod <amacleod@redhat.com>
8377 PR tree-optimization/106514
8378 * value-relation.cc (path_oracle::killing_def) Do not walk the
8379 equivalence set clearing bits.
8381 2022-08-03 Tamar Christina <tamar.christina@arm.com>
8383 * tree-ssa-phiopt.cc (minmax_replacement): Optionally search for the phi
8384 sequence of a three-way conditional.
8385 (replace_phi_edge_with_variable): Support diamonds.
8386 (tree_ssa_phiopt_worker): Detect diamond phi structure for three-way
8388 (strip_bit_not, invert_minmax_code): New.
8390 2022-08-03 Richard Earnshaw <rearnsha@arm.com>
8392 PR rtl-optimization/106187
8393 * alias.h (mems_same_for_tbaa_p): Declare.
8394 * alias.cc (mems_same_for_tbaa_p): New function.
8395 * dse.cc (record_store): Use it instead of open-coding
8397 * cselib.h (cselib_redundant_set_p): Declare.
8398 * cselib.cc: Include alias.h
8399 (cselib_redundant_set_p): New function.
8400 * cfgcleanup.cc: (mark_effect): Use cselib_redundant_set_p instead
8401 of rtx_equal_for_cselib_p.
8402 * postreload.cc (reload_cse_simplify): Use cselib_redundant_set_p.
8403 (reload_cse_noop_set_p): Delete.
8405 2022-08-03 Martin Liska <mliska@suse.cz>
8407 * doc/gcov-dump.texi: Document the new option.
8408 * gcov-dump.cc (main): Parse the new option.
8409 (print_usage): Show the option.
8410 (tag_counters): Sort key:value pairs of TOP N counter.
8412 2022-08-03 Martin Liska <mliska@suse.cz>
8414 * profile.cc (compute_branch_probabilities): Do not collect
8415 stats unless TDF_DETAILS.
8417 2022-08-03 Roger Sayle <roger@nextmovesoftware.com>
8418 Uroš Bizjak <ubizjak@gmail.com>
8421 * config/i386/i386.md (peephole2): New peephole2 to convert
8422 SWI48 moves to/from %rax/%eax where the src is dead to xchg,
8423 when optimizing for minimal size with -Oz.
8425 2022-08-03 Roger Sayle <roger@nextmovesoftware.com>
8427 * config/i386/i386.md (*cmp<dwi>_doubleword): Add a special case
8428 to split comparisons against -1 using AND and CMP -1 instructions.
8430 2022-08-03 Roger Sayle <roger@nextmovesoftware.com>
8432 * config/i386/i386-features.cc (compute_convert_gain): Add gain
8433 for converting suitable TImode shift to a V1TImode shift.
8434 (timode_scalar_chain::convert_insn): Add support for converting
8435 suitable ASHIFT and LSHIFTRT.
8436 (timode_scalar_to_vector_candidate_p): Consider logical shifts
8437 by integer constants that are multiples of 8 to be candidates.
8439 2022-08-03 Roger Sayle <roger@nextmovesoftware.com>
8440 Segher Boessenkool <segher@kernel.crashing.org>
8441 Richard Sandiford <richard.sandiford@arm.com>
8443 * simplify-rtx.cc (simplify_unary_operation_1) <ABS>: Add
8444 optimizations for CLRSB, PARITY, POPCOUNT, SS_ABS and LSHIFTRT
8445 that are all positive to complement the existing FFS and
8446 idempotent ABS simplifications.
8447 <SIGN_EXTEND>: Canonicalize SIGN_EXTEND to ZERO_EXTEND when
8448 val_signbit_known_clear_p is true of the operand.
8449 Simplify sign extensions of SUBREG truncations of operands
8450 that are already suitably (zero) extended.
8451 <ZERO_EXTEND>: Simplify zero extensions of SUBREG truncations
8452 of operands that are already suitably zero extended.
8454 2022-08-02 Andrew MacLeod <amacleod@redhat.com>
8456 PR tree-optimization/106510
8457 * gimple-range-fold.cc (fur_source::register_outgoing_edges):
8458 Check for unsupported statements early.
8460 2022-08-02 Andrew MacLeod <amacleod@redhat.com>
8462 PR tree-optimization/106474
8463 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Query
8464 range of equivalences that may contribute to the range.
8466 2022-08-02 Jose E. Marchesi <jose.marchesi@oracle.com>
8468 * btfout.cc (output_asm_btf_vlen_bytes): Do not use the CHAR
8469 encoding bit in BTF.
8471 2022-08-02 Aldy Hernandez <aldyh@redhat.com>
8473 * gimple-range-fold.cc (fold_using_range::range_of_phi): Remove
8475 (tree_lower_bound): New.
8476 (tree_upper_bound): New.
8477 (fold_using_range::range_of_ssa_name_with_loop_info): Convert to
8479 * gimple-range-fold.h (range_of_ssa_name_with_loop_info): Change
8482 2022-08-02 Richard Biener <rguenther@suse.de>
8484 * tree-ssa-threadbackward.cc
8485 (back_threader_profitability::profitable_path_p): Apply
8486 size constraints to all paths again.
8488 2022-08-02 Aldy Hernandez <aldyh@redhat.com>
8490 * range-op-float.cc (finite_operands_p): New.
8491 (frelop_early_resolve): New.
8492 (default_frelop_fold_range): New.
8493 (class foperator_equal): New.
8494 (class foperator_not_equal): New.
8495 (class foperator_lt): New.
8496 (class foperator_le): New.
8497 (class foperator_gt): New.
8498 (class foperator_ge): New.
8499 (class foperator_unordered): New.
8500 (class foperator_ordered): New.
8501 (class foperator_relop_unknown): New.
8502 (floating_op_table::floating_op_table): Add above classes to
8504 * value-range.h (frange::supports_p): Enable.
8506 2022-08-02 Aldy Hernandez <aldyh@redhat.com>
8508 * tree-core.h (struct tree_ssa_name): Add frange_info and
8510 * value-range-storage.cc (vrange_storage::alloc_slot): Add case
8512 (vrange_storage::set_vrange): Same.
8513 (vrange_storage::get_vrange): Same.
8514 (vrange_storage::fits_p): Same.
8515 (frange_storage_slot::alloc_slot): New.
8516 (frange_storage_slot::set_frange): New.
8517 (frange_storage_slot::get_frange): New.
8518 (frange_storage_slot::fits_p): New.
8519 * value-range-storage.h (class frange_storage_slot): New.
8521 2022-08-02 Aldy Hernandez <aldyh@redhat.com>
8523 * ipa-prop.cc (ipa_compute_jump_functions_for_edge): Limit ranger
8526 2022-08-02 Aldy Hernandez <aldyh@redhat.com>
8528 * value-range.cc (frange::set): Initialize m_props and cleanup.
8530 2022-08-02 Richard Biener <rguenther@suse.de>
8532 PR tree-optimization/106497
8533 * tree-ssa-threadupdate.cc (fwd_jt_path_registry::update_cfg):
8534 Also verify we can copy EDGE_COPY_SRC_JOINER_BLOCK.
8536 2022-08-02 Martin Liska <mliska@suse.cz>
8538 * profile.cc (compute_branch_probabilities): Dump details only
8540 * symtab.cc (symtab_node::dump_base): Do not dump pointer unless
8541 TDF_ADDRESS is used, it makes comparison harder.
8543 2022-08-02 Richard Biener <rguenther@suse.de>
8545 PR tree-optimization/106498
8546 * omp-expand.cc (expand_omp_taskreg): Do not perform virtual
8548 (expand_omp_for): Or here.
8549 (execute_expand_omp): Instead schedule it here together
8550 with CFG cleanup via TODO.
8552 2022-08-02 Richard Biener <rguenther@suse.de>
8555 * dwarf2out.cc (dwarf2out_register_external_die): Adjust
8558 2022-08-02 Richard Biener <rguenther@suse.de>
8560 PR tree-optimization/106495
8561 * tree-ssa-threadbackward.cc
8562 (back_threader_profitability::profitable_path_p): If known_edge
8563 is probably never executed avoid threading.
8565 2022-08-01 David Malcolm <dmalcolm@redhat.com>
8567 * doc/invoke.texi (-Wanalyzer-putenv-of-auto-var): Fix copy&paste
8570 2022-08-01 Roger Sayle <roger@nextmovesoftware.com>
8571 Uroš Bizjak <ubizjak@gmail.com>
8574 * config/i386/i386-features.cc (timode_scalar_chain::convert_insn):
8575 Convert a CONST_SCALAR_INT_P in a REG_EQUAL note into a V1TImode
8578 2022-08-01 H.J. Lu <hjl.tools@gmail.com>
8581 * config/i386/i386.cc (ix86_ifunc_ref_local_ok): New.
8582 (TARGET_IFUNC_REF_LOCAL_OK): Use it.
8584 2022-08-01 Jose E. Marchesi <jose.marchesi@oracle.com>
8587 * ctfc.h (struct ctf_dtdef): Add field linkage.
8588 * ctfc.cc (ctf_add_function): Set ctti_linkage.
8589 * dwarf2ctf.cc (gen_ctf_function_type): Pass a linkage for
8590 function types and subprograms.
8591 * btfout.cc (btf_asm_func_type): Emit linkage information for the
8593 (btf_dtd_emit_preprocess_cb): Propagate the linkage information
8596 2022-08-01 Andrew Stubbs <ams@codesourcery.com>
8597 Jakub Jelinek <jakub@redhat.com>
8599 * omp-simd-clone.cc (simd_clone_adjust): Convert shift_cnt to match
8602 2022-08-01 Sam Feifer <sfeifer@redhat.com>
8604 PR tree-optimization/104992
8605 * match.pd (x / y * y == x): New simplification.
8607 2022-08-01 Aldy Hernandez <aldyh@redhat.com>
8609 * value-range.cc (tree_compare): New.
8610 (frange::set): Make more general.
8611 (frange::normalize_kind): Cleanup and return bool.
8612 (frange::union_): Use normalize_kind return value.
8613 (frange::intersect): Same.
8614 (frange::verify_range): Remove unnecessary else.
8615 * value-range.h (vrp_val_max): Move before frange class.
8616 (vrp_val_min): Same.
8617 (frange::frange): Remove set to m_type.
8619 2022-08-01 Aldy Hernandez <aldyh@redhat.com>
8621 * value-range.cc (vrange::supports_type_p): Use const_tree.
8622 (irange::supports_type_p): Same.
8623 (frange::supports_type_p): Same.
8624 * value-range.h (Value_Range::supports_type_p): Same.
8625 (irange::supports_p): Same.
8627 2022-08-01 Aldy Hernandez <aldyh@redhat.com>
8629 * gimple-range-fold.cc (fold_using_range::range_of_phi): Only
8630 query SCEV for integers.
8631 (fold_using_range::range_of_ssa_name_with_loop_info): Remove
8634 2022-07-31 Roger Sayle <roger@nextmovesoftware.com>
8636 * config/i386/i386.md (define_expand <any_rotate>ti3): For
8637 rotations by 64 bits use new rot[lr]64ti2_doubleword pattern.
8638 (rot[lr]64ti2_doubleword): New post-reload splitter.
8640 2022-07-31 Roger Sayle <roger@nextmovesoftware.com>
8641 H.J. Lu <hjl.tools@gmail.com>
8644 * config/i386/i386-features.cc (timode_check_non_convertible_regs):
8645 Do nothing if REGNO is set in the REGS bitmap, or is a hard reg.
8646 (timode_remove_non_convertible_regs): Update comment.
8647 Call timode_check_non_convertible_reg on all TImode register
8648 DEFs and USEs in each instruction.
8650 2022-07-30 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
8652 * config/xtensa/xtensa.md: Change hard register number used in
8653 the split patterns for indirect sibling call fixups from 10 to 11,
8654 the last free one for the CALL0 ABI.
8656 2022-07-30 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
8658 * config/xtensa/xtensa.cc (xtensa_rtx_costs):
8659 Add new case for IF_THEN_ELSE.
8661 2022-07-29 Andrew Stubbs <ams@codesourcery.com>
8663 * config/gcn/gcn-valu.md (V_INT_noHI): New iterator.
8664 (<expander><mode>3<exec>): Use V_INT_noHI.
8665 (v<expander><mode>3<exec>): Likewise.
8667 2022-07-29 Andrew Stubbs <ams@codesourcery.com>
8669 * config/gcn/gcn.md (one_cmpldi2): New.
8671 2022-07-29 Richard Biener <rguenther@suse.de>
8673 PR tree-optimization/105679
8674 * tree-ssa-threadbackward.cc
8675 (back_threader_profitability::profitable_path_p): Avoid threading
8676 when the entry edge is probably never executed.
8678 2022-07-29 Richard Biener <rguenther@suse.de>
8680 PR tree-optimization/106422
8681 * tree-ssa-threadupdate.cc (fwd_jt_path_registry::update_cfg):
8682 Check whether we can copy thread blocks and cancel the thread if not.
8684 2022-07-29 Jakub Jelinek <jakub@redhat.com>
8686 PR middle-end/106449
8687 * omp-expand.cc (expand_omp_simd): Fix up handling of pointer
8688 iterators in non-rectangular simd loops. Unshare fd->loops[i].n2
8689 or n2 before regimplifying it inside of a condition.
8691 2022-07-29 Jakub Jelinek <jakub@redhat.com>
8693 * omp-expand.cc (expand_omp_for_init_counts, expand_omp_for_init_vars,
8694 extract_omp_for_update_vars, expand_omp_for_ordered_loops,
8695 expand_omp_simd): Don't fold_convert second argument to
8696 fold_build_pointer_plus to sizetype.
8698 2022-07-29 Lulu Cheng <chenglulu@loongson.cn>
8700 * config.in: Regenerate.
8701 * config/loongarch/loongarch.h (ASM_PREFERRED_EH_DATA_FORMAT):
8702 Select the value of the macro definition according to whether
8703 HAVE_AS_EH_FRAME_PCREL_ENCODING_SUPPORT is defined.
8704 * configure: Regenerate.
8705 * configure.ac: Reinstate HAVE_AS_EH_FRAME_PCREL_ENCODING_SUPPORT test.
8707 2022-07-29 Richard Biener <rguenther@suse.de>
8709 * gimple-ssa-warn-restrict.cc (builtin_memref::set_base_and_offset):
8710 Use CONVERT_EXPR_CODE_P.
8712 2022-07-29 Richard Biener <rguenther@suse.de>
8714 * tree-vect-patterns.cc (vect_recog_bool_pattern): Use
8715 get_vectype_for_scalar_type instead of
8716 vect_get_vector_types_for_stmt.
8718 2022-07-28 David Malcolm <dmalcolm@redhat.com>
8721 * doc/invoke.texi: Add -Wanalyzer-putenv-of-auto-var.
8723 2022-07-28 David Malcolm <dmalcolm@redhat.com>
8725 * doc/invoke.texi (-fdiagnostics-show-cwe): Use uref rather than
8727 (Static Analyzer Options): Likewise. Add urefs for all of the
8728 warnings that have associated CWE identifiers.
8730 2022-07-28 Maciej W. Rozycki <macro@embecosm.com>
8732 * doc/implement-c.texi (Floating point implementation): Mention
8733 `-fno-trapping-math' in the context of FENV_ACCESS pragma.
8734 * doc/invoke.texi (Optimize Options): Clarify FENV_ACCESS pragma
8735 implication in the descriptions of `-fno-trapping-math' and
8738 2022-07-28 Maciej W. Rozycki <macro@embecosm.com>
8740 * config/riscv/riscv.md (UNSPECV_FSNVSNAN): New constant.
8741 (QUIET_PATTERN): New int attribute.
8742 (f<quiet_pattern>_quiet<ANYF:mode><X:mode>4): Emit the intended
8743 RTL insns entirely within the preparation statements.
8744 (*f<quiet_pattern>_quiet<ANYF:mode><X:mode>4_default)
8745 (*f<quiet_pattern>_quiet<ANYF:mode><X:mode>4_snan): Remove
8747 (*riscv_fsnvsnan<mode>2): New insn.
8749 2022-07-28 Richard Biener <rguenther@suse.de>
8751 PR middle-end/106457
8752 * tree.cc (array_at_struct_end_p): Handle array objects
8755 2022-07-28 Jakub Jelinek <jakub@redhat.com>
8757 PR tree-optimization/106099
8758 * internal-fn.def (TRAP): New internal fn.
8759 * internal-fn.h (expand_TRAP): Declare.
8760 * internal-fn.cc (expand_TRAP): Define.
8761 * gimple.cc (gimple_build_builtin_unreachable): For BUILT_IN_TRAP,
8762 use internal fn rather than builtin.
8764 2022-07-27 Andrew Carlotti <andrew.carlotti@arm.com>
8766 * doc/loop.texi: Refer to LOOPS_HAVE_RECORDED_EXITS instead.
8768 2022-07-27 WANG Xuerui <i@xen0n.name>
8770 * doc/invoke.texi: Document -m[no-]explicit-relocs for
8773 2022-07-27 Maciej W. Rozycki <macro@embecosm.com>
8775 * config/riscv/riscv.md (stack_protect_set_<mode>): Remove
8776 duplicate backslashes.
8778 2022-07-27 Maciej W. Rozycki <macro@embecosm.com>
8780 * config/riscv/riscv.cc (riscv_rtx_costs) <IF_THEN_ELSE>: New
8783 2022-07-27 Jakub Jelinek <jakub@redhat.com>
8786 * cgraphunit.cc (cgraph_node::assemble_thunks_and_aliases): Don't
8787 output asm thunks for -dx.
8789 2022-07-27 Jakub Jelinek <jakub@redhat.com>
8791 PR middle-end/106332
8792 * opts-common.cc (candidates_list_and_hint): Add gcc_assert
8793 that candidates is not an empty vector.
8795 2022-07-27 Xi Ruoyao <xry111@xry111.site>
8797 * configure.ac (HAVE_AS_EXPLICIT_RELOCS): Define to 1 if the
8798 assembler supports explicit relocation for LoongArch.
8799 * configure: Regenerate.
8800 * config/loongarch/loongarch-opts.h (HAVE_AS_EXPLICIT_RELOCS):
8801 Define to 0 if not defined.
8802 * config/loongarch/genopts/loongarch.opt.in
8803 (TARGET_EXPLICIT_RELOCS): Default to HAVE_AS_EXPLICIT_RELOCS.
8804 * config/loongarch/loongarch.opt: Regenerate.
8806 2022-07-26 Peter Bergner <bergner@linux.ibm.com>
8809 * expr.cc (count_type_elements): Handle OPAQUE_TYPE.
8811 2022-07-26 Lulu Cheng <chenglulu@loongson.cn>
8813 * config/loongarch/loongarch-opts.cc: Modify the output message string
8816 2022-07-26 Martin Liska <mliska@suse.cz>
8818 * doc/tm.texi.in: Fix placement of defmac.
8819 * doc/tm.texi: Copy.
8821 2022-07-26 Martin Liska <mliska@suse.cz>
8823 * doc/tm.texi.in: Fix cross @defmac and @hook.
8824 * doc/tm.texi: Copy.
8826 2022-07-26 Aldy Hernandez <aldyh@redhat.com>
8828 PR tree-optimization/106444
8829 * value-range-pretty-print.cc (vrange_printer::visit): Handle
8831 (vrange_printer::print_irange_bound): Work on wide_int's.
8832 * value-range-pretty-print.h (print_irange_bound): Same.
8833 * value-range.cc (irange::get_nonzero_bits): Handle legacy ranges.
8835 2022-07-26 Richard Biener <rguenther@suse.de>
8837 * tree-ssa-alias.cc (ptr_derefs_may_alias_p): If ptr1
8838 points to a constant continue checking ptr2.
8840 2022-07-26 Andrew Carlotti <Andrew.Carlotti@arm.com>
8842 * config/aarch64/aarch64-builtins.cc
8843 (MODE_d_bf16, MODE_d_f16, MODE_d_f32, MODE_d_f64, MODE_d_s8)
8844 (MODE_d_s16, MODE_d_s32, MODE_d_s64, MODE_d_u8, MODE_d_u16)
8845 (MODE_d_u32, MODE_d_u64, MODE_d_p8, MODE_d_p16, MODE_d_p64)
8846 (MODE_q_bf16, MODE_q_f16, MODE_q_f32, MODE_q_f64, MODE_q_s8)
8847 (MODE_q_s16, MODE_q_s32, MODE_q_s64, MODE_q_u8, MODE_q_u16)
8848 (MODE_q_u32, MODE_q_u64, MODE_q_p8, MODE_q_p16, MODE_q_p64)
8849 (MODE_q_p128): Define macro to map to corresponding mode name.
8850 (QUAL_bf16, QUAL_f16, QUAL_f32, QUAL_f64, QUAL_s8, QUAL_s16)
8851 (QUAL_s32, QUAL_s64, QUAL_u8, QUAL_u16, QUAL_u32, QUAL_u64)
8852 (QUAL_p8, QUAL_p16, QUAL_p64, QUAL_p128): Define macro to map to
8853 corresponding qualifier name.
8854 (LENGTH_d, LENGTH_q): Define macro to map to "" or "q" suffix.
8855 (SIMD_INTR_MODE, SIMD_INTR_QUAL, SIMD_INTR_LENGTH_CHAR): Macro
8856 functions for the above mappings
8857 (VREINTERPRET_BUILTIN2, VREINTERPRET_BUILTINS1, VREINTERPRET_BUILTINS)
8858 (VREINTERPRETQ_BUILTIN2, VREINTERPRETQ_BUILTINS1)
8859 (VREINTERPRETQ_BUILTINS, VREINTERPRET_BUILTIN)
8860 (AARCH64_SIMD_VREINTERPRET_BUILTINS): New macros to create definitions
8861 for all vreinterpret intrinsics
8862 (enum aarch64_builtins): Add vreinterpret function codes
8863 (aarch64_init_simd_intrinsics): New
8864 (handle_arm_neon_h): Improved comment.
8865 (aarch64_general_fold_builtin): Fold vreinterpret calls
8866 * config/aarch64/arm_neon.h
8867 (vreinterpret_p8_f16, vreinterpret_p8_f64, vreinterpret_p8_s8)
8868 (vreinterpret_p8_s16, vreinterpret_p8_s32, vreinterpret_p8_s64)
8869 (vreinterpret_p8_f32, vreinterpret_p8_u8, vreinterpret_p8_u16)
8870 (vreinterpret_p8_u32, vreinterpret_p8_u64, vreinterpret_p8_p16)
8871 (vreinterpret_p8_p64, vreinterpretq_p8_f64, vreinterpretq_p8_s8)
8872 (vreinterpretq_p8_s16, vreinterpretq_p8_s32, vreinterpretq_p8_s64)
8873 (vreinterpretq_p8_f16, vreinterpretq_p8_f32, vreinterpretq_p8_u8)
8874 (vreinterpretq_p8_u16, vreinterpretq_p8_u32, vreinterpretq_p8_u64)
8875 (vreinterpretq_p8_p16, vreinterpretq_p8_p64, vreinterpretq_p8_p128)
8876 (vreinterpret_p16_f16, vreinterpret_p16_f64, vreinterpret_p16_s8)
8877 (vreinterpret_p16_s16, vreinterpret_p16_s32, vreinterpret_p16_s64)
8878 (vreinterpret_p16_f32, vreinterpret_p16_u8, vreinterpret_p16_u16)
8879 (vreinterpret_p16_u32, vreinterpret_p16_u64, vreinterpret_p16_p8)
8880 (vreinterpret_p16_p64, vreinterpretq_p16_f64, vreinterpretq_p16_s8)
8881 (vreinterpretq_p16_s16, vreinterpretq_p16_s32, vreinterpretq_p16_s64)
8882 (vreinterpretq_p16_f16, vreinterpretq_p16_f32, vreinterpretq_p16_u8)
8883 (vreinterpretq_p16_u16, vreinterpretq_p16_u32, vreinterpretq_p16_u64)
8884 (vreinterpretq_p16_p8, vreinterpretq_p16_p64, vreinterpretq_p16_p128)
8885 (vreinterpret_p64_f16, vreinterpret_p64_f64, vreinterpret_p64_s8)
8886 (vreinterpret_p64_s16, vreinterpret_p64_s32, vreinterpret_p64_s64)
8887 (vreinterpret_p64_f32, vreinterpret_p64_u8, vreinterpret_p64_u16)
8888 (vreinterpret_p64_u32, vreinterpret_p64_u64, vreinterpret_p64_p8)
8889 (vreinterpret_p64_p16, vreinterpretq_p64_f64, vreinterpretq_p64_s8)
8890 (vreinterpretq_p64_s16, vreinterpretq_p64_s32, vreinterpretq_p64_s64)
8891 (vreinterpretq_p64_f16, vreinterpretq_p64_f32, vreinterpretq_p64_p128)
8892 (vreinterpretq_p64_u8, vreinterpretq_p64_u16, vreinterpretq_p64_p16)
8893 (vreinterpretq_p64_u32, vreinterpretq_p64_u64, vreinterpretq_p64_p8)
8894 (vreinterpretq_p128_p8, vreinterpretq_p128_p16, vreinterpretq_p128_f16)
8895 (vreinterpretq_p128_f32, vreinterpretq_p128_p64, vreinterpretq_p128_s64)
8896 (vreinterpretq_p128_u64, vreinterpretq_p128_s8, vreinterpretq_p128_s16)
8897 (vreinterpretq_p128_s32, vreinterpretq_p128_u8, vreinterpretq_p128_u16)
8898 (vreinterpretq_p128_u32, vreinterpret_f16_f64, vreinterpret_f16_s8)
8899 (vreinterpret_f16_s16, vreinterpret_f16_s32, vreinterpret_f16_s64)
8900 (vreinterpret_f16_f32, vreinterpret_f16_u8, vreinterpret_f16_u16)
8901 (vreinterpret_f16_u32, vreinterpret_f16_u64, vreinterpret_f16_p8)
8902 (vreinterpret_f16_p16, vreinterpret_f16_p64, vreinterpretq_f16_f64)
8903 (vreinterpretq_f16_s8, vreinterpretq_f16_s16, vreinterpretq_f16_s32)
8904 (vreinterpretq_f16_s64, vreinterpretq_f16_f32, vreinterpretq_f16_u8)
8905 (vreinterpretq_f16_u16, vreinterpretq_f16_u32, vreinterpretq_f16_u64)
8906 (vreinterpretq_f16_p8, vreinterpretq_f16_p128, vreinterpretq_f16_p16)
8907 (vreinterpretq_f16_p64, vreinterpret_f32_f16, vreinterpret_f32_f64)
8908 (vreinterpret_f32_s8, vreinterpret_f32_s16, vreinterpret_f32_s32)
8909 (vreinterpret_f32_s64, vreinterpret_f32_u8, vreinterpret_f32_u16)
8910 (vreinterpret_f32_u32, vreinterpret_f32_u64, vreinterpret_f32_p8)
8911 (vreinterpret_f32_p16, vreinterpret_f32_p64, vreinterpretq_f32_f16)
8912 (vreinterpretq_f32_f64, vreinterpretq_f32_s8, vreinterpretq_f32_s16)
8913 (vreinterpretq_f32_s32, vreinterpretq_f32_s64, vreinterpretq_f32_u8)
8914 (vreinterpretq_f32_u16, vreinterpretq_f32_u32, vreinterpretq_f32_u64)
8915 (vreinterpretq_f32_p8, vreinterpretq_f32_p16, vreinterpretq_f32_p64)
8916 (vreinterpretq_f32_p128, vreinterpret_f64_f16, vreinterpret_f64_f32)
8917 (vreinterpret_f64_p8, vreinterpret_f64_p16, vreinterpret_f64_p64)
8918 (vreinterpret_f64_s8, vreinterpret_f64_s16, vreinterpret_f64_s32)
8919 (vreinterpret_f64_s64, vreinterpret_f64_u8, vreinterpret_f64_u16)
8920 (vreinterpret_f64_u32, vreinterpret_f64_u64, vreinterpretq_f64_f16)
8921 (vreinterpretq_f64_f32, vreinterpretq_f64_p8, vreinterpretq_f64_p16)
8922 (vreinterpretq_f64_p64, vreinterpretq_f64_s8, vreinterpretq_f64_s16)
8923 (vreinterpretq_f64_s32, vreinterpretq_f64_s64, vreinterpretq_f64_u8)
8924 (vreinterpretq_f64_u16, vreinterpretq_f64_u32, vreinterpretq_f64_u64)
8925 (vreinterpret_s64_f16, vreinterpret_s64_f64, vreinterpret_s64_s8)
8926 (vreinterpret_s64_s16, vreinterpret_s64_s32, vreinterpret_s64_f32)
8927 (vreinterpret_s64_u8, vreinterpret_s64_u16, vreinterpret_s64_u32)
8928 (vreinterpret_s64_u64, vreinterpret_s64_p8, vreinterpret_s64_p16)
8929 (vreinterpret_s64_p64, vreinterpretq_s64_f64, vreinterpretq_s64_s8)
8930 (vreinterpretq_s64_s16, vreinterpretq_s64_s32, vreinterpretq_s64_f16)
8931 (vreinterpretq_s64_f32, vreinterpretq_s64_u8, vreinterpretq_s64_u16)
8932 (vreinterpretq_s64_u32, vreinterpretq_s64_u64, vreinterpretq_s64_p8)
8933 (vreinterpretq_s64_p16, vreinterpretq_s64_p64, vreinterpretq_s64_p128)
8934 (vreinterpret_u64_f16, vreinterpret_u64_f64, vreinterpret_u64_s8)
8935 (vreinterpret_u64_s16, vreinterpret_u64_s32, vreinterpret_u64_s64)
8936 (vreinterpret_u64_f32, vreinterpret_u64_u8, vreinterpret_u64_u16)
8937 (vreinterpret_u64_u32, vreinterpret_u64_p8, vreinterpret_u64_p16)
8938 (vreinterpret_u64_p64, vreinterpretq_u64_f64, vreinterpretq_u64_s8)
8939 (vreinterpretq_u64_s16, vreinterpretq_u64_s32, vreinterpretq_u64_s64)
8940 (vreinterpretq_u64_f16, vreinterpretq_u64_f32, vreinterpretq_u64_u8)
8941 (vreinterpretq_u64_u16, vreinterpretq_u64_u32, vreinterpretq_u64_p8)
8942 (vreinterpretq_u64_p16, vreinterpretq_u64_p64, vreinterpretq_u64_p128)
8943 (vreinterpret_s8_f16, vreinterpret_s8_f64, vreinterpret_s8_s16)
8944 (vreinterpret_s8_s32, vreinterpret_s8_s64, vreinterpret_s8_f32)
8945 (vreinterpret_s8_u8, vreinterpret_s8_u16, vreinterpret_s8_u32)
8946 (vreinterpret_s8_u64, vreinterpret_s8_p8, vreinterpret_s8_p16)
8947 (vreinterpret_s8_p64, vreinterpretq_s8_f64, vreinterpretq_s8_s16)
8948 (vreinterpretq_s8_s32, vreinterpretq_s8_s64, vreinterpretq_s8_f16)
8949 (vreinterpretq_s8_f32, vreinterpretq_s8_u8, vreinterpretq_s8_u16)
8950 (vreinterpretq_s8_u32, vreinterpretq_s8_u64, vreinterpretq_s8_p8)
8951 (vreinterpretq_s8_p16, vreinterpretq_s8_p64, vreinterpretq_s8_p128)
8952 (vreinterpret_s16_f16, vreinterpret_s16_f64, vreinterpret_s16_s8)
8953 (vreinterpret_s16_s32, vreinterpret_s16_s64, vreinterpret_s16_f32)
8954 (vreinterpret_s16_u8, vreinterpret_s16_u16, vreinterpret_s16_u32)
8955 (vreinterpret_s16_u64, vreinterpret_s16_p8, vreinterpret_s16_p16)
8956 (vreinterpret_s16_p64, vreinterpretq_s16_f64, vreinterpretq_s16_s8)
8957 (vreinterpretq_s16_s32, vreinterpretq_s16_s64, vreinterpretq_s16_f16)
8958 (vreinterpretq_s16_f32, vreinterpretq_s16_u8, vreinterpretq_s16_u16)
8959 (vreinterpretq_s16_u32, vreinterpretq_s16_u64, vreinterpretq_s16_p8)
8960 (vreinterpretq_s16_p16, vreinterpretq_s16_p64, vreinterpretq_s16_p128)
8961 (vreinterpret_s32_f16, vreinterpret_s32_f64, vreinterpret_s32_s8)
8962 (vreinterpret_s32_s16, vreinterpret_s32_s64, vreinterpret_s32_f32)
8963 (vreinterpret_s32_u8, vreinterpret_s32_u16, vreinterpret_s32_u32)
8964 (vreinterpret_s32_u64, vreinterpret_s32_p8, vreinterpret_s32_p16)
8965 (vreinterpret_s32_p64, vreinterpretq_s32_f64, vreinterpretq_s32_s8)
8966 (vreinterpretq_s32_s16, vreinterpretq_s32_s64, vreinterpretq_s32_f16)
8967 (vreinterpretq_s32_f32, vreinterpretq_s32_u8, vreinterpretq_s32_u16)
8968 (vreinterpretq_s32_u32, vreinterpretq_s32_u64, vreinterpretq_s32_p8)
8969 (vreinterpretq_s32_p16, vreinterpretq_s32_p64, vreinterpretq_s32_p128)
8970 (vreinterpret_u8_f16, vreinterpret_u8_f64, vreinterpret_u8_s8)
8971 (vreinterpret_u8_s16, vreinterpret_u8_s32, vreinterpret_u8_s64)
8972 (vreinterpret_u8_f32, vreinterpret_u8_u16, vreinterpret_u8_u32)
8973 (vreinterpret_u8_u64, vreinterpret_u8_p8, vreinterpret_u8_p16)
8974 (vreinterpret_u8_p64, vreinterpretq_u8_f64, vreinterpretq_u8_s8)
8975 (vreinterpretq_u8_s16, vreinterpretq_u8_s32, vreinterpretq_u8_s64)
8976 (vreinterpretq_u8_f16, vreinterpretq_u8_f32, vreinterpretq_u8_u16)
8977 (vreinterpretq_u8_u32, vreinterpretq_u8_u64, vreinterpretq_u8_p8)
8978 (vreinterpretq_u8_p16, vreinterpretq_u8_p64, vreinterpretq_u8_p128)
8979 (vreinterpret_u16_f16, vreinterpret_u16_f64, vreinterpret_u16_s8)
8980 (vreinterpret_u16_s16, vreinterpret_u16_s32, vreinterpret_u16_s64)
8981 (vreinterpret_u16_f32, vreinterpret_u16_u8, vreinterpret_u16_u32)
8982 (vreinterpret_u16_u64, vreinterpret_u16_p8, vreinterpret_u16_p16)
8983 (vreinterpret_u16_p64, vreinterpretq_u16_f64, vreinterpretq_u16_s8)
8984 (vreinterpretq_u16_s16, vreinterpretq_u16_s32, vreinterpretq_u16_s64)
8985 (vreinterpretq_u16_f16, vreinterpretq_u16_f32, vreinterpretq_u16_u8)
8986 (vreinterpretq_u16_u32, vreinterpretq_u16_u64, vreinterpretq_u16_p8)
8987 (vreinterpretq_u16_p16, vreinterpretq_u16_p64, vreinterpretq_u16_p128)
8988 (vreinterpret_u32_f16, vreinterpret_u32_f64, vreinterpret_u32_s8)
8989 (vreinterpret_u32_s16, vreinterpret_u32_s32, vreinterpret_u32_s64)
8990 (vreinterpret_u32_f32, vreinterpret_u32_u8, vreinterpret_u32_u16)
8991 (vreinterpret_u32_u64, vreinterpret_u32_p8, vreinterpret_u32_p16)
8992 (vreinterpret_u32_p64, vreinterpretq_u32_f64, vreinterpretq_u32_s8)
8993 (vreinterpretq_u32_s16, vreinterpretq_u32_s32, vreinterpretq_u32_s64)
8994 (vreinterpretq_u32_f16, vreinterpretq_u32_f32, vreinterpretq_u32_u8)
8995 (vreinterpretq_u32_u16, vreinterpretq_u32_u64, vreinterpretq_u32_p8)
8996 (vreinterpretq_u32_p16, vreinterpretq_u32_p64, vreinterpretq_u32_p128)
8997 (vreinterpretq_f64_p128, vreinterpretq_p128_f64, vreinterpret_bf16_u8)
8998 (vreinterpret_bf16_u16, vreinterpret_bf16_u32, vreinterpret_bf16_u64)
8999 (vreinterpret_bf16_s8, vreinterpret_bf16_s16, vreinterpret_bf16_s32)
9000 (vreinterpret_bf16_s64, vreinterpret_bf16_p8, vreinterpret_bf16_p16)
9001 (vreinterpret_bf16_p64, vreinterpret_bf16_f16, vreinterpret_bf16_f32)
9002 (vreinterpret_bf16_f64, vreinterpretq_bf16_u8, vreinterpretq_bf16_u16)
9003 (vreinterpretq_bf16_u32, vreinterpretq_bf16_u64, vreinterpretq_bf16_s8)
9004 (vreinterpretq_bf16_s16, vreinterpretq_bf16_s32, vreinterpretq_bf16_s64)
9005 (vreinterpretq_bf16_p8, vreinterpretq_bf16_p16, vreinterpretq_bf16_p64)
9006 (vreinterpretq_bf16_p128, vreinterpretq_bf16_f16)
9007 (vreinterpretq_bf16_f32, vreinterpretq_bf16_f64, vreinterpret_s8_bf16)
9008 (vreinterpret_s16_bf16, vreinterpret_s32_bf16, vreinterpret_s64_bf16)
9009 (vreinterpret_u8_bf16, vreinterpret_u16_bf16, vreinterpret_u32_bf16)
9010 (vreinterpret_u64_bf16, vreinterpret_f16_bf16, vreinterpret_f32_bf16)
9011 (vreinterpret_f64_bf16, vreinterpret_p8_bf16, vreinterpret_p16_bf16)
9012 (vreinterpret_p64_bf16, vreinterpretq_s8_bf16, vreinterpretq_s16_bf16)
9013 (vreinterpretq_s32_bf16, vreinterpretq_s64_bf16, vreinterpretq_u8_bf16)
9014 (vreinterpretq_u16_bf16, vreinterpretq_u32_bf16, vreinterpretq_u64_bf16)
9015 (vreinterpretq_f16_bf16, vreinterpretq_f32_bf16, vreinterpretq_f64_bf16)
9016 (vreinterpretq_p8_bf16, vreinterpretq_p16_bf16, vreinterpretq_p64_bf16)
9017 (vreinterpretq_p128_bf16): Delete
9019 2022-07-26 Andrew Carlotti <andrew.carlotti@arm.com>
9021 * config/aarch64/aarch64-builtins.cc
9022 (aarch64_simd_builtin_std_type): Rename to...
9023 (aarch64_int_or_fp_type): ...this, and allow irrelevant qualifiers.
9024 (aarch64_lookup_simd_builtin_type): Rename to...
9025 (aarch64_simd_builtin_type): ...this. Add const/pointer
9026 support, and extract table lookup to...
9027 (aarch64_lookup_simd_type_in_table): ...this function.
9028 (aarch64_init_crc32_builtins): Update to use aarch64_simd_builtin_type.
9029 (aarch64_init_fcmla_laneq_builtins): Ditto.
9030 (aarch64_init_simd_builtin_functions): Ditto.
9032 2022-07-26 Andrew Carlotti <andrew.carlotti@arm.com>
9034 * config/aarch64/aarch64-builtins.cc
9035 (aarch64_general_gimple_fold_builtin): Add combine.
9037 2022-07-26 Richard Biener <rguenther@suse.de>
9039 PR tree-optimization/106189
9040 * gimple-array-bounds.cc (array_bounds_checker::check_mem_ref):
9041 Divide using offset_ints.
9043 2022-07-26 Lulu Cheng <chenglulu@loongson.cn>
9045 * common/config/loongarch/loongarch-common.cc:
9046 Enable '-fsection-anchors' when O1 and more advanced optimization.
9047 * config/loongarch/genopts/loongarch.opt.in: Add new option
9048 '-mexplicit-relocs', and enable by default.
9049 * config/loongarch/loongarch-protos.h (loongarch_split_move_insn_p):
9050 Delete function declaration.
9051 (loongarch_split_move_insn): Delete function declaration.
9052 (loongarch_split_symbol_type): Add function declaration.
9053 * config/loongarch/loongarch.cc (enum loongarch_address_type):
9054 Add new address type 'ADDRESS_LO_SUM'.
9055 (loongarch_classify_symbolic_expression): New function definitions.
9056 Classify the base of symbolic expression X, given that X appears in
9058 (loongarch_symbol_insns): Add a judgment condition TARGET_EXPLICIT_RELOCS.
9059 (loongarch_split_symbol_type): New function definitions.
9060 Determines whether the symbol load should be split into two instructions.
9061 (loongarch_valid_lo_sum_p): New function definitions.
9062 Return true if a LO_SUM can address a value of mode MODE when the LO_SUM
9063 symbol has type SYMBOL_TYPE.
9064 (loongarch_classify_address): Add handling of 'LO_SUM'.
9065 (loongarch_address_insns): Add handling of 'ADDRESS_LO_SUM'.
9066 (loongarch_signed_immediate_p): Sort code.
9067 (loongarch_12bit_offset_address_p): Return true if address type is ADDRESS_LO_SUM.
9068 (loongarch_const_insns): Add handling of 'HIGH'.
9069 (loongarch_split_move_insn_p): Add the static attribute to the function.
9070 (loongarch_emit_set): New function definitions.
9071 (loongarch_call_tls_get_addr): Add symbol handling when defining TARGET_EXPLICIT_RELOCS.
9072 (loongarch_legitimize_tls_address): Add symbol handling when defining the
9073 TARGET_EXPLICIT_RELOCS macro.
9074 (loongarch_split_symbol): New function definitions. Split symbol.
9075 (loongarch_legitimize_address): Add codes see if the address can split into a high part
9077 (loongarch_legitimize_const_move): Add codes split moves of symbolic constants into
9079 (loongarch_split_move_insn): Delete function definitions.
9080 (loongarch_output_move): Add support for HIGH and LO_SUM.
9081 (loongarch_print_operand_reloc): New function definitions.
9082 Print symbolic operand OP, which is part of a HIGH or LO_SUM in context CONTEXT.
9083 (loongarch_memmodel_needs_release_fence): Sort code.
9084 (loongarch_print_operand): Rearrange alphabetical order and add H and L to support HIGH
9086 (loongarch_print_operand_address): Add handling of 'ADDRESS_LO_SUM'.
9087 (TARGET_MIN_ANCHOR_OFFSET): Define macro to -IMM_REACH/2.
9088 (TARGET_MAX_ANCHOR_OFFSET): Define macro to IMM_REACH/2-1.
9089 * config/loongarch/loongarch.md (movti): Delete the template.
9090 (*movti): Delete the template.
9091 (movtf): Delete the template.
9092 (*movtf): Delete the template.
9093 (*low<mode>): New template of normal symbol low address.
9094 (@tls_low<mode>): New template of tls symbol low address.
9095 (@ld_from_got<mode>): New template load address from got table.
9096 (@ori_l_lo12<mode>): New template.
9097 * config/loongarch/loongarch.opt: Update from loongarch.opt.in.
9098 * config/loongarch/predicates.md: Add support for symbol_type HIGH.
9100 2022-07-26 Lulu Cheng <chenglulu@loongson.cn>
9102 * config/loongarch/constraints.md (a): Delete the constraint.
9103 (b): A constant call not local address.
9104 (h): Delete the constraint.
9105 (t): Delete the constraint.
9106 * config/loongarch/loongarch-opts.cc (loongarch_config_target):
9107 Remove cModel type support other than normal.
9108 * config/loongarch/loongarch-protos.h (enum loongarch_symbol_type):
9109 Add new symbol type 'SYMBOL_PCREL', 'SYMBOL_TLS_IE' and 'SYMBOL_TLS_LE'.
9110 (loongarch_split_symbol): Delete useless function declarations.
9111 (loongarch_split_symbol_type): Delete useless function declarations.
9112 * config/loongarch/loongarch.cc (enum loongarch_address_type):
9113 Delete unnecessary comment information.
9114 (loongarch_symbol_binds_local_p): Modified the judgment order of label
9116 (loongarch_classify_symbol): Return symbol type. If symbol is a label,
9117 or symbol is a local symbol return SYMBOL_PCREL. If is a tls symbol,
9118 return SYMBOL_TLS. If is a not local symbol return SYMBOL_GOT_DISP.
9119 (loongarch_symbolic_constant_p): Add handling of 'SYMBOL_TLS_IE'
9120 'SYMBOL_TLS_LE' and 'SYMBOL_PCREL'.
9121 (loongarch_symbol_insns): Add handling of 'SYMBOL_TLS_IE' 'SYMBOL_TLS_LE'
9123 (loongarch_address_insns): Sort code.
9124 (loongarch_12bit_offset_address_p): Sort code.
9125 (loongarch_14bit_shifted_offset_address_p): Sort code.
9126 (loongarch_call_tls_get_addr): Sort code.
9127 (loongarch_legitimize_tls_address): Sort code.
9128 (loongarch_output_move): Remove schema support for cmodel other than normal.
9129 (loongarch_memmodel_needs_release_fence): Sort code.
9130 (loongarch_print_operand): Sort code.
9131 * config/loongarch/loongarch.h (LARCH_U12BIT_OFFSET_P):
9132 Rename to LARCH_12BIT_OFFSET_P.
9133 (LARCH_12BIT_OFFSET_P): New macro.
9134 * config/loongarch/loongarch.md: Reimplement the function call. Remove schema
9135 support for cmodel other than normal.
9136 * config/loongarch/predicates.md (is_const_call_weak_symbol): Delete this predicate.
9137 (is_const_call_plt_symbol): Delete this predicate.
9138 (is_const_call_global_noplt_symbol): Delete this predicate.
9139 (is_const_call_no_local_symbol): New predicate, determines whether it is a local
9142 2022-07-26 Kewen Lin <linkw@linux.ibm.com>
9145 * config/rs6000/rs6000-p8swap.cc (replace_swapped_aligned_store): Copy
9146 REG_EH_REGION when replacing one store insn having it.
9147 (replace_swapped_aligned_load): Likewise.
9149 2022-07-25 Aldy Hernandez <aldyh@redhat.com>
9151 * Makefile.in (OBJS): Add range-op-float.o.
9152 * range-op.cc (get_float_handler): New.
9153 (range_op_handler::range_op_handler): Save code and type for
9155 (range_op_handler::oeprator bool): Move from header file, and
9156 add support for floats.
9157 (range_op_handler::fold_range): Add support for floats.
9158 (range_op_handler::op1_range): Same.
9159 (range_op_handler::op2_range): Same.
9160 (range_op_handler::lhs_op1_relation): Same.
9161 (range_op_handler::lhs_op2_relation): Same.
9162 (range_op_handler::op1_op2_relation): Same.
9163 * range-op.h (class range_operator_float): New.
9164 (class floating_op_table): New.
9165 * value-query.cc (range_query::get_tree_range): Add case for
9167 * range-op-float.cc: New file.
9169 2022-07-25 Roger Sayle <roger@nextmovesoftware.com>
9170 Uroš Bizjak <ubizjak@gmail.com>
9173 * config/i386/i386-expand.cc (split_double_concat): A new helper
9174 function for setting a double word value from two word values.
9175 * config/i386/i386-protos.h (split_double_concat): Prototype here.
9176 * config/i386/i386.md (zero_extendditi2): New define_insn_and_split.
9177 (*add<dwi>3_doubleword_zext): New define_insn_and_split.
9178 (*sub<dwi>3_doubleword_zext): New define_insn_and_split.
9179 (*concat<mode><dwi>3_1): New define_insn_and_split replacing
9180 previous define_split for implementing DST = (HI<<32)|LO as
9181 pair of move instructions, setting lopart and hipart.
9182 (*concat<mode><dwi>3_2): Likewise.
9183 (*concat<mode><dwi>3_3): Likewise, where HI is zero_extended.
9184 (*concat<mode><dwi>3_4): Likewise, where HI is zero_extended.
9186 2022-07-25 Aldy Hernandez <aldyh@redhat.com>
9188 PR middle-end/106432
9189 * gimple-range.cc (gimple_ranger::range_on_edge): Return false
9190 when the result range type is unsupported.
9192 2022-07-25 Sebastian Huber <sebastian.huber@embedded-brains.de>
9194 * config/rs6000/rtems.h (CPLUSPLUS_CPP_SPEC): Undef.
9196 2022-07-25 Richard Biener <rguenther@suse.de>
9198 PR middle-end/106414
9199 * match.pd (~(x ^ y) -> x == y): Restrict to single bit
9202 2022-07-25 Andre Vieira <andre.simoesdiasvieira@arm.com>
9204 * config/aarch64/aarch64.md (rbit<mode>2): Rename this ...
9205 (@aarch64_rbit<mode>): ... to this and change it in...
9206 (ffs<mode>2,ctz<mode>2): ... here.
9207 (@aarch64_rev16<mode>): New.
9208 * config/aarch64/aarch64-builtins.cc: (aarch64_builtins):
9209 Define the following enum AARCH64_REV16, AARCH64_REV16L,
9210 AARCH64_REV16LL, AARCH64_RBIT, AARCH64_RBITL, AARCH64_RBITLL.
9211 (aarch64_init_data_intrinsics): New.
9212 (aarch64_general_init_builtins): Add call to
9213 aarch64_init_data_intrinsics.
9214 (aarch64_expand_builtin_data_intrinsic): New.
9215 (aarch64_general_expand_builtin): Add call to
9216 aarch64_expand_builtin_data_intrinsic.
9217 * config/aarch64/arm_acle.h (__clz, __clzl, __clzll, __cls, __clsl,
9218 __clsll, __rbit, __rbitl, __rbitll, __rev, __revl, __revll, __rev16,
9219 __rev16l, __rev16ll, __ror, __rorl, __rorll, __revsh): New.
9221 2022-07-25 Martin Liska <mliska@suse.cz>
9223 * doc/extend.texi: Remove trailing whitespaces.
9224 * doc/invoke.texi: Likewise.
9226 2022-07-25 Aldy Hernandez <aldyh@redhat.com>
9228 * value-range-pretty-print.cc (vrange_printer::visit): New.
9229 (vrange_printer::print_frange_prop): New.
9230 * value-range-pretty-print.h (class vrange_printer): Add visit and
9232 * value-range-storage.h (vrange_allocator::alloc_vrange): Handle frange.
9233 (vrange_allocator::alloc_frange): New.
9234 * value-range.cc (vrange::operator=): Handle frange.
9235 (vrange::operator==): Same.
9236 (frange::accept): New.
9238 (frange::normalize_kind): New.
9239 (frange::union_): New.
9240 (frange::intersect): New.
9241 (frange::operator=): New.
9242 (frange::operator==): New.
9243 (frange::supports_type_p): New.
9244 (frange::verify_range): New.
9245 * value-range.h (enum value_range_discriminator): Handle frange.
9246 (class fp_prop): New.
9247 (FP_PROP_ACCESSOR): New.
9248 (class frange_props): New.
9249 (FRANGE_PROP_ACCESSOR): New.
9250 (class frange): New.
9251 (Value_Range::init): Handle frange.
9252 (Value_Range::operator=): Same.
9253 (Value_Range::supports_type_p): Same.
9254 (frange_props::operator==): New.
9255 (frange_props::union_): New.
9256 (frange_props::intersect): New
9257 (frange::frange): New.
9258 (frange::type): New.
9259 (frange::set_varying): New.
9260 (frange::set_undefined): New.
9262 2022-07-25 Peter Bergner <bergner@linux.ibm.com>
9263 Kewen Lin <linkw@linux.ibm.com>
9266 * config/rs6000/rs6000.h (DRIVER_SELF_SPECS): Adjust -mdejagnu-cpu
9267 to filter out all -mtune options.
9269 2022-07-24 Aldy Hernandez <aldyh@redhat.com>
9271 * value-query.cc (range_query::get_value_range): Add assert.
9272 * vr-values.cc (vr_values::range_of_expr): Make sure we don't ICE
9273 on unsupported types in vr_values.
9274 (simplify_using_ranges::simplify): Same.
9276 2022-07-24 Aldy Hernandez <aldyh@redhat.com>
9278 * tree-ssanames.cc (get_nonzero_bits): Return -1 for unsupported
9280 * value-query.cc (get_ssa_name_range_info): Remove precision check.
9282 2022-07-24 Aldy Hernandez <aldyh@redhat.com>
9284 * value-relation.cc (value_relation::set_relation): Remove assert.
9285 (path_oracle::register_relation): Exit when trying to register
9286 same SSA name relations.
9288 2022-07-24 Aldy Hernandez <aldyh@redhat.com>
9290 * gimple-range-cache.cc (ranger_cache::edge_range): Convert to vrange.
9291 (ranger_cache::range_from_dom): Same.
9293 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges): Same.
9295 2022-07-24 Roger Sayle <roger@nextmovesoftware.com>
9296 H.J. Lu <hjl.tools@gmail.com>
9300 * config/i386/i386-features.cc (make_vector_copies): Move from
9301 general_scalar_chain to scalar_chain.
9302 (convert_reg): Likewise.
9303 (convert_insn_common): New scalar_chain method split out from
9304 general_scalar_chain convert_insn.
9305 (convert_registers): Move from general_scalar_chain to
9307 (scalar_chain::convert): Call convert_insn_common before calling
9309 (timode_remove_non_convertible_regs): Iterate until there are
9310 no further changes to the candidates.
9311 * config/i386/i386-features.h (scalar_chain::hash_map): Move
9312 from general_scalar_chain.
9313 (scalar_chain::convert_reg): Likewise.
9314 (scalar_chain::convert_insn_common): New shared method.
9315 (scalar_chain::make_vector_copies): Move from general_scalar_chain.
9316 (scalar_chain::convert_registers): Likewise. No longer virtual.
9317 (general_scalar_chain::hash_map): Delete. Moved to scalar_chain.
9318 (general_scalar_chain::convert_reg): Likewise.
9319 (general_scalar_chain::make_vector_copies): Likewise.
9320 (general_scalar_chain::convert_registers): Delete virtual method.
9321 (timode_scalar_chain::convert_registers): Likewise.
9323 2022-07-23 Immad Mir <mirimmad@outlook.com>
9325 * doc/extend.texi: Add fd_arg, fd_arg_read and fd_arg_write under
9326 "Common Function Attributes" section.
9327 * doc/invoke.texi: Add docs to -Wanalyzer-fd-access-mode-mismatch,
9328 -Wanalyzer-use-after-close, -Wanalyzer-fd-use-without-check that these
9329 warnings may be emitted through usage of three function attributes used
9330 for static analysis of file descriptors namely fd_arg, fd_arg_read and
9333 2022-07-22 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9335 * config/xtensa/xtensa.md (*masktrue_const_bitcmpl):
9336 Add a new insn_and_split pattern, and a few split patterns for
9339 2022-07-22 Richard Sandiford <richard.sandiford@arm.com>
9341 * graphds.cc (graphds_scc): Fix algorithm attribution.
9343 2022-07-22 Richard Biener <rguenther@suse.de>
9345 PR tree-optimization/106403
9346 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Move stored
9347 value valueization after check for IFN_MASKED_STORE or
9350 2022-07-22 Richard Biener <rguenther@suse.de>
9352 PR tree-optimization/106397
9353 * tree-ssa-loop-prefetch.cc (emit_mfence_after_loop): Do
9354 not update SSA form here.
9355 (mark_nontemporal_stores): Return whether we marked any
9356 non-temporal stores and inserted mfence.
9357 (loop_prefetch_arrays): Note when we need to update SSA.
9358 (tree_ssa_prefetch_arrays): Perform required (LC) SSA update
9359 at the end of the pass.
9361 2022-07-22 Richard Biener <rguenther@suse.de>
9363 PR tree-optimization/106387
9364 * tree-vect-stmts.cc (vectorizable_load): Use make_ssa_name
9365 if ptr is not an SSA name.
9367 2022-07-22 liuhongt <hongtao.liu@intel.com>
9370 * config/i386/mmx.md (<code><mode>3): New define_expand, it's
9371 original "<code><mode>3".
9372 (*<code><mode>3): New define_insn, it's original
9373 "<code><mode>3" be extended to handle memory and immediate
9374 operand with ix86_binary_operator_ok. Also adjust define_split
9376 (mmxinsnmode): New mode attribute.
9377 (*mov<mode>_imm): Refactor with mmxinsnmode.
9378 * config/i386/predicates.md
9379 (register_or_x86_64_const_vector_operand): New predicate.
9381 2022-07-22 Will Schmidt <will_schmidt@vnet.ibm.com>
9383 * config/rs6000/altivec.md:
9384 (vstrir_code_<mode>): Rename to...
9385 (vstrir_direct_<mode>): ... this.
9386 (vstrir_p_code_<mode>): Rename to...
9387 (vstrir_p_direct_<mode>): ... this.
9388 (vstril_code_<mode>): Rename to...
9389 (vstril_direct_<mode>): ... this.
9390 (vstril_p_code_<mode>): Rename to...
9391 (vstril_p_direct_<mode>): ... this.
9393 2022-07-22 Will Schmidt <will_schmidt@vnet.ibm.com>
9395 * config/rs6000/rs6000-c.cc: Update comments.
9396 (rs6000_target_modify_macros): Remove bu_mask references.
9397 (rs6000_define_or_undefine_macro): Replace bu_mask reference
9398 with a rs6000_cpu value check.
9399 (rs6000_cpu_cpp_builtins): Remove rs6000_builtin_mask_calculate()
9400 parameter from call to rs6000_target_modify_macros.
9401 * config/rs6000/rs6000-protos.h (rs6000_target_modify_macros,
9402 rs6000_target_modify_macros_ptr): Remove parameter from extern
9404 * config/rs6000/rs6000.cc (rs6000_target_modify_macros_ptr): Remove
9405 parameter from prototype, update calls to this function.
9406 (rs6000_print_builtin_options): Remove prototype, call and function.
9407 (rs6000_builtin_mask_calculate): Remove function.
9408 (rs6000_debug_reg_global): Remove call to rs6000_print_builtin_options.
9409 (rs6000_option_override_internal): Remove rs6000_builtin_mask var
9410 and builtin_mask debug output.
9411 (rs6000_builtin_mask_names): Remove.
9412 (rs6000_pragma_target_parse): Remove prev_bumask, cur_bumask,
9413 diff_bumask references; Update calls to rs6000_target_modify_ptr.
9414 * config/rs6000/rs6000.opt (rs6000_builtin_mask): Remove.
9416 2022-07-21 Sam Feifer <sfeifer@redhat.com>
9418 PR tree-optimization/94920
9419 * match.pd (x >= 0 ? x : 0) + (x <= 0 ? -x : 0): New simplification.
9420 (x <= 0 ? -x : 0): New simplification.
9422 2022-07-21 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
9425 2022-07-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
9427 * tree-ssa-forwprop.cc (simplify_permutation): Use lhs type
9428 instead of TREE_TYPE (arg0) as result type in folding VEC_PERM_EXPR.
9430 2022-07-21 Richard Biener <rguenther@suse.de>
9432 PR tree-optimization/106379
9433 * match.pd (~(a ^ b) -> a == b): New pattern.
9435 2022-07-21 Richard Biener <rguenther@suse.de>
9437 PR tree-optimization/106378
9438 * tree-ssa-dse.cc (initialize_ao_ref_for_dse): Handle
9439 LEN_STORE, add mode to initialize a may-def and handle
9440 MASK_STORE that way.
9441 (dse_optimize_stmt): Query may-defs. Handle internal
9442 functions LEN_STORE and MASK_STORE similar to how
9443 we handle memory builtins but without byte tracking.
9445 2022-07-21 Richard Biener <rguenther@suse.de>
9447 PR tree-optimization/106365
9448 * tree-ssa-sccvn.cc (pd_data::rhs_off): New field determining
9449 the offset to start encoding of RHS from.
9450 (vn_walk_cb_data::vn_walk_cb_data): Initialize it.
9451 (vn_walk_cb_data::push_partial_def): Allow the first partial
9452 definition to be fully providing the def. Offset RHS
9453 before encoding if requested.
9454 (vn_reference_lookup_3): Initialize def_rhs everywhere.
9455 Add support for .MASK_STORE and .LEN_STORE (partial) definitions.
9457 2022-07-21 Richard Biener <rguenther@suse.de>
9459 * tree-ssa-alias.cc (ref_maybe_used_by_call_p_1): Special-case
9460 store internal functions and IFN_MASK_LOAD, IFN_LEN_LOAD
9461 and IFN_MASK_LOAD_LANES.
9462 (call_may_clobber_ref_p_1): Special-case IFN_MASK_STORE,
9463 IFN_LEN_STORE and IFN_MASK_STORE_LANES.
9465 2022-07-20 Alexander Monakov <amonakov@ispras.ru>
9467 PR rtl-optimization/101347
9468 * builtins.cc (expand_builtin) [BUILT_IN_SETJMP_SETUP]: Move
9469 population of nonlocal_goto_handler_labels from here ...
9470 (expand_builtin) [BUILT_IN_SETJMP_RECEIVER]: ... to here.
9471 * rtlanal.cc (remove_node_from_insn_list): Verify that a
9472 duplicate is not present in the remainder of the list.
9474 2022-07-20 Alexander Monakov <amonakov@ispras.ru>
9476 * rtl.h (remove_node_from_expr_list): Remove declaration.
9477 * rtlanal.cc (remove_node_from_expr_list): Remove (no uses).
9479 2022-07-20 Richard Biener <rguenther@suse.de>
9481 * tree-vect-data-refs.cc (bump_vector_ptr): Return an
9482 invariant updated address when the input was invariant.
9484 2022-07-20 liuhongt <hongtao.liu@intel.com>
9486 * passes.def: (Split pass_cse_sincos to pass_expand_powcabs
9487 and pass_cse_sincos, and move pass_cse_sincos after vectorizer).
9488 * timevar.def (TV_TREE_POWCABS): New timevar.
9489 * tree-pass.h (make_pass_expand_powcabs): Split from pass_cse_sincos.
9490 * tree-ssa-math-opts.cc (gimple_expand_builtin_cabs): Ditto.
9491 (class pass_expand_powcabs): Ditto.
9492 (pass_expand_powcabs::execute): Ditto.
9493 (make_pass_expand_powcabs): Ditto.
9494 (pass_cse_sincos::execute): Remove pow/cabs expand part.
9495 (make_pass_cse_sincos): Ditto.
9497 2022-07-20 Richard Biener <richard.guenther@gmail.com>
9498 Hongtao Liu <hongtao.liu@intel.com>
9500 PR tree-optimization/106010
9501 * tree-complex.cc (init_dont_simulate_again): Lower complex
9503 (expand_complex_move): Also expand COMPLEX_CST for rhs.
9505 2022-07-19 Andrew MacLeod <amacleod@redhat.com>
9507 * gimple-range-cache.cc (ranger_cache::range_from_dom): Check
9508 for incoming ranges on join nodes and add to worklist.
9510 2022-07-19 Andrew MacLeod <amacleod@redhat.com>
9512 * gimple-range-cache.cc (ranger_cache::resolve_dom): New.
9513 (ranger_cache::range_from_dom): Put all nodes to be calculated
9514 in the worklist and resolve after the dom walk.
9515 * gimple-range-cache.h (resolve_dom): New prototype.
9517 2022-07-19 Alexander Monakov <amonakov@ispras.ru>
9519 * tree-cfg.cc (gimple_verify_flow_info): Check placement of
9520 returns_twice calls.
9522 2022-07-19 Alexander Monakov <amonakov@ispras.ru>
9524 * cfghooks.cc (duplicate_block): Expand comment.
9525 * tree-cfg.cc (gimple_can_duplicate_bb_p): Reject blocks with
9526 calls that may return twice.
9528 2022-07-19 Alexander Monakov <amonakov@ispras.ru>
9530 * tree-ssa-sink.cc (select_best_block): Punt if selected block
9531 has incoming abnormal edges.
9533 2022-07-19 Martin Liska <mliska@suse.cz>
9535 * doc/extend.texi: Remove trailing :.
9537 2022-07-19 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
9539 * tree-ssa-forwprop.cc (simplify_permutation): Use lhs type
9540 instead of TREE_TYPE (arg0) as result type in folding VEC_PERM_EXPR.
9542 2022-07-19 Richard Biener <rguenther@suse.de>
9544 PR middle-end/106331
9545 * builtins.cc (get_memory_rtx): Compute alignment from
9546 the original address and set MEM_OFFSET to unknown when
9547 we create a MEM_EXPR from the base object of the address.
9549 2022-07-19 Richard Biener <rguenther@suse.de>
9552 * dwarf2out.cc (dwarf2out_register_external_die): Allow
9553 map entry re-use during WPA.
9555 2022-07-19 Roger Sayle <roger@nextmovesoftware.com>
9556 Richard Biener <rguenther@suse.de>
9559 * builtins.cc (fold_builtin_frexp): Call suppress_warning on
9560 COMPOUND_EXPR to silence spurious warning if result isn't used.
9561 (fold_builtin_modf): Likewise.
9562 (do_mpfr_remquo): Likewise.
9564 2022-07-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9566 * config/xtensa/xtensa.cc (xtensa_rtx_costs):
9567 Change the relative cost of '(set (reg) (const_int N))' where
9568 N fits into signed 12-bit from 4 to 0 if optimizing for size.
9569 And use the appropriate macro instead of the bare number 4.
9571 2022-07-18 Andrew MacLeod <amacleod@redhat.com>
9573 PR tree-optimization/106280
9574 * value-relation.cc (dom_oracle::register_relation): Register
9575 transitives only when it is possible for there to be one.
9576 (dom_oracle::set_one_relation): Return NULL if this is an
9579 2022-07-18 Maciej W. Rozycki <macro@embecosm.com>
9581 * doc/invoke.texi (RISC-V Options): Add index references for
9582 `mrelax' and `mriscv-attribute'.
9584 2022-07-18 Maciej W. Rozycki <macro@embecosm.com>
9586 * doc/invoke.texi (Option Summary): Add missing second space
9587 around `-mstack-protector-guard-reg='.
9589 2022-07-18 Maciej W. Rozycki <macro@embecosm.com>
9591 * doc/invoke.texi (Option Summary): Fix `-mno-riscv-attribute'.
9592 (RISC-V Options): Likewise, and `-mriscv-attribute'.
9594 2022-07-18 Claudiu Zissulescu <claziss@gmail.com>
9596 * config/arc/arc-arch.h (arc_tune_attr): Add
9597 ARC_TUNE_ARCHS4X_REL31A variant.
9598 * config/arc/arc.cc (arc_override_options): Tune options for
9600 (arc_sched_issue_rate): Use correct enum.
9601 (arc600_corereg_hazard): Textual change.
9602 (arc_hazard): Add release 310a tunning.
9603 * config/arc/arc.md (tune): Update and take into consideration new
9605 (tune_dspmpy): Likewise.
9606 (tune_store): New attribute.
9607 * config/arc/arc.opt (mtune): New tune option.
9608 * config/arc/arcHS4x.md (hs4x_brcc0, hs4x_brcc1): New cpu units.
9609 (hs4x_brcc_op): New instruction rezervation.
9610 (hs4x_data_store_1_op): Likewise.
9611 * config/arc/arc-cpus.def (hs4x_rel31): New cpu variant.
9612 * config/arc/arc-tables.opt: Regenerate.
9613 * config/arc/t-multilib: Likewise.
9614 * doc/invoke.texi (ARC): Update mcpu and tune sections.
9616 2022-07-18 Richard Biener <rguenther@suse.de>
9618 * tree-loop-distribution.cc (loop_distribution::distribute_loop):
9619 When computing cost-based merging do not disregard builtin
9620 classified partitions in some cases.
9622 2022-07-18 Richard Sandiford <richard.sandiford@arm.com>
9625 * config/arm/arm-builtins.cc (arm_builtin_vectorized_function):
9627 * config/arm/arm-protos.h (arm_builtin_vectorized_function): Delete.
9628 * config/arm/arm.cc (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION):
9630 * config/arm/arm_neon_builtins.def (copysignf): Delete.
9631 * config/arm/iterators.md (nvrint_pattern): New attribute.
9632 * config/arm/neon.md (<NEON_VRINT:nvrint_pattern><VCVTF:mode>2):
9634 (l<NEON_VCVT:nvrint_pattern><su_optab><VCVTF:mode><v_cmp_result>2):
9636 (neon_copysignf<mode>): Rename to...
9637 (copysign<mode>3): ...this.
9639 2022-07-18 Claudiu Zissulescu <claziss@gmail.com>
9641 * config/arc/arc.cc (arc_expand_epilogue): Adjust the frame
9642 pointer first when in interrupts.
9644 2022-07-18 Richard Biener <rguenther@suse.de>
9646 * tree-loop-distribution.cc (copy_loop_before): Add
9647 the ability to replace the original LC PHI defs.
9648 (generate_loops_for_partition): Pass through a flag
9649 whether to redirect original LC PHI defs.
9650 (generate_code_for_partition): Likewise.
9651 (loop_distribution::distribute_loop): Compute the partition
9652 that should provide the LC PHI defs for common reductions
9655 2022-07-18 Richard Ball <richard.ball@arm.com>
9657 * config/aarch64/aarch64.cc (aarch64_evpc_trn): Use std:swap.
9658 (aarch64_evpc_uzp): Likewise.
9659 (aarch64_evpc_zip): Likewise.
9661 2022-07-18 Roger Sayle <roger@nextmovesoftware.com>
9664 * config/i386/i386.md (*ctzsidi2_<s>ext): New insn_and_split
9665 to recognize any_extend:DI of ctz:SI which is implicitly extended.
9666 (*ctzsidi2_<s>ext_falsedep): New define_insn to model a DImode
9667 extended ctz:SI that has preceding xor to break false dependency.
9669 2022-07-18 Roger Sayle <roger@nextmovesoftware.com>
9671 * config/i386/predicates.md (x86_64_const_vector_operand):
9672 Check the operand's mode matches the specified mode argument.
9674 2022-07-18 Roger Sayle <roger@nextmovesoftware.com>
9676 * config/i386/sse.md (kunpckhi): Add UNSPEC_MASKOP unspec.
9677 (kunpcksi): Likewise, add UNSPEC_MASKOP unspec.
9678 (kunpckdi): Likewise, add UNSPEC_MASKOP unspec.
9679 (vec_pack_trunc_qi): Update to specify the now required
9680 UNSPEC_MASKOP unspec.
9681 (vec_pack_trunc_<mode>): Likewise.
9683 2022-07-16 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9685 * config/xtensa/xtensa.md
9686 (*masktrue_const_pow2_minus_one, *masktrue_const_negative_pow2,
9687 *masktrue_const_shifted_mask): If the immediate for bitwise AND is
9688 represented as '-(1 << N)', decrease the lower bound of N from 12
9689 to 1. And the other immediate for conditional branch is now no
9690 longer limited to zero, but also one of some positive integers.
9691 Finally, remove the checks of some conditions, because the comparison
9692 expressions that don't satisfy such checks are determined as
9693 compile-time constants and thus will be optimized away before
9696 2022-07-16 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9698 * config/xtensa/xtensa.cc (xtensa_emit_constantsynth): Remove.
9699 (xtensa_constantsynth_2insn): Change to try all three synthetic
9700 methods and to use the one that fits the immediate value of
9701 the seed into a Narrow Move Immediate instruction "MOVI.N"
9702 when the Code Density Option is configured.
9704 2022-07-15 H.J. Lu <hjl.tools@gmail.com>
9707 * config/i386/i386.cc (ix86_function_ok_for_sibcall): Return
9708 false if callee has indirect_return attribute and caller
9711 2022-07-15 Roger Sayle <roger@nextmovesoftware.com>
9714 * config/i386/i386.md (*andn<dwi>3_doubleword_bmi): Update the
9715 constraints to reflect the output is earlyclobber, unless it is
9716 the same register (pair) as one of the operands.
9718 2022-07-15 David Malcolm <dmalcolm@redhat.com>
9720 * doc/invoke.texi (Static Analyzer Options): Add the new fd
9721 warnings to the initial gccoptlist, and to the list of those
9722 disabled by -fanalyzer-checker=taint.
9724 2022-07-15 Andrew Carlotti <andrew.carlotti@arm.com>
9726 * config/aarch64/aarch64-builtins.cc
9727 (enum aarch64_type_qualifiers): Remove qualifier_internal.
9728 (aarch64_init_simd_builtin_functions): Remove qualifier_internal check.
9730 2022-07-15 Andrew Carlotti <andrew.carlotti@arm.com>
9732 * config/aarch64/aarch64-builtins.cc
9733 (v1di_UP): Add V1DI mode to _UP macros.
9734 * config/aarch64/aarch64-modes.def (VECTOR_MODE): Add V1DI mode.
9735 * config/aarch64/aarch64-simd-builtin-types.def: Use V1DI mode.
9736 * config/aarch64/aarch64-simd.md
9737 (vec_extractv2dfv1df): Replace with...
9738 (vec_extract<mode><V1half>): ...this.
9739 * config/aarch64/aarch64.cc
9740 (aarch64_classify_vector_mode): Add V1DI mode.
9741 * config/aarch64/iterators.md
9742 (VQ_2E, V1HALF, V1half): New.
9743 (nunits): Add V1DI mode.
9745 2022-07-15 Roger Sayle <roger@nextmovesoftware.com>
9748 * config/i386/i386-features.cc (general_scalar_chain::convert_insn):
9749 Fix indentation whitespace.
9750 (timode_scalar_chain::fix_debug_reg_uses): Likewise.
9751 (timode_scalar_chain::convert_insn): Delete dead code.
9752 Update TImode REG_EQUAL_NOTE even if the SET_DEST is already V1TI.
9753 Fix indentation whitespace.
9754 (convertible_comparison_p): Likewise.
9755 (timode_scalar_to_vector_candidate_p): Likewise.
9757 2022-07-15 Aldy Hernandez <aldyh@redhat.com>
9759 * gimple-pretty-print.cc (dump_ssaname_info): Use pp_vrange.
9761 2022-07-15 Aldy Hernandez <aldyh@redhat.com>
9763 * Makefile.in (OBJS): Add value-range-pretty-print.o.
9764 * pretty-print.h (pp_vrange): New.
9765 * value-range.cc (vrange::dump): Call pp version.
9766 (unsupported_range::dump): Move to its own file.
9767 (dump_bound_with_infinite_markers): Same.
9768 (irange::dump): Same.
9769 (irange::dump_bitmasks): Same.
9770 (vrange::debug): Remove.
9771 * value-range.h: Remove virtual designation for dump methods.
9772 Remove dump_bitmasks method.
9773 * value-range-pretty-print.cc: New file.
9774 * value-range-pretty-print.h: New file.
9776 2022-07-15 Aldy Hernandez <aldyh@redhat.com>
9778 * value-range.cc (irange::accept): New.
9779 (unsupported_range::accept): New.
9780 * value-range.h (class vrange_visitor): New.
9781 (class vrange): Add accept method.
9782 (class unsupported_range): Same.
9783 (class Value_Range): Same.
9785 2022-07-15 Jonathan Wakely <jwakely@redhat.com>
9787 * diagnostic-format-json.cc (json_from_location_range): Adjust
9788 to new label_text API.
9789 * diagnostic-format-sarif.cc (sarif_builder::make_location_object):
9791 * diagnostic-show-locus.cc (struct pod_label_text): Likewise.
9792 (layout::print_any_labels): Likewise.
9793 * tree-diagnostic-path.cc (class path_label): Likewise.
9794 (struct event_range): Likewise.
9795 (default_tree_diagnostic_path_printer): Likewise.
9796 (default_tree_make_json_for_path): Likewise.
9798 2022-07-15 konglin1 <lingling.kong@intel.com>
9801 * config/i386/i386-builtin.def (BDESC): Fix [u]comi{ss,sd}
9802 comparison due to intrinsics changed over time.
9803 * config/i386/i386-expand.cc (ix86_ssecom_setcc):
9804 Add unordered check and mode for sse comi codegen.
9805 (ix86_expand_sse_comi): Add unordered check and check a different
9807 (ix86_expand_sse_comi_round):Extract unordered check and mode part
9808 in ix86_ssecom_setcc.
9810 2022-07-15 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
9812 * config/aarch64/aarch64.cc (aarch64_vectorize_vec_perm_const): Use
9813 op_mode instead of vmode in calls to force_reg for op0 and op1.
9815 2022-07-14 H.J. Lu <hjl.tools@gmail.com>
9817 PR tree-optimization/103798
9818 * tree-ssa-forwprop.cc: Include "tree-ssa-strlen.h".
9819 (simplify_builtin_call): Inline memchr with constant strings of
9820 no more than the bytes of a word.
9821 * tree-ssa-strlen.cc (use_in_zero_equality): Make it global.
9822 * tree-ssa-strlen.h (use_in_zero_equality): New.
9824 2022-07-14 Eric Botcazou <ebotcazou@adacore.com>
9826 * gimplify.cc (lookup_tmp_var): Add NOT_GIMPLE_REG boolean parameter
9827 and set DECL_NOT_GIMPLE_REG_P on the variable according to it.
9828 (internal_get_tmp_var): Add NOT_GIMPLE_REG boolean parameter and
9829 pass it in the call to lookup_tmp_var.
9830 (get_formal_tmp_var): Pass false in the call to lookup_tmp_var.
9831 (get_initialized_tmp_var): Likewise.
9832 (prepare_gimple_addressable): Call internal_get_tmp_var instead of
9833 get_initialized_tmp_var with NOT_GIMPLE_REG set to true.
9835 2022-07-14 Martin Liska <mliska@suse.cz>
9837 * doc/gimple.texi: Close properly a deftypefn.
9839 2022-07-14 Martin Liska <mliska@suse.cz>
9841 * doc/gimple.texi: Close properly a deftypefn.
9843 2022-07-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
9845 * config/xtensa/xtensa.md:
9846 In FP constant synthesis split pattern, subcontract to
9847 avoid_constant_pool_reference() as in the case of integer,
9848 because it can handle well too. And cast to int32_t before
9849 calling xtensa_constantsynth() in order to ignore upper 32-bit.
9851 2022-07-13 Aldy Hernandez <aldyh@redhat.com>
9853 * range-op.cc (operator_lt::fold_range): Use nonzero bits.
9855 2022-07-13 Aldy Hernandez <aldyh@redhat.com>
9857 * value-range.cc (irange::copy_to_legacy): Set nonzero mask.
9858 (irange::legacy_intersect): Clear nonzero mask.
9859 (irange::legacy_union): Same.
9860 (irange::invert): Same.
9862 2022-07-13 Richard Biener <rguenther@suse.de>
9864 * tree-ssa-dom.h (record_temporary_equivalences): Remove.
9865 * tree-ssa-dom.cc (dom_jt_state::m_blocks_on_stack): New.
9866 (dom_jt_state::get_blocks_on_stack): Likewise.
9867 (dom_opt_dom_walker::dom_opt_dom_walker): Take dom_jt_state.
9868 (back_propagate_equivalences): Remove dominator bitmap
9869 compute and instead use passed in m_blocks_on_stack.
9870 (record_temporary_equivalences): Likewise.
9871 (record_equivalences_from_incoming_edge): Likewise.
9872 (dom_opt_dom_walker::before_dom_children): Maintain and
9873 pass down blocks on stack.
9874 (dom_opt_dom_walker::after_dom_children): Likewise.
9876 2022-07-13 Andrew Carlotti <andrew.carlotti@arm.com>
9878 * config/aarch64/aarch64-builtins.cc
9879 (aarch64_general_gimple_fold_builtin): Add fixup for invalid GIMPLE.
9881 2022-07-13 Richard Biener <rguenther@suse.de>
9883 PR tree-optimization/106249
9884 * tree-ssa-loop-manip.cc (tree_transform_and_unroll_loop):
9885 Only verify LC SSA of the new_loop if we created it.
9886 Use TODO_update_ssa_nophi for the SSA update after versioning
9889 2022-07-12 Aldy Hernandez <aldyh@redhat.com>
9891 * range-op.cc (unsigned_singleton_p): Remove.
9892 (operator_bitwise_and::remove_impossible_ranges): Remove.
9893 (operator_bitwise_and::fold_range): Set nonzero bits. *
9894 * value-range.cc (irange::get_nonzero_bits): Remove
9895 legacy_mode_p assert.
9896 (irange::dump_bitmasks): Remove legacy_mode_p check.
9898 2022-07-12 Richard Sandiford <richard.sandiford@arm.com>
9901 * predict.h (insn_optimization_type): Declare.
9902 * predict.cc (insn_optimization_type): New function.
9903 * internal-fn.def (IFN_ICEIL, IFN_IFLOOR, IFN_IRINT, IFN_IROUND)
9904 (IFN_LCEIL, IFN_LFLOOR, IFN_LRINT, IFN_LROUND, IFN_LLCEIL)
9905 (IFN_LLFLOOR, IFN_LLRINT, IFN_LLROUND): New internal functions.
9906 * internal-fn.cc (unary_convert_direct): New macro.
9907 (expand_convert_optab_fn): New function.
9908 (expand_unary_convert_optab_fn): New macro.
9909 (direct_unary_convert_optab_supported_p): Likewise.
9910 * optabs.cc (expand_sfix_optab): Pass insn_optimization_type to
9911 convert_optab_handler.
9912 * config/aarch64/aarch64-protos.h
9913 (aarch64_builtin_vectorized_function): Delete.
9914 * config/aarch64/aarch64-builtins.cc
9915 (aarch64_builtin_vectorized_function): Delete.
9916 * config/aarch64/aarch64.cc
9917 (TARGET_VECTORIZE_BUILTIN_VECTORIZED_FUNCTION): Delete.
9918 * config/i386/i386.cc (ix86_optab_supported_p): Handle lround_optab.
9919 * config/i386/i386.md (lround<X87MODEF:mode><SWI248x:mode>2): Remove
9920 optimize_insn_for_size_p test.
9922 2022-07-12 Richard Biener <rguenther@suse.de>
9924 * tree-vect-loop-manip.cc (create_lcssa_for_virtual_phi):
9926 (vect_do_peeling): Do not call it, adjust comment.
9928 2022-07-12 Richard Sandiford <richard.sandiford@arm.com>
9930 * config/aarch64/aarch64-builtins.cc
9931 (aarch64_builtin_vectorized_function): Remove handling of
9932 floor, ceil, trunc, round, nearbyint, sqrt, clz and ctz.
9934 2022-07-11 Andrew MacLeod <amacleod@redhat.com>
9936 PR tree-optimization/106234
9937 * gimple-range-cache.cc (ranger_cache::range_from_dom): Check dominator
9938 cache value before recursively resolving it.
9940 2022-07-11 Roger Sayle <roger@nextmovesoftware.com>
9942 * config/i386/i386-features.h (scalar_chain): Add fields
9943 insns_conv, n_sse_to_integer and n_integer_to_sse to this
9944 parent class, moved from general_scalar_chain.
9945 (scalar_chain::convert_compare): Protected method moved
9946 from general_scalar_chain.
9947 (mark_dual_mode_def): Make protected, not private virtual.
9948 (scalar_chain:convert_op): New private virtual method.
9949 (general_scalar_chain::general_scalar_chain): Simplify constructor.
9950 (general_scalar_chain::~general_scalar_chain): Delete destructor.
9951 (general_scalar_chain): Move insns_conv, n_sse_to_integer and
9952 n_integer_to_sse fields to parent class, scalar_chain.
9953 (general_scalar_chain::mark_dual_mode_def): Delete prototype.
9954 (general_scalar_chain::convert_compare): Delete prototype.
9955 (timode_scalar_chain::compute_convert_gain): Remove simplistic
9956 implementation, convert to a method prototype.
9957 (timode_scalar_chain::mark_dual_mode_def): Delete prototype.
9958 (timode_scalar_chain::convert_op): Prototype new virtual method.
9959 * config/i386/i386-features.cc (scalar_chain::scalar_chain):
9960 Allocate insns_conv and initialize n_sse_to_integer and
9961 n_integer_to_sse fields in constructor.
9962 (scalar_chain::scalar_chain): Free insns_conv in destructor.
9963 (general_scalar_chain::general_scalar_chain): Delete
9964 constructor, now defined in the class declaration.
9965 (general_scalar_chain::~general_scalar_chain): Delete destructor.
9966 (scalar_chain::mark_dual_mode_def): Renamed from
9967 general_scalar_chain::mark_dual_mode_def.
9968 (timode_scalar_chain::mark_dual_mode_def): Delete.
9969 (scalar_chain::convert_compare): Renamed from
9970 general_scalar_chain::convert_compare.
9971 (timode_scalar_chain::compute_convert_gain): New method to
9972 determine the gain from converting a TImode chain to V1TImode.
9973 (timode_scalar_chain::convert_op): New method to convert an
9974 operand from TImode to V1TImode.
9975 (timode_scalar_chain::convert_insn) <case REG>: Only PUT_MODE
9976 on REG_EQUAL notes that were originally TImode (not CONST_INT).
9977 Handle AND, ANDN, XOR, IOR, NOT and COMPARE.
9978 (timode_mem_p): Helper predicate to check where operand is
9979 memory reference with sufficient alignment for TImode STV.
9980 (timode_scalar_to_vector_candidate_p): Use convertible_comparison_p
9981 to check whether COMPARE is convertible. Handle SET_DESTs that
9982 that are REG_P or MEM_P and SET_SRCs that are REG, CONST_INT,
9983 CONST_WIDE_INT, MEM, AND, ANDN, IOR, XOR or NOT.
9985 2022-07-11 Richard Sandiford <richard.sandiford@arm.com>
9987 PR tree-optimization/106250
9988 * tree-vect-loop.cc (vectorizable_reduction): Reinstate final
9989 argument to directly_supported_p.
9991 2022-07-11 Richard Biener <rguenther@suse.de>
9993 * tree-into-ssa.cc (update_ssa): Do not forcefully
9994 re-compute dominance fast queries for TODO_update_ssa_no_phi.
9996 2022-07-11 Richard Biener <rguenther@suse.de>
9998 PR tree-optimization/106228
9999 * tree-vect-data-refs.cc (vect_setup_realignment): Adjust
10000 VUSE compute for the non-loop case.
10002 2022-07-11 Richard Biener <rguenther@suse.de>
10004 * tree-into-ssa.cc (rewrite_mode::REWRITE_UPDATE_REGION): New.
10005 (rewrite_update_dom_walker::rewrite_update_dom_walker): Update.
10006 (rewrite_update_dom_walker::m_in_region_flag): New.
10007 (rewrite_update_dom_walker::before_dom_children): If the region
10008 to update is marked, STOP at exits.
10009 (rewrite_blocks): For REWRITE_UPDATE_REGION mark the region
10011 (dump_update_ssa): Use bitmap_empty_p.
10012 (update_ssa): Likewise. Use REWRITE_UPDATE_REGION when
10013 TODO_update_ssa_no_phi.
10014 * tree-cfgcleanup.cc (cleanup_tree_cfg_noloop): Account
10015 pending update_ssa to the caller.
10017 2022-07-11 Richard Biener <rguenthert@suse.de>
10020 * config/i386/i386-options.cc (ix86_set_current_function):
10021 Rebuild the target optimization node whenever necessary,
10022 not only when the optimization node didn't change.
10024 2022-07-11 Richard Biener <rguenther@suse.de>
10026 PR tree-optimization/106228
10027 * tree-vect-data-refs.cc (vect_setup_realignment): Properly
10028 set a VUSE operand on the emitted load.
10030 2022-07-11 Aldy Hernandez <aldyh@redhat.com>
10032 * gimple-range.cc (gimple_ranger::export_global_ranges): Remove
10033 verification against legacy value_range.
10034 (gimple_ranger::register_inferred_ranges): Same.
10035 (gimple_ranger::export_global_ranges): Rename update_global_range
10037 * tree-core.h (struct range_info_def): Remove.
10038 (struct irange_storage_slot): New.
10039 (struct tree_base): Remove SSA_NAME_ANTI_RANGE_P documentation.
10040 (struct tree_ssa_name): Add vrange_storage support.
10041 * tree-ssanames.cc (range_info_p): New.
10042 (range_info_fits_p): New.
10043 (range_info_alloc): New.
10044 (range_info_free): New.
10045 (range_info_get_range): New.
10046 (range_info_set_range): New.
10047 (set_range_info_raw): Remove.
10048 (set_range_info): Adjust to use vrange_storage.
10049 (set_nonzero_bits): Same.
10050 (get_nonzero_bits): Same.
10051 (duplicate_ssa_name_range_info): Remove overload taking
10053 Rewrite tree overload to use vrange_storage.
10054 (duplicate_ssa_name_fn): Adjust to use vrange_storage.
10055 * tree-ssanames.h (struct range_info_def): Remove.
10056 (set_range_info): Adjust prototype to take vrange.
10057 * tree-vrp.cc (vrp_asserts::remove_range_assertions): Call
10058 duplicate_ssa_name_range_info.
10059 * tree.h (SSA_NAME_ANTI_RANGE_P): Remove.
10060 (SSA_NAME_RANGE_TYPE): Remove.
10061 * value-query.cc (get_ssa_name_range_info): Adjust to use
10063 (update_global_range): Remove.
10064 (get_range_global): Remove as_a<irange>.
10065 * value-query.h (update_global_range): Remove.
10066 * tree-ssa-dom.cc (set_global_ranges_from_unreachable_edges):
10067 Rename update_global_range to set_range_info.
10068 * value-range-storage.cc (vrange_storage::alloc_slot): Remove
10071 2022-07-10 Aldy Hernandez <aldyh@redhat.com>
10073 * value-range.cc (irange::operator=): Call verify_range.
10074 (irange::irange_set): Normalize kind after everything else has
10076 (irange::irange_set_anti_range): Same.
10077 (irange::set): Same.
10078 (irange::verify_range): Disallow nonzero masks for VARYING.
10079 (irange::irange_union): Call verify_range.
10080 Handle nonzero masks better.
10081 (irange::irange_intersect): Same.
10082 (irange::set_nonzero_bits): Calculate mask if either range has an
10084 (irange::intersect_nonzero_bits): Same.
10085 (irange::union_nonzero_bits): Same.
10086 (range_tests_nonzero_bits): New.
10087 (range_tests): Call range_tests_nonzero_bits.
10088 * value-range.h (class irange): Remove set_nonzero_bits method
10090 (irange::varying_compatible_p): Set nonzero mask.
10092 2022-07-10 Xi Ruoyao <xry111@xry111.site>
10094 * config/loongarch/loongarch.md (<any_div>di3_fake): Describe
10095 the sign-extend of result in the RTL template.
10096 (<any_div><mode>3): Adjust for <any_div>di3_fake change.
10098 2022-07-10 Xi Ruoyao <xry111@xry111.site>
10100 * config/loongarch/loongarch.cc (loongarch_check_zero_div_p):
10101 Remove static, for use in the machine description file.
10102 * config/loongarch/loongarch-protos.h:
10103 (loongarch_check_zero_div_p): Add prototype.
10104 * config/loongarch/loongarch.md (enabled): New attr.
10105 (*<optab><mode>3): Add (=r,r,r) and (=&r,0,r) alternatives for
10106 idiv. Conditionally enable the alternatives using
10107 loongarch_check_zero_div_p.
10108 (<optab>di3_fake): Likewise.
10110 2022-07-10 Xi Ruoyao <xry111@xry111.site>
10112 * config/loongarch/loongarch.md (mulsidi3_64bit): Use mulw.d.w
10115 2022-07-09 Aldy Hernandez <aldyh@redhat.com>
10117 * value-range.cc (irange::irange_single_pair_union): Set
10118 VR_VARYING when appropriate.
10120 2022-07-09 Vit Kabele <vit.kabele@sysgo.com>
10122 * stor-layout.cc (finalize_record_size): Extend warning message.
10124 2022-07-09 Sam Feifer <sfeifer@redhat.com>
10126 PR tree-optimization/98304
10127 * match.pd (n - (((n > C1) ? n : C1) & -C2)): New simplification.
10129 2022-07-09 Jeff Law <jeffreyalaw@gmail.com>
10131 * expr.cc (store_expr): Identify trailing NULs in a STRING_CST
10132 initializer and use clear_storage rather than copying the
10133 NULs to the destination array.
10135 2022-07-09 Roger Sayle <roger@nextmovesoftware.com>
10137 * config/i386/i386.md (define_peephole2): Use match_operand of
10138 flags_reg_operand to capture and preserve the mode of FLAGS_REG.
10139 (define_peephole2): Likewise.
10140 (define_peephole2): Likewise...
10142 2022-07-09 Roger Sayle <roger@nextmovesoftware.com>
10144 * config/i386/i386-features.cc (convert_compare): Add support
10145 for *testdi_not_doubleword pattern, "(compare (and (not ...)))"
10146 by generating a pandn followed by ptest.
10147 (convertible_comparison_p): Recognize both *cmpdi_doubleword and
10148 recent *testdi_not_doubleword comparison patterns.
10150 2022-07-09 Tamar Christina <tamar.christina@arm.com>
10152 * config/s390/s390.cc (s390_expand_atomic): Pass false to store_bit_field to
10153 indicate that the value is not undefined.
10155 2022-07-09 Andrew Pinski <apinski@marvell.com>
10157 PR tree-optimization/106087
10158 * tree-ssa-dce.cc (simple_dce_from_worklist): Check
10159 to make sure the statement is only defining one operand.
10161 2022-07-08 Lewis Hyatt <lhyatt@gmail.com>
10163 PR preprocessor/91733
10164 * input.cc (find_end_of_line): New helper function.
10165 (file_cache_slot::get_next_line): Recognize \r as a line ending.
10166 * diagnostic-show-locus.cc (test_escaping_bytes_1): Adapt selftest
10167 since \r will now be interpreted as a line-ending.
10169 2022-07-08 Martin Liska <mliska@suse.cz>
10171 PR sanitizer/106132
10172 * opts.cc (finish_options): Use 2 calls to
10173 report_conflicting_sanitizer_options.
10175 2022-07-08 Richard Biener <rguenther@suse.de>
10177 PR tree-optimization/106226
10178 * tree-vect-loop-manip.cc (vect_do_peeling): Assert that
10179 no SSA update is needed. Move virtual SSA update ...
10180 * tree-vectorizer.cc (pass_vectorize::execute): ... here,
10181 via forced virtual renaming when TODO_update_ssa_only_virtuals
10183 (vect_transform_loops): Return TODO_update_ssa_only_virtuals
10184 when virtual SSA update is required.
10185 (try_vectorize_loop_1): Adjust.
10186 * tree-vect-stmts.cc (vectorizable_simd_clone_call): Allow
10187 virtual renaming if the ABI forces an aggregate return
10188 but the original call did not have a virtual definition.
10190 2022-07-08 Martin Liska <mliska@suse.cz>
10192 * toplev.cc (init_asm_output): Do not init asm_out_file.
10194 2022-07-08 Tamar Christina <tamar.christina@arm.com>
10196 PR tree-optimization/106063
10197 * match.pd: Do not apply pattern after veclower is not supported.
10199 2022-07-08 Thomas Schwinge <thomas@codesourcery.com>
10201 * lto-cgraph.cc (input_offload_tables) <LTO_symtab_edge>: Correct
10204 2022-07-08 Tamar Christina <tamar.christina@arm.com>
10206 * expmed.cc (store_bit_field_1): Add parameter that indicates if value is
10207 still undefined and if so emit a subreg move instead.
10208 (store_integral_bit_field): Likewise.
10209 (store_bit_field): Likewise.
10210 * expr.h (write_complex_part): Likewise.
10211 * expmed.h (store_bit_field): Add new parameter.
10212 * builtins.cc (expand_ifn_atomic_compare_exchange_into_call): Use new
10214 (expand_ifn_atomic_compare_exchange): Likewise.
10215 * calls.cc (store_unaligned_arguments_into_pseudos): Likewise.
10216 * emit-rtl.cc (validate_subreg): Likewise.
10217 * expr.cc (emit_group_store): Likewise.
10218 (copy_blkmode_from_reg): Likewise.
10219 (copy_blkmode_to_reg): Likewise.
10220 (clear_storage_hints): Likewise.
10221 (write_complex_part): Likewise.
10222 (emit_move_complex_parts): Likewise.
10223 (expand_assignment): Likewise.
10224 (store_expr): Likewise.
10225 (store_field): Likewise.
10226 (expand_expr_real_2): Likewise.
10227 * ifcvt.cc (noce_emit_move_insn): Likewise.
10228 * internal-fn.cc (expand_arith_set_overflow): Likewise.
10229 (expand_arith_overflow_result_store): Likewise.
10230 (expand_addsub_overflow): Likewise.
10231 (expand_neg_overflow): Likewise.
10232 (expand_mul_overflow): Likewise.
10233 (expand_arith_overflow): Likewise.
10235 2022-07-08 Haochen Jiang <haochen.jiang@intel.com>
10238 * config/i386/sse.md (sse2_cvtps2pd<mask_name>_1):
10239 Rename from *sse2_cvtps2pd<mask_name>_1.
10240 (vec_unpacks_lo_v4sf): Add handler for memory operand.
10242 2022-07-08 Lulu Cheng <chenglulu@loongson.cn>
10244 * config/loongarch/loongarch.cc (loongarch_compute_frame_info):
10245 Modify fp_sp_offset and gp_sp_offset's calculation method,
10246 when frame->mask or frame->fmask is zero, don't minus UNITS_PER_WORD
10247 or UNITS_PER_FP_REG.
10249 2022-07-07 David Malcolm <dmalcolm@redhat.com>
10251 * diagnostic-format-json.cc (json_from_location_range): Update for
10252 removal of label_text::maybe_free in favor of automatic memory
10254 * diagnostic-format-sarif.cc
10255 (sarif_builder::make_location_object): Likewise.
10256 * diagnostic-show-locus.cc (struct pod_label_text): New.
10257 (class line_label): Convert m_text from label_text to pod_label_text.
10258 (layout::print_any_labels): Move "text" to the line_label.
10259 * tree-diagnostic-path.cc (path_label::get_text): Update for
10260 removal of label_text::maybe_free in favor of automatic memory
10262 (event_range::print): Likewise.
10263 (default_tree_diagnostic_path_printer): Likewise.
10264 (default_tree_make_json_for_path): Likewise.
10266 2022-07-07 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
10268 * statistics.cc (get_function_name): Add check to see if fn is not NULL.
10270 2022-07-07 Richard Biener <rguenther@suse.de>
10272 * tree-into-ssa.cc (iterating_old_ssa_names): New.
10273 (add_new_name_mapping): Grow {new,old}_ssa_names separately
10274 and only when actually needed. Assert we are not growing
10275 the old_ssa_names set when iterating over it.
10276 (update_ssa): Remove old_ssa_names copying and empty_p
10277 query, note we are iterating over it and expect no set changes.
10279 2022-07-07 Thomas Schwinge <thomas@codesourcery.com>
10281 * config/i386/intelmic-mkoffload.cc (generate_host_descr_file)
10282 (prepare_target_image, main): Handle OpenMP 'requires'.
10283 (generate_host_descr_file): Switch to 'GOMP_offload_register_ver',
10284 'GOMP_offload_unregister_ver'.
10286 2022-07-07 Richard Biener <rguenther@suse.de>
10289 * config/i386/i386-builtins.cc (ix86_add_new_builtins): Properly
10292 2022-07-07 Richard Biener <rguenther@suse.de>
10294 * tree-ssa-loop-manip.cc (add_exit_phis_var): Return the
10295 number of LC PHIs inserted.
10296 (add_exit_phis): Return whether any variable required
10297 multiple LC PHI nodes.
10298 (rewrite_into_loop_closed_ssa_1): Use TODO_update_ssa_no_phi
10301 2022-07-07 Richard Biener <rguenther@suse.de>
10303 * tree-ssa-loop-manip.cc (compute_live_loop_exits): Take
10304 the def loop exit block bitmap as argument instead of
10305 re-computing it here.
10306 (add_exit_phis_var): Adjust.
10307 (loop_name_cmp): New function.
10308 (add_exit_phis): Sort variables to insert LC PHI nodes
10309 after definition loop, for each definition loop compute
10310 the exit block bitmap once.
10311 (get_loops_exit): Remove.
10312 (rewrite_into_loop_closed_ssa_1): Do not pre-record
10313 all loop exit blocks into bitmaps. Record loop exits
10316 2022-07-07 Dimitrije Milosevic <Dimitrije.Milosevic@Syrmia.com>
10318 * config/mips/mips.cc (mips_asan_shadow_offset): Reformat
10319 to handle the N32 ABI.
10320 * config/mips/mips.h (SUBTARGET_SHADOW_OFFSET): Remove
10321 the macro, as it is not needed anymore.
10323 2022-07-06 Thomas Schwinge <thomas@codesourcery.com>
10325 * config/gcn/mkoffload.cc (process_obj): Clarify 'target_data' ->
10327 * config/nvptx/mkoffload.cc (process): Likewise.
10329 2022-07-06 Thomas Schwinge <thomas@codesourcery.com>
10331 * omp-general.h (enum omp_requires): Use 'GOMP_REQUIRES_[...]'.
10333 2022-07-05 Andrew MacLeod <amacleod@redhat.com>
10335 * value-relation.cc (relation_to_code): New vector.
10336 (relation_oracle::validate_relation): New.
10337 (set_relation): Allow ssa1 == ssa2 to be registered.
10338 * value-relation.h (validate_relation): New prototype.
10339 (query_relation): Make internal variant protected.
10341 2022-07-05 Roger Sayle <roger@nextmovesoftware.com>
10343 * config/i386/i386.cc (ix86_rtx_costs) <COMPARE>: Provide costs
10344 for double word comparisons and tests (comparisons against zero).
10345 * config/i386/i386.md (*test<mode>_not_doubleword): Split DWI
10346 and;cmp into andn;cmp $0 as a pre-reload splitter.
10347 (*andn<dwi>3_doubleword_bmi): Use <dwi> instead of <mode> in name.
10348 (*<any_or><dwi>3_doubleword): Likewise.
10350 2022-07-05 Roger Sayle <roger@nextmovesoftware.com>
10351 Hongtao Liu <hongtao.liu@intel.com>
10353 * config/i386/i386-builtin.def (__builtin_ia32_palignr128): Change
10354 CODE_FOR_ssse3_palignrti to CODE_FOR_ssse3_palignrv1ti.
10355 * config/i386/i386-expand.cc (expand_vec_perm_palignr): Use V1TImode
10356 and gen_ssse3_palignv1ti instead of TImode.
10357 * config/i386/sse.md (SSESCALARMODE): Delete.
10358 (define_mode_attr ssse3_avx2): Handle V1TImode instead of TImode.
10359 (<ssse3_avx2>_palignr<mode>): Use VIMAX_AVX2_AVX512BW as a mode
10360 iterator instead of SSESCALARMODE.
10361 (ssse3_palignrdi): Optimize cases where operands[3] is 0 or 64,
10362 using a single move instruction (if required).
10364 2022-07-05 Roger Sayle <roger@nextmovesoftware.com>
10365 Uroš Bizjak <ubizjak@gmail.com>
10367 PR rtl-optimization/96692
10368 * config/i386/i386.md (define_split): Split ((A | B) ^ C) ^ D
10369 as (X & ~Y) ^ Z on target BMI when either C or D is A or B.
10371 2022-07-05 Richard Biener <rguenther@suse.de>
10373 PR tree-optimization/106198
10374 * tree-cfgcleanup.cc (repair_loop_structures): Always do a
10375 full LC SSA rewrite but only if any blocks changed loop
10378 2022-07-05 Richard Biener <rguenther@suse.de>
10380 * tree-ssa-loop-manip.cc (find_uses_to_rename_def): Remove.
10381 (find_uses_to_rename_in_loop): Likewise.
10382 (rewrite_into_loop_closed_ssa_1): Remove loop parameter and
10384 (rewrite_into_loop_closed_ssa): Adjust.
10386 2022-07-05 Richard Biener <rguenther@suse.de>
10388 PR tree-optimization/106186
10389 * tree-ssa-propagate.cc (clean_up_loop_closed_phi):
10390 Properly handle virtual PHI nodes.
10392 2022-07-05 Richard Biener <rguenther@suse.de>
10394 PR tree-optimization/106196
10395 * tree-vect-stmts.cc (vect_finish_stmt_generation): Properly
10396 handle aggregate returns of calls for VDEF updates.
10398 2022-07-05 Richard Biener <rguenther@suse.de>
10400 * tree-vect-loop-manip.cc (vect_set_loop_condition_normal):
10403 2022-07-05 Richard Sandiford <richard.sandiford@arm.com>
10405 * tree-vect-patterns.cc (vect_convert_input): Expect the input
10406 type to be signed for optab_vector_mixed_sign. Update the vectype
10407 at the same time as type.
10408 (vect_recog_dot_prod_pattern): Update accordingly. If usdot isn't
10409 available, try sdot instead.
10410 * tree-vect-loop.cc (vect_is_emulated_mixed_dot_prod): New function.
10411 (vect_model_reduction_cost): Model the cost of implementing usdot
10413 (vectorizable_reduction): Likewise. Skip target support test
10414 for lane reductions.
10415 (vect_emulate_mixed_dot_prod): New function.
10416 (vect_transform_reduction): Use it to emulate usdot via sdot.
10418 2022-07-05 Richard Biener <rguenther@suse.de>
10420 PR tree-optimization/106182
10421 * loop-init.cc (fix_loop_structure): Return the number
10422 of newly discovered plus the number of deleted loops.
10423 * tree-cfgcleanup.cc (repair_loop_structures): Adjust
10426 2022-07-04 Aldy Hernandez <aldyh@redhat.com>
10428 * gimple-range-fold.cc
10429 (fold_using_range::range_of_ssa_name_with_loop_info): Restrict the
10430 call to SCEV for irange supported types.
10431 (fold_using_range::range_of_builtin_int_call): Convert to vrange.
10432 * gimple-range.cc (gimple_ranger::prefill_stmt_dependencies): Same.
10433 * tree-ssa-dom.cc (cprop_operand): Same.
10435 2022-07-04 Roger Sayle <roger@nextmovesoftware.com>
10438 * calls.cc (precompute_register_parameters): Allow promotion
10439 of floating point values to be passed in wider integer modes
10440 by calling new convert_float_to_wider_int.
10441 (expand_call): Allow floating point results to be returned in
10442 wider integer modes by calling new convert wider_int_to_float.
10443 * cfgexpand.cc (expand_value_return): Allow backends to promote
10444 a scalar floating point return value to a wider integer mode
10445 by calling new convert_float_to_wider_int.
10446 * expr.cc (convert_float_to_wider_int): New function.
10447 (convert_wider_int_to_float): Likewise.
10448 (expand_expr_real_1) <expand_decl_rtl>: Allow backends to promote
10449 scalar FP PARM_DECLs to wider integer modes, by calling new
10450 convert_wider_int_to_float.
10451 * expr.h (convert_modes): Name arguments for improved documentation.
10452 (convert_float_to_wider_int): Prototype new function here.
10453 (convert_wider_int_to_float): Likewise.
10454 * function.cc (assign_parm_setup_stack): Allow floating point
10455 values to be passed on the stack as wider integer modes by
10456 calling new convert_wider_int_to_float.
10458 2022-07-04 Martin Jambor <mjambor@suse.cz>
10460 PR tree-optimization/105860
10461 * tree-sra.cc (build_reconstructed_reference): Start expr
10462 traversal only just below the outermost union.
10464 2022-07-04 Richard Biener <rguenther@suse.de>
10466 * tree-vect-loop-manip.cc (vect_do_peeling): Revert assert
10467 and update virtual SSA form again. Assert we do so for
10468 a known set of transforms only.
10469 * tree-vectorizer.h (vec_info::any_known_not_updated_vssa): New.
10470 * tree-vect-stmts.cc (vectorizable_load): When vectorizing
10471 using load-lanes allow virtual SSA update.
10473 2022-07-04 Tobias Burnus <tobias@codesourcery.com>
10474 Chung-Lin Tang <cltang@codesourcery.com>
10475 Thomas Schwinge <thomas@codesourcery.com>
10477 * config/gcn/mkoffload.cc (process_asm): Write '#include <stdint.h>'.
10478 (process_obj): Pass omp_requires_mask to GOMP_offload_register_ver.
10479 (main): Ask lto1 to obtain omp_requires_mask and pass it on.
10480 * config/nvptx/mkoffload.cc (process, main): Likewise.
10481 * lto-cgraph.cc (omp_requires_to_name): New.
10482 (input_offload_tables): Save omp_requires_mask.
10483 (output_offload_tables): Read it, check for consistency,
10484 save value for mkoffload.
10485 * omp-low.cc (lower_omp_target): Force output_offloadtables
10486 call for OMP_REQUIRES_TARGET_USED.
10488 2022-07-04 Richard Biener <rguenther@suse.de>
10490 * tree-vect-loop-manip.cc (vect_do_peeling): Assert that
10491 no SSA update is needed instead of updating virtual SSA
10493 * tree-vect-stmts.cc (vectorizable_load): For hoisted
10494 invariant load use the loop entry virtual use.
10495 For emulated gather loads use the virtual use of the
10496 original stmt like vect_finish_stmt_generation would do.
10498 2022-07-04 Eric Botcazou <ebotcazou@adacore.com>
10500 * tree-pretty-print.cc (dump_generic_node) <ARRAY_TYPE>: Add guard
10501 for direct circularity.
10503 2022-07-04 Eric Botcazou <ebotcazou@adacore.com>
10505 * dwarf2out.cc (gen_array_type_die): Use the default lower bound of
10506 the language for vector types.
10508 2022-07-04 Aldy Hernandez <aldyh@redhat.com>
10510 * value-range-storage.cc (irange_storage_slot::set_irange): Set
10511 nonzero bits in irange.
10512 (irange_storage_slot::get_irange): Get nonzero bits from irange.
10513 * value-range.cc (irange::operator=): Set nonzero bits.
10514 (irange::irange_set): Same.
10515 (irange::irange_set_anti_range): Same.
10516 (irange::set): Same.
10517 (irange::verify_range): Same.
10518 (irange::legacy_equal_p): Check nonzero bits.
10519 (irange::equal_p): Same.
10520 (irange::contains_p): Handle nonzero bits.
10521 (irange::irange_union): Same.
10522 (irange::irange_intersect): Same.
10523 (irange::dump): Same.
10524 (irange::set_nonzero_bits): New.
10525 (irange::get_nonzero_bits): New.
10526 (irange::intersect_nonzero_bits): New.
10527 (irange::union_nonzero_bits): New.
10528 (irange::dump_bitmasks): New.
10529 * value-range.h (class irange): Add m_nonzero_mask.
10530 (gt_ggc_mx): Handle nonzero bits.
10532 (irange::set_undefined): Set nonzero bits.
10533 (irange::set_varying): Same.
10534 (irange::normalize_kind): Call set_undefined.
10536 2022-07-04 Richard Biener <rguenther@suse.de>
10538 * tree-ssa-loop-manip.h
10539 (rewrite_virtuals_into_loop_closed_ssa): Remove.
10540 (rewrite_into_loop_closed_ssa_1): Likewise.
10541 * tree-ssa-loop-manip.cc (rewrite_into_loop_closed_ssa_1):
10543 (rewrite_into_loop_closed_ssa): Remove loop overload,
10544 always rewrite virtuals into LC SSA.
10545 (check_loop_closed_ssa_bb): Also check virtuals.
10546 * tree-ssa-dce.cc (remove_dead_phis): Preserve virtual
10547 LC PHIs when in LC SSA.
10548 * gimple-loop-jam.cc (fuse_loops): Do not rewrite into
10549 loop-closed SSA here, but ...
10550 (tree_loop_unroll_and_jam): ... here once.
10551 * tree-if-conv.cc (version_loop_for_if_conversion): Use
10552 the cheaper TODO_update_ssa_no_phi.
10553 * tree-loop-distribution.cc (version_loop_by_alias_check):
10555 * tree-ssa-loop-unswitch.cc (tree_unswitch_single_loop):
10557 * tree-vect-loop-manip.cc (vect_loop_versioning): Likewise.
10558 (tree_unswitch_outer_loop): Do not rewrite virtuals into
10560 * tree-parloops.cc (transform_to_exit_first_loop_alt):
10562 (pass_parallelize_loops::execute): After finishing rewrite
10563 into LC SSA again because we do not maintain it properly.
10565 2022-07-04 Richard Biener <rguenther@suse.de>
10567 PR tree-optimization/106055
10568 * graphite.cc (canonicalize_loop_closed_ssa): Check whether
10571 2022-07-04 Haochen Jiang <haochen.jiang@intel.com>
10574 * config/i386/sse.md (extendv2sfv2df2): New define_expand.
10575 (sse2_cvtps2pd_<mask_name>): Change constraint of operands[1].
10576 (*sse2_cvtps2pd_<mask_name>_1): Rename from extendvsdfv2df2.
10578 2022-07-04 Aldy Hernandez <aldyh@redhat.com>
10580 * ipa-cp.cc (ipcp_vr_lattice::meet_with_1): Use operator!=.
10581 * ipa-prop.cc (struct ipa_vr_ggc_hash_traits): Same.
10582 * tree-ssa-loop-unswitch.cc (struct unswitch_predicate): Use set
10583 with two arguments.
10584 (find_unswitching_predicates_for_bb): Same.
10585 * tree-vrp.cc (range_fold_unary_symbolics_p): Same.
10586 * value-range-equiv.cc (value_range_equiv::equal_p): Use operator==.
10587 * value-range.cc (irange::equal_p): Rename to...
10588 (irange::operator==): ...this.
10589 * value-range.h (irange::set): Remove.
10590 (irange::operator==): Remove.
10591 (irange::set_zero): Use set with two arguments.
10592 * vr-values.cc (vr_values::extract_range_from_binary_expr): Same.
10593 (vr_values::extract_range_from_unary_expr): Same.
10594 (check_for_binary_op_overflow): Same.
10595 (bounds_of_var_in_loop): Same.
10597 2022-07-03 H.J. Lu <hjl.tools@gmail.com>
10600 * config/i386/i386-protos.h (ix86_convert_const_vector_to_integer):
10602 * config/i386/i386.cc (ix86_convert_const_vector_to_integer):
10604 * config/i386/mmx.md (V_16_32_64): New.
10605 (*mov<mode>_imm): New patterns for stores with 16-bit, 32-bit
10606 and 64-bit constant vector.
10607 * config/i386/predicates.md (x86_64_const_vector_operand): New.
10609 2022-07-03 Aldy Hernandez <aldyh@redhat.com>
10611 * gimple-range-cache.cc: Include value-range-storage.h.
10612 * gimple-range-cache.h (class block_range_cache): Add "class" to
10614 * gimple-range-edge.cc
10615 (gimple_outgoing_range::gimple_outgoing_range): Allocate allocator.
10616 (gimple_outgoing_range::~gimple_outgoing_range): Free allocator.
10617 (gimple_outgoing_range::calc_switch_ranges): Dereference allocator.
10618 * gimple-range-edge.h: Add "class" to m_range_allocator.
10619 * gimple-range-infer.cc
10620 (infer_range_manager::infer_range_manager): Allocate allocator.
10621 (infer_range_manager::~infer_range_manager): Free allocator.
10622 (infer_range_manager::get_nonzero): Dereference allocator.
10623 (infer_range_manager::add_range): Same.
10624 * gimple-range-infer.h (class vrange_allocator): Add "class" to
10626 * value-range-storage.h (class vrange_allocator): Move from
10628 (class obstack_vrange_allocator): Same.
10629 (class ggc_vrange_allocator): Same.
10630 (vrange_allocator::alloc_vrange): Same.
10631 (vrange_allocator::alloc_irange): Same.
10632 * value-range.h (class vrange_allocator): Move to value-range-storage.h.
10633 (class obstack_vrange_allocator): Same.
10634 (class ggc_vrange_allocator): Same.
10636 2022-07-03 Aldy Hernandez <aldyh@redhat.com>
10638 * Makefile.in (OBJS): Add value-range-storage.o.
10639 (GTFILES): Add value-range-storage.h.
10640 * gengtype.cc (open_base_files): Add value-range-storage.h.
10641 * value-range-storage.cc: New file.
10642 * value-range-storage.h: New file.
10644 2022-07-03 Xi Ruoyao <xry111@xry111.site>
10645 Lulu Cheng <chenglulu@loongson.cn>
10647 * config/loongarch/loongarch.cc (loongarch_check_zero_div_p):
10648 New static function.
10649 (loongarch_idiv_insns): Use loongarch_check_zero_div_p instead
10650 of TARGET_CHECK_ZERO_DIV.
10651 (loongarch_output_division): Likewise.
10652 * common/config/loongarch/loongarch-common.cc
10653 (TARGET_DEFAULT_TARGET_FLAGS): Remove unneeded hook.
10654 * doc/invoke.texi: Update to match the new behavior.
10656 2022-07-03 Ian Lance Taylor <iant@golang.org>
10658 * tree-ssa-dse.cc (dse_optimize_stmt): Only delete a trapping
10659 statement if -fdelete-dead-exceptions.
10661 2022-07-02 Tim Lange <mail@tim-lange.me>
10664 * doc/invoke.texi: Added Wanalyzer-allocation-size.
10666 2022-07-02 Immad Mir <mirimmad17@gmail.com>
10669 * Makefile.in (ANALYZER_OBJS): Add sm-fd.o.
10670 * doc/invoke.texi: Add -Wanalyzer-fd-double-close, -Wanalyzer-fd-leak,
10671 -Wanalyzer-fd-access-mode-mismatch, -Wanalyzer-fd-use-without-check,
10672 -Wanalyzer-fd-use-after-close.
10674 2022-07-01 Aldy Hernandez <aldyh@redhat.com>
10676 * wide-int.h (struct trailing_wide_ints): Add m_num_elements.
10677 (trailing_wide_ints::set_precision): Add num_elements argument.
10678 (trailing_wide_ints::extra_size): Same.
10680 2022-07-01 Uroš Bizjak <ubizjak@gmail.com>
10682 * config/i386/i386.md (*andn<mode>3_doubleword_bmi):
10683 Use "r" constraint for operand 1.
10685 2022-07-01 Richard Biener <rguenther@suse.de>
10687 * tree-into-ssa.cc (rewrite_update_dom_walker::before_dom_children):
10688 Do not look at interesting_blocks which is a copy of
10690 (update_ssa): Do not initialize it.
10691 (pass_build_ssa::execute): Set interesting_blocks to NULL
10692 after releasing it.
10694 2022-07-01 Richard Biener <rguenther@suse.de>
10696 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Revert
10697 back to using maybe_ne (off, -1).
10699 2022-07-01 Richard Biener <rguenther@suse.de>
10701 * tree-ssa-sccvn.cc (vn_nary_op_insert_into): Make
10702 checking dominance check conditional on flag_checking.
10704 2022-07-01 Tobias Burnus <tobias@codesourcery.com>
10706 * config/gcn/gcn-protos.h (print_operand_address): Remove register
10707 keyword on 'rtx addr' argument.
10709 2022-07-01 Roger Sayle <roger@nextmovesoftware.com>
10710 Uroš Bizjak <ubizjak@gmail.com>
10712 * config/i386/i386.md (*andn<mode>3_doubleword_bmi): Add constraints
10713 to post-reload define_insn_and_split.
10715 2022-07-01 Jakub Jelinek <jakub@redhat.com>
10717 PR middle-end/106144
10718 * wide-int.cc (wi::shifted_mask): If end >= prec, return right after
10719 emitting element for shift or if shift is 0 first element after start.
10720 (wide_int_cc_tests): Add tests for equivalency of wi::mask and
10721 wi::shifted_mask with 0 start.
10723 2022-07-01 Roger Sayle <roger@nextmovesoftware.com>
10726 * config/i386/i386.md (peephole2): Avoid generating pop %esp
10727 when optimizing for size.
10729 2022-07-01 Roger Sayle <roger@nextmovesoftware.com>
10730 Uroš Bizjak <ubizjak@gmail.com>
10732 * config/i386/i386.md (general_szext_operand): Add TImode
10733 support using x86_64_hilo_general_operand predicate.
10734 (*cmp<dwi>_doubleword): Use x86_64_hilo_general_operand predicate.
10735 (*add<dwi>3_doubleword): Improved optimization of zero addition.
10736 (and<mode>3): Use SDWIM mode iterator to add support for double
10737 word bit-wise AND in TImode. Use force_reg when double word
10738 immediate operand isn't x86_64_hilo_general_operand.
10739 (and<dwi>3_doubleword): Generalized from anddi3_doubleword and
10740 converted into a post-reload splitter.
10741 (*andndi3_doubleword): Previous define_insn deleted.
10742 (*andn<mode>3_doubleword_bmi): New define_insn_and_split for
10743 TARGET_BMI that splits post-reload.
10744 (*andn<mode>3_doubleword): New define_insn_and_split for
10745 !TARGET_BMI, that lowers/splits before reload.
10746 (<any_or><mode>3): Use SDWIM mode iterator to add suppport for
10747 double word bit-wise XOR and bit-wise IOR in TImode. Use
10748 force_reg when double word immediate operand isn't
10749 x86_64_hilo_general_operand.
10750 (*<any_or>di3_doubleword): Generalized from <any_or>di3_doubleword.
10751 (one_cmpl<mode>2): Use SDWIM mode iterator to add support for
10752 double word bit-wise NOT in TImode.
10753 (one_cmpl<dwi>2_doubleword): Generalize from one_cmpldi2_doubleword
10754 and converted into a post-reload splitter.
10756 2022-07-01 Eric Botcazou <ebotcazou@adacore.com>
10758 PR middle-end/105874
10759 * expr.cc (expand_expr_real_1) <normal_inner_ref>: Force
10760 EXPAND_MEMORY for the expansion of the inner reference only
10761 in the usual cases where a memory reference is required.
10763 2022-07-01 Richard Biener <rguenther@suse.de>
10765 PR tree-optimization/106131
10766 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Force alias-set
10767 zero when offsetting the read looking through an aggregate
10770 2022-07-01 Martin Liska <mliska@suse.cz>
10772 PR tree-optimization/106126
10773 * gimple-if-to-switch.cc (struct condition_info): Save
10775 (find_conditions): Parse all BBs.
10776 (pass_if_to_switch::execute): Allow only side effects for first
10779 2022-07-01 Haochen Jiang <haochen.jiang@intel.com>
10781 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_AVX512F_UNSET):
10782 Add OPTION_MASK_ISA2_AVX512BW_UNSET, remove
10783 OPTION_MASK_ISA2_AVX512BF16_UNSET and
10784 OPTION_MASK_ISA2_AVX512FP16_UNSET.
10786 2022-06-30 Joseph Myers <joseph@codesourcery.com>
10789 * lto-wrapper.cc (find_option): Add argument start.
10790 (merge_and_complain): Loop over existing_opt_index and
10791 existing_opt2_index for Xassembler check. Update calls to
10793 (find_and_merge_options): Add argument first to determine whether
10794 to merge options with those passed in *opts.
10795 (run_gcc): Update calls to find_and_merge_options.
10797 2022-06-30 Aldy Hernandez <aldyh@redhat.com>
10799 * gimple-range-cache.cc (block_range_cache::block_range_cache):
10800 Rename vrange_allocator to obstack_vrange_allocator.
10801 (ssa_global_cache::ssa_global_cache): Same.
10802 * gimple-range-edge.h (class gimple_outgoing_range): Same.
10803 * gimple-range-infer.h (class infer_range_manager): Same.
10804 * value-range.h (class vrange_allocator): Make abstract.
10805 (class obstack_vrange_allocator): Inherit from vrange_allocator.
10806 (class ggc_vrange_allocator): New.
10808 2022-06-30 Roger Sayle <roger@nextmovesoftware.com>
10809 Uroš Bizjak <ubizjak@gmail.com>
10811 * config/i386/i386.md (swap_mode): Rename from *swap<mode> to
10812 provide gen_swapsi.
10813 (<any_rotate>di3): Handle !TARGET_64BIT rotations by 32 bits
10814 via new gen_<insn>32di2_doubleword below.
10815 (<anyrotate>32di2_doubleword): New define_insn_and_split
10816 that splits after reload as either a pair of move instructions
10817 or an xchgl (using gen_swapsi).
10819 2022-06-30 Richard Biener <rguenther@suse.de>
10821 * domwalk.h (dom_walker::dom_walker): Update comment to
10822 reflect reality and new special argument value for
10824 * domwalk.cc (dom_walker::dom_walker): Recognize -1
10827 (rewrite_update_dom_walker::rewrite_update_dom_walker): Tell
10828 dom_walker to not use RPO.
10830 2022-06-30 Martin Liska <mliska@suse.cz>
10832 * tree-ssa-dom.cc (pass_dominator::execute): Remove m_ranger as
10835 2022-06-30 Andrew MacLeod <amacleod@redhat.com>
10837 PR tree-optimization/106114
10838 * gimple-range-fold.cc (fold_using_range::relation_fold_and_or): Check
10839 statement operands instead of GORI cache.
10841 2022-06-30 Antoni Boucher <bouanto@zoho.com>
10844 * config/i386/sse.md: Fix asm generation.
10846 2022-06-29 Sergei Trofimovich <siarheit@google.com>
10849 * system.h: Introduce INCLUDE_PTHREAD_H macros to include <pthread.h>.
10851 2022-06-29 Joseph Myers <joseph@codesourcery.com>
10853 * config/nios2/nios2.cc (nios2_load_pic_address): Use gen_rtx_MEM
10854 not gen_const_mem for UNSPEC_PIC_CALL_SYM.
10856 2022-06-29 Richard Biener <rguenther@suse.de>
10858 PR rtl-optimization/106082
10859 * combine.cc (distribute_notes): Preserve notes when
10860 they indicate a call doesn't perform a non-local goto.
10862 2022-06-29 Richard Biener <rguenther@suse.de>
10864 PR tree-optimization/106112
10865 * tree-ssa-sccvn.cc (valueized_wider_op): Properly extend
10866 a constant operand according to its type.
10868 2022-06-29 Martin Liska <mliska@suse.cz>
10870 * doc/invoke.texi: Remove removed evrp-mode.
10872 2022-06-29 Lulu Cheng <chenglulu@loongson.cn>
10875 * config/loongarch/loongarch.cc (loongarch_build_integer):
10876 Remove undefined behavior from code.
10878 2022-06-28 Dimitar Dimitrov <dimitar@dinux.eu>
10880 * doc/sourcebuild.texi: Document new no_alignment_constraints
10881 effective target check.
10883 2022-06-28 Aldy Hernandez <aldyh@redhat.com>
10885 * Makefile.in: Remove gimple-ssa-evrp.o and gimple-ssa-evrp-analyze.o.
10886 * flag-types.h (enum evrp_mode): Remove.
10887 * params.opt: Remove --param=evrp-mode.
10888 * tree-vrp.cc (make_pass_early_vrp): New.
10889 (pass_vrp::execute): Call early VRP instance.
10890 * gimple-ssa-evrp-analyze.cc: Removed.
10891 * gimple-ssa-evrp-analyze.h: Removed.
10892 * gimple-ssa-evrp.cc: Removed.
10894 2022-06-28 Alexandre Oliva <oliva@adacore.com>
10896 * Makefile.in (TFLAGS): New.
10897 (GCC_FOR_TARGET): Add TFLAGS.
10898 (FLAGS_TO_PASS): Pass TFLAGS down.
10900 2022-06-28 Richard Biener <rguenther@suse.de>
10902 * tree-ssa-loop-split.cc (fix_loop_bb_probability): Do not
10905 2022-06-28 Richard Biener <rguenther@suse.de>
10907 * tree-into-ssa.cc (insert_updated_phi_nodes_for): Use
10908 mark_block_for_update.
10909 (update_ssa): Adjust.
10911 2022-06-28 Xi Ruoyao <xry111@xry111.site>
10914 * config/loongarch/loongarch.h (REG_CLASS_CONTENTS): Exclude
10915 $r13 from SIBCALL_REGS.
10916 * config/loongarch/loongarch.cc (loongarch_regno_to_class):
10917 Change $r13 to JIRL_REGS.
10919 2022-06-28 Aldy Hernandez <aldyh@redhat.com>
10921 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Remove
10922 irange::supports_p.
10924 2022-06-28 Richard Biener <rguenther@suse.de>
10926 PR middle-end/106053
10927 * match.pd ((T)a == (T)b): Avoid folding away sign
10928 changes in a comparison if we'd truncate to a boolean.
10930 2022-06-28 Kewen Lin <linkw@linux.ibm.com>
10932 * config/rs6000/rs6000.md (*rotl<mode>3_insert_4): Replace mode
10933 iterator GPR with SImode, adjust the condition and output template,
10935 (*rotlsi3_insert_4): ... this.
10937 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10939 * genmatch.cc: Add "final" and "override" to various vfunc
10940 implementations, removing redundant "virtual" as appropriate.
10941 * gensupport.cc: Likewise.
10942 * gimple-range-cache.h: Likewise.
10943 * ipa-icf-gimple.h: Likewise.
10944 * ipa-icf.h: Likewise.
10945 * read-md.h: Likewise.
10946 * read-rtl-function.cc: Likewise.
10947 * tree-ssa-loop-ch.cc: Likewise.
10948 * tree-ssa-sccvn.cc: Likewise.
10950 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10952 * config/i386/i386-features.h: Add "final" and "override" to
10953 scalar_chain vfunc implementations as appropriate.
10955 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10957 * tree-switch-conversion.h: Add "final" and "override" to cluster
10958 vfunc implementations as appropriate.
10960 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10962 * expr.cc: Add "final" and "override" to op_by_pieces_d vfunc
10963 implementations as appropriate.
10965 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10967 * compare-elim.cc: Add "final" and "override" to dom_walker vfunc
10968 implementations, removing redundant "virtual" as appropriate.
10969 * gimple-ssa-strength-reduction.cc: Likewise.
10970 * ipa-prop.cc: Likewise.
10971 * rtl-ssa/blocks.cc: Likewise.
10972 * tree-into-ssa.cc: Likewise.
10973 * tree-ssa-dom.cc: Likewise.
10974 * tree-ssa-math-opts.cc: Likewise.
10975 * tree-ssa-phiopt.cc: Likewise.
10976 * tree-ssa-propagate.cc: Likewise.
10977 * tree-ssa-sccvn.cc: Likewise.
10978 * tree-ssa-strlen.cc: Likewise.
10979 * tree-ssa-uncprop.cc: Likewise.
10981 2022-06-27 David Malcolm <dmalcolm@redhat.com>
10983 * adjust-alignment.cc: Add "final" and "override" to opt_pass
10984 vfunc implementations, removing redundant "virtual" as
10986 * asan.cc: Likewise.
10987 * auto-inc-dec.cc: Likewise.
10988 * auto-profile.cc: Likewise.
10989 * bb-reorder.cc: Likewise.
10990 * cfgcleanup.cc: Likewise.
10991 * cfgexpand.cc: Likewise.
10992 * cfgrtl.cc: Likewise.
10993 * cgraphbuild.cc: Likewise.
10994 * combine-stack-adj.cc: Likewise.
10995 * combine.cc: Likewise.
10996 * compare-elim.cc: Likewise.
10997 * config/i386/i386-features.cc: Likewise.
10998 * coroutine-passes.cc: Likewise.
10999 * cprop.cc: Likewise.
11000 * cse.cc: Likewise.
11001 * dce.cc: Likewise.
11002 * df-core.cc: Likewise.
11003 * dse.cc: Likewise.
11004 * dwarf2cfi.cc: Likewise.
11005 * early-remat.cc: Likewise.
11006 * except.cc: Likewise.
11007 * final.cc: Likewise.
11008 * function.cc: Likewise.
11009 * fwprop.cc: Likewise.
11010 * gcse.cc: Likewise.
11011 * gimple-harden-conditionals.cc: Likewise.
11012 * gimple-if-to-switch.cc: Likewise.
11013 * gimple-isel.cc: Likewise.
11014 * gimple-laddress.cc: Likewise.
11015 * gimple-loop-interchange.cc: Likewise.
11016 * gimple-loop-jam.cc: Likewise.
11017 * gimple-loop-versioning.cc: Likewise.
11018 * gimple-low.cc: Likewise.
11019 * gimple-ssa-backprop.cc: Likewise.
11020 * gimple-ssa-evrp.cc: Likewise.
11021 * gimple-ssa-isolate-paths.cc: Likewise.
11022 * gimple-ssa-nonnull-compare.cc: Likewise.
11023 * gimple-ssa-split-paths.cc: Likewise.
11024 * gimple-ssa-store-merging.cc: Likewise.
11025 * gimple-ssa-strength-reduction.cc: Likewise.
11026 * gimple-ssa-warn-access.cc: Likewise.
11027 * gimple-ssa-warn-alloca.cc: Likewise.
11028 * gimple-ssa-warn-restrict.cc: Likewise.
11029 * gimple-warn-recursion.cc: Likewise.
11030 * graphite.cc: Likewise.
11031 * ifcvt.cc: Likewise.
11032 * init-regs.cc: Likewise.
11033 * ipa-comdats.cc: Likewise.
11034 * ipa-cp.cc: Likewise.
11035 * ipa-devirt.cc: Likewise.
11036 * ipa-fnsummary.cc: Likewise.
11037 * ipa-free-lang-data.cc: Likewise.
11038 * ipa-icf.cc: Likewise.
11039 * ipa-inline.cc: Likewise.
11040 * ipa-modref.cc: Likewise.
11041 * ipa-profile.cc: Likewise.
11042 * ipa-pure-const.cc: Likewise.
11043 * ipa-reference.cc: Likewise.
11044 * ipa-split.cc: Likewise.
11045 * ipa-sra.cc: Likewise.
11046 * ipa-visibility.cc: Likewise.
11047 * ipa.cc: Likewise.
11048 * ira.cc: Likewise.
11049 * jump.cc: Likewise.
11050 * loop-init.cc: Likewise.
11051 * lower-subreg.cc: Likewise.
11052 * mode-switching.cc: Likewise.
11053 * modulo-sched.cc: Likewise.
11054 * multiple_target.cc: Likewise.
11055 * omp-expand.cc: Likewise.
11056 * omp-low.cc: Likewise.
11057 * omp-oacc-kernels-decompose.cc: Likewise.
11058 * omp-oacc-neuter-broadcast.cc: Likewise.
11059 * omp-offload.cc: Likewise.
11060 * omp-simd-clone.cc: Likewise.
11061 * passes.cc: Likewise.
11062 * postreload-gcse.cc: Likewise.
11063 * postreload.cc: Likewise.
11064 * predict.cc: Likewise.
11065 * recog.cc: Likewise.
11066 * ree.cc: Likewise.
11067 * reg-stack.cc: Likewise.
11068 * regcprop.cc: Likewise.
11069 * reginfo.cc: Likewise.
11070 * regrename.cc: Likewise.
11071 * reorg.cc: Likewise.
11072 * sancov.cc: Likewise.
11073 * sanopt.cc: Likewise.
11074 * sched-rgn.cc: Likewise.
11075 * stack-ptr-mod.cc: Likewise.
11076 * store-motion.cc: Likewise.
11077 * tracer.cc: Likewise.
11078 * trans-mem.cc: Likewise.
11079 * tree-call-cdce.cc: Likewise.
11080 * tree-cfg.cc: Likewise.
11081 * tree-cfgcleanup.cc: Likewise.
11082 * tree-complex.cc: Likewise.
11083 * tree-eh.cc: Likewise.
11084 * tree-emutls.cc: Likewise.
11085 * tree-if-conv.cc: Likewise.
11086 * tree-into-ssa.cc: Likewise.
11087 * tree-loop-distribution.cc: Likewise.
11088 * tree-nrv.cc: Likewise.
11089 * tree-object-size.cc: Likewise.
11090 * tree-parloops.cc: Likewise.
11091 * tree-predcom.cc: Likewise.
11092 * tree-profile.cc: Likewise.
11093 * tree-sra.cc: Likewise.
11094 * tree-ssa-ccp.cc: Likewise.
11095 * tree-ssa-copy.cc: Likewise.
11096 * tree-ssa-dce.cc: Likewise.
11097 * tree-ssa-dom.cc: Likewise.
11098 * tree-ssa-dse.cc: Likewise.
11099 * tree-ssa-forwprop.cc: Likewise.
11100 * tree-ssa-ifcombine.cc: Likewise.
11101 * tree-ssa-loop-ch.cc: Likewise.
11102 * tree-ssa-loop-im.cc: Likewise.
11103 * tree-ssa-loop-ivcanon.cc: Likewise.
11104 * tree-ssa-loop-prefetch.cc: Likewise.
11105 * tree-ssa-loop-split.cc: Likewise.
11106 * tree-ssa-loop-unswitch.cc: Likewise.
11107 * tree-ssa-loop.cc: Likewise.
11108 * tree-ssa-math-opts.cc: Likewise.
11109 * tree-ssa-phiopt.cc: Likewise.
11110 * tree-ssa-phiprop.cc: Likewise.
11111 * tree-ssa-pre.cc: Likewise.
11112 * tree-ssa-reassoc.cc: Likewise.
11113 * tree-ssa-sccvn.cc: Likewise.
11114 * tree-ssa-sink.cc: Likewise.
11115 * tree-ssa-strlen.cc: Likewise.
11116 * tree-ssa-structalias.cc: Likewise.
11117 * tree-ssa-uncprop.cc: Likewise.
11118 * tree-ssa-uninit.cc: Likewise.
11119 * tree-ssanames.cc: Likewise.
11120 * tree-stdarg.cc: Likewise.
11121 * tree-switch-conversion.cc: Likewise.
11122 * tree-tailcall.cc: Likewise.
11123 * tree-vect-generic.cc: Likewise.
11124 * tree-vectorizer.cc: Likewise.
11125 * tree-vrp.cc: Likewise.
11126 * tsan.cc: Likewise.
11127 * ubsan.cc: Likewise.
11128 * var-tracking.cc: Likewise.
11129 * vtable-verify.cc: Likewise.
11130 * web.cc: Likewise.
11132 2022-06-27 Andrew Stubbs <ams@codesourcery.com>
11134 * config/gcn/gcn.md (*movbi): Remove assembler bug workarounds.
11136 (movdi_symbol_save_scc): Likewise.
11138 2022-06-27 Roger Sayle <roger@nextmovesoftware.com>
11140 PR rtl-optimization/7061
11141 * config/i386/i386.md (*highpartdisi2): New define_insn_and_split.
11143 2022-06-27 Roger Sayle <roger@nextmovesoftware.com>
11145 PR tree-optimization/94026
11146 * match.pd (((X << C1) & C2) eq/ne C3): New simplification.
11147 (((X >> C1) & C2) eq/ne C3): Likewise.
11149 2022-06-27 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
11151 * statistics.cc: Include tree.h.
11152 (get_function_name): New function.
11153 (statistics_fini_pass_2): Call get_function_name instead of
11154 current_function_name.
11155 (statistics_counter_event): Call get_function_name instead of
11157 (statistics_histogram_event): Likewise.
11159 2022-06-27 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11161 * config/xtensa/xtensa.md:
11162 Suppress unnecessary emitting nop insn in the split patterns for
11163 integer/FP constant synthesis, and add new peephole2 pattern that
11164 folds such synthesized additions.
11166 2022-06-25 Aldy Hernandez <aldyh@redhat.com>
11168 * tree-ssa-dom.cc (dom_jt_state): Pass ranger to constructor
11170 (dom_jt_state::push): Remove m_evrp.
11171 (dom_jt_state::pop): Same.
11172 (dom_jt_state::record_ranges_from_stmt): Remove.
11173 (dom_jt_state::register_equiv): Remove updating of evrp ranges.
11174 (class dom_jt_simplifier): Pass ranger to constructor.
11175 Inherit from hybrid_jt_simplifier.
11176 (dom_jt_simplifier::simplify): Convert to ranger.
11177 (pass_dominator::execute): Same.
11178 (all_uses_feed_or_dominated_by_stmt): New.
11179 (dom_opt_dom_walker::set_global_ranges_from_unreachable_edges): New.
11180 (dom_opt_dom_walker::before_dom_children): Call
11181 set_global_ranges_from_unreachable_edges.
11182 Do not call record_ranges_from_stmt.
11183 (dom_opt_dom_walker::after_dom_children): Remove evrp use.
11184 (cprop_operand): Use int_range<> instead of value_range.
11185 (dom_opt_dom_walker::fold_cond): New.
11186 (dom_opt_dom_walker::optimize_stmt): Pass ranger to
11188 Use fold_cond() instead of vrp_visit_cond_stmt().
11189 * tree-ssa-threadedge.cc (jt_state::register_equivs_stmt): Do not
11190 pass state to simplifier.
11191 * vr-values.h (class vr_values): Make fold_cond public.
11193 2022-06-25 Jeff Law <jeffreyalaw@gmail.com>
11195 * common/config/tilegx/tilegx-common.cc: Removed.
11196 * common/config/tilepro/tilepro-common.cc: Removed.
11197 * config.gcc: Remove tilegx and tilepro entries.
11198 * config/tilegx/constraints.md: Removed.
11199 * config/tilegx/feedback.h: Removed.
11200 * config/tilegx/linux.h: Removed.
11201 * config/tilegx/mul-tables.cc: Removed.
11202 * config/tilegx/predicates.md: Removed.
11203 * config/tilegx/sync.md: Removed.
11204 * config/tilegx/t-tilegx: Removed.
11205 * config/tilegx/tilegx-builtins.h: Removed.
11206 * config/tilegx/tilegx-c.cc: Removed.
11207 * config/tilegx/tilegx-generic.md: Removed.
11208 * config/tilegx/tilegx-modes.def: Removed.
11209 * config/tilegx/tilegx-multiply.h: Removed.
11210 * config/tilegx/tilegx-opts.h: Removed.
11211 * config/tilegx/tilegx-protos.h: Removed.
11212 * config/tilegx/tilegx.cc: Removed.
11213 * config/tilegx/tilegx.h: Removed.
11214 * config/tilegx/tilegx.md: Removed.
11215 * config/tilegx/tilegx.opt: Removed.
11216 * config/tilepro/constraints.md: Removed.
11217 * config/tilepro/feedback.h: Removed.
11218 * config/tilepro/gen-mul-tables.cc: Removed.
11219 * config/tilepro/linux.h: Removed.
11220 * config/tilepro/mul-tables.cc: Removed.
11221 * config/tilepro/predicates.md: Removed.
11222 * config/tilepro/t-tilepro: Removed.
11223 * config/tilepro/tilepro-builtins.h: Removed.
11224 * config/tilepro/tilepro-c.cc: Removed.
11225 * config/tilepro/tilepro-generic.md: Removed.
11226 * config/tilepro/tilepro-modes.def: Removed.
11227 * config/tilepro/tilepro-multiply.h: Removed.
11228 * config/tilepro/tilepro-protos.h: Removed.
11229 * config/tilepro/tilepro.cc: Removed.
11230 * config/tilepro/tilepro.h: Removed.
11231 * config/tilepro/tilepro.md: Removed.
11232 * config/tilepro/tilepro.opt: Removed.
11233 * configure.ac: Remove tilegx and tilepro entries.
11234 * configure: Rebuilt.
11235 * doc/extend.texi: Remove tilegx and tilepro entries.
11236 * doc/install.texi: Remove tilegx and tilepro entries.
11237 * doc/invoke.texi: Remove tilegx and tilepro entries.
11238 * doc/md.texi: Remove tilegx and tilepro entries.
11240 2022-06-25 Roger Sayle <roger@nextmovesoftware.com>
11241 Richard Biener <rguenther@suse.de>
11243 * regcprop.cc (pass_cprop_hardreg::execute): Perform a third
11244 iteration over each basic block that was updated by the second
11247 2022-06-24 Jason Merrill <jason@redhat.com>
11251 * doc/invoke.texi: Document changes.
11253 2022-06-24 Iain Buclaw <ibuclaw@gdcproject.org>
11255 * config/tilepro/gen-mul-tables.cc (tilegx_emit): Adjust loop
11256 condition to avoid overflow.
11258 2022-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
11260 * config/rs6000/aix71.h (MASK_PPC_GPOPT, MASK_PPC_GFXOPT): Replace with
11261 OPTION_MASK_PPC_GPOPT, OPTION_MASK_PPC_GFXOPT.
11262 * config/rs6000/darwin.h (MASK_PPC_GFXOPT): Replace with
11263 OPTION_MASK_PPC_GFXOPT.
11264 * config/rs6000/darwin64-biarch.h (MASK_PPC_GFXOPT): Same.
11265 * config/rs6000/default64.h (MASK_PPC_GPOPT, MASK_PPC_GFXOPT): Replace with
11266 OPTION_MASK_PPC_GPOPT, OPTION_MASK_PPC_GFXOPT.
11267 * config/rs6000/rs6000-c.cc: Update comment.
11268 * config/rs6000/rs6000-cpus.def: Update RS6000_CPU macro calls.
11269 * config/rs6000/rs6000.cc (rs6000_darwin_file_start): Replace
11270 MASK_PPC_GPOPT with OPTION_MASK_PPC_GPOPT.
11271 (rs6000_builtin_mask_names): Replace MASK_PPC_GFXOPT, MASK_POPCNTB
11272 with OPTION_MASK_PPC_GFXOPT, OPTION_MASK_POPCNTB.
11273 * config/rs6000/rs6000.h: (MASK_P8_VECTOR, MASK_P9_VECTOR,
11274 MASK_P9_MISC, MASK_POPCNTB, MASK_POPCNTD, MASK_PPC_GFXOPT,
11275 MASK_PPC_GPOPT, MASK_RECIP_PRECISION, MASK_SOFT_FLOAT,
11276 MASK_VSX, MASK_POWER10, MASK_P10_FUSION): Delete.
11278 2022-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
11280 * config/rs6000/aix71.h (TARGET_DEFAULT): Replace MASK_MFCRF with
11282 * config/rs6000/darwin.h (TARGET_DEFAULT): Replace MASK_MULTIPLE with
11283 OPTION_MASK_MULTIPLE.
11284 * config/rs6000/darwin64-biarch.h (TARGET_DEFAULT): Same.
11285 * config/rs6000/default64.h (TARGET_DEFAULT): Replace MASK_MFCRF with
11287 * config/rs6000/eabi.h (TARGET_DEFAULT): Replace MASK_EABI with
11289 * config/rs6000/eabialtivec.h (TARGET_DEFAULT): Same.
11290 * config/rs6000/linuxaltivec.h (TARGET_DEFAULT): Replace
11291 MASK_ALTIVEC with OPTION_MASK_ALTIVEC.
11292 * config/rs6000/rs6000-cpus.def (MASK_ALTIVEC, MASK_CMPB,
11293 MASK_CRYPTO, MASK_DFP, MASK_DIRECT_MOVE, MASK_DLMZB, MASK_EABI,
11294 MASK_FLOAT128_KEYWORD, MASK_FLOAT128_HW, MASK_FPRND,
11295 MASK_P8_FUSION, MASK_HARD_FLOAT, MASK_HTM, MASK_ISEL, MASK_MFCRF,
11296 MASK_MMA, MASK_MULHW, MASK_MULTIPLE, MASK_NO_UPDATE):
11298 OPTION_MASK_ALTIVEC, OPTION_MASK_CMPB, OPTION_MASK_CRYPTO,
11299 OPTION_MASK_DFP, OPTION_MASK_DIRECT_MOVE, OPTION_MASK_DLMZB,
11300 OPTION_MASK_EABI, OPTION_MASK_FLOAT128_KEYWORD,
11301 OPTION_MASK_FLOAT128_HW, OPTION_MASK_FPRND, OPTION_MASK_P8_FUSION,
11302 OPTION_MASK_HARD_FLOAT, OPTION_MASK_HTM, OPTION_MASK_ISEL,
11303 OPTION_MASK_MFCRF, OPTION_MASK_MMA, OPTION_MASK_MULHW,
11304 OPTION_MASK_MULTIPLE, OPTION_MASK_NO_UPDATE.
11305 * config/rs6000/rs6000.cc (rs6000_darwin_file_start): Replace
11306 MASK_MFCRF, MASK_ALTIVEC with OPTION_MASK_MFCRF, OPTION_MASK_ALTIVEC.
11307 * config/rs6000/rs6000.h (TARGET_DEFAULT): Replace MASK_MULTIPLE
11308 with OPTION_MASK_MULTIPLE.
11309 (MASK_ALTIVEC, MASK_CMPB, MASK_CRYPTO, MASK_DFP,
11310 MASK_DIRECT_MOVE, MASK_DLMZB, MASK_EABI, MASK_FLOAT128_KEYWORD,
11311 MASK_FLOAT128_HW, MASK_FPRND, MASK_P8_FUSION, MASK_HARD_FLOAT,
11312 MASK_HTM, MASK_ISEL, MASK_MFCRF, MASK_MMA, MASK_MULHW,
11313 MASK_MULTIPLE, MASK_NO_UPDATE): Delete.
11314 * config/rs6000/vxworks.h (TARGET_DEFAULT): Replace MASK_EABI
11315 with OPTION_MASK_EABI.
11317 2022-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
11319 * config/rs6000/rs6000.cc (RS6000_BTM_ALTIVEC, RS6000_BTM_CMPB,
11320 RS6000_BTM_VSX, RS6000_BTM_FRE, RS6000_BTM_P8_VECTOR,
11321 RS6000_BTM_P9_VECTOR, RS6000_BTM_P9_MISC, RS6000_BTM_MODULO,
11322 RS6000_BTM_CRYPTO, RS6000_BTM_HTM): Replace with OPTION_MASK_ALTIVEC,
11323 OPTION_MASK_CMPB, OPTION_MASK_VSX, OPTION_MASK_POPCNTB,
11324 OPTION_MASK_P8_VECTOR, OPTION_MASK_P9_VECTOR, OPTION_MASK_P9_MISC,
11325 OPTION_MASK_MODULO, OPTION_MASK_CRYPTO, OPTION_MASK_HTM.
11326 * config/rs6000/rs6000.h (RS6000_BTM_MODULO, RS6000_BTM_ALTIVEC,
11327 RS6000_BTM_CMPB, RS6000_BTM_VSX, RS6000_BTM_P8_VECTOR,
11328 RS6000_BTM_P9_VECTOR, RS6000_BTM_P9_MISC, RS6000_BTM_CRYPTO,
11329 RS6000_BTM_HTM, RS6000_BTM_FRE): Remove.
11331 2022-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
11333 * config/rs6000/rs6000-c.cc: Update comments.
11334 * config/rs6000/rs6000.cc (RS6000_BTM_FRES, RS6000_BTM_FRSQRTE,
11335 RS6000_BTM_FRSQRTES, RS6000_BTM_POPCNTD, RS6000_BTM_CELL,
11336 RS6000_BTM_64BIT, RS6000_BTM_POWERPC64, RS6000_BTM_DFP,
11337 RS6000_BTM_HARD_FLOAT,RS6000_BTM_LDBL128, RS6000_BTM_FLOAT128,
11338 RS6000_BTM_FLOAT128_HW, RS6000_BTM_MMA, RS6000_BTM_P10): Replace
11339 with OPTION_MASK_PPC_GFXOPT, OPTION_MASK_PPC_GFXOPT,
11340 OPTION_MASK_POPCNTB, OPTION_MASK_POPCNTD,
11341 OPTION_MASK_FPRND, MASK_64BIT, MASK_POWERPC64,
11342 OPTION_MASK_DFP, OPTION_MASK_SOFT_FLOAT, OPTION_MASK_MULTIPLE,
11343 OPTION_MASK_FLOAT128_KEYWORD, OPTION_MASK_FLOAT128_HW,
11344 OPTION_MASK_MMA, OPTION_MASK_POWER10.
11345 * config/rs6000/rs6000.h (RS6000_BTM_FRES, RS6000_BTM_FRSQRTE,
11346 RS6000_BTM_FRSQRTES, RS6000_BTM_POPCNTD, RS6000_BTM_CELL,
11347 RS6000_BTM_DFP, RS6000_BTM_HARD_FLOAT, RS6000_BTM_LDBL128,
11348 RS6000_BTM_64BIT, RS6000_BTM_POWERPC64, RS6000_BTM_FLOAT128,
11349 RS6000_BTM_FLOAT128_HW, RS6000_BTM_MMA, RS6000_BTM_P10): Delete.
11351 2022-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
11353 * config/rs6000/rs6000.h (RS6000_BTM_COMMON, RS6000_BTM_ALWAYS,
11354 MASK_REGNAMES, OPTION_MASK_REGNAMES, MASK_PROTOTYPE,
11355 OPTION_MASK_PROTOTYPE, MASK_UPDATE, OPTION_MASK_UPDATE): Remove.
11357 2022-06-24 Richard Biener <rguenther@suse.de>
11359 PR middle-end/106070
11360 * match.pd (a != b ? a : b): Fix translation of
11361 operand_equal_for_comparison_p.
11363 2022-06-24 Jan Hubicka <jh@suse.cz>
11366 * tree-ssa-alias.cc (stmt_kills_ref_p): Check for external throw.
11368 2022-06-24 Martin Liska <mliska@suse.cz>
11370 PR middle-end/106059
11371 * profile-count.h: *= and /= operators need to modify this
11374 2022-06-24 Roger Sayle <roger@nextmovesoftware.com>
11375 Uroš Bizjak <ubizjak@gmail.com>
11378 * config/i386/i386.md (*<any_or>di3_doubleword): Split after
11379 reload. Use rtx_equal_p to avoid creating memory-to-memory moves,
11380 and emit NOTE_INSN_DELETED if operand[2] is zero (i.e. with -O0).
11382 2022-06-24 Alexandre Oliva <oliva@adacore.com>
11384 * common.opt (nostdlib++): New.
11385 * doc/invoke.texi (-nostdlib++): Document it.
11387 2022-06-24 Alexandre Oliva <oliva@adacore.com>
11389 * doc/sourcebuild.texi (Environment attributes): Document
11392 2022-06-23 David Malcolm <dmalcolm@redhat.com>
11394 * common.opt (fdiagnostics-show-rules): New option.
11395 * diagnostic-format-json.cc (diagnostic_output_format_init_json):
11396 Fix up context->show_rules.
11397 * diagnostic-format-sarif.cc
11398 (diagnostic_output_format_init_sarif): Likewise.
11399 * diagnostic-metadata.h (diagnostic_metadata::rule): New class.
11400 (diagnostic_metadata::precanned_rule): New class.
11401 (diagnostic_metadata::add_rule): New.
11402 (diagnostic_metadata::get_num_rules): New.
11403 (diagnostic_metadata::get_rule): New.
11404 (diagnostic_metadata::m_rules): New field.
11405 * diagnostic.cc (diagnostic_initialize): Initialize show_rules.
11406 (print_any_rules): New.
11407 (diagnostic_report_diagnostic): Call it.
11408 * diagnostic.h (diagnostic_context::show_rules): New field.
11409 * doc/invoke.texi (-fno-diagnostics-show-rules): New option.
11410 * opts.cc (common_handle_option): Handle
11411 OPT_fdiagnostics_show_rules.
11412 * toplev.cc (general_init): Set up global_dc->show_rules.
11414 2022-06-23 Martin Liska <mliska@suse.cz>
11417 * ubsan.cc (sanitize_unreachable_fn): Change order of calls
11418 in order to initialize UBSAN built-ins.
11420 2022-06-23 Martin Liska <mliska@suse.cz>
11423 * ipa-icf.cc (sem_item_optimizer::filter_removed_items):
11424 Skip variables with body_removed.
11426 2022-06-23 liuhongt <hongtao.liu@intel.com>
11428 * config/i386/sse.md:(sse4_2_pcmpestr): Replace REGNO with
11430 (sse4_2_pcmpistr): Ditto.
11432 2022-06-23 Xionghu Luo <xionghuluo@tencent.com>
11434 * cgraph.cc (cgraph_edge::redirect_call_stmt_to_callee): Fix
11436 * tree-ssa-loop-ivopts.cc (struct iv_cand): Likewise.
11437 * tree-switch-conversion.h: Likewise.
11439 2022-06-22 Jason Merrill <jason@redhat.com>
11442 * common.opt: Add -funreachable-traps.
11443 * doc/invoke.texi (-funreachable-traps): Document it.
11444 * opts.cc (finish_options): Enable at -O0 or -Og.
11445 * tree.cc (build_common_builtin_nodes): Add __builtin_trap.
11446 (builtin_decl_unreachable, build_builtin_unreachable): New.
11447 * tree.h: Declare them.
11448 * ubsan.cc (sanitize_unreachable_fn): Factor out.
11449 (ubsan_instrument_unreachable): Use
11450 gimple_build_builtin_unreachable.
11451 * ubsan.h (sanitize_unreachable_fn): Declare.
11452 * gimple.cc (gimple_build_builtin_unreachable): New.
11453 * gimple.h: Declare it.
11454 * builtins.cc (expand_builtin_unreachable): Add assert.
11455 (fold_builtin_0): Call build_builtin_unreachable.
11456 * sanopt.cc: Don't run for just SANITIZE_RETURN
11457 or SANITIZE_UNREACHABLE when trapping.
11458 * cgraphunit.cc (walk_polymorphic_call_targets): Use new
11459 unreachable functions.
11460 * gimple-fold.cc (gimple_fold_call)
11461 (gimple_get_virt_method_for_vtable)
11462 * ipa-fnsummary.cc (redirect_to_unreachable)
11463 * ipa-prop.cc (ipa_make_edge_direct_to_target)
11464 (ipa_impossible_devirt_target)
11465 * ipa.cc (walk_polymorphic_call_targets)
11466 * tree-cfg.cc (pass_warn_function_return::execute)
11467 (execute_fixup_cfg)
11468 * tree-ssa-loop-ivcanon.cc (remove_exits_and_undefined_stmts)
11470 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt):
11473 2022-06-22 Richard Sandiford <richard.sandiford@arm.com>
11475 PR tree-optimization/106019
11476 * tree-data-ref.cc (dr_may_alias_p): Try using the
11477 innermost_loop_behavior to disambiguate non-loop queries.
11479 2022-06-22 Palmer Dabbelt <palmer@rivosinc.com>
11481 * doc/invoke.texi (RISC-V): Document -mtune=thead-c906.
11483 2022-06-22 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11485 * config/xtensa/xtensa.md (bswapsi2_internal):
11486 Enlarge the buffer that is obviously smaller than the template
11487 string given to sprintf().
11489 2022-06-21 Roger Sayle <roger@nextmovesoftware.com>
11490 Marek Polacek <polacek@redhat.com>
11491 Segher Boessenkool <segher@kernel.crashing.org>
11492 Kewen Lin <linkw@linux.ibm.com>
11495 * config/rs6000/rs6000.md (rotl<mode>3_insert_3): Check that
11496 exact_log2 doesn't return -1 (or zero).
11497 (plus_xor): New code iterator.
11498 (*rotl<mode>3_insert_3_<code>): New define_insn_and_split.
11500 2022-06-21 Nathan Sidwell <nathan@acm.org>
11502 * doc/invoke.texi (C++ Modules): Remove language-linkage
11503 as missing feature.
11505 2022-06-21 Arjun Shankar <arjun@redhat.com>
11507 PR tree-optimization/94899
11508 * match.pd (X + C < Y + C -> (signed) X < (signed) Y, if C is
11509 0x80000000): New simplification.
11511 2022-06-21 Jakub Jelinek <jakub@redhat.com>
11513 PR rtl-optimization/106032
11514 * ifcvt.cc (noce_try_sign_mask): Punt if !t_unconditional, and
11515 t may_trap_or_fault_p, even if it is cheap.
11517 2022-06-21 Jakub Jelinek <jakub@redhat.com>
11519 PR middle-end/106030
11520 * expr.cc (expand_cond_expr_using_cmove): Pass NULL_RTX instead of
11521 temp to expand_operands if mode has been promoted.
11523 2022-06-21 Xionghu Luo <xionghuluo@tencent.com>
11526 * gimple-if-to-switch.cc (find_conditions): Don't skip the first
11529 2022-06-21 Siddhesh Poyarekar <siddhesh@gotplt.org>
11531 PR tree-optimization/105736
11532 * tree-object-size.cc (addr_object_size): Return size_unknown
11533 when object offset computation returns an error.
11535 2022-06-20 H.J. Lu <hjl.tools@gmail.com>
11538 * config/i386/i386.cc (ix86_function_ok_for_sibcall): Return
11539 false if PIC register is used when calling ifunc functions.
11541 2022-06-20 Richard Biener <rguenther@suse.de>
11543 PR middle-end/106027
11544 * fold-const.cc (fold_to_nonsharp_ineq_using_bound): Use the
11545 type of the prevailing comparison for the new comparison type.
11546 (fold_binary_loc): Use proper types for the A < X && A + 1 > Y
11547 to A < X && A >= Y folding.
11549 2022-06-20 Kewen Lin <linkw@linux.ibm.com>
11551 PR tree-optimization/105940
11552 * tree-vect-loop.cc (vect_analyze_loop_2): Add new parameter
11553 slp_done_for_suggested_uf and adjust with it accordingly.
11554 (vect_analyze_loop_1): Add new variable slp_done_for_suggested_uf,
11555 pass it down to vect_analyze_loop_2 for the initial analysis and
11556 applying suggested unroll factor.
11557 (vect_is_simple_reduction): Add parameter slp and adjust with it.
11558 (vect_analyze_scalar_cycles_1): Add parameter slp and pass down.
11559 (vect_analyze_scalar_cycles): Likewise.
11561 2022-06-20 Martin Liska <mliska@suse.cz>
11563 * bb-reorder.cc (find_traces_1_round): Add operators / and * and
11565 (better_edge_p): Likewise.
11566 * cfgloop.cc (find_subloop_latch_edge_by_profile): Likewise.
11567 * cfgloopmanip.cc (scale_loop_profile): Likewise.
11568 * cfgrtl.cc (force_nonfallthru_and_redirect): Likewise.
11569 * cgraph.cc (cgraph_edge::maybe_hot_p): Likewise.
11570 * config/sh/sh.cc (expand_cbranchdi4): Likewise.
11571 * dojump.cc (do_compare_rtx_and_jump): Likewise.
11572 * final.cc (compute_alignments): Likewise.
11573 * ipa-cp.cc (update_counts_for_self_gen_clones): Likewise.
11574 (decide_about_value): Likewise.
11575 * ipa-inline-analysis.cc (do_estimate_edge_time): Likewise.
11576 * loop-unroll.cc (unroll_loop_runtime_iterations): Likewise.
11577 * modulo-sched.cc (sms_schedule): Likewise.
11578 * omp-expand.cc (extract_omp_for_update_vars): Likewise.
11579 (expand_omp_ordered_sink): Likewise.
11580 (expand_omp_for_ordered_loops): Likewise.
11581 (expand_omp_for_static_nochunk): Likewise.
11582 * predict.cc (maybe_hot_count_p): Likewise.
11583 (probably_never_executed): Likewise.
11584 (set_even_probabilities): Likewise.
11585 (handle_missing_profiles): Likewise.
11586 (expensive_function_p): Likewise.
11587 * profile-count.h: Likewise.
11588 * profile.cc (compute_branch_probabilities): Likewise.
11589 * stmt.cc (emit_case_dispatch_table): Likewise.
11590 * symtab-thunks.cc (expand_thunk): Likewise.
11591 * tree-ssa-loop-manip.cc (tree_transform_and_unroll_loop): Likewise.
11592 * tree-ssa-sink.cc (select_best_block): Likewise.
11593 * tree-switch-conversion.cc (switch_decision_tree::analyze_switch_statement): Likewise.
11594 (switch_decision_tree::balance_case_nodes): Likewise.
11595 (switch_decision_tree::emit_case_nodes): Likewise.
11596 * tree-vect-loop.cc (scale_profile_for_vect_loop): Likewise.
11598 2022-06-20 yulong <shiyulong@iscas.ac.cn>
11600 * config/riscv/riscv-builtins.cc (RISCV_ATYPE_VOID_PTR): New.
11601 * config/riscv/riscv-cmo.def (RISCV_BUILTIN): Changed the FUNCTION_TYPE
11603 * config/riscv/riscv-ftypes.def (0): Remove unused.
11606 2022-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11608 * config/xtensa/xtensa.cc (xtensa_is_insn_L32R_p):
11609 Consider relaxed MOVI instructions as L32R.
11611 2022-06-19 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11613 * config/xtensa/xtensa.cc (xtensa_emit_move_sequence):
11614 Use can_create_pseudo_p(), instead of using individual
11615 reload_in_progress and reload_completed.
11616 (xtensa_expand_block_set_small_loop): Use xtensa_simm8x256(),
11617 the existing predicate function.
11618 (xtensa_is_insn_L32R_p, gen_int_relational, xtensa_emit_sibcall):
11619 Use the standard RTX code predicate macros such as MEM_P,
11620 SYMBOL_REF_P and/or CONST_INT_P.
11621 * config/xtensa/xtensa.md: Avoid using numeric literals to determine
11622 if callee-saved register, at the split patterns for indirect sibcall
11625 2022-06-18 Jakub Jelinek <jakub@redhat.com>
11627 * common.opt (flag_sanitize_trap): New variable.
11628 (fsanitize-trap=, fsanitize-trap): New options.
11629 (fsanitize-undefined-trap-on-error): Change into deprecated alias
11630 for -fsanitize-trap=all.
11631 * opts.h (struct sanitizer_opts_s): Add can_trap member.
11632 * opts.cc (finish_options): Complain about unsupported
11633 -fsanitize-trap= options.
11634 (sanitizer_opts): Add can_trap values to all entries.
11635 (get_closest_sanitizer_option): Ignore -fsanitize-trap=
11636 options which have can_trap false.
11637 (parse_sanitizer_options): Add support for -fsanitize-trap=.
11638 For -fsanitize-trap=all, enable
11639 SANITIZE_UNDEFINED | SANITIZE_UNDEFINED_NONDEFAULT. Disallow
11640 -fsanitize-trap=vptr here.
11641 (common_handle_option): Handle OPT_fsanitize_trap_ and
11642 OPT_fsanitize_trap.
11643 * sanopt.cc (maybe_optimize_ubsan_null_ifn): Check
11644 flag_sanitize_trap & SANITIZE_{NULL,ALIGNMENT} instead of
11645 flag_sanitize_undefined_trap_on_error.
11646 * gcc.cc (sanitize_spec_function): Use
11647 flag_sanitize & ~flag_sanitize_trap instead of flag_sanitize
11648 and drop use of flag_sanitize_undefined_trap_on_error in
11649 "undefined" handling.
11650 * ubsan.cc (ubsan_instrument_unreachable): Use
11651 flag_sanitize_trap & SANITIZE_??? instead of
11652 flag_sanitize_undefined_trap_on_error.
11653 (ubsan_expand_bounds_ifn, ubsan_expand_null_ifn,
11654 ubsan_expand_objsize_ifn, ubsan_expand_ptr_ifn,
11655 ubsan_build_overflow_builtin, instrument_bool_enum_load,
11656 ubsan_instrument_float_cast, instrument_nonnull_arg,
11657 instrument_nonnull_return, instrument_builtin): Likewise.
11658 * doc/invoke.texi (-fsanitize-trap=, -fsanitize-trap): Document.
11659 (-fsanitize-undefined-trap-on-error): Document as deprecated
11660 alias of -fsanitize-trap.
11662 2022-06-18 Jakub Jelinek <jakub@redhat.com>
11664 PR middle-end/105998
11665 * varasm.cc (narrowing_initializer_constant_valid_p): Check
11666 SCALAR_INT_MODE_P instead of INTEGRAL_MODE_P, also break on
11667 ! INTEGRAL_TYPE_P and do the same check also on op{0,1}'s type.
11669 2022-06-18 Roger Sayle <roger@nextmovesoftware.com>
11671 PR tree-optimization/105835
11672 * match.pd (convert (mult zero_one_valued_p@1 INTEGER_CST@2)):
11673 Narrow integer multiplication by a zero_one_valued_p operand.
11674 (convert (cond @1 INTEGER_CST@2 INTEGER_CST@3)): Push integer
11675 conversions inside COND_EXPR where both data operands are
11678 2022-06-18 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11680 * config/xtensa/constraints.md (Y):
11681 Change to include integer constants until reload begins.
11682 * config/xtensa/predicates.md (move_operand): Ditto.
11683 * config/xtensa/xtensa.cc (xtensa_emit_move_sequence):
11684 Change to allow storing integer constants into litpool only after
11687 2022-06-17 Uroš Bizjak <ubizjak@gmail.com>
11690 * config/alpha/alpha-protos.h (alpha_store_data_bypass_p): New.
11691 * config/alpha/alpha.cc (alpha_store_data_bypass_p): New function.
11692 (alpha_store_data_bypass_p_1): Ditto.
11693 * config/alpha/ev4.md: Use alpha_store_data_bypass_p instead
11694 of generic store_data_bypass_p.
11695 (ev4_ist_c): Remove insn reservation.
11697 2022-06-17 Uroš Bizjak <ubizjak@gmail.com>
11700 * config/i386/i386.cc (ix86_function_arg): Assert that
11701 the mode of pointer argumet is equal to ptr_mode, not Pmode.
11703 2022-06-17 Uroš Bizjak <ubizjak@gmail.com>
11706 * config/i386/sse.md (vpmov splitter): Use (match_dup ...)
11707 instead of REGNO comparisons in combine splitter.
11709 2022-06-17 Segher Boessenkool <segher@kernel.crashing.org>
11711 * config/rs6000/rs6000.cc (rs6000_invalid_conversion): Correct some
11714 2022-06-17 Kito Cheng <kito.cheng@sifive.com>
11716 * config/riscv/bitmanip.md: Supress warning.
11718 2022-06-17 Richard Earnshaw <rearnsha@arm.com>
11721 * config/arm/arm.cc (arm_print_operand, case 'V'): Use UINTVAL.
11722 Clear bits in the mask above bit 31.
11724 2022-06-17 Richard Earnshaw <rearnsha@arm.com>
11726 * config/arm/mve.md (*mve_mov<mode>): Re-order constraints
11727 to avoid spilling trivial literals to the constant pool.
11729 2022-06-16 David Malcolm <dmalcolm@redhat.com>
11731 * gimple-ssa-warn-access.cc (warn_string_no_nul): Add
11732 auto_diagnostic_group to group any warning with its note.
11733 (maybe_warn_for_bound): Likewise.
11734 (check_access): Likewise.
11735 (warn_dealloc_offset): Likewise.
11736 (pass_waccess::maybe_warn_memmodel): Likewise.
11737 (pass_waccess::maybe_check_dealloc_call): Likewise.
11738 (pass_waccess::warn_invalid_pointer): Likewise.
11739 (pass_waccess::check_dangling_stores): Likewise.
11741 2022-06-16 Jason Merrill <jason@redhat.com>
11743 * opts.cc (common_handle_option) [OPT_fsanitize_]: Set
11744 opts_set->x_flag_sanitize.
11746 2022-06-16 Jason Merrill <jason@redhat.com>
11748 * flags.h (issue_strict_overflow_warning): Comment #endif.
11750 2022-06-16 Andrew MacLeod <amacleod@redhat.com>
11752 * gimple-range-cache.cc (ranger_cache::apply_inferred_ranges): If name
11753 was invaraint before, clear the invariant bit.
11754 * gimple-range-gori.cc (gori_map::set_range_invariant): Add a flag.
11755 * gimple-range-gori.h (gori_map::set_range_invariant): Adjust prototype.
11757 2022-06-16 Andrew MacLeod <amacleod@redhat.com>
11759 * tree-ssa-propagate.cc (before_dom_children): Call value_of_stmt.
11761 2022-06-16 Jakub Jelinek <jakub@redhat.com>
11763 PR tree-optimization/105983
11764 * match.pd (y == XXX_MIN || x < y -> x <= y - 1,
11765 y != XXX_MIN && x >= y -> x > y - 1): Use :cs instead of :s
11766 on non-equality comparisons.
11768 2022-06-16 Jakub Jelinek <jakub@redhat.com>
11770 PR tree-optimization/105984
11771 * match.pd (__builtin_mul_overflow_p (x, cst, (stype) 0) ->
11772 x > stype_max / cst || x < stype_min / cst): fold_convert @1
11773 to TREE_TYPE (@0) just once and test for negative divisor
11774 also on that folded constant instead of on @1.
11776 2022-06-16 Jakub Jelinek <jakub@redhat.com>
11778 PR middle-end/105951
11779 * tree-ssa-ccp.cc (optimize_atomic_bit_test_and,
11780 optimize_atomic_op_fetch_cmp_0): Remember gimple_call_fn (call)
11781 as last argument to the internal functions.
11782 * builtins.cc (expand_ifn_atomic_bit_test_and): Adjust for the
11783 extra call argument to ifns. If expand_atomic_fetch_op fails for the
11784 lhs == NULL_TREE case, fall through into the optab code with
11785 gen_reg_rtx (mode) as target. If second expand_atomic_fetch_op
11786 fails, construct a CALL_EXPR and expand that.
11787 (expand_ifn_atomic_op_fetch_cmp_0): Adjust for the extra call argument
11788 to ifns. If expand_atomic_fetch_op fails, construct a CALL_EXPR and
11791 2022-06-16 Haochen Gui <guihaoc@gcc.gnu.org>
11794 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_builtin): Enable
11795 gimple folding for RS6000_BIF_VCMPEQUT, RS6000_BIF_VCMPNET,
11796 RS6000_BIF_CMPGE_1TI, RS6000_BIF_CMPGE_U1TI, RS6000_BIF_VCMPGTUT,
11797 RS6000_BIF_VCMPGTST, RS6000_BIF_CMPLE_1TI, RS6000_BIF_CMPLE_U1TI.
11798 * config/rs6000/vector.md (VEC_IC): New mode iterator. Add support
11799 for new Power10 V1TI instructions.
11800 (vec_cmp<mode><mode>): Set mode iterator to VEC_IC.
11801 (vec_cmpu<mode><mode>): Likewise.
11802 (vector_nlt<mode>): Set mode iterator to VEC_IC.
11803 (vector_nltv1ti): Remove.
11804 (vector_gtu<mode>): Set mode iterator to VEC_IC.
11805 (vector_gtuv1ti): Remove.
11806 (vector_nltu<mode>): Set mode iterator to VEC_IC.
11807 (vector_nltuv1ti): Remove.
11808 (vector_geu<mode>): Set mode iterator to VEC_IC.
11809 (vector_ngt<mode>): Likewise.
11810 (vector_ngtv1ti): Remove.
11811 (vector_ngtu<mode>): Set mode iterator to VEC_IC.
11812 (vector_ngtuv1ti): Remove.
11813 (vector_gtu_<mode>_p): Set mode iterator to VEC_IC.
11814 (vector_gtu_v1ti_p): Remove.
11815 (vrotl<mode>3): Set mode iterator to VEC_IC. Emit insns for V1TI.
11816 (vrotlv1ti3): Remove.
11817 (vashr<mode>3): Set mode iterator to VEC_IC. Emit insns for V1TI.
11818 (vashrv1ti3): Remove.
11820 2022-06-16 Martin Liska <mliska@suse.cz>
11822 * gengtype-state.cc (read_a_state_token): Do not skip extra
11823 character after escaped sequence.
11825 2022-06-16 Martin Liska <mliska@suse.cz>
11828 * spellcheck.cc (test_find_closest_string): Add new test.
11829 * spellcheck.h (class best_match): Prefer a difference in
11830 trailing sign symbol.
11832 2022-06-16 liuhongt <hongtao.liu@intel.com>
11834 PR tree-optimization/53533
11835 * match.pd: Simplify (B * v + C) * D -> BD * v + CD and
11836 (v + B) * C + D -> C * v + BCD when B,C,D are all INTEGER_CST,
11837 and there's no overflow or !TYPE_OVERFLOW_UNDEFINED.
11839 2022-06-15 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11841 * config/xtensa/xtensa.md (DSC): New split pattern and mode iterator.
11843 2022-06-15 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11845 * config/xtensa/predicates.md (reload_operand):
11847 * config/xtensa/xtensa.md: New peephole2 pattern.
11849 2022-06-15 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11851 * config/xtensa/xtensa.md (*round_up_to_even):
11852 New insn-and-split pattern.
11853 (*signed_ge_zero): Ditto.
11855 2022-06-15 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11857 * config/xtensa/xtensa-protos.h (xtensa_prepare_expand_call,
11858 xtensa_emit_sibcall): New prototypes.
11859 (xtensa_expand_epilogue): Add new argument that specifies whether
11860 or not sibling call.
11861 * config/xtensa/xtensa.cc (TARGET_FUNCTION_OK_FOR_SIBCALL):
11862 New macro definition.
11863 (xtensa_prepare_expand_call): New function in order to share
11865 (xtensa_emit_sibcall, xtensa_function_ok_for_sibcall):
11867 (xtensa_expand_epilogue): Add new argument sibcall_p and use it
11868 for sibling call handling.
11869 * config/xtensa/xtensa.md (call, call_value):
11870 Use xtensa_prepare_expand_call.
11871 (call_internal, call_value_internal):
11872 Add the condition in order to be disabled if sibling call.
11873 (sibcall, sibcall_value, sibcall_epilogue): New expansions.
11874 (sibcall_internal, sibcall_value_internal): New insn patterns,
11875 and split ones in order to take care of the indirect sibcalls.
11877 2022-06-15 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
11879 * doc/invoke.texi: Document -mextra-l32r-costs= option.
11881 2022-06-15 David Malcolm <dmalcolm@redhat.com>
11884 * doc/invoke.texi: Add -fno-analyzer-undo-inlining.
11885 * tree-diagnostic-path.cc (default_tree_diagnostic_path_printer):
11886 Extend -fdiagnostics-path-format=separate-events so that with
11887 -fdiagnostics-show-path-depths it prints fndecls as well as stack
11890 2022-06-15 David Malcolm <dmalcolm@redhat.com>
11892 * value-relation.h: Add "final" and "override" to relation_oracle
11893 vfunc implementations as appropriate.
11895 2022-06-15 Richard Sandiford <richard.sandiford@arm.com>
11897 PR middle-end/105975
11898 Revert everything apart from the expand_fn_using_insn and
11899 expand_direct_optab_fn changes from:
11900 * internal-fn.def (DEF_INTERNAL_INSN_FN): New macro.
11901 (GOMP_SIMT_ENTER_ALLOC, GOMP_SIMT_EXIT, GOMP_SIMT_LANE)
11902 (GOMP_SIMT_LAST_LANE, GOMP_SIMT_ORDERED_PRED, GOMP_SIMT_VOTE_ANY)
11903 (GOMP_SIMT_XCHG_BFLY, GOMP_SIMT_XCHG_IDX): Use it.
11904 * internal-fn.h (direct_internal_fn_info::directly_mapped): New
11906 (direct_internal_fn_info::vectorizable): Reduce to 1 bit.
11907 (direct_internal_fn_p): Also return true for internal functions
11908 that map directly to instructions defined target-insns.def.
11909 (direct_internal_fn): Adjust comment accordingly.
11910 * internal-fn.cc (direct_insn, optab1, optab2, vectorizable_optab1)
11911 (vectorizable_optab2): New local macros.
11912 (not_direct): Initialize directly_mapped.
11913 (mask_load_direct, load_lanes_direct, mask_load_lanes_direct)
11914 (gather_load_direct, len_load_direct, mask_store_direct)
11915 (store_lanes_direct, mask_store_lanes_direct, vec_cond_mask_direct)
11916 (vec_cond_direct, scatter_store_direct, len_store_direct)
11917 (vec_set_direct, unary_direct, binary_direct, ternary_direct)
11918 (cond_unary_direct, cond_binary_direct, cond_ternary_direct)
11919 (while_direct, fold_extract_direct, fold_left_direct)
11920 (mask_fold_left_direct, check_ptrs_direct): Use the macros above.
11921 (expand_GOMP_SIMT_ENTER_ALLOC, expand_GOMP_SIMT_EXIT): Delete
11922 (expand_GOMP_SIMT_LANE, expand_GOMP_SIMT_LAST_LANE): Likewise;
11923 (expand_GOMP_SIMT_ORDERED_PRED, expand_GOMP_SIMT_VOTE_ANY): Likewise.
11924 (expand_GOMP_SIMT_XCHG_BFLY, expand_GOMP_SIMT_XCHG_IDX): Likewise.
11925 (direct_internal_fn_types): Handle functions that map to instructions
11926 defined in target-insns.def.
11927 (direct_internal_fn_types): Likewise.
11928 (direct_internal_fn_supported_p): Likewise.
11929 (internal_fn_expanders): Likewise.
11930 (expand_fn_using_insn): New function,
11931 split out and adapted from...
11932 (expand_direct_optab_fn): ...here.
11933 (expand_GOMP_SIMT_ENTER_ALLOC): Use it.
11934 (expand_GOMP_SIMT_EXIT): Likewise.
11935 (expand_GOMP_SIMT_LANE): Likewise.
11936 (expand_GOMP_SIMT_LAST_LANE): Likewise.
11937 (expand_GOMP_SIMT_ORDERED_PRED): Likewise.
11938 (expand_GOMP_SIMT_VOTE_ANY): Likewise.
11939 (expand_GOMP_SIMT_XCHG_BFLY): Likewise.
11940 (expand_GOMP_SIMT_XCHG_IDX): Likewise.
11942 2022-06-15 Richard Earnshaw <rearnsha@arm.com>
11945 * config/arm/arm.cc (gen_cpymem_ldrd_strd): Rename low_reg and hi_reg
11946 to first_reg and second_reg respectively. Initialize them correctly
11947 when generating big-endian code.
11949 2022-06-15 Richard Earnshaw <rearnsha@arm.com>
11952 * config/arm/arm.cc (arm_bfi_1_p): Use UINTVAL instead of XUINT.
11954 2022-06-15 Richard Biener <rguenther@suse.de>
11956 PR tree-optimization/105971
11957 * tree-ssa-alias.cc (refs_may_alias_p_2): Put bail-out for
11958 FUNCTION_DECL and LABEL_DECL refs after decl-decl disambiguation
11959 to leak less surprising alias results.
11961 2022-06-15 Richard Biener <rguenther@suse.de>
11963 PR tree-optimization/105969
11964 * gimple-ssa-sprintf.cc (get_origin_and_offset_r): Avoid division
11965 by zero in overflow check.
11967 2022-06-15 Richard Sandiford <richard.sandiford@arm.com>
11969 PR tree-optimization/105254
11970 PR tree-optimization/105940
11972 * config/aarch64/aarch64.cc
11973 (aarch64_vector_costs::determine_suggested_unroll_factor): Take a
11974 loop_vec_info as argument. Restrict the unroll factor to values
11975 that divide the VF.
11976 (aarch64_vector_costs::finish_cost): Update call accordingly.
11978 2022-06-15 Richard Sandiford <richard.sandiford@arm.com>
11980 * read-rtl.cc (find_int): Substitute symbolic constants
11981 before converting the string to an integer.
11983 2022-06-15 Roger Sayle <roger@nextmovesoftware.com>
11984 Richard Biener <rguenther@suse.de>
11986 * match.pd (convert (lshift @1 INTEGER_CST@2)): Narrow integer
11987 left shifts by a constant when the result is truncated, and the
11988 shift constant is well-defined.
11989 * tree-vect-patterns.cc (vect_recog_rotate_pattern): Add
11990 support for rotations of signed integer types, by lowering
11991 using unsigned vector shifts.
11993 2022-06-15 liuhongt <hongtao.liu@intel.com>
11996 * config/i386/sse.md (*avx_cmp<mode>3_ltint_not): Force_reg
11999 2022-06-14 Surya Kumari Jangala <jskumari@linux.ibm.com>
12001 PR rtl-optimization/105041
12002 * regrename.cc (check_new_reg_p): Use nregs value from du chain.
12004 2022-06-14 Segher Boessenkool <segher@kernel.crashing.org>
12006 * config/rs6000/vsx.md (VS_scalar): Delete.
12007 (rest of file): Adjust.
12009 2022-06-14 Jan Hubicka <hubicka@ucw.cz>
12012 * ipa-prop.cc (ipa_load_from_parm_agg): Punt on volatile loads.
12014 2022-06-14 Philipp Tomsich <philipp.tomsich@vrull.eu>
12016 * config/riscv/bitmanip.md: Add split to handle opportunities
12017 for slli + sh[123]add.uw
12019 2022-06-14 Philipp Tomsich <philipp.tomsich@vrull.eu>
12021 * config/riscv/predicates.md (consecutive_bits_operand):
12022 Implement new predicate.
12024 2022-06-14 Richard Biener <rguenther@suse.de>
12026 PR tree-optimization/105946
12027 * tree-ssa-uninit.cc (maybe_warn_pass_by_reference):
12028 Do not look at arguments not specified in the function call.
12030 2022-06-14 Richard Biener <rguenther@suse.de>
12032 PR middle-end/105965
12033 * match.pd (view_convert CONSTRUCTOR): Handle single-element
12036 2022-06-14 Eric Botcazou <ebotcazou@adacore.com>
12038 * warning-control.cc (copy_warning) [generic version]: Do not erase
12039 the warning data of the destination location when the no-warning
12040 bit is not set on the source.
12041 (copy_warning) [tree version]: Return early if TO is equal to FROM.
12042 (copy_warning) [gimple version]: Likewise.
12044 2022-06-14 Kewen Lin <linkw@linux.ibm.com>
12046 PR tree-optimization/105940
12047 * tree-vect-loop.cc (vect_analyze_loop_2): Move the place of
12048 applying suggested_unroll_factor after start_over.
12050 2022-06-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12052 * config/xtensa/predicates.md (shifted_mask_operand):
12054 * config/xtensa/xtensa.md (*andsi3_const_pow2_minus_one):
12055 New insn-and-split pattern.
12056 (*andsi3_const_negative_pow2, *andsi3_const_shifted_mask,
12057 *masktrue_const_pow2_minus_one, *masktrue_const_negative_pow2,
12058 *masktrue_const_shifted_mask): Ditto.
12060 2022-06-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12062 * config/xtensa/xtensa.md (*masktrue_bitcmpl): New insn pattern.
12064 2022-06-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12066 * config/xtensa/xtensa-protos.h (xtensa_emit_branch):
12067 Remove the first argument.
12068 (xtensa_emit_bit_branch): Remove it because now called only from the
12069 output statement of *bittrue insn pattern.
12070 * config/xtensa/xtensa.cc (gen_int_relational): Remove the last
12071 argument 'p_invert', and make so that the condition is reversed by
12073 (xtensa_expand_conditional_branch): Share the common path, and remove
12074 condition inversion code.
12075 (xtensa_emit_branch, xtensa_emit_movcc): Simplify by removing the
12076 "false side" pattern.
12077 (xtensa_emit_bit_branch): Remove it because of the abovementioned
12078 reason, and move the function body to *bittrue insn pattern.
12079 * config/xtensa/xtensa.md (*bittrue): Transplant the output
12080 statement from removed xtensa_emit_bit_branch().
12081 (*bfalse, *ubfalse, *bitfalse, *maskfalse): Remove the "false side"
12084 2022-06-14 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12086 * config/xtensa/predicates.md (logical_shift_operator,
12087 xtensa_shift_per_byte_operator): New predicates.
12088 * config/xtensa/xtensa-protos.h (xtensa_shlrd_which_direction):
12090 * config/xtensa/xtensa.cc (xtensa_shlrd_which_direction):
12091 New helper function for funnel shift patterns.
12092 * config/xtensa/xtensa.md (ior_op): New code iterator.
12093 (*ashlsi3_1): Replace with new split pattern.
12094 (*shift_per_byte): Unify *ashlsi3_3x, *ashrsi3_3x and *lshrsi3_3x.
12095 (*shift_per_byte_omit_AND_0, *shift_per_byte_omit_AND_1):
12096 New insn-and-split patterns that redirect to *xtensa_shift_per_byte,
12097 in order to omit unnecessary bitwise AND operation.
12098 (*shlrd_reg_<code>, *shlrd_const_<code>, *shlrd_per_byte_<code>,
12099 *shlrd_per_byte_<code>_omit_AND):
12100 New insn patterns for funnel shifts.
12102 2022-06-13 Jason Merrill <jason@redhat.com>
12104 * tree-cfg.cc (pass_warn_function_return::execute): Also check
12107 2022-06-13 Maciej W. Rozycki <macro@embecosm.com>
12109 * config/riscv/riscv.md (length): Remove the explicit setting
12112 2022-06-13 H.J. Lu <hjl.tools@gmail.com>
12114 * common/config/i386/cpuinfo.h (get_available_features): Require
12115 AVX for F16C and VAES.
12117 2022-06-13 Uroš Bizjak <ubizjak@gmail.com>
12120 * config/i386/predicates.md (register_no_elim_operand):
12121 Return true for subreg of a memory operand.
12123 2022-06-13 Richard Sandiford <richard.sandiford@arm.com>
12125 * internal-fn.def (DEF_INTERNAL_INSN_FN): New macro.
12126 (GOMP_SIMT_ENTER_ALLOC, GOMP_SIMT_EXIT, GOMP_SIMT_LANE)
12127 (GOMP_SIMT_LAST_LANE, GOMP_SIMT_ORDERED_PRED, GOMP_SIMT_VOTE_ANY)
12128 (GOMP_SIMT_XCHG_BFLY, GOMP_SIMT_XCHG_IDX): Use it.
12129 * internal-fn.h (direct_internal_fn_info::directly_mapped): New
12131 (direct_internal_fn_info::vectorizable): Reduce to 1 bit.
12132 (direct_internal_fn_p): Also return true for internal functions
12133 that map directly to instructions defined target-insns.def.
12134 (direct_internal_fn): Adjust comment accordingly.
12135 * internal-fn.cc (direct_insn, optab1, optab2, vectorizable_optab1)
12136 (vectorizable_optab2): New local macros.
12137 (not_direct): Initialize directly_mapped.
12138 (mask_load_direct, load_lanes_direct, mask_load_lanes_direct)
12139 (gather_load_direct, len_load_direct, mask_store_direct)
12140 (store_lanes_direct, mask_store_lanes_direct, vec_cond_mask_direct)
12141 (vec_cond_direct, scatter_store_direct, len_store_direct)
12142 (vec_set_direct, unary_direct, binary_direct, ternary_direct)
12143 (cond_unary_direct, cond_binary_direct, cond_ternary_direct)
12144 (while_direct, fold_extract_direct, fold_left_direct)
12145 (mask_fold_left_direct, check_ptrs_direct): Use the macros above.
12146 (expand_GOMP_SIMT_ENTER_ALLOC, expand_GOMP_SIMT_EXIT): Delete
12147 (expand_GOMP_SIMT_LANE, expand_GOMP_SIMT_LAST_LANE): Likewise;
12148 (expand_GOMP_SIMT_ORDERED_PRED, expand_GOMP_SIMT_VOTE_ANY): Likewise.
12149 (expand_GOMP_SIMT_XCHG_BFLY, expand_GOMP_SIMT_XCHG_IDX): Likewise.
12150 (direct_internal_fn_types): Handle functions that map to instructions
12151 defined in target-insns.def.
12152 (direct_internal_fn_types): Likewise.
12153 (direct_internal_fn_supported_p): Likewise.
12154 (internal_fn_expanders): Likewise.
12156 2022-06-13 Richard Sandiford <richard.sandiford@arm.com>
12158 * internal-fn.cc (expand_fn_using_insn): New function,
12159 split out and adapted from...
12160 (expand_direct_optab_fn): ...here.
12161 (expand_GOMP_SIMT_ENTER_ALLOC): Use it.
12162 (expand_GOMP_SIMT_EXIT): Likewise.
12163 (expand_GOMP_SIMT_LANE): Likewise.
12164 (expand_GOMP_SIMT_LAST_LANE): Likewise.
12165 (expand_GOMP_SIMT_ORDERED_PRED): Likewise.
12166 (expand_GOMP_SIMT_VOTE_ANY): Likewise.
12167 (expand_GOMP_SIMT_XCHG_BFLY): Likewise.
12168 (expand_GOMP_SIMT_XCHG_IDX): Likewise.
12170 2022-06-13 Jakub Jelinek <jakub@redhat.com>
12172 * omp-expand.cc (expand_omp_target): Remap user provided
12173 device clause arguments, -1 to -2 and -2 to -3, either
12174 at compile time if constant, or at runtime.
12176 2022-06-13 Eric Botcazou <ebotcazou@adacore.com>
12178 * common.opt (finstrument-functions): Set explicit value.
12179 (-finstrument-functions-once): New option.
12180 * doc/invoke.texi (Program Instrumentation Options): Document it.
12181 * gimplify.cc (build_instrumentation_call): New static function.
12182 (gimplify_function_tree): Call it to emit the instrumentation calls
12183 if -finstrument-functions[-once] is specified.
12185 2022-06-13 Eric Botcazou <ebotcazou@adacore.com>
12187 * dwarf2out.cc (output_one_line_info_table): Initialize prev_addr.
12188 * gimple.h (gimple_set_location): Do not copy warning data from
12189 the previous location when it is UNKNOWN_LOCATION.
12190 * optabs.cc (expand_widen_pattern_expr): Always set oprnd{1,2}.
12192 2022-06-13 Jakub Jelinek <jakub@redhat.com>
12195 * config/i386/i386.md (*ashl<dwi>3_doubleword_mask,
12196 *<insn><dwi>3_doubleword_mask): Use operands[3] masked with
12197 (<MODE_SIZE> * BITS_PER_UNIT) - 1 as AND operand instead of
12198 operands[3] unmodified.
12200 2022-06-12 Simon Wright <simon@pushface.org>
12203 * config/darwin-driver.cc (darwin_find_version_from_kernel): If the OS
12204 version is darwin20 (macOS 11) or greater, truncate the version to the
12207 2022-06-12 Mark Mentovai <mark@mentovai.com>
12209 * config/darwin-c.cc: Make -mmacosx-version-min more future-proof.
12211 2022-06-12 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
12214 * config/aarch64/aarch64-sve-builtins-base.cc: Include ssa.h.
12215 (svld1rq_impl::fold): Define.
12216 * config/aarch64/aarch64.cc (expand_vec_perm_d): Define new members
12217 op_mode and op_vec_flags.
12218 (aarch64_evpc_reencode): Initialize newd.op_mode and
12220 (aarch64_evpc_sve_dup): New function.
12221 (aarch64_expand_vec_perm_const_1): Gate existing calls to
12222 aarch64_evpc_* functions under d->vmode == d->op_mode,
12223 and call aarch64_evpc_sve_dup.
12224 (aarch64_vectorize_vec_perm_const): Remove assert
12225 d->vmode != d->op_mode, and initialize d.op_mode and d.op_vec_flags.
12226 * tree-cfg.cc (verify_gimple_assign_ternary): Allow different
12227 vector types for lhs and rhs in VEC_PERM_EXPR if rhs3 is
12230 2022-06-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12232 * config/xtensa/xtensa-protos.h (xtensa_constantsynth):
12234 * config/xtensa/xtensa.cc (xtensa_emit_constantsynth,
12235 xtensa_constantsynth_2insn, xtensa_constantsynth_rtx_SLLI,
12236 xtensa_constantsynth_rtx_ADDSUBX, xtensa_constantsynth):
12237 New backend functions that process the abovementioned logic.
12238 (xtensa_emit_move_sequence): Revert the previous changes.
12239 * config/xtensa/xtensa.md: New split patterns for integer
12240 and floating-point, as the frontend part.
12242 2022-06-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12244 * config/xtensa/xtensa.cc (xtensa_rtx_costs): Correct wrong case
12245 for ABS and NEG, add missing case for BSWAP and CLRSB, and
12246 double the costs for integer divisions using libfuncs if
12247 optimizing for speed, in order to take advantage of fast constant
12248 division by multiplication.
12249 (TARGET_INSN_COST): New macro definition.
12250 (xtensa_is_insn_L32R_p, xtensa_insn_cost): New functions for
12251 calculating relative costs of a RTL insns, for both of speed and
12253 * config/xtensa/xtensa.md (return, nop, trap): Correct values of
12254 the attribute "length" that depends on TARGET_DENSITY.
12255 (define_asm_attributes, blockage, frame_blockage): Add missing
12257 * config/xtensa/xtensa.opt (-mextra-l32r-costs=): New machine-
12258 dependent option, however, preparatory work for now.
12260 2022-06-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12262 * config/xtensa/xtensa.cc (xtensa_expand_block_set_small_loop):
12263 Pass through the block length / loop count conditions if
12264 zero-overhead looping is configured and active,
12266 2022-06-11 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12268 * config/xtensa/xtensa.md (mulsidi3, umulsidi3):
12269 Split into individual signedness, in order to use libcall
12270 "__umulsidi3" but not the other.
12271 (<u>mulhisi3): Merge into one by using code iterator.
12272 (<u>mulsidi3, mulhisi3, umulhisi3): Remove.
12274 2022-06-11 Michael Meissner <meissner@linux.ibm.com>
12276 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Do
12277 not generate block copies with vector pair instructions if we are
12278 tuning for power10.
12280 2022-06-10 Roger Sayle <roger@nextmovesoftware.com>
12282 PR rtl-optimization/7061
12283 * expr.cc (emit_group_store): For groups that consist of a single
12284 scalar integer register that hold a complex mode value, use
12285 gen_lowpart to generate a SUBREG to "view_convert" to the complex
12286 mode. For modes of different sizes, first convert to an integer
12287 mode of the appropriate size.
12289 2022-06-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12291 * config/xtensa/xtensa.md (clrsbsi2): New insn pattern.
12293 2022-06-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12295 * config/xtensa/xtensa.md (*andsi3_bitcmpl):
12296 New insn_and_split pattern.
12298 2022-06-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12300 * config/xtensa/xtensa.md (one_cmplsi2):
12301 Rearrange as an insn_and_split pattern.
12303 2022-06-09 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
12305 * config/xtensa/xtensa.md (bswaphi2): New insn pattern.
12307 2022-06-09 Segher Boessenkool <segher@kernel.crashing.org>
12309 * config/rs6000/rs6000.md (FP_ISA3): Delete.
12310 (float<QHI:mode><FP_ISA3:mode>2): Rename to...
12311 (float<QHI:mode><SFDF:mode>2): ... this. Adjust.
12312 (*float<QHI:mode><FP_ISA3:mode>2_internal): Rename to...
12313 (*float<QHI:mode><SFDF:mode>2_internal): ... this. Adjust.
12314 (floatuns<QHI:mode><FP_ISA3:mode>2): Rename to...
12315 (floatuns<QHI:mode><SFDF:mode>2): ... this. Adjust.
12316 (*floatuns<QHI:mode><FP_ISA3:mode>2_internal): Rename to...
12317 (*floatuns<QHI:mode><SFDF:mode>2_internal): ... this. Adjust.
12319 2022-06-09 Maciej W. Rozycki <macro@embecosm.com>
12321 * config/riscv/riscv.md
12322 (*f<quiet_pattern>_quiet<ANYF:mode><X:mode>4_default)
12323 (*f<quiet_pattern>_quiet<ANYF:mode><X:mode>4_snan): Emit a tab
12324 rather than space with FSFLAGS.
12326 2022-06-09 Tobias Burnus <tobias@codesourcery.com>
12328 * omp-offload.cc (omp_discover_declare_target_tgt_fn_r,
12329 omp_discover_declare_target_fn_r): Don't walk reverse-offload
12332 2022-06-09 Jakub Jelinek <jakub@redhat.com>
12334 * doc/invoke.texi (-Waddress): Fix a typo in small example.
12335 Fix typos inptr_t -> intptr_t and uinptr_t -> uintptr_t.
12337 2022-06-09 Cui,Lili <lili.cui@intel.com>
12340 * config/i386/x86-tune-costs.h (skylake_cost): Raise the gpr load cost
12341 from 4 to 6 and gpr store cost from 6 to 8. Change SSE loads and
12342 unaligned loads cost from {6, 6, 6, 10, 20} to {8, 8, 8, 8, 16}.
12343 (icelake_cost): Ditto.
12344 (alderlake_cost): Raise the gpr store cost from 6 to 8 and SSE loads,
12345 stores and unaligned stores cost from {6, 6, 6, 10, 15} to
12348 2022-06-09 Haochen Gui <guihaoc@gcc.gnu.org>
12350 * config/rs6000/rs6000.md (define_split for bswapdi load): Merge shift
12351 and ior insns to one rotate and mask insn.
12352 (define_split for bswapdi register): Likewise.
12354 2022-06-08 Roger Sayle <roger@nextmovesoftware.com>
12356 PR middle-end/105874
12357 * expr.cc (expand_expr_real_1) <normal_inner_ref>: New local
12358 variable tem_modifier for calculating the expand_modifier enum to
12359 use for expanding tem. If tem is a VAR_DECL, use EXPAND_MEMORY.
12361 2022-06-08 Max Filippov <jcmvbkbc@gmail.com>
12364 * config/xtensa/xtensa.md (movdi): Rename 'first' and 'second'
12365 to 'lowpart' and 'highpart' so that they match 'gen_lowpart' and
12366 'gen_highpart' bitwise semantics and fix order of highpart and
12367 lowpart depending on target endianness.
12369 2022-06-08 Chung-Ju Wu <jasonwucj@gmail.com>
12371 * config/arm/arm-cpus.in (star-mc1): New cpu.
12372 * config/arm/arm-tables.opt: Regenerate.
12373 * config/arm/arm-tune.md: Regenerate.
12374 * doc/invoke.texi: Update docs.
12376 2022-06-08 liuhongt <hongtao.liu@intel.com>
12380 * config/i386/i386.md (*movsi_internal): Change alternative
12382 (*movdi_internal): Ditto.
12383 * config/i386/sse.md (vec_set<mode>_0): Change alternative *r
12385 (*vec_extractv4sf_mem): Ditto.
12386 (*vec_extracthf): Ditto.
12388 2022-06-07 Richard Earnshaw <rearnsha@arm.com>
12391 * config/arm/arm.cc (arm_bfi_1_p): New function.
12392 (arm_bfi_p): New function.
12393 (arm_rtx_costs_internal): Add costs for BFI idioms.
12394 (arm_print_operand [case 'V']): Format output for BFI/BFC masks.
12395 * config/arm/constraints.md (Dj): New constraint.
12396 * config/arm/arm.md (arm_andsi3_insn): Add alternative to use BFC.
12397 (insv_zero): Convert to an insn with a split.
12398 (*bfi, *bfi_alt1, *bfi_alt2, *bfi_alt3): New patterns.
12400 2022-06-07 liuhongt <hongtao.liu@intel.com>
12403 * config/i386/sse.md (ssse3_palignrdi): Change alternative 2
12406 2022-06-07 Roger Sayle <roger@nextmovesoftware.com>
12408 PR middle-end/105853
12410 * calls.cc (load_register_parameters): Call store_constructor
12411 and int_expr_size directly instead of expanding via expand_expr.
12412 * expr.cc (static void store_constructor): Don't prototype here.
12413 (static HOST_WIDE_INT int_expr_size): Likewise.
12414 (store_constructor): No longer static.
12415 (int_expr_size): Likewise, no longer static.
12416 * expr.h (store_constructor): Prototype here.
12417 (int_expr_size): Prototype here.
12419 2022-06-07 Jan Beulich <jbeulich@suse.com>
12422 2022-06-03 Jan Beulich <jbeulich@suse.com>
12424 * Makefile.in (ORIGINAL_OBJCOPY_FOR_TARGET): New.
12425 * configure.ac: Check for objcopy, producing
12426 ORIGINAL_OBJCOPY_FOR_TARGET.
12427 * configure: Update accordingly.
12428 * exec-tool.in (ORIGINAL_OBJCOPY_FOR_TARGET): New.
12431 2022-06-07 Jakub Jelinek <jakub@redhat.com>
12433 * tree.h (OMP_CLAUSE_LINEAR_OLD_LINEAR_MODIFIER): Define.
12434 * tree-pretty-print.cc (dump_omp_clause) <case OMP_CLAUSE_LINEAR>:
12435 Adjust clause printing style depending on
12436 OMP_CLAUSE_LINEAR_OLD_LINEAR_MODIFIER.
12438 2022-06-07 Jan Beulich <jbeulich@suse.com>
12440 * config/i386/i386-builtin.def (__builtin_ia32_psadbw256):
12442 * config/i386/i386-builtin-types.def: New function type
12443 (V4DI, V32QI, V32QI).
12444 * config/i386/i386-expand.cc (ix86_expand_args_builtin): Handle
12445 V4DI_FTYPE_V32QI_V32QI.
12447 2022-06-07 Jan Beulich <jbeulich@suse.com>
12449 * config/i386/i386.cc (ix86_attr_length_vex_default): Take REX.B
12450 into account for reg-only insns.
12452 2022-06-07 Roger Sayle <roger@nextmovesoftware.com>
12454 * config/i386/i386.cc (ix86_rtx_costs): Add a new case for
12455 IF_THEN_ELSE, and provide costs for TARGET_XOP's vpcmov and
12456 TARGET_CMOVE's (scalar integer) conditional moves.
12457 * config/i386/sse.md (define_split): Recognize XOP's vpcmov
12458 from its equivalent (canonical) pxor;pand;pxor sequence.
12460 2022-06-07 Kewen Lin <linkw@linux.ibm.com>
12462 * machmode.def (VECTOR_MODES_WITH_PREFIX): Update document for
12465 2022-06-06 Andrew Stubbs <ams@codesourcery.com>
12467 * config.in: Regenerate.
12468 * configure: Regenerate.
12469 * configure.ac: Reinstate HAVE_GAS_ARM_EXTENDED_ARCH test.
12471 2022-06-04 Roger Sayle <roger@nextmovesoftware.com>
12473 PR middle-end/95126
12474 * calls.cc (load_register_parameters): When loading a suitable
12475 immediate_const_ctor_p VAR_DECL into a single word_mode register,
12476 construct it directly in a pseudo rather than read it (by parts)
12478 * expr.cc (int_expr_size): Make tree argument a const_tree.
12479 (immediate_const_ctor_p): Helper predicate. Return true for
12480 simple constructors that may be materialized in a register.
12481 (expand_expr_real_1) [VAR_DECL]: When expanding a constant
12482 VAR_DECL with a suitable immediate_const_ctor_p constructor
12483 use store_constructor to materialize it directly in a pseudo.
12484 * expr.h (immediate_const_ctor_p): Prototype here.
12485 * varasm.cc (initializer_constant_valid_for_bitfield_p): Change
12486 VALUE argument from tree to const_tree.
12487 * varasm.h (initializer_constant_valid_for_bitfield_p): Update
12490 2022-06-04 Jakub Jelinek <jakub@redhat.com>
12493 * config/i386/i386.md (*ashl<dwi>3_doubleword_mask,
12494 *<insn><dwi>3_doubleword_mask): If top bit of mask is clear, but lower
12495 bits of mask aren't all set, use operands[2] mode for the AND
12496 operation instead of always SImode.
12498 2022-06-03 Jakub Jelinek <jakub@redhat.com>
12500 PR middle-end/30314
12501 PR middle-end/105777
12502 * match.pd (__builtin_mul_overflow_p (x, cst, (stype) 0) ->
12503 x > stype_max / cst || x < stype_min / cst): New simplification.
12505 2022-06-03 Aldy Hernandez <aldyh@redhat.com>
12507 * gimple-range-cache.cc (ranger_cache::range_from_dom): Use
12509 * gimple-range.cc (gimple_ranger::register_inferred_ranges): Same.
12510 * value-range.h (Value_Range::Value_Range): Implement copy
12511 constructor for Value_Range.
12513 2022-06-03 Aldy Hernandez <aldyh@redhat.com>
12515 * value-range.h (struct vrange_traits): Remove.
12516 (is_a): Rewrite without vrange_traits.
12519 2022-06-03 Aldy Hernandez <aldyh@redhat.com>
12521 * value-range.cc (vrange::contains_p): Implement.
12522 (vrange::type): Return void.
12523 (vrange::supports_type_p): Implement.
12524 (irange::fits_p): Same.
12525 (vrange::set_undefined): Same.
12526 (irange::set_nonnegative): Same.
12527 (vrange::set_varying): Same.
12528 (vrange::union_): Same.
12529 (unsupported_range::set): Move to vrange.
12530 (unsupported_range::type): Move to vrange.
12531 (vrange::intersect): Implement for varying and undefined.
12532 (vrange::zero_p): Implement.
12533 (unsupported_range::supports_type_p): Move to vrange.
12534 (vrange::nonzero_p): Implement.
12535 (unsupported_range::set_undefined): Move to vrange.
12536 (unsupported_range::set_varying): Same.
12537 (unsupported_range::dump): Same.
12538 (unsupported_range::union_): Same. Implement for varying and
12540 (unsupported_range::intersect): Move to vrange.
12541 (unsupported_range::zero_p): Same.
12542 (unsupported_range::nonzero_p): Same.
12543 (unsupported_range::set_nonzero): Same.
12544 (unsupported_range::set_zero): Same.
12545 (unsupported_range::set_nonnegative): Same.
12546 (unsupported_range::fits_p): Same.
12547 * value-range.h (class vrange): Remove abstract markers for most
12549 (class unsupported_range): Remove most methods as they will now be
12550 inherited from vrange.
12552 2022-06-03 Aldy Hernandez <aldyh@redhat.com>
12554 * gimple-range-edge.cc (gimple_outgoing_range_stmt_p): Adjust for
12555 an object level supports_type_p for irange and a static
12556 Value_Range::supports_type_p.
12557 * gimple-range-fold.cc (fold_using_range::range_of_range_op): Same.
12558 (fold_using_range::range_of_address): Same.
12559 (fold_using_range::range_of_builtin_call): Same.
12560 * gimple-range-fold.h (gimple_range_type): Same.
12561 (gimple_range_ssa_p): Same.
12562 * gimple-range-path.cc (path_range_query::internal_range_of_expr):
12564 (path_range_query::range_of_stmt): Same.
12565 (path_range_query::add_to_imports): Same.
12566 * gimple-range.cc (gimple_ranger::range_on_edge): Same.
12567 (gimple_ranger::export_global_ranges): Same.
12568 * gimple-ssa-evrp-analyze.cc
12569 (evrp_range_analyzer::record_ranges_from_phis): Same.
12570 * range-op.cc (range_operator::wi_fold): Same.
12571 (range_operator::fold_range): Same.
12572 * tree-ssa-loop-ch.cc (entry_loop_condition_is_static): Same.
12573 * tree-ssa-loop-unswitch.cc (struct unswitch_predicate): Same.
12574 (evaluate_control_stmt_using_entry_checks): Same.
12575 * tree-ssa-threadedge.cc
12576 (hybrid_jt_simplifier::compute_ranges_from_state): Same.
12577 * tree-vrp.cc (supported_types_p): Same.
12578 * value-query.cc (range_query::value_of_expr): Same.
12579 (range_query::value_on_edge): Same.
12580 (range_query::value_of_stmt): Same.
12581 (range_query::get_tree_range): Same.
12582 (get_range_global): Same.
12583 (global_range_query::range_of_expr): Same.
12584 * value-range-equiv.h (class value_range_equiv): Same.
12585 * value-range.cc (irange::supports_type_p): Same.
12586 (unsupported_range::supports_type_p): Same.
12587 * value-range.h (enum value_range_discriminator): Same.
12588 (Value_Range::init): Same.
12589 (Value_Range::supports_type_p): Same.
12590 (irange::supports_type_p): Same.
12591 (irange::supports_p): Same.
12592 (vrange::supports_type_p): Same.
12593 (vrange_allocator::alloc_vrange): Same.
12595 2022-06-03 Jan Beulich <jbeulich@suse.com>
12597 * Makefile.in (ORIGINAL_OBJCOPY_FOR_TARGET): New.
12598 * configure.ac: Check for objcopy, producing
12599 ORIGINAL_OBJCOPY_FOR_TARGET.
12600 * configure: Update accordingly.
12601 * exec-tool.in (ORIGINAL_OBJCOPY_FOR_TARGET): New.
12604 2022-06-03 Jan Beulich <jbeulich@suse.com>
12606 * config/i386/mmx.md (mmx_psadbw): Convert to expander.
12607 (*mmx_psadbw): New. Mark as commutative.
12608 * config/i386/sse.md (<sse2_avx2>_psadbw): Convert to expander.
12609 (*<sse2_avx2>_psadbw): New. Mark as commutative.
12611 2022-06-03 Alexandre Oliva <oliva@adacore.com>
12613 PR tree-optimization/105665
12614 PR tree-optimization/100810
12615 * tree-ssa-loop-ivopts.cc
12616 (ssa_name_maybe_undef_p, ssa_name_set_maybe_undef): New.
12617 (ssa_name_any_use_dominates_bb_p, mark_ssa_maybe_undefs): New.
12618 (find_ssa_undef): Check precomputed flag and intervening uses.
12619 (tree_ssa_iv_optimize): Call mark_ssa_maybe_undefs.
12621 2022-06-02 David Malcolm <dmalcolm@redhat.com>
12623 * Makefile.in (OBJS): Add tree-diagnostic-client-data-hooks.o and
12624 tree-logical-location.o.
12625 (OBJS-libcommon): Add diagnostic-format-sarif.o; reorder.
12626 (CFLAGS-tree-diagnostic-client-data-hooks.o): Add TARGET_NAME.
12627 * common.opt (fdiagnostics-format=): Add sarif-stderr and sarif-file.
12628 (sarif-stderr, sarif-file): New enum values.
12629 * diagnostic-client-data-hooks.h: New file.
12630 * diagnostic-format-sarif.cc: New file.
12631 * diagnostic-path.h (enum diagnostic_event::verb): New enum.
12632 (enum diagnostic_event::noun): New enum.
12633 (enum diagnostic_event::property): New enum.
12634 (struct diagnostic_event::meaning): New struct.
12635 (diagnostic_event::get_logical_location): New vfunc.
12636 (diagnostic_event::get_meaning): New vfunc.
12637 (simple_diagnostic_event::get_logical_location): New vfunc impl.
12638 (simple_diagnostic_event::get_meaning): New vfunc impl.
12639 * diagnostic.cc: Include "diagnostic-client-data-hooks.h".
12640 (diagnostic_initialize): Initialize m_client_data_hooks.
12641 (diagnostic_finish): Clean up m_client_data_hooks.
12642 (diagnostic_event::meaning::dump_to_pp): New.
12643 (diagnostic_event::meaning::maybe_get_verb_str): New.
12644 (diagnostic_event::meaning::maybe_get_noun_str): New.
12645 (diagnostic_event::meaning::maybe_get_property_str): New.
12646 (get_cwe_url): Make non-static.
12647 (diagnostic_output_format_init): Handle
12648 DIAGNOSTICS_OUTPUT_FORMAT_SARIF_STDERR and
12649 DIAGNOSTICS_OUTPUT_FORMAT_SARIF_FILE.
12650 * diagnostic.h (enum diagnostics_output_format): Add
12651 DIAGNOSTICS_OUTPUT_FORMAT_SARIF_STDERR and
12652 DIAGNOSTICS_OUTPUT_FORMAT_SARIF_FILE.
12653 (class diagnostic_client_data_hooks): New forward decl.
12654 (class logical_location): New forward decl.
12655 (diagnostic_context::m_client_data_hooks): New field.
12656 (diagnostic_output_format_init_sarif_stderr): New decl.
12657 (diagnostic_output_format_init_sarif_file): New decl.
12658 (get_cwe_url): New decl.
12659 * doc/invoke.texi (-fdiagnostics-format=): Add sarif-stderr and
12661 * doc/sourcebuild.texi (Scan a particular file): Add
12662 scan-sarif-file and scan-sarif-file-not.
12663 * langhooks-def.h (lhd_get_sarif_source_language): New decl.
12664 (LANG_HOOKS_GET_SARIF_SOURCE_LANGUAGE): New macro.
12665 (LANG_HOOKS_INITIALIZER): Add
12666 LANG_HOOKS_GET_SARIF_SOURCE_LANGUAGE.
12667 * langhooks.cc (lhd_get_sarif_source_language): New.
12668 * langhooks.h (lang_hooks::get_sarif_source_language): New field.
12669 * logical-location.h: New file.
12670 * plugin.cc (struct for_each_plugin_closure): New.
12671 (for_each_plugin_cb): New.
12672 (for_each_plugin): New.
12673 * plugin.h (for_each_plugin): New decl.
12674 * tree-diagnostic-client-data-hooks.cc: New file.
12675 * tree-diagnostic.cc: Include "diagnostic-client-data-hooks.h".
12676 (tree_diagnostics_defaults): Populate m_client_data_hooks.
12677 * tree-logical-location.cc: New file.
12678 * tree-logical-location.h: New file.
12680 2022-06-02 David Malcolm <dmalcolm@redhat.com>
12682 * common.opt (fdiagnostics-format=): Add json-stderr and json-file
12684 (DIAGNOSTICS_OUTPUT_FORMAT_JSON): Rename to...
12685 (DIAGNOSTICS_OUTPUT_FORMAT_JSON_STDERR): ...this.
12686 (diagnostics_output_format): Add json-stderr and json-file.
12687 * diagnostic-format-json.cc (json_flush_to_file): New.
12688 (json_final_cb): Convert to...
12689 (json_flush_to_file): ...this, ...
12690 (json_stderr_final_cb): ...this, and...
12691 (json_file_final_cb): ...this.
12692 (diagnostic_output_format_init): Move to diagnostic.cc.
12693 (json_output_base_file_name): New.
12694 (diagnostic_output_format_init_json): New.
12695 (diagnostic_output_format_init_json_stderr): New.
12696 (diagnostic_output_format_init_json_file): New.
12697 * diagnostic.cc (diagnostic_output_format_init): Move here from
12698 diagnostic-format-json.cc; update for changes to enum.
12699 * diagnostic.h (enum diagnostics_output_format): Rename
12700 DIAGNOSTICS_OUTPUT_FORMAT_JSON to
12701 DIAGNOSTICS_OUTPUT_FORMAT_JSON_STDERR, and add
12702 DIAGNOSTICS_OUTPUT_FORMAT_JSON_FILE.
12703 (diagnostic_output_format_init): Add base_file_name param.
12704 (diagnostic_output_format_init_json_stderr): New decl.
12705 (diagnostic_output_format_init_json_file): New dec.
12706 * doc/invoke.texi (-fdiagnostics-format=): Add "json-stderr" and
12707 "json-file". Rewrite so that the existing "json" is a synonym of
12709 * gcc.cc (driver_handle_option): Pass dump_base_name to
12710 diagnostic_output_format_init.
12711 * opts.cc (common_handle_option): Likewise.
12713 2022-06-02 David Malcolm <dmalcolm@redhat.com>
12715 * json.cc (string::print): Fix escaping of '\'.
12717 2022-06-02 Philipp Tomsich <philipp.tomsich@vrull.eu>
12719 * config/riscv/riscv.cc (riscv_build_integer_1): Rewrite value as
12720 (-1 << 31) for the single-bit case, when operating on (1 << 31)
12722 * config/riscv/riscv.h (SINGLE_BIT_MASK_OPERAND): Allow for
12723 any single-bit value, moving the special case for (1 << 31) to
12724 riscv_build_integer_1 (in riscv.c).
12726 2022-06-02 Roger Sayle <roger@nextmovesoftware.com>
12729 * config/i386/sse.md (V_128_256):Add V1TI and V2TI.
12730 (define_mode_attr avxsizesuffix): Add support for V1TI and V2TI.
12732 2022-06-02 Jakub Jelinek <jakub@redhat.com>
12735 * config/i386/i386.md (*ashl<dwi>3_doubleword_mask): Remove :SI
12736 from AND and its operands and just verify operands[2] has HImode,
12737 SImode or for TARGET_64BIT DImode. Allow operands[3] to be a mask
12738 with all low 6 (64-bit) or 5 (32-bit) bits set and in that case
12739 just throw away the masking. Use force_reg before calling
12741 (*ashl<dwi>3_doubleword_mask_1): Allow operands[3] to be a mask
12742 with all low 6 (64-bit) or 5 (32-bit) bits set and in that case
12743 just throw away the masking.
12744 (*ashl<mode>3_doubleword): Rename to ...
12745 (ashl<mode>3_doubleword): ... this.
12746 (*ashl<mode>3_mask): Remove :SI from AND and its operands and just
12747 verify operands[2] has HImode, SImode or for TARGET_64BIT DImode.
12748 Use force_reg before calling gen_lowpart.
12749 (*<insn><mode>3_mask): Likewise.
12750 (*<insn><dwi>3_doubleword_mask): Likewise. Allow operands[3] to be
12751 a mask with all low 6 (64-bit) or 5 (32-bit) bits set and in that
12752 case just throw away the masking. Use force_reg before calling
12754 (*<insn><dwi>3_doubleword_mask_1): Allow operands[3] to be a mask
12755 with all low 6 (64-bit) or 5 (32-bit) bits set and in that case just
12756 throw away the masking.
12757 (*<insn><mode>3_doubleword): Rename to ...
12758 (<insn><mode>3_doubleword): ... this.
12759 (*<insn><mode>3_mask): Remove :SI from AND and its operands and just
12760 verify operands[2] has HImode, SImode or for TARGET_64BIT DImode.
12761 Use force_reg before calling gen_lowpart.
12762 (splitter after it): Remove :SI from AND and its operands and just
12763 verify operands[2] has HImode, SImode or for TARGET_64BIT DImode.
12764 (*<btsc><mode>_mask, *<btsc><mode>_mask): Remove :SI from AND and its
12765 operands and just verify operands[1] has HImode, SImode or for
12766 TARGET_64BIT DImode. Use force_reg before calling gen_lowpart.
12767 (*jcc_bt<mode>_mask_1): New define_insn_and_split pattern.
12768 * config/i386/i386.cc (ix86_rtx_costs): For ZERO_EXTRACT with
12769 ZERO_EXTEND QI->SI in last operand ignore the cost of the ZERO_EXTEND.
12771 2022-06-02 Richard Biener <rguenther@suse.de>
12773 PR tree-optimization/101668
12774 * tree-vect-slp.cc (vect_build_slp_tree_1): Allow BIT_FIELD_REFs
12775 for vector types with compatible lane types.
12776 (vect_build_slp_tree_2): Deal with this.
12777 (vect_add_slp_permutation): Adjust. Emit lowpart/concat
12778 special cases without VEC_PERM.
12779 (vectorizable_slp_permutation): Select the operand vector
12780 type and relax requirements. Handle identity permutes
12781 with mismatching operand types.
12782 * optabs-query.cc (can_vec_perm_const_p): Only allow variable
12783 permutes for op_mode == mode.
12785 2022-06-02 Richard Biener <rguenther@suse.de>
12787 PR tree-optimization/105802
12788 * tree-ssa-loop-unswitch.cc (find_unswitching_predicates_for_bb):
12789 Make sure to also compute the range in the type of the switch index.
12791 2022-06-01 David Seifert <soap@gentoo.org>
12794 * configure: Regenerate.
12796 2022-06-01 H.J. Lu <hjl.tools@gmail.com>
12798 PR rtl-optimization/105638
12799 * df-core.cc (df_find_single_def_src): Moved and renamed from
12800 find_single_def_src in loop-iv.cc. Change the argument to rtx
12801 and use rtx_equal_p. Return null for partial or conditional
12803 * df.h (df_find_single_def_src): New prototype.
12804 * dse.cc (record_store): Use the constant source if the source
12805 register is set only once.
12806 * loop-iv.cc (find_single_def_src): Moved to df-core.cc.
12807 (replace_single_def_regs): Replace find_single_def_src with
12808 df_find_single_def_src.
12810 2022-06-01 Wilco Dijkstra <wilco.dijkstra@arm.com>
12812 * config/aarch64/aarch64.opt (explicit_tune_core): Rename to
12814 (explicit_arch): Rename to selected_arch.
12815 (x_aarch64_override_tune_string): Remove.
12816 (aarch64_ra_sign_key): Add as TargetVariable so it gets saved/restored.
12817 (aarch64_override_tune_string): Add Save so it gets saved/restored.
12818 * config/aarch64/aarch64.h (aarch64_architecture_version): Remove.
12819 * config/aarch64/aarch64.cc (aarch64_architecture_version): Remove.
12820 (processor): Remove archtecture_version field.
12821 (selected_arch): Remove global.
12822 (selected_cpu): Remove global.
12823 (selected_tune): Remove global.
12824 (aarch64_ra_sign_key): Move global to aarch64.opt so it is saved.
12825 (aarch64_override_options_internal): Use aarch64_get_tune_cpu.
12826 (aarch64_override_options): Further simplify code to only set
12827 selected_arch and selected_tune globals.
12828 (aarch64_option_save): Remove now that target options are saved.
12829 (aarch64_option_restore): Remove redundant target option restores.
12830 * config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Use
12832 * config/aarch64/aarch64-opts.h (aarch64_key_type): Add, moved from...
12833 * config/aarch64/aarch64-protos.h (aarch64_key_type): Remove.
12834 (aarch64_ra_sign_key): Remove.
12836 2022-06-01 Jakub Jelinek <jakub@redhat.com>
12838 PR middle-end/30314
12839 * match.pd (__builtin_mul_overflow_p (x, cst, (utype) 0) ->
12840 x > ~(utype)0 / cst): New simplification.
12842 2022-06-01 Richard Biener <rguenther@suse.de>
12844 PR tree-optimization/105786
12845 * tree-loop-distribution.cc
12846 (loop_distribution::transform_reduction_loop): Only do strlen
12847 replacement for integer type reductions.
12849 2022-06-01 Jakub Jelinek <jakub@redhat.com>
12851 PR tree-optimization/105770
12852 * tree-ssa-loop-unswitch.cc (find_unswitching_predicates_for_bb): Cast
12853 CASE_LOW and CASE_HIGH to TREE_TYPE (idx) before comparisons with idx.
12855 2022-06-01 Aldy Hernandez <aldyh@redhat.com>
12857 * gimple-range-cache.cc (ssa_block_ranges::dump): Convert to vrange.
12858 (sbr_vector::sbr_vector): Same.
12859 (sbr_vector::grow): Same.
12860 (sbr_vector::set_bb_range): Same.
12861 (sbr_vector::get_bb_range): Same.
12862 (sbr_sparse_bitmap::sbr_sparse_bitmap): Same.
12863 (sbr_sparse_bitmap::set_bb_range): Same.
12864 (sbr_sparse_bitmap::get_bb_range): Same.
12865 (block_range_cache::set_bb_range): Same.
12866 (block_range_cache::get_bb_range): Same.
12867 (block_range_cache::dump): Same.
12868 (ssa_global_cache::get_global_range): Same.
12869 (ssa_global_cache::set_global_range): Same.
12870 (ssa_global_cache::clear): Same.
12871 (ssa_global_cache::dump): Same.
12872 (ranger_cache::get_global_range): Same.
12873 (ranger_cache::set_global_range): Same.
12874 (ranger_cache::range_of_def): Same.
12875 (ranger_cache::entry_range): Same.
12876 (ranger_cache::exit_range): Same.
12877 (ranger_cache::edge_range): Same.
12878 (ranger_cache::range_of_expr): Same.
12879 (ranger_cache::range_on_edge): Same.
12880 (ranger_cache::block_range): Same.
12881 (ranger_cache::propagate_cache): Same.
12882 (ranger_cache::fill_block_cache): Same.
12883 (ranger_cache::range_from_dom): Same.
12884 * gimple-range-cache.h: Same.
12885 * gimple-range-edge.cc (gimple_outgoing_range::get_edge_range):
12887 (gimple_outgoing_range::switch_edge_range): Same.
12888 (gimple_outgoing_range::edge_range_p): Same.
12889 * gimple-range-edge.h: Same.
12890 * gimple-range-fold.cc (fur_source::get_operand): Same.
12891 (fur_source::get_phi_operand): Same.
12892 (fur_edge::get_operand): Same.
12893 (fur_edge::get_phi_operand): Same.
12894 (fur_stmt::get_operand): Same.
12895 (fur_stmt::get_phi_operand): Same.
12896 (fur_list::fur_list): Same.
12897 (fur_list::get_operand): Same.
12898 (fur_list::get_phi_operand): Same.
12899 (fold_range): Same.
12900 (adjust_imagpart_expr): Same.
12901 (adjust_realpart_expr): Same.
12902 (gimple_range_adjustment): Same.
12903 (fold_using_range::fold_stmt): Same.
12904 (fold_using_range::range_of_range_op): Same.
12905 (fold_using_range::range_of_address): Same.
12906 (fold_using_range::range_of_phi): Same.
12907 (fold_using_range::range_of_call): Same.
12908 (fold_using_range::range_of_builtin_call): Same.
12909 (fold_using_range::range_of_builtin_int_call): Same.
12910 (fold_using_range::range_of_cond_expr): Same.
12911 (fur_source::register_outgoing_edges): Same.
12912 * gimple-range-fold.h (fold_range): Same.
12913 (gimple_range_type): Same.
12914 (gimple_range_ssa_p): Same.
12915 * gimple-range-gori.cc (gimple_range_calc_op1): Same.
12916 (gimple_range_calc_op2): Same.
12917 (gori_compute::compute_operand_range_switch): Same.
12918 (gori_compute::compute_operand_range): Same.
12919 (gori_compute::logical_combine): Same.
12920 (gori_compute::compute_logical_operands): Same.
12921 (gori_compute::compute_operand1_range): Same.
12922 (gori_compute::compute_operand2_range): Same.
12923 (gori_compute::compute_operand1_and_operand2_range): Same.
12924 (gori_compute::outgoing_edge_range_p): Same.
12925 (gori_compute::condexpr_adjust): Same.
12926 * gimple-range-gori.h (gimple_range_calc_op1): Same.
12927 (gimple_range_calc_op2): Same.
12928 * gimple-range-path.cc (path_range_query::get_cache): Same.
12929 (path_range_query::set_cache): Same.
12930 (path_range_query::range_on_path_entry): Same.
12931 (path_range_query::internal_range_of_expr): Same.
12932 (path_range_query::range_of_expr): Same.
12933 (path_range_query::ssa_range_in_phi): Same.
12934 (path_range_query::range_defined_in_block): Same.
12935 (path_range_query::compute_ranges_in_phis): Same.
12936 (path_range_query::compute_ranges_in_block): Same.
12937 (path_range_query::add_to_imports): Same.
12938 (path_range_query::range_of_stmt): Same.
12939 * gimple-range-path.h: Same.
12940 * gimple-range-infer.cc (gimple_infer_range::add_range): Same.
12941 (gimple_infer_range::~side_effect_manager): Same.
12942 (gimple_infer_range::get_nonzero): Same.
12943 (gimple_infer_range::maybe_adjust_range): Same.
12944 (gimple_infer_range::add_range): Same.
12945 * gimple-range-infer.h: Same.
12946 * gimple-range-tests.cc: Same.
12947 * gimple-range-trace.cc (range_tracer::trailer): Same.
12948 (debug_seed_ranger): Same.
12949 * gimple-range-trace.h: Same.
12950 * gimple-range.cc (gimple_ranger::range_of_expr): Same.
12951 (gimple_ranger::range_on_entry): Same.
12952 (gimple_ranger::range_on_exit): Same.
12953 (gimple_ranger::range_on_edge): Same.
12954 (gimple_ranger::fold_range_internal): Same.
12955 (gimple_ranger::range_of_stmt): Same.
12956 (gimple_ranger::prefill_name): Same.
12957 (gimple_ranger::prefill_stmt_dependencies): Same.
12958 (gimple_ranger::export_global_ranges): Same.
12959 (gimple_ranger::dump_bb): Same.
12960 * gimple-range.h: Same.
12961 * gimple-ssa-warn-access.cc (check_nul_terminated_array): Same.
12962 (memmodel_to_uhwi): Same.
12963 * tree-ssa-loop-niter.cc (refine_value_range_using_guard): Same.
12964 (determine_value_range): Same.
12965 (record_nonwrapping_iv): Same.
12966 (infer_loop_bounds_from_signedness): Same.
12967 (scev_var_range_cant_overflow): Same.
12968 * tree-ssa-threadedge.cc (hybrid_jt_simplifier::simplify): Same.
12969 * value-query.cc (range_query::range_on_edge): Same.
12970 (range_query::range_of_stmt): Same.
12971 (range_query::value_of_expr): Same.
12972 (range_query::value_on_edge): Same.
12973 (range_query::value_of_stmt): Same.
12974 (range_query::get_tree_range): Same.
12975 (update_global_range): Same.
12976 (get_range_global): Same.
12977 (gimple_range_global): Same.
12978 (global_range_query::range_of_expr): Same.
12979 (range_query::query_relation): Same.
12980 * value-query.h (gimple_range_global): Same.
12981 (update_global_range): Same.
12982 * vr-values.cc (vr_values::range_of_expr): Same.
12983 (bounds_of_var_in_loop): Same.
12984 (simplify_using_ranges::vrp_visit_cond_stmt): Same.
12985 * vr-values.h (class vr_values): Same.
12986 * tree-ssa-loop-unswitch.cc (unswitch_predicate): Same.
12988 2022-06-01 Aldy Hernandez <aldyh@redhat.com>
12990 * gimple-range-cache.cc (sbr_vector::sbr_vector): Adjust for
12992 (sbr_vector::grow): Same.
12993 (sbr_vector::set_bb_range): Same.
12994 (sbr_sparse_bitmap::sbr_sparse_bitmap): Same.
12995 (sbr_sparse_bitmap::set_bb_range): Same.
12996 (block_range_cache::~block_range_cache): Same.
12997 (block_range_cache::set_bb_range): Same.
12998 (ssa_global_cache::ssa_global_cache): Same.
12999 (ssa_global_cache::~ssa_global_cache): Same.
13000 (ssa_global_cache::set_global_range): Same.
13001 * gimple-range-cache.h (block_range_cache): Same.
13002 (ssa_global_cache): Same.
13003 * gimple-range-edge.cc
13004 (gimple_outgoing_range::calc_switch_ranges): Same.
13005 * gimple-range-edge.h (gimple_outgoing_range): Same.
13006 * gimple-range-infer.cc (infer_range_manager::get_nonzero):
13008 (infer_range_manager::add_range): Same.
13009 * gimple-range-infer.h (class infer_range_manager): Same.
13010 * value-range.h (class irange_allocator): Rename to...
13011 (class vrange_allocator): ...this.
13012 (irange_allocator::irange_allocator): New.
13013 (vrange_allocator::vrange_allocator): New.
13014 (irange_allocator::~irange_allocator): New.
13015 (vrange_allocator::~vrange_allocator): New.
13016 (irange_allocator::get_memory): Rename to...
13017 (vrange_allocator::alloc): ...this.
13018 (vrange_allocator::alloc_vrange): Rename from...
13019 (irange_allocator::allocate): ...this.
13020 (vrange_allocator::alloc_irange): New.
13022 2022-06-01 Aldy Hernandez <aldyh@redhat.com>
13024 * gimple-range-edge.cc (gimple_outgoing_range_stmt_p): Adjust for
13025 vrange and convert range_op_handler function calls to use the
13026 identically named object.
13027 * gimple-range-fold.cc (gimple_range_operand1): Same.
13028 (gimple_range_operand2): Same.
13029 (fold_using_range::fold_stmt): Same.
13030 (fold_using_range::range_of_range_op): Same.
13031 (fold_using_range::range_of_builtin_ubsan_call): Same.
13032 (fold_using_range::relation_fold_and_or): Same.
13033 (fur_source::register_outgoing_edges): Same.
13034 * gimple-range-fold.h (gimple_range_handler): Remove.
13035 * gimple-range-gori.cc (gimple_range_calc_op1): Adjust for vrange.
13036 (gimple_range_calc_op2): Same.
13037 (range_def_chain::get_def_chain): Same.
13038 (gori_compute::compute_operand_range): Same.
13039 (gori_compute::condexpr_adjust): Same.
13040 * gimple-range.cc (gimple_ranger::prefill_name): Same.
13041 (gimple_ranger::prefill_stmt_dependencies): Same.
13042 * range-op.cc (get_bool_state): Same.
13043 (class operator_equal): Add using clause.
13044 (class operator_not_equal): Same.
13045 (class operator_lt): Same.
13046 (class operator_le): Same.
13047 (class operator_gt): Same.
13048 (class operator_ge): Same.
13049 (class operator_plus): Same.
13050 (class operator_minus): Same.
13051 (class operator_mult): Same.
13052 (class operator_exact_divide): Same.
13053 (class operator_lshift): Same.
13054 (class operator_rshift): Same.
13055 (class operator_cast): Same.
13056 (class operator_logical_and): Same.
13057 (class operator_bitwise_and): Same.
13058 (class operator_logical_or): Same.
13059 (class operator_bitwise_or): Same.
13060 (class operator_bitwise_xor): Same.
13061 (class operator_trunc_mod): Same.
13062 (class operator_logical_not): Same.
13063 (class operator_bitwise_not): Same.
13064 (class operator_cst): Same.
13065 (class operator_identity): Same.
13066 (class operator_unknown): Same.
13067 (class operator_abs): Same.
13068 (class operator_negate): Same.
13069 (class operator_addr_expr): Same.
13070 (class pointer_or_operator): Same.
13071 (operator_plus::op1_range): Adjust for vrange.
13072 (operator_minus::op1_range): Same.
13073 (operator_mult::op1_range): Same.
13074 (operator_cast::op1_range): Same.
13075 (operator_bitwise_not::fold_range): Same.
13076 (operator_negate::fold_range): Same.
13077 (range_op_handler): Rename to...
13078 (get_handler): ...this.
13079 (range_op_handler::range_op_handler): New.
13080 (range_op_handler::fold_range): New.
13081 (range_op_handler::op1_range): New.
13082 (range_op_handler::op2_range): New.
13083 (range_op_handler::lhs_op1_relation): New.
13084 (range_op_handler::lhs_op2_relation): New.
13085 (range_op_handler::op1_op2_relation): New.
13086 (range_cast): Adjust for vrange.
13087 * range-op.h (range_op_handler): Remove function.
13088 (range_cast): Adjust for vrange.
13089 (class range_op_handler): New.
13090 (get_bool_state): Adjust for vrange.
13091 (empty_range_varying): Same.
13092 (relop_early_resolve): Same.
13093 * tree-data-ref.cc (compute_distributive_range): Same.
13094 * tree-vrp.cc (get_range_op_handler): Remove.
13095 (range_fold_binary_symbolics_p): Use range_op_handler class
13096 instead of get_range_op_handler.
13097 (range_fold_unary_symbolics_p): Same.
13098 (range_fold_binary_expr): Same.
13099 (range_fold_unary_expr): Same.
13100 * value-query.cc (range_query::get_tree_range): Adjust for vrange.
13102 2022-06-01 Aldy Hernandez <aldyh@redhat.com>
13104 * gimple-range-fold.h (gimple_range_type): Check type before
13105 calling supports_type_p.
13106 * gimple-range-path.cc (path_range_query::range_of_stmt): Same.
13107 * value-query.cc (range_query::get_tree_range): Same.
13108 * value-range.cc (Value_Range::lower_bound): New.
13109 (Value_Range::upper_bound): New.
13110 (Value_Range::dump): New.
13111 * value-range.h (class Value_Range): New.
13112 (irange::supports_type_p): Do not check if type is non-zero.
13114 2022-06-01 Aldy Hernandez <aldyh@redhat.com>
13116 * value-range-equiv.cc (value_range_equiv::set): New.
13117 * value-range-equiv.h (class value_range_equiv): Make set method
13119 Remove default bitmap argument from set method.
13120 * value-range.cc (vrange::contains_p): New.
13121 (vrange::singleton_p): New.
13122 (vrange::operator=): New.
13123 (vrange::operator==): New.
13124 (irange::fits_p): Move to .cc file.
13125 (irange::set_nonnegative): New.
13126 (unsupported_range::unsupported_range): New.
13127 (unsupported_range::set): New.
13128 (unsupported_range::type): New.
13129 (unsupported_range::set_undefined): New.
13130 (unsupported_range::set_varying): New.
13131 (unsupported_range::dump): New.
13132 (unsupported_range::union_): New.
13133 (unsupported_range::intersect): New.
13134 (unsupported_range::zero_p): New.
13135 (unsupported_range::nonzero_p): New.
13136 (unsupported_range::set_nonzero): New.
13137 (unsupported_range::set_zero): New.
13138 (unsupported_range::set_nonnegative): New.
13139 (unsupported_range::fits_p): New.
13140 (irange::set): Call irange::set_undefined.
13141 (irange::verify_range): Check discriminator field.
13142 (irange::dump): Dump [irange] marker.
13143 (irange::debug): Move to...
13144 (vrange::debug): ...here.
13145 (dump_value_range): Accept vrange.
13147 * value-range.h (enum value_range_discriminator): New.
13148 (class vrange): New.
13149 (class unsupported_range): New.
13150 (struct vrange_traits): New.
13153 (class irange): Inherit from vrange.
13154 (dump_value_range): Adjust for vrange.
13155 (irange::kind): Rename to...
13156 (vrange::kind): ...this.
13157 (irange::varying_p): Rename to...
13158 (vrange::varying_p): ...this.
13159 (irange::undefined_p): Rename to...
13160 (vrange::undefined_p): ...this.
13161 (irange::irange): Set discriminator.
13162 (irange::union_): Convert to irange before passing to irange
13164 (irange::intersect): Same.
13165 (vrange::supports_type_p): New.
13166 * vr-values.cc (vr_values::extract_range_from_binary_expr): Pass
13167 NULL bitmap argument to value_range_equiv::set.
13168 (vr_values::extract_range_basic): Same.
13170 2022-06-01 Richard Biener <rguenther@suse.de>
13172 PR tree-optimization/105763
13173 * tree-ssa-loop-unswitch.cc (find_unswitching_predicates_for_bb):
13174 Check gimple_range_ssa_p.
13176 2022-05-31 Jason Merrill <jason@redhat.com>
13178 * Makefile.in (TAGS): Look at libcpp/*.cc.
13180 2022-05-31 Christophe Lyon <christophe.lyon@arm.com>
13182 * config/aarch64/aarch64.cc (aarch64_gimplify_va_arg_expr):
13183 Prefix mode names with E_.
13185 2022-05-31 Alan Modra <amodra@gmail.com>
13187 * dwarf2out.cc (gen_namelist_decl): Adjust to suit correct
13188 spelling of DW_AT_namelist_item.
13190 2022-05-31 Jakub Jelinek <jakub@redhat.com>
13192 * omp-low.cc (build_outer_var_ref): For code == OMP_CLAUSE_ALLOCATE
13193 allow var to be private in the outer context.
13194 (lower_private_allocate): Pass OMP_CLAUSE_ALLOCATE as last argument
13195 to build_outer_var_ref.
13197 2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
13199 * config/i386/i386.cc (ix86_modes_tieable_p): Allow SCmode to be
13200 tieable with DImode on TARGET_64BIT, and SCmode tieable with
13201 V2SFmode, and DCmode with V2DFmode.
13203 2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
13205 PR rtl-optimization/101617
13206 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Add a
13207 special case (indicated by negate_cc_compare_p) to generate a
13208 -1/0 mask using neg;sbb.
13209 * config/i386/i386.md (x86_neg<mode>_ccc): New define_expand
13210 to generate an *x86_neg<mode>_ccc instruction.
13211 (x86_mov<mode>cc_0_m1_neg): Likewise, a new define_expand to
13212 generate a *x86_mov<mode>cc_0_m1_neg instruction.
13214 2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
13216 * rtlanal.cc (rtx_cost) <MULT>: Treat FMA, SS_MULT, US_MULT,
13217 SMUL_HIGHPART and UMUL_HIGHPART as having the same cost as MULT.
13218 <DIV>: Likewise, SS_DIV and US_DIV have the same default as DIV.
13220 2022-05-30 Roger Sayle <roger@nextmovesoftware.com>
13223 * config/i386/i386-expand.cc (ix86_expand_branch): Don't decompose
13224 DI mode equality/inequality using XOR here. Instead generate a
13225 COMPARE for doubleword modes (DImode on !TARGET_64BIT or TImode).
13226 * config/i386/i386-features.cc (gen_gpr_to_xmm_move_src): Use
13227 gen_rtx_SUBREG when NUNITS is 1, i.e. for TImode to V1TImode.
13228 (general_scalar_chain::convert_compare): New function to convert
13229 scalar equality/inequality comparison into vector operations.
13230 (general_scalar_chain::convert_insn) [COMPARE]: Refactor. Call
13231 new convert_compare helper method.
13232 (convertible_comparion_p): Update to match doubleword COMPARE
13233 of two register, memory or integer constant operands.
13234 * config/i386/i386-features.h (general_scalar_chain::convert_compare):
13235 Prototype/declare member function here.
13236 * config/i386/i386.md (cstore<mode>4): Change mode to SDWIM, but
13237 only allow new doubleword modes for EQ and NE operators.
13238 (*cmp<dwi>_doubleword): New define_insn_and_split, to split a
13239 doubleword comparison into a pair of XORs followed by an IOR to
13240 set the (zero) flags register, optimizing the XORs if possible.
13241 * config/i386/sse.md (V_AVX): Include V1TI and V2TI in mode
13242 iterator; V_AVX is (currently) only used by ptest.
13243 (sse4_1 mode attribute): Update to support V1TI and V2TI.
13245 2022-05-30 Uroš Bizjak <ubizjak@gmail.com>
13247 * config/i386/i386.md: Remove constraints when used with
13248 const_int_operand, const0_operand, const_1_operand, constm1_operand,
13249 const8_operand, const128_operand, const248_operand, const123_operand,
13250 const2367_operand, const1248_operand, const359_operand,
13251 const_4_or_8_to_11_operand, const48_operand, const_0_to_1_operand,
13252 const_0_to_3_operand, const_0_to_4_operand, const_0_to_5_operand,
13253 const_0_to_7_operand, const_0_to_15_operand, const_0_to_31_operand,
13254 const_0_to_63_operand, const_0_to_127_operand, const_0_to_255_operand,
13255 const_0_to_255_mul_8_operand, const_1_to_31_operand,
13256 const_1_to_63_operand, const_2_to_3_operand, const_4_to_5_operand,
13257 const_4_to_7_operand, const_6_to_7_operand, const_8_to_9_operand,
13258 const_8_to_11_operand, const_8_to_15_operand, const_10_to_11_operand,
13259 const_12_to_13_operand, const_12_to_15_operand, const_14_to_15_operand,
13260 const_16_to_19_operand, const_16_to_31_operand, const_20_to_23_operand,
13261 const_24_to_27_operand and const_28_to_31_operand.
13262 * config/i386/mmx.md: Ditto.
13263 * config/i386/sse.md: Ditto.
13264 * config/i386/subst.md: Ditto.
13265 * config/i386/sync.md: Ditto.
13267 2022-05-30 Jan Beulich <jbeulich@suse.com>
13269 * config/i386/i386.md (bmi2_umul<mode><dwi>3_1): Correct MEM_P()
13272 2022-05-30 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
13274 * config/arm/arm.cc (arm_vectorize_vec_perm_const): Adjust prototype.
13276 2022-05-29 Iain Sandoe <iain@sandoe.co.uk>
13279 * config/darwin.h: Move versions-specific handling of multiply_defined
13280 from SUBTARGET_DRIVER_SELF_SPECS to LINK_SPEC.
13282 2022-05-29 Eric Gallager <egallager@gcc.gnu.org>
13285 * doc/sourcebuild.texi: Add entries for the c++tools,
13286 gotools, libbacktrace, libcc1, libcody, liboffloadmic,
13287 and libsanitizer directories. Remove entry for boehm-gc.
13288 Fix alphabetization for libquadmath.
13290 2022-05-28 Joel Holdsworth <jholdsworth@nvidia.com>
13292 * config/avr/avr-mcus.def: Add device definitions.
13293 * doc/avr-mmcu.texi: Corresponding changes.
13294 * config/avr/gen-avr-mmcu-texi.cc: Added support for avr
13296 * config/avr/gen-avr-mmcu-specs.cc: Prevent -mmcu=avr* flags
13297 from leaking into cc1.
13299 2022-05-28 Vladimir Makarov <vmakarov@gcc.gnu.org>
13302 * config/sh/sh.cc (sh_register_move_cost): Avoid cost "2" (which
13303 is special) for various scenarios.
13305 2022-05-28 Iain Sandoe <iain@sandoe.co.uk>
13307 * config/darwin.h (REAL_LIBGCC_SPEC): Update the comment block
13308 describing this macro.
13310 2022-05-27 Richard Biener <rguenther@suse.de>
13312 * tree-dfa.cc (get_ref_base_and_extent): Avoid shift.
13314 2022-05-27 Martin Jambor <mjambor@suse.cz>
13317 * ipa-prop.cc (propagate_controlled_uses): Check type of the
13318 constant before adding a LOAD reference.
13320 2022-05-27 Jakub Jelinek <jakub@redhat.com>
13322 * tree-core.h (enum omp_clause_code): Rename OMP_CLAUSE_TO_DECLARE
13323 to OMP_CLAUSE_ENTER.
13324 * tree.h (OMP_CLAUSE_ENTER_TO): Define.
13325 * tree.cc (omp_clause_num_ops, omp_clause_code_name): Rename
13326 OMP_CLAUSE_TO_DECLARE to OMP_CLAUSE_ENTER.
13327 * tree-pretty-print.cc (dump_omp_clause): Handle OMP_CLAUSE_ENTER
13328 instead of OMP_CLAUSE_TO_DECLARE, if OMP_CLAUSE_ENTER_TO, print
13329 "to" instead of "enter".
13330 * tree-nested.cc (convert_nonlocal_omp_clauses,
13331 convert_local_omp_clauses): Handle OMP_CLAUSE_ENTER instead of
13332 OMP_CLAUSE_TO_DECLARE.
13334 2022-05-27 Richard Biener <rguenther@suse.de>
13336 PR tree-optimization/105726
13337 * gimple-ssa-warn-restrict.cc (builtin_memref::set_base_and_offset):
13338 Constrain array-of-flexarray case more.
13340 2022-05-27 Jakub Jelinek <jakub@redhat.com>
13342 PR sanitizer/105729
13343 * fold-const.cc (fold_unary_loc): Don't optimize (X &) ((Y *) z + w)
13344 to (X &) z + w if -fsanitize=null during GENERIC folding.
13346 2022-05-27 Roger Sayle <roger@nextmovesoftware.com>
13348 * match.pd (match_zero_one_valued_p): New predicate.
13349 (mult @0 @1): Use zero_one_valued_p for optimization to the
13350 expression "bit_and @0 @1".
13351 (bit_and (negate zero_one_valued_p@0) @1): Optimize to MULT_EXPR.
13352 (plus @0 (mult (minus @1 @0) zero_one_valued_p@2)): New transform.
13353 (minus @0 (mult (minus @0 @1) zero_one_valued_p@2)): Likewise.
13354 (bit_xor @0 (mult (bit_xor @0 @1) zero_one_valued_p@2)): Likewise.
13355 Remove three redundant transforms obsoleted by the three above.
13357 2022-05-27 Roger Sayle <roger@nextmovesoftware.com>
13359 * config/i386/i386.md (*test<mode>_not): New define_insn_and_split
13360 to split a combined "and;cmp" sequence into "not;test".
13362 2022-05-27 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13364 * config/xtensa/xtensa.md (bswapsi2): New expansion pattern.
13365 (bswapsi2_internal): Revise the template and condition, and add
13366 detection code for preceding the same insn in order to omit a
13367 "SSAI 8" instruction of the latter.
13368 (bswapdi2): Suppress built-in insn expansion with the corresponding
13369 library call when optimizing for size.
13371 2022-05-27 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13373 * config/xtensa/xtensa-protos.h
13374 (xtensa_expand_block_set_unrolled_loop,
13375 xtensa_expand_block_set_small_loop): New prototypes.
13376 * config/xtensa/xtensa.cc (xtensa_sizeof_MOVI,
13377 xtensa_expand_block_set_unrolled_loop,
13378 xtensa_expand_block_set_small_loop): New functions.
13379 * config/xtensa/xtensa.md (setmemsi): New expansion pattern.
13380 * config/xtensa/xtensa.opt (mlongcalls): Add target mask.
13382 2022-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13384 * config/xtensa/xtensa.cc (xtensa_expand_block_move):
13385 Make instruction counting more accurate, and simplify emitting insns.
13387 2022-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13389 * config/xtensa/constraints.md (M, O): Use the macro.
13390 * config/xtensa/predicates.md (addsubx_operand, extui_fldsz_operand,
13391 sext_fldsz_operand): Ditto.
13392 * config/xtensa/xtensa.cc (xtensa_simm8, xtensa_simm8x256,
13393 xtensa_simm12b, xtensa_uimm8, xtensa_uimm8x2, xtensa_uimm8x4,
13394 xtensa_mask_immediate, smalloffset_mem_p, printx, xtensa_call_save_reg,
13395 xtensa_expand_prologue): Ditto.
13396 * config/xtensa/xtensa.h (FUNCTION_ARG_REGNO_P): Ditto.
13398 2022-05-26 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
13400 * config/xtensa/predicates.md (extui_fldsz_operand): Simplify.
13401 * config/xtensa/xtensa.cc (xtensa_mask_immediate, print_operand):
13404 2022-05-25 Andrew MacLeod <amacleod@redhat.com>
13406 * gimple-range-cache.cc: Adjust comments.
13407 * gimple-range-infer.cc: Adjust comments.
13408 * gimple-range-infer.h: Adjust comments.
13409 * gimple-range.cc: Adjust comments.
13411 2022-05-25 Andrew MacLeod <amacleod@redhat.com>
13413 * Makefile.in (OBJS): Use gimple-range-infer.o.
13414 * gimple-range-cache.cc (ranger_cache::fill_block_cache): Change msg.
13415 (ranger_cache::range_from_dom): Rename var side_effect to infer.
13416 (ranger_cache::apply_inferred_ranges): Rename from apply_side_effects.
13417 * gimple-range-cache.h: Include gimple-range-infer.h.
13418 (class ranger_cache): Adjust prototypes, use infer_range_manager.
13419 * gimple-range-infer.cc: Rename from gimple-range-side-effects.cc.
13420 (gimple_infer_range::*): Rename from stmt_side_effects.
13421 (infer_range_manager::*): Rename from side_effect_manager.
13422 * gimple-range-side-effect.cc: Rename.
13423 * gimple-range-side-effect.h: Rename.
13424 * gimple-range-infer.h: Rename from gimple-range-side-effects.h.
13425 (class gimple_infer_range): Rename from stmt_side_effects.
13426 (class infer_range_manager): Rename from side_effect_manager.
13427 * gimple-range.cc (gimple_ranger::register_inferred_ranges): Rename
13428 from register_side_effects.
13429 * gimple-range.h (register_inferred_ranges): Adjust prototype.
13430 * range-op.h: Adjust comment.
13431 * tree-vrp.cc (rvrp_folder::pre_fold_bb): Use register_inferred_ranges.
13432 (rvrp_folder::post_fold_bb): Use register_inferred_ranges.
13434 2022-05-25 Simon Cook <simon.cook@embecosm.com>
13436 * config/riscv/arch-canonicalize: Only add mafd extension if
13437 base was rv32/rv64g.
13439 2022-05-25 Tobias Burnus <tobias@codesourcery.com>
13441 * doc/invoke.texi (AMD GCN Options): Add gfx908/gfx90a.
13443 2022-05-25 Jakub Jelinek <jakub@redhat.com>
13445 PR sanitizer/105714
13446 * asan.cc (has_stmt_been_instrumented_p): For assignments which
13447 are both stores and loads, return true only if both destination
13448 and source have been instrumented.
13450 2022-05-25 Martin Liska <mliska@suse.cz>
13451 Richard Biener <rguenther@suse.de>
13453 * dbgcnt.def (DEBUG_COUNTER): Add loop_unswitch counter.
13454 * params.opt (max-unswitch-level): Remove.
13455 * doc/invoke.texi (max-unswitch-level): Likewise.
13456 * tree-cfg.cc (gimple_lv_add_condition_to_bb): Support not
13457 gimplified expressions.
13458 * tree-ssa-loop-unswitch.cc (struct unswitch_predicate): New.
13459 (tree_may_unswitch_on): Rename to ...
13460 (find_unswitching_predicates_for_bb): ... this and handle
13462 (get_predicates_for_bb): Likewise.
13463 (set_predicates_for_bb): Likewise.
13464 (init_loop_unswitch_info): Likewise.
13465 (tree_ssa_unswitch_loops): Prepare stuff before calling
13466 tree_unswitch_single_loop.
13467 (tree_unswitch_single_loop): Rework the function using
13468 pre-computed predicates and with a per original loop cost model.
13470 (add_predicate_to_path): Likewise.
13471 (find_range_for_lhs): Likewise.
13472 (simplify_using_entry_checks): Rename to ...
13473 (evaluate_control_stmt_using_entry_checks): ... this, handle
13474 switch statements and improve simplifications using ranger.
13475 (simplify_loop_version): Rework using
13476 evaluate_control_stmt_using_entry_checks.
13477 (evaluate_bbs): New.
13478 (evaluate_loop_insns_for_predicate): Likewise.
13479 (tree_unswitch_loop): Adjust to allow switch statements and
13480 pass in the edge to unswitch.
13481 (clean_up_after_unswitching): New.
13482 (pass_tree_unswitch::execute): Pass down fun.
13484 2022-05-24 Eugene Rozenfeld <erozen@microsoft.com>
13486 * tree-vect-loop-manip.cc (vect_do_peeling): Save/restore profile
13487 counts for the epilog loop.
13489 2022-05-24 Martin Sebor <msebor@redhat.com>
13490 Richard Biener <rguenther@suse.de>
13492 PR middle-end/105604
13493 * gimple-ssa-sprintf.cc (set_aggregate_size_and_offset): Add comments.
13494 (get_origin_and_offset_r): Remove null handling. Handle variable array
13496 (get_origin_and_offset): Handle null argument here. Simplify.
13497 (alias_offset): Update comment.
13498 * pointer-query.cc (field_at_offset): Update comment. Handle members
13499 of variable-length types.
13501 2022-05-24 Prathamesh Kulkarni <prathamesh.kulkarni@linaro.org>
13503 * target.def (vec_perm_const): Define new parameter op_mode and
13505 * doc/tm.texi: Regenerate.
13506 * config/aarch64/aarch64.cc (aarch64_vectorize_vec_perm_const): Adjust
13507 vec_perm_const hook to add new parameter op_mode and return false
13508 if result and operand modes do not match.
13509 * config/arm/arm.cc (arm_vectorize_vec_perm_const): Likewise.
13510 * config/gcn/gcn.cc (gcn_vectorize_vec_perm_const): Likewise.
13511 * config/ia64/ia64.cc (ia64_vectorize_vec_perm_const): Likewise.
13512 * config/mips/mips.cc (mips_vectorize_vec_perm_const): Likewise.
13513 * config/rs6000/rs6000.cc (rs6000_vectorize_vec_perm_const): Likewise
13514 * config/s390/s390.cc (s390_vectorize_vec_perm_const): Likewise.
13515 * config/sparc/sparc.cc (sparc_vectorize_vec_perm_const): Likewise.
13516 * config/i386/i386-expand.cc (ix86_vectorize_vec_perm_const): Likewise.
13517 * config/i386/i386-expand.h (ix86_vectorize_vec_perm_const): Adjust
13519 * config/i386/sse.md (ashrv4di3): Adjust call to vec_perm_const hook.
13520 (ashrv2di3): Likewise.
13521 * optabs.cc (expand_vec_perm_const): Likewise.
13522 * optabs-query.h (can_vec_perm_const_p): Adjust prototype.
13523 * optabs-query.cc (can_vec_perm_const_p): Define new parameter
13524 op_mode and pass it to vec_perm_const hook.
13525 (can_mult_highpart_p): Adjust call to can_vec_perm_const_p.
13526 * match.pd (vec_perm X Y CST): Likewise.
13527 * tree-ssa-forwprop.cc (simplify_vector_constructor): Likewise.
13528 * tree-vect-data-refs.cc (vect_grouped_store_supported): Likewise.
13529 (vect_grouped_load_supported): Likewise.
13530 (vect_shift_permute_load_chain): Likewise.
13531 * tree-vect-generic.cc (lower_vec_perm): Likewise.
13532 * tree-vect-loop-manip.cc (interleave_supported_p): Likewise.
13533 * tree-vect-loop.cc (have_whole_vector_shift): Likewise.
13534 * tree-vect-patterns.cc (vect_recog_rotate_pattern): Likewise.
13535 * tree-vect-slp.cc (can_duplicate_and_interleave_p): Likewise.
13536 (vect_transform_slp_perm_load): Likewise.
13537 (vectorizable_slp_permutation): Likewise.
13538 * tree-vect-stmts.cc (perm_mask_for_reverse): Likewise.
13539 (vectorizable_bswap): Likewise.
13540 (scan_store_can_perm_p): Likewise.
13541 (vect_gen_perm_mask_checked): Likewise.
13543 2022-05-24 H.J. Lu <hjl.tools@gmail.com>
13546 * config/i386/i386.opt: Remove Undocumented.
13547 * doc/invoke.texi: Document -mcet-switch.
13549 2022-05-24 Andrew Stubbs <ams@codesourcery.com>
13551 * config.gcc (amdgcn): Accept --with-arch=gfx908 and gfx90a.
13552 * config/gcn/gcn-opts.h (enum gcn_isa): New.
13553 (TARGET_GCN3): Use enum gcn_isa.
13554 (TARGET_GCN3_PLUS): Likewise.
13555 (TARGET_GCN5): Likewise.
13556 (TARGET_GCN5_PLUS): Likewise.
13557 (TARGET_CDNA1): New.
13558 (TARGET_CDNA1_PLUS): New.
13559 (TARGET_CDNA2): New.
13560 (TARGET_CDNA2_PLUS): New.
13561 (TARGET_M0_LDS_LIMIT): New.
13562 (TARGET_PACKED_WORK_ITEMS): New.
13563 * config/gcn/gcn.cc (gcn_isa): Change to enum gcn_isa.
13564 (gcn_option_override): Recognise CDNA ISA variants.
13565 (gcn_omp_device_kind_arch_isa): Support gfx90a.
13566 (gcn_expand_prologue): Make m0 init optional.
13567 Add support for packed work items.
13568 (output_file_start): Support gfx90a.
13569 (gcn_hsa_declare_function_name): Support gfx90a metadata.
13570 * config/gcn/gcn.h (TARGET_CPU_CPP_BUILTINS):Add __CDNA1__ and
13572 * config/gcn/gcn.md (<su>mulsi3_highpart): Use TARGET_GCN5_PLUS.
13573 (<su>mulsi3_highpart_imm): Likewise.
13574 (<su>mulsidi3): Likewise.
13575 (<su>mulsidi3_imm): Likewise.
13576 * config/gcn/gcn.opt (gpu_type): Add gfx90a.
13577 * config/gcn/mkoffload.cc (EF_AMDGPU_MACH_AMDGCN_GFX90a): New.
13578 (main): Support gfx90a.
13579 * config/gcn/t-gcn-hsa: Add gfx90a multilib.
13580 * config/gcn/t-omp-device: Add gfx90a isa.
13582 2022-05-24 Andrew Stubbs <ams@codesourcery.com>
13584 * config.in: Regenerate.
13585 * config/gcn/gcn-hsa.h (X_FIJI): Delete.
13593 (NO_XNACK): New macro.
13594 (NO_SRAM_ECC): New macro.
13595 (SRAMOPT): Keep only v4 variant.
13596 (HSACO3_SELECT_OPT): Delete.
13597 (DRIVER_SELF_SPECS): Delete.
13598 (ASM_SPEC): Remove LLVM 9 support.
13599 * config/gcn/gcn-valu.md
13600 (gather<mode>_insn_2offsets<exec>): Remove assembler bug workaround.
13601 (scatter<mode>_insn_2offsets<exec_scatter>): Likewise.
13602 * config/gcn/gcn.cc (output_file_start): Remove LLVM 9 support.
13603 (print_operand_address): Remove assembler bug workaround.
13604 * config/gcn/mkoffload.cc (EF_AMDGPU_XNACK_V3): Delete.
13605 (EF_AMDGPU_SRAM_ECC_V3): Delete.
13606 (SET_XNACK_ON): Delete v3 variants.
13607 (SET_XNACK_OFF): Delete v3 variants.
13608 (TEST_XNACK): Delete v3 variants.
13609 (SET_SRAM_ECC_ON): Delete v3 variants.
13610 (SET_SRAM_ECC_ANY): Delete v3 variants.
13611 (SET_SRAM_ECC_OFF): Delete v3 variants.
13612 (SET_SRAM_ECC_UNSUPPORTED): Delete v3 variants.
13613 (TEST_SRAM_ECC_ANY): Delete v3 variants.
13614 (TEST_SRAM_ECC_ON): Delete v3 variants.
13615 (copy_early_debug_info): Remove v3 support.
13616 (main): Remove v3 support.
13617 * configure: Regenerate.
13618 * configure.ac: Replace all GCN feature checks with a version check.
13620 2022-05-24 Roger Sayle <roger@nextmovesoftware.com>
13622 * config/i386/i386.md (peephole2): Convert xor;neg;adc;neg,
13623 i.e. a double word negation of a zero extended operand, to
13626 2022-05-24 Roger Sayle <roger@nextmovesoftware.com>
13628 PR tree-optimization/105668
13629 * config/i386/i386-expand.cc (ix86_expand_sse_movcc): Support
13630 V1TImode, just like V2DImode.
13631 * config/i386/sse.md (vcond_mask_<mode><sseintvecmodelower>):
13632 Use VI_128 mode iterator instead of VI124_128 to include V2DI.
13633 (vcond_mask_v2div2di): Delete.
13634 (vcond_mask_v1tiv1ti): New define_expand.
13636 2022-05-24 Roger Sayle <roger@nextmovesoftware.com>
13638 * genpreds.cc (write_lookup_constraint_1): Avoid generating a call
13639 to strncmp for strings of length one.
13641 2022-05-24 ShiYulong <shiyulong@iscas.ac.cn>
13643 * config/riscv/predicates.md (imm5_operand): Add a new operand type for
13644 prefetch instructions.
13645 * config/riscv/riscv-builtins.cc (AVAIL): Add new AVAILs for CMO ISA
13647 (RISCV_ATYPE_SI): New.
13648 (RISCV_ATYPE_DI): New.
13649 * config/riscv/riscv-ftypes.def (0): New.
13651 * config/riscv/riscv.md (riscv_clean_<mode>): New.
13652 (riscv_flush_<mode>): New.
13653 (riscv_inval_<mode>): New.
13654 (riscv_zero_<mode>): New.
13656 (riscv_prefetchi_<mode>): New.
13657 * config/riscv/riscv-cmo.def: New file.
13659 2022-05-24 ShiYulong <shiyulong@iscas.ac.cn>
13661 * common/config/riscv/riscv-common.cc: Add zicbom, zicboz, zicbop extensions.
13662 * config/riscv/riscv-opts.h (MASK_ZICBOZ): New.
13663 (MASK_ZICBOM): New.
13664 (MASK_ZICBOP): New.
13665 (TARGET_ZICBOZ): New.
13666 (TARGET_ZICBOM): New.
13667 (TARGET_ZICBOP): New.
13668 * config/riscv/riscv.opt (riscv_zicmo_subext): New.
13670 2022-05-24 David Malcolm <dmalcolm@redhat.com>
13672 * tree-vect-slp-patterns.cc: Add "final" and "override" to
13673 vect_pattern::build impls as appropriate.
13675 2022-05-24 David Malcolm <dmalcolm@redhat.com>
13677 * ipa-cp.cc: Add "final" and "override" to call_summary_base vfunc
13678 implementations, removing redundant "virtual" as appropriate.
13679 * ipa-fnsummary.h: Likewise.
13680 * ipa-modref.cc: Likewise.
13681 * ipa-param-manipulation.cc: Likewise.
13682 * ipa-profile.cc: Likewise.
13683 * ipa-prop.h: Likewise.
13684 * ipa-pure-const.cc: Likewise.
13685 * ipa-reference.cc: Likewise.
13686 * ipa-sra.cc: Likewise.
13687 * symbol-summary.h: Likewise.
13688 * symtab-thunks.cc: Likewise.
13690 2022-05-24 Martin Liska <mliska@suse.cz>
13693 2022-05-24 Martin Liska <mliska@suse.cz>
13695 * expmed.cc (emit_store_flag_1): Mitigate -Wmaybe-uninitialized
13698 2022-05-24 Martin Liska <mliska@suse.cz>
13700 * expmed.cc (emit_store_flag_1): Mitigate -Wmaybe-uninitialized
13703 2022-05-24 Bruno Haible <bruno@clisp.org>
13706 * doc/install.texi (Configuration): Add more details about --with-zstd.
13707 Document --with-zstd-include and --with-zstd-lib
13709 2022-05-24 Richard Biener <rguenther@suse.de>
13711 PR middle-end/105711
13712 * expmed.cc (extract_bit_field_as_subreg): Add op0_mode parameter
13714 (extract_bit_field_1): Pass down the mode of op0 to
13715 extract_bit_field_as_subreg.
13717 2022-05-24 Vineet Gupta <vineetg@rivosinc.com>
13719 * config/riscv/riscv.cc: (struct riscv_tune_param): Add
13721 (rocket_tune_info): Add default fmv_cost 8.
13722 (sifive_7_tune_info): Ditto.
13723 (thead_c906_tune_info): Ditto.
13724 (optimize_size_tune_info): Ditto.
13725 (riscv_register_move_cost): Use fmv_cost for int<->fp moves.
13727 2022-05-24 Jakub Jelinek <jakub@redhat.com>
13730 * omp-builtins.def (BUILT_IN_GOMP_TASKWAIT_DEPEND_NOWAIT): New
13732 * gimplify.cc (gimplify_omp_task): Diagnose taskwait with nowait
13733 clause but no depend clauses.
13734 * omp-expand.cc (expand_taskwait_call): Use
13735 BUILT_IN_GOMP_TASKWAIT_DEPEND_NOWAIT rather than
13736 BUILT_IN_GOMP_TASKWAIT_DEPEND if nowait clause is present.
13738 2022-05-24 Richard Biener <rguenther@suse.de>
13740 PR tree-optimization/100221
13741 * tree-ssa-dse.cc (contains_phi_arg): New function.
13742 (dse_classify_store): Postpone PHI defs that feed another PHI in defs.
13744 2022-05-24 Richard Biener <rguenther@suse.de>
13746 PR tree-optimization/105629
13747 * tree-ssa-phiopt.cc (spaceship_replacement): Allow
13748 a sign-extending conversion.
13750 2022-05-24 Kewen Lin <linkw@linux.ibm.com>
13753 * config/rs6000/rs6000-p8swap.cc (union_defs): Assert def_insn can't
13755 (union_uses): Skip debug use_insn.
13757 2022-05-23 Vineet Gupta <vineetg@rivosinc.com>
13759 * config/riscv/predicates.md (const_0_operand): Remove
13761 * config/riscv/riscv.cc (riscv_rtx_costs): Add check for
13763 * config/riscv/riscv.h (TARGET_SUPPORTS_WIDE_INT): New define.
13765 2022-05-23 Mayshao <mayshao-oc@zhaoxin.com>
13767 * common/config/i386/cpuinfo.h (get_zhaoxin_cpu): Detect
13768 the specific type of Zhaoxin CPU, and return Zhaoxin CPU name.
13769 (cpu_indicator_init): Handle Zhaoxin processors.
13770 * common/config/i386/i386-common.cc: Add lujiazui.
13771 * common/config/i386/i386-cpuinfo.h (enum processor_vendor): Add
13773 (enum processor_types): Add ZHAOXIN_FAM7H.
13774 (enum processor_subtypes): Add ZHAOXIN_FAM7H_LUJIAZUI.
13775 * config.gcc: Add lujiazui.
13776 * config/i386/cpuid.h (signature_SHANGHAI_ebx): Add
13777 Signatures for zhaoxin
13778 (signature_SHANGHAI_ecx): Ditto.
13779 (signature_SHANGHAI_edx): Ditto.
13780 * config/i386/driver-i386.cc (host_detect_local_cpu): Let
13781 -march=native recognize lujiazui processors.
13782 * config/i386/i386-c.cc (ix86_target_macros_internal): Add lujiazui.
13783 * config/i386/i386-options.cc (m_LUJIAZUI): New_definition.
13784 * config/i386/i386.h (enum processor_type): Ditto.
13785 * config/i386/i386.md: Add lujiazui.
13786 * config/i386/x86-tune-costs.h (struct processor_costs): Add
13788 * config/i386/x86-tune-sched.cc (ix86_issue_rate): Add lujiazui.
13789 (ix86_adjust_cost): Ditto.
13790 * config/i386/x86-tune.def (X86_TUNE_SCHEDULE): Add lujiazui Tunnings.
13791 (X86_TUNE_PARTIAL_REG_DEPENDENCY): Ditto.
13792 (X86_TUNE_SSE_PARTIAL_REG_DEPENDENCY): Ditto.
13793 (X86_TUNE_SSE_PARTIAL_REG_FP_CONVERTS_DEPENDENCY): Ditto.
13794 (X86_TUNE_SSE_PARTIAL_REG_CONVERTS_DEPENDENCY): Ditto.
13795 (X86_TUNE_MOVX): Ditto.
13796 (X86_TUNE_MEMORY_MISMATCH_STALL): Ditto.
13797 (X86_TUNE_FUSE_CMP_AND_BRANCH_32): Ditto.
13798 (X86_TUNE_FUSE_CMP_AND_BRANCH_64): Ditto.
13799 (X86_TUNE_FUSE_CMP_AND_BRANCH_SOFLAGS): Ditto.
13800 (X86_TUNE_FUSE_ALU_AND_BRANCH): Ditto.
13801 (X86_TUNE_ACCUMULATE_OUTGOING_ARGS): Ditto.
13802 (X86_TUNE_USE_LEAVE): Ditto.
13803 (X86_TUNE_PUSH_MEMORY): Ditto.
13804 (X86_TUNE_LCP_STALL): Ditto.
13805 (X86_TUNE_USE_INCDEC): Ditto.
13806 (X86_TUNE_INTEGER_DFMODE_MOVES): Ditto.
13807 (X86_TUNE_OPT_AGU): Ditto.
13808 (X86_TUNE_PREFER_KNOWN_REP_MOVSB_STOSB): Ditto.
13809 (X86_TUNE_MISALIGNED_MOVE_STRING_PRO_EPILOGUES): Ditto.
13810 (X86_TUNE_USE_SAHF): Ditto.
13811 (X86_TUNE_USE_BT): Ditto.
13812 (X86_TUNE_AVOID_FALSE_DEP_FOR_BMI): Ditto.
13813 (X86_TUNE_ONE_IF_CONV_INSN): Ditto.
13814 (X86_TUNE_AVOID_MFENCE): Ditto.
13815 (X86_TUNE_EXPAND_ABS): Ditto.
13816 (X86_TUNE_USE_SIMODE_FIOP): Ditto.
13817 (X86_TUNE_USE_FFREEP): Ditto.
13818 (X86_TUNE_EXT_80387_CONSTANTS): Ditto.
13819 (X86_TUNE_SSE_UNALIGNED_LOAD_OPTIMAL): Ditto.
13820 (X86_TUNE_SSE_UNALIGNED_STORE_OPTIMAL): Ditto.
13821 (X86_TUNE_SSE_TYPELESS_STORES): Ditto.
13822 (X86_TUNE_SSE_LOAD0_BY_PXOR): Ditto.
13823 * doc/extend.texi: Add details about lujiazui.
13824 * doc/invoke.texi: Add details about lujiazui.
13825 * config/i386/lujiazui.md: Introduce lujiazui cpu and include new md file.
13827 2022-05-23 Martin Liska <mliska@suse.cz>
13829 * config/tilepro/gen-mul-tables.cc (ARRAY_SIZE): Add new macro.
13831 2022-05-23 Richard Biener <rguenther@suse.de>
13833 * tree-ssa-forwprop.cc (forward_propagate_into_cond): Remove.
13834 (pass_forwprop::execute): Do not propagate into COND_EXPR conditions.
13836 2022-05-23 Richard Biener <rguenther@suse.de>
13838 * gimple-expr.cc (is_gimple_condexpr): Remove.
13839 * gimple-expr.h (is_gimple_condexpr): Likewise.
13840 * gimplify.cc (gimplify_expr): Remove is_gimple_condexpr usage.
13841 * tree-if-conv.cc (set_bb_predicate): Likewie.
13842 (add_to_predicate_list): Likewise.
13843 (gen_phi_arg_condition): Likewise.
13844 (predicate_scalar_phi): Likewise.
13845 (predicate_statements): Likewise.
13847 2022-05-23 Richard Biener <rguenther@suse.de>
13849 * gimple-expr.cc (is_gimple_condexpr): Equate to is_gimple_val.
13850 * gimplify.cc (gimplify_pure_cond_expr): Gimplify the condition
13852 * gimple-fold.cc (valid_gimple_rhs_p): Simplify.
13853 * tree-cfg.cc (verify_gimple_assign_ternary): Likewise.
13854 * gimple-loop-interchange.cc (loop_cand::undo_simple_reduction):
13855 Build the condition of the COND_EXPR separately.
13856 * tree-ssa-loop-im.cc (move_computations_worker): Likewise.
13857 * tree-vect-generic.cc (expand_vector_condition): Likewise.
13858 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
13860 * vr-values.cc (simplify_using_ranges::simplify): Likewise.
13861 * tree-vect-patterns.cc: Add comment indicating we are
13862 building invalid COND_EXPRs and why.
13863 * omp-expand.cc (expand_omp_simd): Gimplify the condition
13864 to the COND_EXPR separately.
13865 (expand_omp_atomic_cas): Note part that should be unreachable
13867 * tree-ssa-forwprop.cc (forward_propagate_into_cond): Adjust
13868 condition for valid replacements.
13869 * tree-if-conv.cc (predicate_bbs): Simulate previous
13870 re-folding of the condition in folded COND_EXPRs which
13871 is necessary because of unfolded GIMPLE_CONDs in the IL
13872 as in for example gcc.dg/fold-bopcond-1.c.
13873 * gimple-range-gori.cc (gori_compute::condexpr_adjust):
13874 Handle that the comparison is now in the def stmt of
13875 the select operand. Required by gcc.dg/pr104526.c.
13877 2022-05-23 Tobias Burnus <tobias@codesourcery.com>
13880 * langhooks-def.h (lhd_omp_array_size): New.
13881 (LANG_HOOKS_OMP_ARRAY_SIZE): Define.
13882 (LANG_HOOKS_DECLS): Add it.
13883 * langhooks.cc (lhd_omp_array_size): New.
13884 * langhooks.h (struct lang_hooks_for_decls): Add hook.
13885 * omp-low.cc (scan_sharing_clauses, lower_omp_target):
13886 Handle GOMP_MAP_FIRSTPRIVATE for array descriptors.
13888 2022-05-23 Roger Sayle <roger@nextmovesoftware.com>
13890 * config/i386/i386.cc (ix86_rtx_costs) <case AND>: Split from
13891 XOR/IOR case. Account for two instructions for double-word
13892 operations. In case of vector pandn, account for single
13893 instruction. Likewise for integer andn with TARGET_BMI.
13894 <case NOT>: Vector NOT requires more than 1 instruction (pxor).
13895 <case NEG>: Double-word negation requires 3 instructions.
13897 2022-05-23 Tsukasa OI <research_trasio@irq.a4lg.com>
13899 * common/config/riscv/riscv-common.cc (riscv_supported_std_ext):
13900 Fix "K" extension prefix to be placed before "J".
13901 * config/riscv/arch-canonicalize: Likewise.
13903 2022-05-23 liuhongt <hongtao.liu@intel.com>
13905 * config/i386/x86-tune-costs.h (skylake_cost): Increase gpr
13906 <-> mask cost from 5 to 6.
13907 (icelake_cost): Ditto.
13909 2022-05-20 Wilco Dijkstra <wilco.dijkstra@arm.com>
13911 * config/aarch64/aarch64.md
13912 (and_<SHIFT:optab><mode>3_compare0): Support rotate left.
13913 (and_<SHIFT:optab>si3_compare0_uxtw): Likewise.
13914 (<LOGICAL:optab>_<SHIFT:optab><mode>3): Likewise.
13915 (<LOGICAL:optab>_<SHIFT:optab>si3_uxtw): Likewise.
13916 (one_cmpl_<optab><mode>2): Likewise.
13917 (<LOGICAL:optab>_one_cmpl_<SHIFT:optab><mode>3): Likewise.
13918 (<LOGICAL:optab>_one_cmpl_<SHIFT:optab>sidi_uxtw): New pattern.
13919 (eor_one_cmpl_<SHIFT:optab><mode>3_alt): Support rotate left.
13920 (eor_one_cmpl_<SHIFT:optab>sidi3_alt_ze): Likewise.
13921 (and_one_cmpl_<SHIFT:optab><mode>3_compare0): Likewise.
13922 (and_one_cmpl_<SHIFT:optab>si3_compare0_uxtw): Likewise.
13923 (and_one_cmpl_<SHIFT:optab><mode>3_compare0_no_reuse): Likewise.
13924 (and_<SHIFT:optab><mode>3nr_compare0): Likewise.
13925 (*<optab>si3_insn_uxtw): Use SHIFT_no_rotate.
13926 (rolsi3_insn_uxtw): New pattern.
13927 * config/aarch64/iterators.md (SHIFT): Add rotate left.
13928 (SHIFT_no_rotate): Add new iterator.
13929 (SHIFT:shift): Print rotate left as ror.
13930 (is_rotl): Add test for left rotate.
13932 2022-05-20 Wilco Dijkstra <wilco.dijkstra@arm.com>
13934 * config.gcc (aarch64*-*-*): Simplify --with-cpu and --with-arch
13935 processing. Add support for architectural extensions.
13936 * config/aarch64/aarch64.h (TARGET_CPU_DEFAULT): Remove
13937 AARCH64_CPU_DEFAULT_FLAGS.
13938 (TARGET_CPU_NBITS): Remove.
13939 (TARGET_CPU_MASK): Remove.
13940 * config/aarch64/aarch64.cc (AARCH64_CPU_DEFAULT_FLAGS): Remove define.
13941 (get_tune_cpu): Assert CPU is always valid.
13942 (get_arch): Assert architecture is always valid.
13943 (aarch64_override_options): Cleanup CPU selection code and simplify logic.
13944 (aarch64_option_restore): Remove unnecessary checks on tune.
13946 2022-05-20 David Malcolm <dmalcolm@redhat.com>
13948 * config/aarch64/aarch64-sve-builtins-base.cc: Replace uses of
13949 "FINAL" and "OVERRIDE" with "final" and "override".
13950 * config/aarch64/aarch64-sve-builtins-functions.h: Likewise.
13951 * config/aarch64/aarch64-sve-builtins-shapes.cc: Likewise.
13952 * config/aarch64/aarch64-sve-builtins-sve2.cc: Likewise.
13953 * diagnostic-path.h: Likewise.
13954 * digraph.cc: Likewise.
13955 * gcc-rich-location.h: Likewise.
13956 * gimple-array-bounds.cc: Likewise.
13957 * gimple-loop-versioning.cc: Likewise.
13958 * gimple-range-cache.cc: Likewise.
13959 * gimple-range-cache.h: Likewise.
13960 * gimple-range-fold.cc: Likewise.
13961 * gimple-range-fold.h: Likewise.
13962 * gimple-range-tests.cc: Likewise.
13963 * gimple-range.h: Likewise.
13964 * gimple-ssa-evrp.cc: Likewise.
13965 * input.cc: Likewise.
13966 * json.h: Likewise.
13967 * read-rtl-function.cc: Likewise.
13968 * tree-complex.cc: Likewise.
13969 * tree-diagnostic-path.cc: Likewise.
13970 * tree-ssa-ccp.cc: Likewise.
13971 * tree-ssa-copy.cc: Likewise.
13972 * tree-vrp.cc: Likewise.
13973 * value-query.h: Likewise.
13974 * vr-values.h: Likewise.
13976 2022-05-20 Marcel Vollweiler <marcel@codesourcery.com>
13978 * omp-low.cc (omp_runtime_api_call): Added target_memcpy_async and
13979 target_memcpy_rect_async to omp_runtime_apis array.
13981 2022-05-20 Christophe Lyon <christophe.lyon@arm.com>
13983 * doc/sourcebuild.texi (Decimal floating point attributes): Document
13984 dfp_bid effective-target.
13986 2022-05-20 Christophe Lyon <christophe.lyon@arm.com>
13988 * config/aarch64/aarch64.cc
13989 (aarch64_split_128bit_move): Handle DFP modes.
13990 (aarch64_mode_valid_for_sched_fusion_p): Likewise.
13991 (aarch64_classify_address): Likewise.
13992 (aarch64_legitimize_address_displacement): Likewise.
13993 (aarch64_reinterpret_float_as_int): Likewise.
13994 (aarch64_float_const_zero_rtx_p): Likewise.
13995 (aarch64_can_const_movi_rtx_p): Likewise.
13996 (aarch64_anchor_offset): Likewise.
13997 (aarch64_secondary_reload): Likewise.
13998 (aarch64_rtx_costs): Likewise.
13999 (aarch64_legitimate_constant_p): Likewise.
14000 (aarch64_gimplify_va_arg_expr): Likewise.
14001 (aapcs_vfp_sub_candidate): Likewise.
14002 (aarch64_vfp_is_call_or_return_candidate): Likewise.
14003 (aarch64_output_scalar_simd_mov_immediate): Likewise.
14004 (aarch64_gen_adjusted_ldpstp): Likewise.
14005 (aarch64_scalar_mode_supported_p): Accept DFP modes if enabled.
14006 * config/aarch64/aarch64.md
14007 (movsf_aarch64): Use SFD iterator and rename into
14009 (movdf_aarch64): Use DFD iterator and rename into
14011 (movtf_aarch64): Use TFD iterator and rename into
14013 (split pattern for move TF mode): Use TFD iterator.
14014 * config/aarch64/iterators.md
14015 (GPF_TF_F16_MOV): Add DFP modes.
14016 (SFD, DFD, TFD): New iterators.
14017 (GPF_TF): Add DFP modes.
14018 (TX, DX, DX2): Likewise.
14020 2022-05-20 Christophe Lyon <christophe.lyon@arm.com>
14022 * configure: Regenerate.
14024 2022-05-19 Roger Sayle <roger@nextmovesoftware.com>
14026 PR middle-end/98865
14027 * expr.cc (expand_expr_real_2) [MULT_EXPR]: Expand X*Y as X&Y
14028 when both X and Y are [0, 1], X*Y as X&-Y when Y is [0,1] and
14029 likewise X*Y as -X&Y when X is [0,1] using tree_nonzero_bits.
14031 2022-05-19 Will Schmidt <will_schmidt@vnet.ibm.com>
14033 * config/rs6000/rs6000-builtins.def: Rephrase
14034 to remove RS6000_BTC_SPECIAL from comment.
14035 * config/rs6000/rs6000.h (RS6000_BTC_UNARY, RS6000_BTC_BINARY,
14036 RS6000_BTC_TERNARY, RS6000_BTC_QUATERNARY,
14037 RS6000_BTC_QUINARY, RS6000_BTC_SENARY, RS6000_BTC_OPND_MASK,
14038 RS6000_BTC_SPECIAL, RS6000_BTC_PREDICATE, RS6000_BTC_ABS,
14039 RS6000_BTC_DST, RS6000_BTC_TYPE_MASK, RS6000_BTC_MISC,
14040 RS6000_BTC_CONST, RS6000_BTC_PURE, RS6000_BTC_FP,
14041 RS6000_BTC_QUAD, RS6000_BTC_PAIR, RS6000_BTC_QUADPAIR,
14042 RS6000_BTC_ATTR_MASK, RS6000_BTC_SPR, RS6000_BTC_VOID,
14043 RS6000_BTC_CR, RS6000_BTC_OVERLOADED, RS6000_BTC_GIMPLE,
14044 RS6000_BTC_MISC_MASK, RS6000_BTC_MEM, RS6000_BTC_SAT,
14045 RS6000_BTM_ALWAYS): Delete.
14047 2022-05-19 Richard Biener <rguenther@suse.de>
14049 * omp-expand.cc (expand_omp_atomic_cas): Do not short-cut
14050 computation of the new value.
14052 2022-05-19 Richard Biener <rguenther@suse.de>
14054 * tree-ssa-pre.cc (get_or_alloc_expression_id): Remove.
14055 (add_to_value): Use get_expression_id.
14056 (bitmap_insert_into_set): Likewise.
14057 (bitmap_value_insert_into_set): Likewise.
14059 2022-05-19 David Malcolm <dmalcolm@redhat.com>
14061 * doc/invoke.texi (-fanalyzer-checker=): Add
14062 -Wanalyzer-va-list-leak and -Wanalyzer-va-list-use-after-va-end to
14063 the list of analyzer warnings disabled by
14064 -fanalyzer-checker=taint.
14066 2022-05-19 Jakub Jelinek <jakub@redhat.com>
14069 * cfgexpand.cc (expand_debug_expr): For VAR_DECL, punt for
14070 global vars without symtab node even when they have DECL_RTL
14073 2022-05-19 Jakub Jelinek <jakub@redhat.com>
14076 * pointer-query.cc (gimple_parm_array_size): Return NULL if var
14077 doesn't have pointer or reference type.
14079 2022-05-18 Marek Polacek <polacek@redhat.com>
14082 * doc/invoke.texi: Document -Wenum-int-mismatch.
14084 2022-05-18 Uros Bizjak <ubizjak@gmail.com>
14086 * config/i386/gnu-user-common.h (defined): Only define
14087 TARGET_CAN_SPLIT_STACK for glibc targets.
14088 * config/i386/gnu.h (defined): Ditto.
14090 2022-05-18 Roger Sayle <roger@nextmovesoftware.com>
14092 * config/i386/i386.cc (ix86_rtx_costs) [MULT]: When mode size
14093 is wider than word_mode, a multiplication costs three word_mode
14094 multiplications and two word_mode additions.
14096 2022-05-18 Roger Sayle <roger@nextmovesoftware.com>
14098 * config/i386/i386.md (define_split): Split *andsi_1
14099 and *andn_si_ccno after reload with -Oz.
14101 2022-05-18 Frederik Harwath <frederik@codesourcery.com>
14103 * graphite-scop-detection.cc (scop_detection::can_represent_loop):
14104 Output reason for failure to dump file.
14105 (scop_detection::harmful_loop_in_region): Likewise.
14106 (scop_detection::graphite_can_represent_expr): Likewise.
14107 (scop_detection::stmt_has_simple_data_refs_p): Likewise.
14108 (scop_detection::stmt_simple_for_scop_p): Likewise.
14109 (print_sese_loop_numbers): New function.
14110 (scop_detection::add_scop): Use from here.
14112 2022-05-18 liuhongt <hongtao.liu@intel.com>
14114 PR middle-end/103462
14115 * match.pd (bitwise_induction_p): New match.
14116 * tree-scalar-evolution.cc (gimple_bitwise_induction_p):
14118 (analyze_and_compute_bitwise_induction_effect): New function.
14119 (enum bit_op_kind): New enum.
14120 (final_value_replacement_loop): Enhanced to handle bitwise
14123 2022-05-18 Haochen Gui <guihaoc@gcc.gnu.org>
14126 * config/rs6000/rs6000.md (*subfsi3_carry_in_xx_64): New.
14128 2022-05-18 liuhongt <hongtao.liu@intel.com>
14131 * config/i386/i386.md (*bmi2_bzhi_zero_extendsidi_4): New
14134 2022-05-18 liuhongt <hongtao.liu@intel.com>
14137 * config/i386/i386-expand.cc (ix86_expand_branch): Use ptest
14138 for QImode when code is EQ or NE.
14139 * config/i386/i386.md (cbranchoi4): New expander.
14141 2022-05-18 Peter Bergner <bergner@linux.ibm.com>
14142 Segher Boessenkool <segher@kernel.crashing.org>
14145 * config/rs6000/mma.md (mma_<vv>, mma_<avv>, mma_<pv>, mma_<apv>,
14146 mma_<vvi4i4i8>, mma_<avvi4i4i8>, mma_<vvi4i4i2>, mma_<avvi4i4i2>,
14147 mma_<vvi4i4>, mma_<avvi4i4>, mma_<pvi4i2>, mma_<apvi4i2>,
14148 mma_<vvi4i4i4>, mma_<avvi4i4i4>): Replace "wa" constraints with "v,?wa".
14149 Update other operands accordingly.
14151 2022-05-17 Marek Polacek <polacek@redhat.com>
14153 * godump.cc (go_output_typedef): Use the DECL_INITIAL of the TREE_VALUE.
14155 2022-05-17 Pat Haugen <pthaugen@linux.ibm.com>
14158 * config/rs6000/rs6000-call.cc (rs6000_function_arg_advance_1): Bump
14159 register count when not splitting IEEE 128-bit Complex.
14161 2022-05-17 Tobias Burnus <tobias@codesourcery.com>
14163 * omp-low.cc (check_omp_nesting_restrictions): Skip warning for
14164 target inside target if inner is reverse offload.
14166 2022-05-17 Tobias Burnus <tobias@codesourcery.com>
14168 * config/gcn/mkoffload.cc (process_obj): Revert: Use ARRAY_SIZE.
14169 * config/nvptx/mkoffload.cc (process): Likewise.
14171 2022-05-17 Andrew MacLeod <amacleod@redhat.com>
14173 * Makefile.in (OBJS): Add gimple-range-side-effect.o.
14174 * gimple-range-cache.cc (non_null_ref::non_null_ref): Delete.
14175 (non_null_ref::~non_null_ref): Delete.
14176 (non_null_ref::set_nonnull): Delete.
14177 (non_null_ref::non_null_deref_p): Delete.
14178 (non_null_ref::process_name): Delete.
14179 (ranger_cache::ranger_cache): Initialize m_exit object.
14180 (ranger_cache::fill_block_cache): Use m_exit object intead of nonnull.
14181 (ranger_cache::range_from_dom): Use side_effect class and m_exit object.
14182 (ranger_cache::update_to_nonnull): Delete.
14183 (non_null_loadstore): Delete.
14184 (ranger_cache::block_apply_nonnull): Delete.
14185 (ranger_cache::apply_side_effects): New.
14186 * gimple-range-cache.h (class non_null_ref): Delete.
14187 (non_null_ref::adjust_range): Delete.
14188 (class ranger_cache): Adjust prototypes, add side effect manager.
14189 * gimple-range-path.cc (path_range_query::range_defined_in_block): Use
14190 side effect manager for queries.
14191 (path_range_query::adjust_for_non_null_uses): Ditto.
14192 * gimple-range-path.h (class path_range_query): Delete non_null_ref.
14193 * gimple-range-side-effect.cc: New.
14194 * gimple-range-side-effect.h: New.
14195 * gimple-range.cc (gimple_ranger::gimple_ranger): Update contructor.
14196 (gimple_ranger::range_of_expr): Check def block for override value.
14197 (gimple_ranger::range_on_entry): Don't scan dominators for non-null.
14198 (gimple_ranger::range_on_edge): Check for outgoing side-effects.
14199 (gimple_ranger::register_side_effects): Call apply_side_effects.
14200 (enable_ranger): Update contructor.
14201 * gimple-range.h (class gimple_ranger): Update prototype.
14202 (enable_ranger): Update prototype.
14203 * tree-vrp.cc (execute_ranger_vrp): Invoke without immediate-use flag.
14205 2022-05-17 Giuliano Belinassi <gbelinassi@suse.de>
14208 * targhooks.cc (default_print_patchable_function_entry_1): Handle COMDAT case.
14209 * varasm.cc (switch_to_comdat_section): New
14210 (handle_vtv_comdat_section): Call switch_to_comdat_section.
14211 * varasm.h: Declare switch_to_comdat_section.
14213 2022-05-17 Richard Biener <rguenther@suse.de>
14215 * cfgloopmanip.cc (duplicate_loop_body_to_header_edge): Do
14216 not clear bb->aux of the copied blocks.
14218 2022-05-17 Andrew MacLeod <amacleod@redhat.com>
14220 PR tree-optimization/105458
14221 * value-relation.cc (path_oracle::register_relation): Merge, then check
14224 2022-05-17 Uroš Bizjak <ubizjak@gmail.com>
14228 * config/i386/i386.md: Remove constraints when used with
14229 const_int_operand, const0_operand, const_1_operand, constm1_operand,
14230 const8_operand, const128_operand, const248_operand, const123_operand,
14231 const2367_operand, const1248_operand, const359_operand,
14232 const_4_or_8_to_11_operand, const48_operand, const_0_to_1_operand,
14233 const_0_to_3_operand, const_0_to_4_operand, const_0_to_5_operand,
14234 const_0_to_7_operand, const_0_to_15_operand, const_0_to_31_operand,
14235 const_0_to_63_operand, const_0_to_127_operand, const_0_to_255_operand,
14236 const_0_to_255_mul_8_operand, const_1_to_31_operand,
14237 const_1_to_63_operand, const_2_to_3_operand, const_4_to_5_operand,
14238 const_4_to_7_operand, const_6_to_7_operand, const_8_to_9_operand,
14239 const_8_to_11_operand, const_8_to_15_operand, const_10_to_11_operand,
14240 const_12_to_13_operand, const_12_to_15_operand, const_14_to_15_operand,
14241 const_16_to_19_operand, const_16_to_31_operand, const_20_to_23_operand,
14242 const_24_to_27_operand and const_28_to_31_operand.
14243 * config/i386/mmx.md: Ditto.
14244 * config/i386/sse.md: Ditto.
14245 * config/i386/subst.md: Ditto.
14246 * config/i386/sync.md: Ditto.
14248 2022-05-17 Thomas Schwinge <thomas@codesourcery.com>
14250 * diagnostic.cc: Don't advise to call 'abort' instead of
14252 * system.h: Advise to call 'internal_error' instead of 'abort' or
14255 2022-05-17 Frederik Harwath <frederik@codesourcery.com>
14257 * graphite-sese-to-poly.cc (build_poly_sr_1): Fix a typo and
14258 a reference to a variable which does not exist.
14259 * graphite-isl-ast-to-gimple.cc (gsi_insert_earliest): Fix typo
14262 2022-05-17 Frederik Harwath <frederik@codesourcery.com>
14264 * graphite-sese-to-poly.cc (isl_id_for_ssa_name): Rename to ...
14265 (isl_id_for_parameter): ... this new function name.
14266 (build_scop_context): Adjust function use.
14268 2022-05-17 Tobias Burnus <tobias@codesourcery.com>
14271 * config/gcn/t-omp-device (arch): Add 'amdgcn' besides existing 'gcn'.
14272 * config/gcn/gcn.cc (gcn_omp_device_kind_arch_isa): Likewise.
14274 2022-05-17 Jakub Jelinek <jakub@redhat.com>
14276 * tree-core.h (enum omp_clause_depend_kind): Add
14277 OMP_CLAUSE_DEPEND_INOUTSET.
14278 * tree-pretty-print.cc (dump_omp_clause): Handle
14279 OMP_CLAUSE_DEPEND_INOUTSET.
14280 * gimplify.cc (gimplify_omp_depend): Likewise.
14281 * omp-low.cc (lower_depend_clauses): Likewise.
14283 2022-05-17 Jakub Jelinek <jakub@redhat.com>
14286 * config/i386/sse.md (vec_cmpeqv2div2di, vec_cmpeqv1tiv1ti): Use
14287 andv4si3 only for EQ, for NE use iorv4si3 instead.
14289 2022-05-17 Richard Biener <rguenther@suse.de>
14291 PR tree-optimization/105618
14292 * tree-ssa-sink.cc (statement_sink_location): For virtual
14293 PHI uses ignore those defining the used virtual operand.
14295 2022-05-17 Jakub Jelinek <jakub@redhat.com>
14297 * tree-ssa-forwprop.cc (simplify_bitfield_ref): Spelling fix:
14298 hanlde -> handle. Fix up comment formatting.
14300 2022-05-17 liuhongt <hongtao.liu@intel.com>
14303 * config/i386/sse.md (*vec_concatv4si): Extend to ..
14304 (*vec_concat<mode>): .. V16QI and V8HImode.
14305 (*vec_concatv16qi_permt2): New pre_reload define_insn_and_split.
14306 (*vec_concatv8hi_permt2): Ditto.
14308 2022-05-17 liuhongt <hongtao.liu@intel.com>
14310 PR tree-optimization/105591
14311 * tree-ssa-forwprop.cc (simplify_bitfield_ref): Clamp
14312 vec_perm_expr index.
14314 2022-05-16 Jason Merrill <jason@redhat.com>
14317 * attribs.cc (decl_attributes): Fix broken typedefs here.
14319 2022-05-16 David Malcolm <dmalcolm@redhat.com>
14322 * Makefile.in (ANALYZER_OBJS): Add analyzer/varargs.o.
14323 * doc/invoke.texi: Add -Wanalyzer-va-arg-type-mismatch,
14324 -Wanalyzer-va-list-exhausted, -Wanalyzer-va-list-leak, and
14325 -Wanalyzer-va-list-use-after-va-end.
14327 2022-05-16 Richard Biener <rguenther@suse.de>
14329 * gimple-match.h (gimple_build): Move code_helper overloads ...
14330 * gimple-fold.h (gimple_build): ... here.
14331 (gimple_build): Transition to new worker API. Provide
14332 overloads from sequence-based API.
14333 (gimple_convert): Likewise.
14334 (gimple_convert_to_ptrofftype): Likewise.
14335 (gimple_build_vector_from_val): Likewise.
14336 (gimple_build_vector): Likewise.
14337 (gimple_build_round_up): Likewise.
14338 * gimple-fold.cc (gimple_build_insert_seq): New helper.
14339 (gimple_build): Use it. Transition combined_fn and code_helper
14341 (gimple_convert): Transition to new worker API.
14342 (gimple_convert_to_ptrofftype): Likewise.
14343 (gimple_build_vector_from_val): Likewise.
14344 (gimple_build_vector): Likewise.
14345 (gimple_build_round_up): Likewise.
14347 2022-05-16 Richard Biener <rguenther@suse.de>
14349 * gimple-match.h (code_helper): Move class ...
14350 * tree.h (code_helper): ... here.
14352 2022-05-16 Martin Liska <mliska@suse.cz>
14354 * opts-global.cc (write_langs): Add comment.
14356 2022-05-16 Eric Botcazou <ebotcazou@adacore.com>
14358 * dwarf2out.cc (loc_list_from_tree_1) <TRUTH_NOT_EXPR>: Do a logical
14359 instead of a bitwise negation.
14360 <COND_EXPR>: Swap the operands if the condition is TRUTH_NOT_EXPR.
14362 2022-05-16 Martin Liska <mliska@suse.cz>
14364 * attribs.cc (diag_attr_exclusions): Use ARRAY_SIZE.
14365 (decls_mismatched_attributes): Likewise.
14366 * builtins.cc (c_strlen): Likewise.
14367 * cfg.cc (DEF_BASIC_BLOCK_FLAG): Likewise.
14368 * common/config/aarch64/aarch64-common.cc (aarch64_option_init_struct): Likewise.
14369 * config/aarch64/aarch64-builtins.cc (aarch64_lookup_simd_builtin_type): Likewise.
14370 (aarch64_init_simd_builtin_types): Likewise.
14371 (aarch64_init_builtin_rsqrt): Likewise.
14372 * config/aarch64/aarch64.cc (is_madd_op): Likewise.
14373 * config/arm/arm-builtins.cc (arm_lookup_simd_builtin_type): Likewise.
14374 (arm_init_simd_builtin_types): Likewise.
14375 * config/avr/gen-avr-mmcu-texi.cc (mcus[ARRAY_SIZE): Likewise.
14376 (c_prefix): Likewise.
14378 * config/c6x/c6x.cc (N_SAVE_ORDER): Likewise.
14379 * config/darwin-c.cc (darwin_register_frameworks): Likewise.
14380 * config/gcn/mkoffload.cc (process_obj): Likewise.
14381 * config/i386/i386-builtins.cc (get_builtin_code_for_version): Likewise.
14382 (fold_builtin_cpu): Likewise.
14383 * config/m32c/m32c.cc (PUSHM_N): Likewise.
14384 * config/nvptx/mkoffload.cc (process): Likewise.
14385 * config/rs6000/driver-rs6000.cc (host_detect_local_cpu): Likewise.
14386 * config/s390/s390.cc (NR_C_MODES): Likewise.
14387 * config/tilepro/gen-mul-tables.cc (find_sequences): Likewise.
14388 (create_insn_code_compression_table): Likewise.
14389 * config/vms/vms.cc (NBR_CRTL_NAMES): Likewise.
14390 * diagnostic-format-json.cc (json_from_expanded_location): Likewise.
14391 * dwarf2out.cc (ARRAY_SIZE): Likewise.
14392 * genhooks.cc (emit_documentation): Likewise.
14393 (emit_init_macros): Likewise.
14394 * gimple-ssa-sprintf.cc (format_floating): Likewise.
14395 * gimple-ssa-warn-access.cc (memmodel_name): Likewise.
14396 * godump.cc (keyword_hash_init): Likewise.
14397 * hash-table.cc (hash_table_higher_prime_index): Likewise.
14398 * input.cc (for_each_line_table_case): Likewise.
14399 * ipa-free-lang-data.cc (free_lang_data): Likewise.
14400 * ipa-inline.cc (sanitize_attrs_match_for_inline_p): Likewise.
14401 * optc-save-gen.awk: Likewise.
14402 * spellcheck.cc (test_metric_conditions): Likewise.
14403 * tree-vect-slp-patterns.cc (sizeof): Likewise.
14404 (ARRAY_SIZE): Likewise.
14405 * tree.cc (build_common_tree_nodes): Likewise.
14407 2022-05-16 Martin Liska <mliska@suse.cz>
14409 * opts-global.cc (write_langs): Allocate at least one byte.
14411 2022-05-16 Richard Biener <rguenther@suse.de>
14413 * match.pd (A cmp B ? A : B -> min/max): New patterns
14414 carried over from fold_cond_expr_with_comparison.
14416 2022-05-16 liuhongt <hongtao.liu@intel.com>
14419 * config/i386/i386-expand.cc
14420 (expand_vec_perm_pslldq_psrldq_por): Fail when (d->perm[i] ==
14421 d->perm[i-1] + 1) && d->perm[i] == nelt && start != -1.
14423 2022-05-15 Uroš Bizjak <ubizjak@gmail.com>
14425 * config/i386/i386.md: Remove constraints when used with
14426 const_int_operand, const0_operand, const_1_operand, constm1_operand,
14427 const8_operand, const128_operand, const248_operand, const123_operand,
14428 const2367_operand, const1248_operand, const359_operand,
14429 const_4_or_8_to_11_operand, const48_operand, const_0_to_1_operand,
14430 const_0_to_3_operand, const_0_to_4_operand, const_0_to_5_operand,
14431 const_0_to_7_operand, const_0_to_15_operand, const_0_to_31_operand,
14432 const_0_to_63_operand, const_0_to_127_operand, const_0_to_255_operand,
14433 const_0_to_255_mul_8_operand, const_1_to_31_operand,
14434 const_1_to_63_operand, const_2_to_3_operand, const_4_to_5_operand,
14435 const_4_to_7_operand, const_6_to_7_operand, const_8_to_9_operand,
14436 const_8_to_11_operand, const_8_to_15_operand, const_10_to_11_operand,
14437 const_12_to_13_operand, const_12_to_15_operand, const_14_to_15_operand,
14438 const_16_to_19_operand, const_16_to_31_operand, const_20_to_23_operand,
14439 const_24_to_27_operand and const_28_to_31_operand.
14440 * config/i386/mmx.md: Ditto.
14441 * config/i386/sse.md: Ditto.
14442 * config/i386/subst.md: Ditto.
14443 * config/i386/sync.md: Ditto.
14445 2022-05-13 Roger Sayle <roger@nextmovesoftware.com>
14446 Uroš Bizjak <ubizjak@gmail.com>
14448 * config/i386/sse.md (vec_cmpeqv2div2di): Enable for TARGET_SSE2.
14449 For !TARGET_SSE4_1, expand as a V4SI vector comparison, followed
14450 by a pshufd and pand.
14451 (vec_cmpeqv1tiv1ti): New define_expand implementing V1TImode
14452 vector equality as a V2DImode vector comparison (see above),
14453 followed by a pshufd and pand.
14455 2022-05-13 Roger Sayle <roger@nextmovesoftware.com>
14457 PR tree-optimization/83907
14458 * tree-ssa-strlen.cc (handle_builtin_memset): Record a strinfo
14459 for memset with an constant char value.
14460 (handle_store): Improved handling of stores with a first byte
14461 of zero, but not storing_all_zeros_p.
14463 2022-05-13 Philipp Tomsich <philipp.tomsich@vrull.eu>
14464 Manolis Tsamis <manolis.tsamis@vrull.eu>
14466 * config/riscv/riscv.h (CLZ_DEFINED_VALUE_AT_ZERO): Implement.
14467 (CTZ_DEFINED_VALUE_AT_ZERO): Same.
14468 * doc/sourcebuild.texi: add documentation for RISC-V specific
14469 test target keywords
14471 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14473 PR tree-optimization/105597
14474 * range-op.cc (operator_minus::lhs_op1_relation): Use op1 instead
14475 of the lhs and make sure it is not undefined.
14477 2022-05-13 Sebastian Pop <spop@amazon.com>
14480 * config/aarch64/aarch64-protos.h (atomic_ool_names): Increase dimension
14482 * config/aarch64/aarch64.cc (aarch64_atomic_ool_func): Call
14483 memmodel_from_int and handle MEMMODEL_SYNC_*.
14484 (DEF0): Add __aarch64_*_sync functions.
14486 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14488 * gimple-range-fold.cc (fold_using_range::range_of_phi): Use new VREL_*
14490 * gimple-range-path.cc (maybe_register_phi_relation): Ditto.
14491 * range-op.cc (*::lhs_op1_relation): Return relation_kind, and use
14492 new VREL enumerated values.
14493 (*::lhs_op2_relation): Ditto.
14494 (*::op1_op2_relation): Ditto.
14495 (*::fold_range): Use new VREL enumerated values.
14496 (minus_op1_op2_relation_effect): Ditto.
14497 (range_relational_tests): Ditto.
14498 * range-op.h (fold_range, op1_range, op2_range): Use VREL_VARYING.
14499 (lhs_op1_relation, lhs_op2_relation, op1_op2_relation): Return
14501 (*_op1_op2_relation): Return relation_kind.
14502 (relop_early_resolve): Use VREL_UNDEFINED.
14503 * value-query.cc (range_query::query_relation): Use VREL_VARYING.
14504 * value-relation.cc (VREL_LAST): Change enumerated value.
14505 (vrel_range_assert): Delete.
14506 (print_relation): Remove range assert.
14507 (rr_negate_table): Adjust table to use new enumerated values..
14508 (relation_negate): Remove range assert.
14509 (rr_swap_table): Adjust.
14510 (relation_swap): Remove range assert.
14511 (rr_intersect_table): Adjust.
14512 (relation_intersect): Remove range assert.
14513 (rr_union_table): Adjust.
14514 (relation_union): Remove range assert.
14515 (rr_transitive_table): Adjust.
14516 (relation_transitive): Remove range assert.
14517 (equiv_oracle::query_relation): Use new VREL enumerated values.
14518 (equiv_oracle::register_relation): Ditto.
14519 (relation_oracle::register_stmt): Ditto.
14520 (dom_oracle::set_one_relation): Ditto.
14521 (dom_oracle::register_transitives): Ditto.
14522 (dom_oracle::query_relation): Ditto.
14523 (path_oracle::register_relation): Ditto.
14524 (path_oracle::query_relation): Ditto.
14525 * value-relation.h (enum relation_kind_t): New relation_kind.
14526 (*_op1_op2_relation): Adjust prototypes.
14528 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14530 * gimple-range-edge.cc (calc_switch_ranges): Check union return value.
14531 * value-range.cc (irange::legacy_verbose_union_): Add return value.
14532 (irange::irange_single_pair_union): New.
14533 (irange::irange_union): Add return value.
14534 * value-range.h (class irange): Adjust prototypes.
14536 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14538 * value-range.cc (irange::legacy_verbose_intersect): Add return value.
14539 (irange::irange_contains_p): New.
14540 (irange::irange_intersect): Add return value.
14541 * value-range.h (class irange): Adjust prototypes.
14543 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14545 * gimple-range-cache.cc (ranger_cache::get_global_range): Return the
14546 had_global value instead.
14548 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14550 PR tree-optimization/104547
14551 * gimple-range-fold.cc (fold_using_range::range_of_range_op): Add
14552 the op1/op2 relation to the relation call.
14553 * range-op.cc (*::lhs_op1_relation): Add param.
14554 (*::lhs_op2_relation): Ditto.
14555 (operator_minus::lhs_op1_relation): New.
14556 (range_relational_tests): Add relation param.
14557 * range-op.h (lhs_op1_relation, lhs_op2_relation): Adjust prototype.
14559 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14561 * gimple-range.cc (gimple_ranger::register_side_effects): First check
14562 if the DEF should be exported as a global.
14563 * tree-vrp.cc (rvrp_folder::pre_fold_bb): Process PHI side effects,
14564 which will export globals.
14565 (execute_ranger_vrp): Remove call to export_global_ranges.
14567 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14569 * value-relation.cc (path_oracle::reset_path): Clear killing_defs.
14571 2022-05-13 Andrew MacLeod <amacleod@redhat.com>
14573 * gimple-range-cache.cc (ranger_cache::ranger_cache): Start with
14575 (ranger_cache::entry_range): Add rfd_mode parameter.
14576 (ranger_cache::exit_range): Ditto.
14577 (ranger_cache::edge_range): New. Incorporate from range_on_edge.
14578 (ranger_cache::range_of_expr): Adjust call to entry_range.
14579 (ranger_cache::range_on_edge): Split to edge_range and call.
14580 (ranger_cache::fill_block_cache): Always invoke range_from_dom.
14581 (ranger_cache::range_from_dom): Make reentrant, add search mode, handle
14582 mutiple predecessors.
14583 (ranger_cache::update_to_nonnull): Adjust call to exit_range.
14584 * gimple-range-cache.h (ranger_cache): Add enum rfd_mode. Adjust
14587 2022-05-13 Alexandre Oliva <oliva@adacore.com>
14589 * gimple-harden-conditionals.cc: Include sbitmap.h.
14590 (pass_harden_conditional_branches::execute): Skip new blocks.
14591 (pass_harden_compares::execute): Likewise.
14593 2022-05-13 Richard Earnshaw <rearnsha@arm.com>
14596 * config/arm/mve.md (*movmisalign<mode>_mve_store): Use
14597 mve_memory_operand.
14598 (*movmisalign<mode>_mve_load): Likewise.
14599 * config/arm/vec-common.md (movmisalign<mode>): Convert to generator
14601 (@movmisalign<mode>): ... thus. Use generic predicates and then
14602 rework operands if they are not valid. For MVE rework to a
14603 narrower element size if the alignment is not high enough.
14605 2022-05-13 Richard Earnshaw <rearnsha@arm.com>
14607 * config/arm/arm.cc (mve_vector_mem_operand): Allow SP_REGNUM
14608 when there is no write-back. Fix use when strict is true.
14610 2022-05-13 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
14612 * config/xtensa/xtensa.h (TARGET_HAS_NO_HW_DIVIDE): New macro
14615 2022-05-13 Takayuki 'January June' Suwa <jjsuwa_sys3175@yahoo.co.jp>
14617 * config/xtensa/xtensa.md (extvsi, extvsi_internal, extzvsi,
14618 extzvsi_internal): Rename from extv, extv_internal, extzv and
14619 extzv_internal, respectively.
14621 2022-05-13 Eric Botcazou <ebotcazou@adacore.com>
14623 * tree-sra.cc (sra_modify_assign): Check that scalar storage order
14624 is the same on the LHS and RHS before rewriting one with the model
14627 2022-05-13 Richard Biener <rguenther@suse.de>
14629 * gimple-fold.cc (gimple_build): Adjust for new
14631 * gimple-fold.h (gimple_build): New main APIs with
14632 iterator, insert direction and iterator update.
14633 (gimple_build): New forwarder template.
14634 (clear_padding_type_may_have_padding_p): Remove.
14635 (clear_type_padding_in_mask): Likewise.
14636 (arith_overflowed_p): Likewise.
14637 * fold-const.h (clear_padding_type_may_have_padding_p): Declare.
14638 (clear_type_padding_in_mask): Likewise.
14639 (arith_overflowed_p): Likewise.
14640 * tree-vect-generic.cc (gimplify_build3): Use main gimple_build API.
14641 (gimplify_build2): Likewise.
14642 (gimplify_build1): Likewise.
14643 * ubsan.cc (ubsan_expand_ptr_ifn): Likewise, avoid extra
14645 * gengtype.cc (open_base_files): Re-order includes.
14646 * builtins.cc: Re-order gimple-fold.h include.
14647 * calls.cc: Likewise.
14648 * cgraphbuild.cc: Likewise.
14649 * cgraphunit.cc: Likewise.
14650 * config/rs6000/rs6000-builtin.cc: Likewise.
14651 * config/rs6000/rs6000-call.cc: Likewise.
14652 * config/rs6000/rs6000.cc: Likewise.
14653 * config/s390/s390.cc: Likewise.
14654 * expr.cc: Likewise.
14655 * fold-const.cc: Likewise.
14656 * function-tests.cc: Likewise.
14657 * gimple-match-head.cc: Likewise.
14658 * gimple-range-fold.cc: Likewise.
14659 * gimple-ssa-evrp-analyze.cc: Likewise.
14660 * gimple-ssa-evrp.cc: Likewise.
14661 * gimple-ssa-sprintf.cc: Likewise.
14662 * gimple-ssa-warn-access.cc: Likewise.
14663 * gimplify.cc: Likewise.
14664 * graphite-isl-ast-to-gimple.cc: Likewise.
14665 * ipa-cp.cc: Likewise.
14666 * ipa-devirt.cc: Likewise.
14667 * ipa-prop.cc: Likewise.
14668 * omp-low.cc: Likewise.
14669 * pointer-query.cc: Likewise.
14670 * range-op.cc: Likewise.
14671 * tree-cfg.cc: Likewise.
14672 * tree-if-conv.cc: Likewise.
14673 * tree-inline.cc: Likewise.
14674 * tree-object-size.cc: Likewise.
14675 * tree-ssa-ccp.cc: Likewise.
14676 * tree-ssa-dom.cc: Likewise.
14677 * tree-ssa-forwprop.cc: Likewise.
14678 * tree-ssa-ifcombine.cc: Likewise.
14679 * tree-ssa-loop-ivcanon.cc: Likewise.
14680 * tree-ssa-math-opts.cc: Likewise.
14681 * tree-ssa-pre.cc: Likewise.
14682 * tree-ssa-propagate.cc: Likewise.
14683 * tree-ssa-reassoc.cc: Likewise.
14684 * tree-ssa-sccvn.cc: Likewise.
14685 * tree-ssa-strlen.cc: Likewise.
14686 * tree-ssa.cc: Likewise.
14687 * value-pointer-equiv.cc: Likewise.
14688 * vr-values.cc: Likewise.
14690 2022-05-13 Alexandre Oliva <oliva@adacore.com>
14692 PR rtl-optimization/105455
14693 * gimple-harden-conditionals.cc (insert_check_and_trap): Set
14694 probabilities for newly-conditional edges.
14696 2022-05-13 liuhongt <hongtao.liu@intel.com>
14698 PR tree-optimization/102583
14699 * tree-ssa-forwprop.cc (simplify_bitfield_ref): Extended to a
14700 contiguous stride in the VEC_PERM_EXPR.
14702 2022-05-12 Richard Biener <rguenther@suse.de>
14704 PR rtl-optimization/105577
14705 * dse.cc (rest_of_handle_dse): Make sure to purge dead EH
14706 edges before running fast DCE via df_analyze.
14708 2022-05-12 Richard Biener <rguenther@suse.de>
14710 PR tree-optimization/105562
14711 * tree-ssa-sccvn.cc (vn_reference_lookup_3): Disambiguate
14712 against all CLOBBER defs if there's not an obvious must-alias
14713 and we are not doing redundant store elimination.
14714 (vn_walk_cb_data::redundant_store_removal_p): New field.
14715 (vn_reference_lookup_pieces): Initialize it.
14716 (vn_reference_lookup): Add argument to specify if we are
14717 doing redundant store removal.
14718 (eliminate_dom_walker::eliminate_stmt): Specify we do.
14719 * tree-ssa-sccvn.h (vn_reference_lookup): Adjust.
14721 2022-05-12 Haochen Jiang <haochen.jiang@intel.com>
14724 * config/i386/sse.md (vi1avx2const): New define_mode_attr.
14725 (pxor/pcmpeqb/pmovmskb/cmp 0xffff to ptest splitter):
14726 New define_split pattern.
14728 2022-05-12 Jakub Jelinek <jakub@redhat.com>
14730 * gimplify.cc (gimplify_omp_depend): Don't build_fold_addr_expr
14731 if null_pointer_node.
14732 (gimplify_scan_omp_clauses): Likewise.
14733 * tree-pretty-print.cc (dump_omp_clause): Print null_pointer_node
14736 2022-05-11 Patrick Palka <ppalka@redhat.com>
14738 * tree.h (TREE_VEC_BEGIN): Define.
14739 (TREE_VEC_END): Correct 'length' member access.
14740 (class tree_vec_range): Define.
14742 2022-05-11 Segher Boessenkool <segher@kernel.crashing.org>
14744 * config/rs6000/rs6000.md: Use wa instead of <Fv>.
14746 2022-05-11 Segher Boessenkool <segher@kernel.crashing.org>
14748 * config/rs6000/rs6000.md: Use d instead of <Ff>.
14750 2022-05-11 Segher Boessenkool <segher@kernel.crashing.org>
14752 * config/rs6000/constraints.md (register_constraint "f"): Use
14753 RS6000_CONSTRAINT_d.
14754 * config/rs6000/rs6000.cc (rs6000_debug_reg_global): Do not handle
14755 RS6000_CONSTRAINT_f.
14756 (rs6000_init_hard_regno_mode_ok): Ditto. Reorder and simplify a bit.
14757 * config/rs6000/rs6000.h (r6000_reg_class_enum): Delete
14758 RS6000_CONSTRAINT_d.
14760 2022-05-11 Richard Biener <rguenther@suse.de>
14762 * gimple-fold.h (gimple_build): Use variadic template
14763 functions for the gimple_build API forwarders without
14764 location_t argument.
14766 2022-05-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
14768 PR middle-end/70090
14769 * asan.cc (initialize_sanitizer_builtins): Register
14770 __builtin_dynamic_object_size if necessary.
14772 2022-05-11 Richard Biener <rguenther@suse.de>
14774 PR rtl-optimization/105559
14775 * cfgrtl.cc (delete_insn_and_edges): Only perform search to BB_END
14776 for non-debug insns.
14778 2022-05-11 Richard Biener <rguenther@suse.de>
14780 * generic-match-head.cc: Include tree-eh.h.
14781 * match.pd ((cond ...) cmp X): New simplification inspired
14782 by fold_binary_op_with_conditional_arg.
14783 (eq/ne (cmp ...) true/false): Likewise.
14785 2022-05-11 Thomas Schwinge <thomas@codesourcery.com>
14787 * doc/install.texi: Don't document '--with-hsa-runtime',
14788 '--with-hsa-runtime-include', '--with-hsa-runtime-lib'.
14790 2022-05-11 Martin Liska <mliska@suse.cz>
14793 * doc/install.texi: Document the configure option --with-zstd.
14795 2022-05-11 Martin Liska <mliska@suse.cz>
14797 * common/config/i386/cpuinfo.h (has_cpu_feature): Directly
14798 compute index in cpu_features2.
14799 (set_cpu_feature): Likewise.
14800 * config/i386/i386-builtins.cc (fold_builtin_cpu): Also remove
14801 loop for cpu_features2 and use NOP_EXPRs.
14803 2022-05-11 Richard Biener <rguenther@suse.de>
14805 PR bootstrap/105551
14806 * opts.cc (finish_options): Also disable var-tracking if
14807 !DWARF2_DEBUGGING_INFO.
14809 2022-05-11 liuhongt <hongtao.liu@intel.com>
14812 * config/i386/sse.md (*vec_set<mode>_0_zero_extendhi): New
14813 pre_reload define_insn_and_split.
14814 (*vec_setv2di_0_zero_extendhi_1): Ditto.
14815 (*vec_set<mode>_0_zero_extendsi): Ditto.
14816 (*vec_setv2di_0_zero_extendsi_1): Ditto.
14817 (ssewvecmode): New mode attr.
14818 (ssewvecmodelower): Ditto.
14819 (ssepackmodelower): Ditto.
14821 2022-05-11 Kewen Lin <linkw@linux.ibm.com>
14823 * config/rs6000/constraints.md (register constraint v): Use
14824 rs6000_constraints[RS6000_CONSTRAINT_v] instead of ALTIVEC_REGS.
14826 2022-05-11 Martin Liska <mliska@suse.cz>
14829 * config/riscv/riscv.opt: Remove Separate from
14830 -msmall-data-limit=.
14831 * optc-gen.awk: Report error for the described situation.
14832 * gcc.cc: Use Separate syntax.
14833 * opts.cc (gen_command_line_string): Change option name.
14835 2022-05-11 Haochen Gui <guihaoc@gcc.gnu.org>
14838 * match.pd (minmax): Skip constant folding for fmin/fmax when both
14839 arguments are sNaN or one is sNaN and another is NaN.
14841 2022-05-10 Jakub Jelinek <jakub@redhat.com>
14843 PR tree-optimization/105528
14844 * gimple-isel.cc (gimple_expand_vec_set_expr): After gsi_remove
14845 set *gsi to gsi_for_stmt (ass_stmt). Fix up function comment.
14847 2022-05-10 Segher Boessenkool <segher@kernel.crashing.org>
14849 * doc/md.texi (Defining Mode Iterators): Correct example replacement
14852 2022-05-10 Maciej W. Rozycki <macro@embecosm.com>
14854 * config/riscv/riscv.md (UNSPEC_FMIN, UNSPEC_FMAX): New
14856 (fmin<mode>3, fmax<mode>3): New insns.
14858 2022-05-10 Richard Biener <rguenther@suse.de>
14860 * tree-if-conv.cc (fold_build_cond_expr): Use
14861 match-and-simplify to simplify the condition.
14862 (ifcvt_follow_ssa_use_edges): Remove.
14863 (predicate_scalar_phi): Use follow_all_ssa_edges.
14865 2022-05-10 Thomas Schwinge <thomas@codesourcery.com>
14867 PR middle-end/100400
14868 * omp-oacc-kernels-decompose.cc
14869 (visit_loops_in_gang_single_region) <GIMPLE_OMP_FOR>: Explicitly
14870 call 'internal_error'.
14872 2022-05-10 Richard Biener <rguenther@suse.de>
14874 PR middle-end/105537
14875 * toplev.cc (process_options): Move flag_var_tracking
14877 * opts.cc (finish_options): ... here.
14879 2022-05-10 Martin Liska <mliska@suse.cz>
14881 * basic-block.h (struct basic_block_d): Use void *
14883 * cfgloop.h: Likewise.
14884 * cgraph.h: Likewise.
14885 * gengtype-state.cc (state_ident_by_name): Likewise.
14886 (record_type): Likewise.
14887 (read_state_already_seen_type): Likewise.
14888 * gengtype.cc (dump_type): Likewise.
14889 (input_file_by_name): Likewise.
14891 * ggc-common.cc (ggc_cleared_alloc_ptr_array_two_args): Likewise.
14892 * ipa-utils.h (struct ipa_dfs_info): Likewise.
14893 * plugin.cc (htab_hash_plugin): Likewise.
14895 2022-05-10 Richard Biener <rguenther@suse.de>
14897 * flags.h (dwarf_debuginfo_p): Add opts argument, guard
14898 API with !GENERATOR_FILE.
14899 * opts.cc (global_options): Poison.
14900 (global_options_set): Likewise.
14901 (finish_options): Refer to options via opts.
14903 2022-05-10 Uroš Bizjak <ubizjak@gmail.com>
14904 Roger Sayle <roger@nextmovesoftware.com>
14906 * config/i386/i386.md (*testqi_1_maybe_si): Prefer shorter SImode
14907 alternative when optimizing for size and the immediate operand is
14908 const_0_to_127_operand.
14909 (*andqi_2_maybe_si): Likewise.
14910 * config/i386/predicates.md (const_0_to_127_operand): New predicate.
14912 2022-05-10 Roger Sayle <roger@nextmovesoftware.com>
14914 * tree-chrec.cc (chrec_apply): Attempt to fold the linear chrec
14915 "{a, +, a} (x-1)" as "a*x", as the number of loop iterations, x-1,
14918 2022-05-10 Eric Botcazou <ebotcazou@adacore.com>
14921 * config/sparc/sparc.cc (sparc_vectorize_vec_perm_const): Return
14922 true only for 8-byte vector modes.
14924 2022-05-10 Siddhesh Poyarekar <siddhesh@gotplt.org>
14926 PR middle-end/70090
14927 * ubsan.cc (ubsan_expand_objsize_ifn): Allow non-constant SIZE.
14928 (instrument_object_size): Get dynamic object size expression.
14930 2022-05-10 Jiufu Guo <guojiufu@linux.ibm.com>
14932 PR preprocessor/101168
14933 * config/rs6000/rs6000-c.cc (rs6000_macro_to_expand):
14934 Avoid empty identifier.
14936 2022-05-09 Qing Zhao <qing.zhao@oracle.com>
14939 * config/i386/i386.cc (zero_call_used_regno_mode): use V2SImode
14940 as a generic MMX mode instead of V4HImode.
14941 (zero_all_mm_registers): Use SET to zero instead of MOV for
14942 zeroing scratch registers.
14943 (ix86_zero_call_used_regs): Likewise.
14945 2022-05-09 liuhongt <hongtao.liu@intel.com>
14948 * config/i386/i386-expand.cc
14949 (expand_vec_perm_pslldq_psrldq_por): New function.
14950 (ix86_expand_vec_perm_const_1): Try
14951 expand_vec_perm_pslldq_psrldq_por for both 3-instruction and
14952 4/5-instruction sequence.
14954 2022-05-09 Martin Liška <mliska@suse.cz>
14956 * basic-block.h (STATIC_ASSERT): Use normal STATIC_ASSERT.
14957 * system.h (STATIC_ASSERT): Define as static_assert for C++
14958 and fallback to array index in C.
14960 2022-05-09 Richard Biener <rguenther@suse.de>
14962 PR tree-optimization/105517
14963 * tree-ssa-sccvn.cc (vn_reference_lookup): Make sure the accumulated
14964 offset can be represented in the POINTER_PLUS_EXPR IL.
14965 (vn_reference_insert): Likewise.
14966 * poly-int.h (sext_hwi): Add poly version of sext_hwi.
14968 2022-05-09 Richard Biener <rguenther@suse.de>
14970 * match.pd: Remove #if GIMPLE guards around ! using patterns.
14972 2022-05-09 liuhongt <hongtao.liu@intel.com>
14975 * config/i386/sse.md (*sse4_1_<code>v2qiv2di2<mask_name>_1):
14977 (*sse4_1_zero_extendv2qiv2di2_2): New pre_reload
14978 define_insn_and_split.
14980 2022-05-09 Alex Coplan <alex.coplan@arm.com>
14982 * cgraph.cc (cgraph_node::get_create): Don't dump if the current
14983 symtab state is PARSING.
14985 2022-05-09 Martin Liska <mliska@suse.cz>
14987 * system.h (LIKELY): Define.
14988 (UNLIKELY): Likewise.
14989 * domwalk.cc (sort_bbs_postorder): Use {,UN}LIKELY
14991 * dse.cc (set_position_unneeded): Likewise.
14992 (set_all_positions_unneeded): Likewise.
14993 (any_positions_needed_p): Likewise.
14994 (all_positions_needed_p): Likewise.
14995 * expmed.cc (flip_storage_order): Likewise.
14996 * genmatch.cc (dt_simplify::gen_1): Likewise.
14997 * ggc-common.cc (gt_pch_save): Likewise.
14998 * print-rtl.cc: Likewise.
14999 * rtl-iter.h (T>::array_type::~array_type): Likewise.
15000 (T>::next): Likewise.
15001 * rtl-ssa/internals.inl: Likewise.
15002 * rtl-ssa/member-fns.inl: Likewise.
15003 * rtlanal.cc (T>::add_subrtxes_to_queue): Likewise.
15004 (rtx_properties::try_to_add_dest): Likewise.
15005 * rtlanal.h (growing_rtx_properties::repeat): Likewise.
15006 (vec_rtx_properties_base::~vec_rtx_properties_base): Likewise.
15007 * simplify-rtx.cc (simplify_replace_fn_rtx): Likewise.
15008 * sort.cc (likely): Likewise.
15009 (mergesort): Likewise.
15010 * wide-int.h (wi::eq_p): Likewise.
15011 (wi::ltu_p): Likewise.
15012 (wi::cmpu): Likewise.
15013 (wi::bit_and): Likewise.
15014 (wi::bit_and_not): Likewise.
15015 (wi::bit_or): Likewise.
15016 (wi::bit_or_not): Likewise.
15017 (wi::bit_xor): Likewise.
15018 (wi::add): Likewise.
15019 (wi::sub): Likewise.
15021 2022-05-09 Kito Cheng <kito.cheng@sifive.com>
15023 * config/riscv/arch-canonicalize: Handle g correctly.
15025 2022-05-07 Marek Polacek <polacek@redhat.com>
15029 * tree-core.h (struct attribute_spec): Update comment for HANDLER.
15031 2022-05-06 Jason Merrill <jason@redhat.com>
15033 * diagnostic-color.cc: Add fnname and targs color entries.
15034 * doc/invoke.texi: Document them.
15036 2022-05-06 Jason Merrill <jason@redhat.com>
15038 * vec.h (vec::iterate): Fix comment.
15040 2022-05-06 Michael Meissner <meissner@linux.ibm.com>
15043 * config/rs6000/rs6000.cc (rs6000_can_inline_p): Ignore -mpower8-fusion
15044 and -mpower10-fusion options for inlining purposes.
15046 2022-05-06 Christophe Lyon <christophe.lyon@arm.com>
15048 * config/aarch64/iterators.md (GPF_TF_F16): Delete.
15050 2022-05-06 Marcel Vollweiler <marcel@codesourcery.com>
15052 * omp-low.cc (omp_runtime_api_call): Added target_is_accessible to
15053 omp_runtime_apis array.
15055 2022-05-06 Hafiz Abid Qadeer <abidh@codesourcery.com>
15057 * omp-low.cc (omp_maybe_offloaded_ctx): New prototype.
15058 (scan_sharing_clauses): Check a restriction on allocate clause.
15060 2022-05-05 Sandra Loosemore <sandra@codesourcery.com>
15062 * gimplify.cc (gimplify_omp_for): Update messages for SCHEDULED
15063 and ORDERED clause conflict errors. Add check for GRAINSIZE and
15064 NUM_TASKS on TASKLOOP.
15066 2022-05-05 Martin Liska <mliska@suse.cz>
15068 * genautomata.cc (create_composed_state): Remove dead code.
15069 * graphite-poly.cc (print_pdrs): Likewise.
15070 * lto-wrapper.cc (run_gcc): Likewise.
15071 * tree-switch-conversion.cc (switch_decision_tree::balance_case_nodes):
15074 2022-05-05 Martin Liska <mliska@suse.cz>
15076 * tree-profile.cc (gimple_gen_ic_profiler): Prefix names with
15078 (gimple_gen_time_profiler): Likewise.
15080 2022-05-05 Martin Liska <mliska@suse.cz>
15082 * value-prof.cc (stream_out_histogram_value): Remove sanity
15085 2022-05-05 Richard Biener <rguenther@suse.de>
15087 PR tree-optimization/104162
15088 * tree-ssa-sccvn.cc (vn_reference_lookup): Handle
15089 &MEM[_1 + 5].a[i] like a POINTER_PLUS_EXPR if the offset
15091 (vn_reference_insert): Likewise.
15093 2022-05-05 Richard Biener <rguenther@suse.de>
15095 * tree-core.h (tree_real_cst::real_cst_ptr): Remove pointer
15096 to real_value field.
15097 (tree_real_cst::value): Add real_value field.
15098 * tree.h (TREE_REAL_CST_PTR): Adjust.
15099 * tree.cc (build_real): Remove separate allocation.
15100 * tree-streamer-in.cc (unpack_ts_real_cst_value_fields):
15103 2022-05-05 Richard Biener <rguenther@suse.de>
15105 * tree-ssa-ifcombine.cc (bb_no_side_effects_p): Avoid executing
15106 divisions with undefined overflow unconditionally.
15107 (pass_tree_ifcombine::execute): Rewrite stmts with undefined
15108 overflow to defined.
15110 2022-05-05 Richard Biener <rguenther@suse.de>
15112 PR tree-optimization/105484
15113 * gimple-isel.cc (gimple_expand_vec_set_expr): Clean EH, return
15114 whether the CFG changed.
15115 (gimple_expand_vec_exprs): When the CFG changed, clean it up.
15117 2022-05-05 Richard Biener <rguenther@suse.de>
15119 PR tree-optimization/104595
15120 * tree-vect-patterns.cc (vect_recog_bool_pattern): For
15121 COND_EXPR do not fail if check_bool_pattern returns false.
15123 2022-05-04 Tobias Burnus <tobias@codesourcery.com>
15125 * omp-low.cc (lower_omp_target): Fix use_device_{addr,ptr} with list
15126 item that is in an outer data-sharing clause.
15128 2022-05-04 Richard Biener <rguenther@suse.de>
15130 PR tree-optimization/104658
15131 * tree-vect-slp.cc (vect_slp_convert_to_external): Do not
15132 create VECTOR_BOOLEAN_TYPE_P extern defs. Reset the vector
15133 type on nodes we promote.
15134 (vectorizable_bb_reduc_epilogue): Deal with externalized
15136 * tree-vect-stmts.cc (vect_maybe_update_slp_op_vectype): Do
15137 not allow VECTOR_BOOLEAN_TYPE_P extern defs.
15139 2022-05-04 Richard Biener <rguenther@suse.de>
15141 PR tree-optimization/103116
15142 * tree-vect-stmts.cc (get_group_load_store_type): Handle the
15143 case we need peeling for gaps even though GROUP_GAP is zero.
15145 2022-05-04 Martin Liska <mliska@suse.cz>
15147 * gengtype-state.cc (read_a_state_token): Remove dead code.
15148 * ipa-profile.cc (ipa_profile_read_summary_section): Likewise.
15150 2022-05-04 Richard Biener <rguenther@suse.de>
15152 * gimplify.cc (gimplify_init_constructor): First gimplify,
15153 then simplify the result to a VECTOR_CST.
15155 2022-05-04 Jakub Jelinek <jakub@redhat.com>
15157 * genconditions.cc (write_conditions): Append a { nullptr, -1 }
15158 element at the end of insn_conditions.
15159 (write_writer): Use ARRAY_SIZE (insn_conditions) - 1 instead of
15160 ARRAY_SIZE (insn_conditions).
15162 2022-05-04 Christophe Lyon <christophe.lyon@arm.com>
15165 * config/arm/arm.cc (simd_valid_immediate): Exit when input is a
15166 vector of booleans and MVE is not enabled.
15168 2022-05-04 Richard Biener <rguenther@suse.de>
15171 * tree-cfgcleanup.cc (move_debug_stmts_from_forwarder):
15172 Move debug stmts to the predecessor if moving to the
15173 destination is not possible.
15174 (remove_forwarder_block): Adjust.
15175 (remove_forwarder_block_with_phi): Likewise.
15177 2022-05-03 Roger Sayle <roger@nextmovesoftware.com>
15179 PR tree-optimization/102950
15180 * range-op.cc (wi_optimize_signed_bitwise_op): New function to
15181 determine bounds of bitwise operations on signed types.
15182 (operator_bitwise_and::wi_fold): Call the above function.
15183 (operator_bitwise_or::wi_fold): Likewise.
15184 (operator_bitwise_xor::wi_fold): Likewise. Additionally, the
15185 result can't be zero if the operands can't be equal.
15187 2022-05-03 Iain Sandoe <iain@sandoe.co.uk>
15189 * config/darwin.cc (darwin_label_is_anonymous_local_objc_name): Make
15190 protocol class methods linker-visible.
15192 2022-05-03 Uroš Bizjak <ubizjak@gmail.com>
15195 * config/i386/sse.md (*vec_extract<mode>_0_mem): New pre-reload
15196 define_insn_and_split pattern.
15198 2022-05-03 Richard Biener <rguenther@suse.de>
15200 PR middle-end/105083
15201 * tree-scalar-evolution.cc (scev_initialize): Verify we
15202 have appropriate loop state.
15203 * tree-ssa-dce.cc (perform_tree_ssa_dce): Re-order SCEV and
15204 loop init and finalization.
15206 2022-05-03 Richard Biener <rguenther@suse.de>
15208 PR middle-end/105461
15209 * opts.cc (finish_options): Match the condition to
15210 disable flag_var_tracking to that of process_options.
15212 2022-05-03 Richard Biener <rguenther@suse.de>
15214 * opts.cc: #undef OPTIONS_SET_P.
15215 (finish_options): Use opts_set instead of OPTIONS_SET_P.
15217 2022-05-03 Richard Biener <rguenther@suse.de>
15219 PR tree-optimization/105394
15220 * tree-vect-generic.cc (expand_vector_condition): Adjust
15221 comp_width for non-integer mode masks as well.
15223 2022-05-03 Marcel Vollweiler <marcel@codesourcery.com>
15225 * omp-low.cc (omp_runtime_api_call): Added get_mapped_ptr to
15226 omp_runtime_apis array.
15228 2022-05-02 Richard Biener <rguenther@suse.de>
15230 * tree-scalar-evolution.cc (expression_expensive_p):
15231 Never consider mismatched calls as cheap.
15233 2022-05-02 Richard Biener <rguenther@suse.de>
15235 PR tree-optimization/104240
15236 * tree-vect-slp.cc (op1_op0_map): New.
15237 (vect_get_operand_map): Handle compares.
15238 (vect_build_slp_tree_1): Support swapped operands for
15241 2022-05-02 Jakub Jelinek <jakub@redhat.com>
15244 * cfgexpand.cc (expand_debug_expr): Don't make_decl_rtl_for_debug
15245 if there is no symtab node for the VAR_DECL.
15247 2022-05-02 Sebastian Huber <sebastian.huber@embedded-brains.de>
15249 * gcov-io.cc (gcov_rewrite): Clear the file error status.
15251 2022-05-02 Richard Biener <rguenther@suse.de>
15253 PR tree-optimization/105437
15254 * tree-vect-slp.cc (vect_schedule_slp_node): Handle the
15255 case where last_stmt alters control flow.
15257 2022-05-02 Richard Biener <rguenther@suse.de>
15259 * dojump.cc (do_jump): Use CASE_CONVERT.
15260 * tree-ssa-dom.cc (edge_info::derive_equivalences): Likewise.
15262 2022-05-02 Jakub Jelinek <jakub@redhat.com>
15264 * system.h: Include initializer_list.
15266 2022-05-01 Segher Boessenkool <segher@kernel.crashing.org>
15268 * config/rs6000/constraints.md (Y constraint): Fix comment.
15270 2022-05-01 Aldy Hernandez <aldyh@redhat.com>
15272 * tree-ssanames.cc (set_range_info): Denormalize VR_VARYING to
15273 VR_RANGE before passing a piecewise range to set_range_info_raw.
15275 2022-04-30 Patrick Palka <ppalka@redhat.com>
15277 * gengtype.cc (adjust_field_tree_exp): Remove.
15278 (adjust_field_type): Don't handle the "tree_exp" special attribute.
15279 * tree-core.h (struct tree_exp): Remove "special" and "desc"
15280 attributes. Add "length" attribute.
15282 2022-04-29 Martin Jambor <mjambor@suse.cz>
15285 * cgraph.cc (cgraph_node::remove): Release body of the node this
15286 is clone_of if appropriate.
15288 2022-04-29 Uroš Bizjak <ubizjak@gmail.com>
15291 * config/i386/i386.md (adcl/neg -> sbb peephole): New peephole2.
15293 2022-04-29 Richard Biener <rguenther@suse.de>
15295 * gimple-expr.cc (is_gimple_condexpr): Adjust comment.
15296 (canonicalize_cond_expr_cond): Move here from gimple.cc,
15297 allow both COND_EXPR and GIMPLE_COND forms.
15298 * gimple-expr.h (canonicalize_cond_expr_cond): Declare.
15299 * gimple.cc (canonicalize_cond_expr_cond): Remove here.
15300 * gimple.h (canonicalize_cond_expr_cond): Likewise.
15301 * gimple-loop-versioning.cc (loop_versioning::version_loop):
15302 Use is_gimple_condexpr_for_cond.
15303 * tree-parloops.cc (gen_parallel_loop): Likewise.
15304 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Check for
15305 a proper cond expr after canonicalize_cond_expr_cond.
15306 Use is_gimple_condexpr_for_cond where appropriate.
15307 * tree-ssa-loop-manip.cc (determine_exit_conditions): Likewise.
15308 * tree-vect-loop-manip.cc (slpeel_add_loop_guard): Likewise.
15310 2022-04-29 Richard Biener <rguenther@suse.de>
15312 * gimple-iterator.h (gsi_after_labels): Add overload for
15314 (gsi_start_1): Rename to gsi_start and take a reference.
15315 (gsi_last_1): Likewise.
15316 * gimple-iterator.cc (gsi_for_stmt): Use gsi_start.
15317 * omp-low.cc (lower_rec_input_clauses): Likewise.
15318 (lower_omp_scan): Likewise.
15320 2022-04-29 Richard Biener <rguenther@suse.de>
15322 PR tree-optimization/105431
15323 * tree-ssa-math-opts.cc (powi_as_mults_1): Make n unsigned.
15324 (powi_as_mults): Use absu_hwi.
15325 (gimple_expand_builtin_powi): Remove now pointless n != -n
15328 2022-04-29 Aldy Hernandez <aldyh@redhat.com>
15330 * range-op.cc (empty_range_varying): Move to range-op.h.
15331 (range_true): Move to range.h.
15332 (range_false): Same.
15333 (range_true_and_false): Same.
15334 (enum bool_range_state): Move to range-op.h.
15335 (relop_early_resolve): Same.
15336 (operator_equal::op1_op2_relation): Abstract code to...
15337 (equal_op1_op2_relation): ...here.
15338 (operator_not_equal::op1_op2_relation): Abstract code to...
15339 (not_equal_op1_op2_relation): ...here.
15340 (operator_lt::op1_op2_relation): Abstract code to...
15341 (lt_op1_op2_relation): ...here.
15342 (operator_le::op1_op2_relation): Abstract code to...
15343 (le_op1_op2_relation): ...here.
15344 (operator_gt::op1_op2_relation): Abstract code to...
15345 (gt_op1_op2_relation): ...here.
15346 (operator_ge::op1_op2_relation): Abstract code to...
15347 (ge_op1_op2_relation): ...here.
15348 (class range_op_table): Move to range-op.h.
15349 * range-op.h (equal_op1_op2_relation): Moved from range-op.cc.
15350 (not_equal_op1_op2_relation): Same.
15351 (lt_op1_op2_relation): Same.
15352 (le_op1_op2_relation): Same.
15353 (gt_op1_op2_relation): Same.
15354 (ge_op1_op2_relation): Same.
15355 (enum bool_range_state): Same.
15356 (get_bool_state): Same.
15357 (empty_range_varying): Same.
15358 (relop_early_resolve): Same.
15359 (class range_op_table): Same.
15360 * range.h (range_true): Same.
15361 (range_false): Same.
15362 (range_true_and_false): Same.
15364 2022-04-29 Aldy Hernandez <aldyh@redhat.com>
15366 * gimple-fold.cc (size_must_be_zero_p): Use reference
15368 * gimple-ssa-evrp-analyze.cc
15369 (evrp_range_analyzer::record_ranges_from_incoming_edge): Rename
15370 intersect to legacy_verbose_intersect.
15371 * ipa-cp.cc (ipcp_vr_lattice::meet_with_1): Use reference instead
15373 * tree-ssa-dom.cc (dom_jt_simplifier::simplify): Use value_range
15374 instead of value_range_equiv.
15375 * tree-vrp.cc (extract_range_from_plus_minus_expr): Use reference
15376 instead of pointer.
15377 (find_case_label_range): Same.
15378 * value-range-equiv.cc (value_range_equiv::intersect): Rename to...
15379 (value_range_equiv::legacy_verbose_intersect): ...this.
15380 (value_range_equiv::union_): Rename to...
15381 (value_range_equiv::legacy_verbose_union_): ...this.
15382 * value-range-equiv.h (class value_range_equiv): Rename union and
15383 intersect to legacy_verbose_{intersect,union}.
15384 * value-range.cc (irange::union_): Rename to...
15385 (irange::legacy_verbose_union_): ...this.
15386 (irange::intersect): Rename to...
15387 (irange::legacy_verbose_intersect): ...this.
15388 * value-range.h (irange::union_): Rename union_ to
15389 legacy_verbose_union.
15390 (irange::intersect): Rename intersect to legacy_verbose_intersect.
15391 * vr-values.cc (vr_values::update_value_range): Same.
15392 (vr_values::extract_range_for_var_from_comparison_expr): Same.
15393 (vr_values::extract_range_from_cond_expr): Rename union_ to
15394 legacy_verbose_union.
15395 (vr_values::extract_range_from_phi_node): Same.
15397 2022-04-29 Aldy Hernandez <aldyh@redhat.com>
15399 * gimple-ssa-evrp-analyze.cc
15400 (evrp_range_analyzer::set_ssa_range_info): Use *range_info methods
15402 * gimple-ssa-sprintf.cc (try_substitute_return_value): Same.
15403 * ipa-prop.cc (ipcp_update_vr): Same.
15404 * tree-inline.cc (remap_ssa_name): Same.
15405 * tree-ssa-copy.cc (fini_copy_prop): Same.
15406 * tree-ssa-math-opts.cc (optimize_spaceship): Same.
15407 * tree-ssa-phiopt.cc (replace_phi_edge_with_variable): Same.
15408 * tree-ssa-pre.cc (insert_into_preds_of_block): Same.
15409 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt): Same.
15410 * tree-ssa-strlen.cc (set_strlen_range): Same.
15411 (strlen_pass::handle_builtin_string_cmp): Same.
15412 * tree-ssanames.cc (set_range_info): Make static.
15413 (duplicate_ssa_name_range_info): Make static and add a new variant
15414 calling the static.
15415 * tree-ssanames.h (set_range_info): Remove version taking wide ints.
15416 (duplicate_ssa_name_range_info): Remove version taking a
15417 range_info_def and replace with a version taking SSA names.
15418 * tree-vect-loop-manip.cc (vect_gen_vector_loop_niters): Use *range_info methods
15420 (vect_do_peeling): Same.
15421 * tree-vrp.cc (vrp_asserts::remove_range_assertions): Same.
15422 * vr-values.cc (simplify_truth_ops_using_ranges): Same.
15424 2022-04-29 Aldy Hernandez <aldyh@redhat.com>
15426 * value-range.h (irange::irange): Use set_undefined.
15428 2022-04-29 Aldy Hernandez <aldyh@redhat.com>
15430 * gimple-range-cache.h (non_null_ref::adjust_range): Do not use
15431 irange::intersect (wide_int, wide_int).
15432 * gimple-range-fold.cc (adjust_pointer_diff_expr): Same.
15433 (adjust_imagpart_expr): Same.
15434 * value-range.h (irange::intersect (wide_int, wide_int)): Make
15437 2022-04-29 Richard Biener <rguenther@suse.de>
15439 PR tree-optimization/104322
15440 * tree-vect-loop.cc (vectorizable_reduction): Remove dead code.
15442 2022-04-29 Richard Biener <rguenther@suse.de>
15444 PR middle-end/105376
15445 * tree.cc (build_real): Special case dconst* arguments
15446 for decimal floating point types.
15448 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15450 * doc/gcov.texi (Profiling and Test Coverage in Freestanding
15451 Environments): New section.
15453 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15455 * doc/gcov-tool.texi: Document merge-stream subcommand.
15456 * doc/invoke.texi (fprofile-info-section): Mention merge-stream
15457 subcommand of gcov-tool.
15458 * gcov-tool.cc (gcov_profile_merge_stream): Declare.
15459 (print_merge_stream_usage_message): New.
15460 (merge_stream_usage): Likewise.
15461 (do_merge_stream): Likewise.
15462 (print_usage): Call print_merge_stream_usage_message().
15463 (main): Call do_merge_stream() to execute merge-stream subcommand.
15465 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15467 * gcov-io.cc (gcov_file_error): New enum.
15468 (gcov_var): Use gcov_file_error enum for the error member.
15469 (gcov_open): Use GCOV_FILE_NO_ERROR.
15470 (gcov_close): Use GCOV_FILE_WRITE_ERROR.
15471 (gcov_write): Likewise.
15472 (gcov_write_unsigned): Likewise.
15473 (gcov_write_string): Likewise.
15474 (gcov_read_bytes): Set error code if EOF is reached.
15475 (gcov_read_counter): Use GCOV_FILE_COUNTER_OVERFLOW.
15477 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15479 * gcov-io.cc (GCOV_MODE_STDIN): Define.
15480 (gcov_position): For gcov-tool, return calculated position if file is
15482 (gcov_open): For gcov-tool, use stdin if filename is NULL.
15483 (gcov_close): For gcov-tool, do not close stdin.
15484 (gcov_read_bytes): For gcov-tool, update position if file is stdin.
15485 (gcov_sync): For gcov-tool, discard input if file is stdin.
15487 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15489 * doc/invoke.texi (fprofile-info-section): Mention
15490 __gcov_filename_to_gcfn(). Use "freestanding" to match with C11
15491 standard language. Fix minor example code issues.
15492 * gcov-io.h (GCOV_FILENAME_MAGIC): Define and document.
15494 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15496 * gcov-io.cc (gcov_seek): Make it static.
15497 * gcov-io.h (struct gcov_summary): Do not mention gcov_seek().
15499 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15501 * gcov-tool.cc (gcov_do_dump): Add mode parameter.
15502 (gcov_output_files): Open files for reading and writing.
15504 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15506 * gcov-io.cc (gcov_open): Always use the mode parameter.
15507 * gcov-io.h (gcov_open): Declare it unconditionally.
15509 2022-04-28 Sebastian Huber <sebastian.huber@embedded-brains.de>
15511 * gcov-tool.cc (gcov_profile_merge): Adjust return type.
15512 (profile_merge): Allow merging of directories which contain no profile
15515 2022-04-28 David Malcolm <dmalcolm@redhat.com>
15517 * doc/invoke.texi (-fdump-analyzer-feasibility): Mention the
15520 2022-04-28 Jakub Jelinek <jakub@redhat.com>
15522 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Create a
15523 temporary only if out overlaps compare_op, not when it overlaps
15526 2022-04-28 Jakub Jelinek <jakub@redhat.com>
15529 * cgraph.cc (cgraph_node::verify_node): Don't verify
15530 semantic_interposition flag against
15531 opt_for_fn (decl, flag_semantic_interposition) for aliases in lto1.
15533 2022-04-28 Jakub Jelinek <jakub@redhat.com>
15536 * config/i386/i386.cc (ix86_gimplify_va_arg): Mark va_arg_tmp
15537 temporary TREE_ADDRESSABLE before trying to gimplify ADDR_EXPR
15540 2022-04-28 Jonathan Wakely <jwakely@redhat.com>
15542 * doc/install.texi (Configuration): Remove misleading text
15543 around LE PowerPC Linux multilibs.
15545 2022-04-28 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
15548 * doc/install.texi (Tools/packages necessary for building GCC)
15549 (GDC): Document libphobos requirement.
15550 (Host/target specific installation notes for GCC, *-*-solaris2*):
15551 Document libphobos and GDC specifics.
15553 2022-04-28 Richard Biener <rguenther@suse.de>
15555 PR tree-optimization/105219
15556 * tree-vect-loop.cc (vect_transform_loop): Disable
15557 special code narrowing the vectorized epilogue max
15558 iterations when peeling for alignment or gaps was in effect.
15560 2022-04-28 Xi Ruoyao <xry111@mengyan1223.wang>
15562 * config/loongarch/loongarch.cc
15563 (loongarch_flatten_aggregate_field): Ignore empty fields for
15566 2022-04-27 Lulu Cheng <chenglulu@loongson.cn>
15568 * config/loongarch/loongarch.md: Add fdiv define_expand template,
15569 then generate floating-point division and floating-point reciprocal
15572 2022-04-27 Lulu Cheng <chenglulu@loongson.cn>
15574 * config/loongarch/loongarch.md: Add '(clobber (mem:BLK (scratch)))'
15575 to PLV instruction templates.
15577 2022-04-27 Richard Biener <rguenther@suse.de>
15579 PR middle-end/104492
15580 * gimple-ssa-warn-access.cc
15581 (pass_waccess::warn_invalid_pointer): Exclude equality compare
15582 diagnostics for all kind of invalidations.
15583 (pass_waccess::check_dangling_uses): Fix post-dominator query.
15584 (pass_waccess::check_pointer_uses): Likewise.
15586 2022-04-27 Andreas Krebbel <krebbel@linux.ibm.com>
15589 * config/s390/s390-protos.h (s390_function_arg_vector): Remove
15591 * config/s390/s390.cc (s390_single_field_struct_p): New function.
15592 (s390_function_arg_vector): Invoke s390_single_field_struct_p.
15593 (s390_function_arg_float): Likewise.
15595 2022-04-27 Jakub Jelinek <jakub@redhat.com>
15597 PR sanitizer/105396
15598 * asan.cc (asan_redzone_buffer::emit_redzone_byte): Handle the case
15599 where offset is bigger than off but smaller than m_prev_offset + 32
15600 bits by pushing one or more 0 bytes. Sink the
15601 m_shadow_bytes.safe_push (value); flush_if_full (); statements from
15602 all cases to the end of the function.
15604 2022-04-27 Kewen Lin <linkw@linux.ibm.com>
15607 * config/rs6000/rs6000-builtins.def (NEG_V2DI): Move to [power8-vector]
15610 2022-04-26 Thomas Schwinge <thomas@codesourcery.com>
15612 * config/gcn/gcn.cc (gcn_print_lds_decl): Make "gang-private
15613 data-share memory exhausted" error more verbose.
15615 2022-04-26 Martin Liska <mliska@suse.cz>
15618 * lto-wrapper.cc (print_lto_docs_link): Use global_dc.
15619 (run_gcc): Parse OPT_fdiagnostics_urls_.
15620 (main): Initialize global_dc.
15622 2022-04-26 Jakub Jelinek <jakub@redhat.com>
15624 PR rtl-optimization/105314
15625 * ifcvt.cc (noce_try_store_flag_mask): Don't require that the non-zero
15626 operand is equal to if_info->x, instead use the non-zero operand
15627 as one of the operands of AND with if_info->x as target.
15629 2022-04-26 Jakub Jelinek <jakub@redhat.com>
15631 PR tree-optimization/105374
15632 * tree-ssa-reassoc.cc (eliminate_redundant_comparison): Punt if
15633 !fold_convertible_p rather than assuming fold_convert must succeed.
15635 2022-04-26 Jakub Jelinek <jakub@redhat.com>
15638 * config/i386/i386.cc (ix86_veclibabi_svml, ix86_veclibabi_acml): Pass
15639 el_mode == DFmode ? double_type_node : float_type_node instead of
15640 TREE_TYPE (type_in) as first arguments to mathfn_built_in.
15642 2022-04-25 David Malcolm <dmalcolm@redhat.com>
15645 * gimple-fold.cc (gimple_fold_builtin_memory_op): Explicitly set
15646 the location of new_stmt in all places that don't already set it,
15647 whether explicitly, or via a call to gsi_replace.
15649 2022-04-25 Paul A. Clarke <pc@us.ibm.com>
15651 * doc/extend.texi (Other Builtins): Correct reference to 'modff'.
15653 2022-04-25 Andrew MacLeod <amacleod@redhat.com>
15655 PR tree-optimization/105276
15656 * gimple-range.cc (gimple_ranger::prefill_stmt_dependencies): Include
15657 existing global range with calculated value.
15659 2022-04-25 Richard Biener <rguenther@suse.de>
15661 PR tree-optimization/105368
15662 * tree-ssa-math-opts.cc (powi_cost): Use absu_hwi.
15664 2022-04-25 Richard Biener <rguenther@suse.de>
15666 PR tree-optimization/100810
15667 * tree-ssa-loop-ivopts.cc (struct iv_cand): Add involves_undefs flag.
15668 (find_ssa_undef): New function.
15669 (add_candidate_1): Avoid adding derived candidates with
15670 undefined SSA names and mark the original ones.
15671 (determine_group_iv_cost_generic): Reject rewriting
15672 uses with a different IV when that involves undefined SSA names.
15674 2022-04-25 Steven G. Kargl <kargl@gcc.gnu.org>
15677 * config/freebsd.h: Define TARGET_LIBC_HAS_FUNCTION to be
15678 bsd_libc_has_function.
15679 * targhooks.cc (bsd_libc_has_function): New function.
15680 Expand the supported math functions to inclue C99 libm.
15681 * targhooks.h (bsd_libc_has_function): New Prototype.
15683 2022-04-25 Richard Biener <rguenther@suse.de>
15685 PR rtl-optimization/105231
15686 * combine.cc (distribute_notes): Assert that a REG_EH_REGION
15687 with landing pad > 0 is from i3. Put any REG_EH_REGION note
15688 on i3 or drop it if the insn can not trap.
15689 (try_combine): Ensure that we can merge REG_EH_REGION notes
15690 with non-call exceptions. Ensure we are not splitting a
15691 trapping part of an insn with non-call exceptions when there
15692 is any REG_EH_REGION note to preserve.
15694 2022-04-25 Hongyu Wang <hongyu.wang@intel.com>
15697 * config/i386/avx512fintrin.h (_mm512_scalef_round_pd):
15698 Add parentheses for parameters and djust format.
15699 (_mm512_mask_scalef_round_pd): Ditto.
15700 (_mm512_maskz_scalef_round_pd): Ditto.
15701 (_mm512_scalef_round_ps): Ditto.
15702 (_mm512_mask_scalef_round_ps): Ditto.
15703 (_mm512_maskz_scalef_round_ps): Ditto.
15704 (_mm_scalef_round_sd): Use _mm_undefined_pd.
15705 (_mm_scalef_round_ss): Use _mm_undefined_ps.
15706 (_mm_mask_scalef_round_sd): New macro.
15707 (_mm_mask_scalef_round_ss): Ditto.
15708 (_mm_maskz_scalef_round_sd): Ditto.
15709 (_mm_maskz_scalef_round_ss): Ditto.
15711 2022-04-23 Jakub Jelinek <jakub@redhat.com>
15714 * config/i386/i386-expand.cc (ix86_expand_int_movcc): Handle
15715 op0 == cst1 ? op0 : op3 like op0 == cst1 ? cst1 : op3 for the non-cmov
15718 2022-04-22 Segher Boessenkool <segher@kernel.crashing.org>
15721 * config/rs6000/rs6000.md (pack<mode> for FMOVE128): New expander.
15722 (pack<mode> for FMOVE128): Rename and split the insn_and_split to...
15723 (pack<mode>_hard for FMOVE128): ... this...
15724 (pack<mode>_soft for FMOVE128): ... and this.
15726 2022-04-22 Paul A. Clarke <pc@us.ibm.com>
15728 * doc/extend.texi: Correct "This" to "These".
15730 2022-04-22 Jakub Jelinek <jakub@redhat.com>
15732 PR rtl-optimization/105333
15733 * rtlanal.cc (replace_rtx): Use simplify_subreg or
15734 simplify_unary_operation if CONST_SCALAR_INT_P rather than just
15737 2022-04-21 Segher Boessenkool <segher@kernel.crashing.org>
15741 * config/rs6000/rs6000.md (zero_extendqi<mode>2 for EXTQI): Disparage
15742 the "Z" alternatives in {l,st}{f,xs}iwzx.
15743 (zero_extendhi<mode>2 for EXTHI): Ditto.
15744 (zero_extendsi<mode>2 for EXTSI): Ditto.
15745 (*movsi_internal1): Ditto.
15746 (*mov<mode>_internal1 for QHI): Ditto.
15747 (movsd_hardfloat): Ditto.
15749 2022-04-21 Martin Liska <mliska@suse.cz>
15751 * configure.ac: Enable compressed debug sections for mold
15753 * configure: Regenerate.
15755 2022-04-21 Jakub Jelinek <jakub@redhat.com>
15758 * emit-rtl.cc (emit_copy_of_insn_after): Don't call mark_jump_label
15761 2022-04-20 Richard Biener <rguenther@suse.de>
15763 PR tree-optimization/104912
15764 * tree-vect-loop-manip.cc (vect_loop_versioning): Split
15765 the cost model check to a separate BB to make sure it is
15766 checked first and not combined with other version checks.
15768 2022-04-20 Richard Biener <rguenther@suse.de>
15770 PR tree-optimization/105312
15771 * gimple-isel.cc (gimple_expand_vec_cond_expr): Query both
15772 VCOND and VCONDU for EQ and NE.
15774 2022-04-20 Jan Hubicka <hubicka@ucw.cz>
15777 * ipa-modref-tree.cc (modref_access_node::closer_pair_p): Use
15778 poly_offset_int to avoid overflow.
15779 (modref_access_node::update2): likewise.
15781 2022-04-20 Jakub Jelinek <jakub@redhat.com>
15784 * cgraph.cc (cgraph_node::create): Set node->semantic_interposition
15785 to opt_for_fn (decl, flag_semantic_interposition).
15786 * cgraphclones.cc (cgraph_node::create_clone): Copy over
15787 semantic_interposition flag.
15789 2022-04-19 Sergei Trofimovich <siarheit@google.com>
15791 PR gcov-profile/105282
15792 * value-prof.cc (stream_out_histogram_value): Allow negative counts
15793 on HIST_TYPE_INDIR_CALL.
15795 2022-04-19 Jakub Jelinek <jakub@redhat.com>
15798 * config/sparc/sparc.cc (epilogue_renumber): If ORIGINAL_REGNO,
15799 use gen_raw_REG instead of gen_rtx_REG and copy over also
15800 ORIGINAL_REGNO. Use return 0; instead of /* fallthrough */.
15802 2022-04-19 Richard Biener <rguenther@suse.de>
15804 PR tree-optimization/104010
15805 PR tree-optimization/103941
15806 * tree-vect-slp.cc (vect_bb_slp_scalar_cost): When
15807 we run into stmts in patterns continue walking those
15808 for uses outside of the vectorized region instead of
15809 marking the lane live.
15811 2022-04-18 Hans-Peter Nilsson <hp@axis.com>
15813 * doc/install.texi <CRIS>: Remove references to removed websites and
15814 adjust for cris-*-elf being the only remaining toolchain.
15816 2022-04-18 Hans-Peter Nilsson <hp@axis.com>
15818 * doc/invoke.texi <CRIS>: Remove references to options for removed
15819 subtarget cris-axis-linux-gnu and tweak wording accordingly.
15821 2022-04-16 Gerald Pfeifer <gerald@pfeifer.com>
15823 * doc/install.texi (Specific): Adjust mingw-w64 download link.
15825 2022-04-15 Hongyu Wang <hongyu.wang@intel.com>
15827 * config/i386/smmintrin.h: Correct target pragma from sse4.1
15828 and sse4.2 to crc32 for crc32 intrinsics.
15830 2022-04-14 Indu Bhagat <indu.bhagat@oracle.com>
15833 * ctfc.cc (ctf_dvd_ignore_insert): New function.
15834 (ctf_dvd_ignore_lookup): Likewise.
15835 (ctf_add_variable): Keep track of non-defining decl DIEs.
15836 (new_ctf_container): Initialize the new hash-table.
15837 (ctfc_delete_container): Empty hash-table.
15838 * ctfc.h (struct ctf_container): Add new hash-table.
15839 (ctf_dvd_ignore_lookup): New declaration.
15840 (ctf_add_variable): Add additional argument.
15841 * ctfout.cc (ctf_dvd_preprocess_cb): Skip adding CTF variable
15842 record for non-defining decl for which a defining decl exists
15844 (ctf_preprocess): Defer updating the number of global objts
15846 (output_ctf_header): Use ctfc_vars_list_count as some CTF
15847 variables may not make it to the final output.
15848 (output_ctf_vars): Likewise.
15849 * dwarf2ctf.cc (gen_ctf_variable): Skip generating CTF variable
15850 if this is known to be a non-defining decl DIE.
15852 2022-04-14 Indu Bhagat <indu.bhagat@oracle.com>
15854 * ctfc.h (struct ctf_container): Introduce a new member.
15855 * ctfout.cc (ctf_list_add_ctf_vars): Use it instead of static
15858 2022-04-14 Jakub Jelinek <jakub@redhat.com>
15861 * simplify-rtx.cc (simplify_const_binary_operation): For shifts
15862 or rotates by VOIDmode constant integer shift count use word_mode
15863 for the operand if int_mode is narrower than word.
15865 2022-04-14 Robin Dapp <rdapp@linux.ibm.com>
15867 * config/s390/s390.cc (s390_get_sched_attrmask): Add z16.
15868 (s390_get_unit_mask): Likewise.
15869 (s390_is_fpd): Likewise.
15870 (s390_is_fxd): Likewise.
15871 * config/s390/s390.h (s390_tune_attr): Set max tune level to z16.
15872 * config/s390/s390.md (z900,z990,z9_109,z9_ec,z10,z196,zEC12,z13,z14,z15):
15874 (z900,z990,z9_109,z9_ec,z10,z196,zEC12,z13,z14,z15,z16):
15876 * config/s390/3931.md: New file.
15878 2022-04-13 Richard Sandiford <richard.sandiford@arm.com>
15880 PR tree-optimization/105254
15881 * config/aarch64/aarch64.cc
15882 (aarch64_vector_costs::determine_suggested_unroll_factor): Take a
15883 loop_vec_info as argument. Restrict the unroll factor to values
15884 that divide the VF.
15885 (aarch64_vector_costs::finish_cost): Update call accordingly.
15887 2022-04-13 Richard Biener <rguenther@suse.de>
15889 PR tree-optimization/105263
15890 * tree-ssa-reassoc.cc (try_special_add_to_ops): Do not consume
15891 negates in multiplication chains with DFP.
15893 2022-04-13 Jakub Jelinek <jakub@redhat.com>
15895 PR middle-end/105253
15896 * tree.cc (tree_builtin_call_types_compatible_p): If PROP_gimple,
15897 use useless_type_conversion_p checks instead of TYPE_MAIN_VARIANT
15898 comparisons or tree_nop_conversion_p checks.
15900 2022-04-13 Hongyu Wang <hongyu.wang@intel.com>
15903 * config/i386/i386-expand.cc (ix86_expand_cmpxchg_loop):
15904 Add missing set to target_val at pause label.
15906 2022-04-13 Jakub Jelinek <jakub@redhat.com>
15909 * attribs.cc (decl_attributes): Don't set
15910 DECL_FUNCTION_SPECIFIC_TARGET if target_option_default_node is
15913 2022-04-13 Richard Biener <rguenther@suse.de>
15915 PR tree-optimization/105250
15916 * fold-const.cc (fold_convertible_p): Revert
15917 r12-7979-geaaf77dd85c333, instead check for size equality
15918 of the vector types involved.
15920 2022-04-13 Richard Biener <rguenther@suse.de>
15923 2022-04-13 Richard Biener <rguenther@suse.de>
15925 PR tree-optimization/104912
15926 * tree-vect-loop-manip.cc (vect_loop_versioning): Split
15927 the cost model check to a separate BB to make sure it is
15928 checked first and not combined with other version checks.
15930 2022-04-13 Richard Biener <rguenther@suse.de>
15932 PR tree-optimization/104912
15933 * tree-vect-loop-manip.cc (vect_loop_versioning): Split
15934 the cost model check to a separate BB to make sure it is
15935 checked first and not combined with other version checks.
15937 2022-04-13 Jakub Jelinek <jakub@redhat.com>
15939 * tree-scalar-evolution.cc (expression_expensive_p): Fix a comment typo.
15941 2022-04-12 Antoni Boucher <bouanto@zoho.com>
15944 * reginfo.cc: New functions (clear_global_regs_cache,
15945 reginfo_cc_finalize) to avoid an issue where compiling the same
15946 code multiple times gives an error about assigning the same
15947 register to 2 global variables.
15948 * rtl.h: New function (reginfo_cc_finalize).
15949 * toplev.cc: Call it.
15951 2022-04-12 Antoni Boucher <bouanto@zoho.com>
15954 * toplev.cc: Call the new function tree_cc_finalize in
15956 * tree.cc: New functions (clear_nonstandard_integer_type_cache
15957 and tree_cc_finalize) to clear the cache of non-standard integer
15958 types to avoid having issues with some optimizations of
15959 bitcast where the SSA_NAME will have a size of a cached
15960 integer type that should have been invalidated, causing a
15961 comparison of integer constant to fail.
15962 * tree.h: New function (tree_cc_finalize).
15964 2022-04-12 Thomas Schwinge <thomas@codesourcery.com>
15967 * config/nvptx/nvptx.h (ASM_SPEC): Don't set.
15968 * config/nvptx/nvptx.opt (misa): Adjust comment.
15970 2022-04-12 Thomas Schwinge <thomas@codesourcery.com>
15973 2022-03-03 Tom de Vries <tdevries@suse.de>
15975 * config/nvptx/nvptx.h (ASM_SPEC): Add %{misa=sm_30:--no-verify}.
15977 2022-04-12 Thomas Schwinge <thomas@codesourcery.com>
15980 2022-03-31 Tom de Vries <tdevries@suse.de>
15982 * config/nvptx/nvptx.h (ASM_SPEC): Use "-m sm_35" for -misa=sm_30.
15984 2022-04-12 Richard Biener <rguenther@suse.de>
15987 * tree-ssa-dce.cc (mark_stmt_if_obviously_necessary): Do not
15988 include local escaped memory as obviously necessary stores.
15990 2022-04-12 Richard Biener <rguenther@suse.de>
15992 PR tree-optimization/105235
15993 * tree-ssa-math-opts.cc (execute_cse_conv_1): Clean EH and
15994 return whether the CFG changed.
15995 (execute_cse_sincos_1): Adjust.
15997 2022-04-12 Przemyslaw Wirkus <Przemyslaw.Wirkus@arm.com>
16000 * config/arm/t-aprofile (MULTI_ARCH_OPTS_A): Remove Armv9-a options.
16001 (MULTI_ARCH_DIRS_A): Remove Armv9-a diretories.
16002 (MULTILIB_REQUIRED): Don't require Armv9-a libraries.
16003 (MULTILIB_MATCHES): Treat Armv9-a as equivalent to Armv8-a.
16004 (MULTILIB_REUSE): Remove remap rules for Armv9-a.
16005 * config/arm/t-multilib (v9_a_nosimd_variants): Delete.
16006 (MULTILIB_MATCHES): Remove mappings for v9_a_nosimd_variants.
16008 2022-04-12 Richard Biener <rguenther@suse.de>
16010 PR tree-optimization/105232
16011 * tree.cc (component_ref_size): Bail out for too large
16012 or non-constant sizes.
16014 2022-04-12 Richard Biener <rguenther@suse.de>
16016 PR tree-optimization/105226
16017 * tree-vect-loop-manip.cc (vect_loop_versioning): Verify
16018 we can split the exit of an outer loop we choose to version.
16020 2022-04-12 Jakub Jelinek <jakub@redhat.com>
16022 * config/i386/i386-expand.cc (ix86_emit_i387_sinh, ix86_emit_i387_cosh,
16023 ix86_emit_i387_tanh, ix86_emit_i387_asinh, ix86_emit_i387_acosh,
16024 ix86_emit_i387_atanh, ix86_emit_i387_log1p, ix86_emit_i387_round,
16025 ix86_emit_swdivsf, ix86_emit_swsqrtsf,
16026 ix86_expand_atomic_fetch_op_loop, ix86_expand_cmpxchg_loop):
16028 * config/i386/i386.cc (warn_once_call_ms2sysv_xlogues): Likewise.
16030 2022-04-12 Jakub Jelinek <jakub@redhat.com>
16033 * config/i386/i386-expand.cc (ix86_emit_i387_log1p): Call
16034 do_pending_stack_adjust.
16036 2022-04-12 Jakub Jelinek <jakub@redhat.com>
16038 PR rtl-optimization/105211
16039 * builtins.cc (expand_builtin_int_roundingfn_2): If mathfn_built_in_1
16040 fails for TREE_TYPE (arg), retry it with
16041 TREE_VALUE (TYPE_ARG_TYPES (TREE_TYPE (fndecl))) and if even that
16042 fails, emit call normally.
16044 2022-04-12 Andreas Krebbel <krebbel@linux.ibm.com>
16046 * common/config/s390/s390-common.cc: Rename PF_ARCH14 to PF_Z16.
16047 * config.gcc: Add z16 as march/mtune switch.
16048 * config/s390/driver-native.cc (s390_host_detect_local_cpu):
16049 Recognize z16 with -march=native.
16050 * config/s390/s390-opts.h (enum processor_type): Rename
16051 PROCESSOR_ARCH14 to PROCESSOR_3931_Z16.
16052 * config/s390/s390.cc (PROCESSOR_ARCH14): Rename to ...
16053 (PROCESSOR_3931_Z16): ... throughout the file.
16054 (s390_processor processor_table): Add z16 as cpu string.
16055 * config/s390/s390.h (enum processor_flags): Rename PF_ARCH14 to
16057 (TARGET_CPU_ARCH14): Rename to ...
16058 (TARGET_CPU_Z16): ... this.
16059 (TARGET_CPU_ARCH14_P): Rename to ...
16060 (TARGET_CPU_Z16_P): ... this.
16061 (TARGET_ARCH14): Rename to ...
16062 (TARGET_Z16): ... this.
16063 (TARGET_ARCH14_P): Rename to ...
16064 (TARGET_Z16_P): ... this.
16065 * config/s390/s390.md (cpu_facility): Rename arch14 to z16 and
16066 check TARGET_Z16 instead of TARGET_ARCH14.
16067 * config/s390/s390.opt: Add z16 to processor_type.
16068 * doc/invoke.texi: Document z16 and arch14.
16070 2022-04-12 chenglulu <chenglulu@loongson.cn>
16072 * config/loongarch/loongarch.cc: Fix bug for
16073 tmpdir-g++.dg-struct-layout-1/t033.
16075 2022-04-11 Peter Bergner <bergner@linux.ibm.com>
16078 * config/rs6000/rs6000.cc (rs6000_sibcall_aix): Handle pcrel sibcalls
16079 to longcall functions.
16081 2022-04-11 Jason Merrill <jason@redhat.com>
16083 * ipa-free-lang-data.cc (free_lang_data_in_decl): Fix typos.
16085 2022-04-11 Segher Boessenkool <segher@kernel.crashing.org>
16089 * config/rs6000/rs6000.md (unpack<mode>_nodm): Add m,r,i alternative.
16091 2022-04-11 Jakub Jelinek <jakub@redhat.com>
16093 PR tree-optimization/105218
16094 * tree-ssa-phiopt.cc (value_replacement): If middle_bb has
16095 more than one predecessor or phi's bb more than 2 predecessors,
16096 reset phi result uses instead of adding a debug temp.
16098 2022-04-11 Kito Cheng <kito.cheng@sifive.com>
16101 * config.gcc: Pass -misa-spec to arch-canonicalize and
16102 multilib-generator.
16103 * config/riscv/arch-canonicalize: Adding -misa-spec option.
16104 (SUPPORTED_ISA_SPEC): New.
16105 (arch_canonicalize): New argument `isa_spec`.
16106 Handle multiple ISA spec versions.
16107 * config/riscv/multilib-generator: Adding -misa-spec option.
16109 2022-04-11 Kito Cheng <kito.cheng@sifive.com>
16111 * config/riscv/arch-canonicalize: Add TODO item.
16112 (IMPLIED_EXT): Sync.
16113 (arch_canonicalize): Checking until no change.
16115 2022-04-11 Tamar Christina <tamar.christina@arm.com>
16118 * tree-vect-stmts.cc (vectorizable_condition): Prevent cond swap when
16121 2022-04-11 Jason Merrill <jason@redhat.com>
16124 * pointer-query.cc (compute_objsize_r) [POINTER_PLUS_EXPR]: Require
16127 2022-04-11 Jakub Jelinek <jakub@redhat.com>
16129 PR tree-optimization/104639
16130 * tree-ssa-phiopt.cc: Include tree-ssa-propagate.h.
16131 (value_replacement): Optimize (x != cst1 ? x : cst2) != cst3
16134 2022-04-11 Jeff Law <jeffreyalaw@gmail.com>
16136 * config/bfin/bfin.md (rol_one): Fix pattern to indicate the
16137 sign bit of the source ends up in CC.
16139 2022-04-09 Jan Hubicka <hubicka@ucw.cz>
16142 * cgraphunit.cc (cgraph_node::analyze): update semantic_interposition
16145 2022-04-09 Jan Hubicka <hubicka@ucw.cz>
16147 * ipa-modref.cc (ipa_merge_modref_summary_after_inlining): Propagate
16148 nondeterministic and side_effects flags.
16150 2022-04-08 Andre Vieira <andre.simoesdiasvieira@arm.com>
16153 * config.gcc: Shift ext_mask by TARGET_CPU_NBITS.
16154 * config/aarch64/aarch64.h (TARGET_CPU_NBITS): New macro.
16155 (TARGET_CPU_MASK): Likewise.
16156 (TARGET_CPU_DEFAULT): Use TARGET_CPU_NBITS.
16157 * config/aarch64/aarch64.cc (aarch64_get_tune_cpu): Use TARGET_CPU_MASK.
16158 (aarch64_get_arch): Likewise.
16159 (aarch64_override_options): Use TARGET_CPU_NBITS.
16161 2022-04-08 Richard Biener <rguenther@suse.de>
16163 PR tree-optimization/105198
16164 * tree-predcom.cc (find_looparound_phi): Check whether
16165 the found memory location of the entry value is clobbered
16166 inbetween the value we want to use and loop entry.
16168 2022-04-08 Jakub Jelinek <jakub@redhat.com>
16170 PR tree-optimization/105189
16171 * fold-const.cc (make_range_step): Fix up handling of
16172 (unsigned) x +[low, -] ranges for signed x if low fits into
16175 2022-04-08 Richard Biener <rguenther@suse.de>
16177 PR tree-optimization/105175
16178 * tree-vect-stmts.cc (vectorizable_operation): Suppress
16179 -Wvector-operation-performance if using emulated vectors.
16180 * tree-vect-generic.cc (expand_vector_piecewise): Do not diagnose
16181 -Wvector-operation-performance when suppressed.
16182 (expand_vector_parallel): Likewise.
16183 (expand_vector_comparison): Likewise.
16184 (expand_vector_condition): Likewise.
16185 (lower_vec_perm): Likewise.
16186 (expand_vector_conversion): Likewise.
16188 2022-04-07 Tamar Christina <tamar.christina@arm.com>
16191 * config/aarch64/aarch64-builtins.cc (handle_arm_acle_h): New.
16192 (aarch64_general_init_builtins): Move LS64 code.
16193 * config/aarch64/aarch64-c.cc (aarch64_pragma_aarch64): Support
16195 * config/aarch64/aarch64-protos.h (handle_arm_acle_h): New.
16196 * config/aarch64/arm_acle.h: Add pragma GCC aarch64 "arm_acle.h".
16198 2022-04-07 Richard Biener <rguenther@suse.de>
16199 Jan Hubicka <hubicka@ucw.cz>
16202 * tree-ssa-alias.h (ptr_deref_may_alias_global_p,
16203 ref_may_alias_global_p, ref_may_alias_global_p,
16204 stmt_may_clobber_global_p, pt_solution_includes_global): Add
16205 bool parameters indicating whether escaped locals should be
16207 * tree-ssa-structalias.cc (pt_solution_includes_global):
16208 When the new escaped_nonlocal_p flag is true also consider
16209 pt->vars_contains_escaped.
16210 * tree-ssa-alias.cc (ptr_deref_may_alias_global_p):
16211 Pass down new escaped_nonlocal_p flag.
16212 (ref_may_alias_global_p): Likewise.
16213 (stmt_may_clobber_global_p): Likewise.
16214 (ref_may_alias_global_p_1): Likewise. For decls also
16215 query the escaped solution if true.
16216 (ref_may_access_global_memory_p): Remove.
16217 (modref_may_conflict): Use ref_may_alias_global_p with
16218 escaped locals considered global.
16219 (ref_maybe_used_by_stmt_p): Adjust.
16220 * ipa-fnsummary.cc (points_to_local_or_readonly_memory_p):
16222 * tree-ssa-dse.cc (dse_classify_store): Likewise.
16223 * trans-mem.cc (thread_private_new_memory): Likewise, but
16224 consider escaped locals global.
16225 * tree-ssa-dce.cc (mark_stmt_if_obviously_necessary): Likewise.
16227 2022-04-07 Richard Biener <rguenther@suse.de>
16229 PR tree-optimization/105185
16230 * tree-ssa-sccvn.cc (visit_reference_op_call): Simplify
16231 modref query again.
16233 2022-04-07 Tamar Christina <tamar.christina@arm.com>
16236 * config/aarch64/aarch64-simd.md
16237 (aarch64_reduc_plus_internal<mode>): Fix RTL and rename to...
16238 (reduc_plus_scal_<mode>): ... This.
16239 (reduc_plus_scal_v4sf): Moved.
16240 (aarch64_reduc_plus_internalv2si): Fix RTL and rename to...
16241 (reduc_plus_scal_v2si): ... This.
16243 2022-04-07 Jakub Jelinek <jakub@redhat.com>
16245 PR tree-optimization/102586
16246 * langhooks.h (struct lang_hooks_for_types): Add classtype_as_base
16248 * langhooks-def.h (LANG_HOOKS_CLASSTYPE_AS_BASE): Define.
16249 (LANG_HOOKS_FOR_TYPES_INITIALIZER): Add it.
16250 * gimple-fold.cc (clear_padding_type): Use ftype instead of
16251 TREE_TYPE (field) some more. For artificial FIELD_DECLs without
16252 name try the lang_hooks.types.classtype_as_base langhook and
16253 if it returns non-NULL, use that instead of ftype for recursive call.
16255 2022-04-07 Jakub Jelinek <jakub@redhat.com>
16257 PR tree-optimization/105150
16258 * tree.cc (tree_builtin_call_types_compatible_p): New function.
16259 (get_call_combined_fn): Use it.
16261 2022-04-07 Richard Biener <rguenther@suse.de>
16263 PR middle-end/105165
16264 * tree-complex.cc (expand_complex_asm): Sorry for asm goto
16267 2022-04-07 liuhongt <hongtao.liu@intel.com>
16269 * config/i386/sse.md (<sse2_avx2>_andnot<mode>3_mask):
16271 (<sse>_andnot<mode>3<mask_name>): Disable V*HFmode patterns
16273 (<code><mode>3<mask_name>): Ditto.
16274 (*<code><mode>3<mask_name>): Ditto.
16275 (VFB_128_256): Adjust condition of V8HF/V16HFmode according to
16280 2022-04-06 Jakub Jelinek <jakub@redhat.com>
16282 PR rtl-optimization/104985
16283 * combine.cc (struct undo): Add where.regno member.
16284 (do_SUBST_MODE): Rename to ...
16285 (subst_mode): ... this. Change first argument from rtx * into int,
16286 operate on regno_reg_rtx[regno] and save regno into where.regno.
16287 (SUBST_MODE): Remove.
16288 (try_combine): Use subst_mode instead of SUBST_MODE, change first
16289 argument from regno_reg_rtx[whatever] to whatever. For UNDO_MODE, use
16290 regno_reg_rtx[undo->where.regno] instead of *undo->where.r.
16291 (undo_to_marker): For UNDO_MODE, use regno_reg_rtx[undo->where.regno]
16292 instead of *undo->where.r.
16293 (simplify_set): Use subst_mode instead of SUBST_MODE, change first
16294 argument from regno_reg_rtx[whatever] to whatever.
16296 2022-04-06 Jakub Jelinek <jakub@redhat.com>
16299 * config/sh/sh.opt (mdiv=): Add Save.
16301 2022-04-06 Martin Liska <mliska@suse.cz>
16304 * common.opt: Document properly based on what it does.
16305 * gcc.cc (display_help): Unify with what we have in common.opt.
16306 * opts.cc (common_handle_option): Do not print undocumented
16309 2022-04-06 Xi Ruoyao <xry111@mengyan1223.wang>
16311 * config/mips/mips.cc (mips_fpr_return_fields): Ignore
16312 cxx17_empty_base_field_p fields and set an indicator.
16313 (mips_return_in_msb): Adjust for mips_fpr_return_fields change.
16314 (mips_function_value_1): Inform psABI change about C++17 empty
16317 2022-04-06 Jakub Jelinek <jakub@redhat.com>
16319 PR tree-optimization/105150
16320 * gimple.cc (gimple_builtin_call_types_compatible_p): Use
16321 builtin_decl_explicit here...
16322 (gimple_call_builtin_p, gimple_call_combined_fn): ... rather than
16325 2022-04-06 Richard Biener <rguenther@suse.de>
16327 PR tree-optimization/105173
16328 * tree-ssa-reassoc.cc (find_insert_point): Get extra
16329 insert_before output argument and compute it.
16330 (insert_stmt_before_use): Adjust.
16331 (rewrite_expr_tree): Likewise.
16333 2022-04-06 Richard Biener <rguenther@suse.de>
16336 * ipa-modref-tree.cc (modref_access_node::get_ao_ref ): Bail
16337 out for non-pointer arguments.
16339 2022-04-06 Richard Biener <rguenther@suse.de>
16341 PR tree-optimization/105163
16342 * tree-ssa-reassoc.cc (repropagate_negates): Avoid propagating
16345 2022-04-06 Jakub Jelinek <jakub@redhat.com>
16347 PR tree-optimization/105150
16348 * gimple.cc (gimple_call_builtin_p, gimple_call_combined_fn):
16349 For BUILT_IN_NORMAL calls, call gimple_builtin_call_types_compatible_p
16350 preferrably on builtin_decl_explicit decl rather than fndecl.
16351 * tree-ssa-strlen.cc (valid_builtin_call): Don't call
16352 gimple_builtin_call_types_compatible_p here.
16354 2022-04-06 Richard Sandiford <richard.sandiford@arm.com>
16356 PR tree-optimization/103761
16357 * tree-vect-stmts.cc (check_load_store_for_partial_vectors): Replace
16358 the ncopies parameter with an slp_node parameter. Calculate the
16359 number of vectors based on it and vectype. Rename lambda to
16360 group_memory_nvectors.
16361 (vectorizable_store, vectorizable_load): Update calls accordingly.
16363 2022-04-06 Martin Liska <mliska@suse.cz>
16365 * doc/invoke.texi: Document it.
16367 2022-04-06 Richard Biener <rguenther@suse.de>
16369 PR tree-optimization/105148
16370 * tree-ssa-loop-ivopts.cc (idx_record_use): Walk raw operands
16371 2 and 3 of ARRAY_REFs.
16373 2022-04-06 Roger Sayle <roger@nextmovesoftware.com>
16375 * config/i386/sse.md (ANDNOT_MODE): New mode iterator for TF and V1TI.
16376 (*andnottf3): Replace with...
16377 (*andnot<mode>3): New define_insn using ANDNOT_MODE.
16379 2022-04-06 Richard Biener <rguenther@suse.de>
16381 PR tree-optimization/105142
16382 * gimple-fold.h (maybe_fold_and_comparisons): Add defaulted
16383 basic-block parameter.
16384 (maybe_fold_or_comparisons): Likewise.
16385 * gimple-fold.cc (follow_outer_ssa_edges): New.
16386 (maybe_fold_comparisons_from_match_pd): Use follow_outer_ssa_edges
16387 when an outer condition basic-block is specified.
16388 (and_comparisons_1, and_var_with_comparison,
16389 and_var_with_comparison_1, or_comparisons_1,
16390 or_var_with_comparison, or_var_with_comparison_1): Receive and pass
16391 down the outer condition basic-block.
16392 * tree-ssa-ifcombine.cc (ifcombine_ifandif): Pass down the
16393 basic-block of the outer condition.
16395 2022-04-06 Kewen Lin <linkw@linux.ibm.com>
16398 * config/rs6000/rs6000.cc (rs6000_maybe_emit_maxc_minc): Support more
16399 comparison codes UNLT/UNLE/UNGT/UNGE.
16401 2022-04-05 David Malcolm <dmalcolm@redhat.com>
16403 * doc/extend.texi (Common Function Attributes): Document that
16404 'access' does not imply 'nonnull'.
16406 2022-04-05 Uroš Bizjak <ubizjak@gmail.com>
16409 * config/i386/mmx.md (*movv2qi_internal):
16410 Change insn mode of alternative 5 to HF for TARGET_AVX512FP16.
16412 2022-04-05 Richard Sandiford <richard.sandiford@arm.com>
16414 * config/aarch64/aarch64.md (aarch64_cpymemdi): Turn into a
16415 define_expand and turn operands 0 and 1 from REGs to MEMs.
16416 (*aarch64_cpymemdi): New pattern.
16417 (aarch64_setmemdi): Turn into a define_expand and turn operand 0
16418 from a REG to a MEM.
16419 (*aarch64_setmemdi): New pattern.
16420 * config/aarch64/aarch64.cc (aarch64_expand_cpymem_mops): Use
16421 copy_to_mode_reg on all three registers. Replace the original
16422 MEM addresses rather than creating wild reads and writes.
16423 (aarch64_expand_setmem_mops): Likewise for the size and for the
16424 destination memory and address.
16426 2022-04-05 Richard Sandiford <richard.sandiford@arm.com>
16429 * config/aarch64/aarch64-protos.h (aarch64_simd_switcher): New class.
16430 * config/aarch64/aarch64-sve-builtins.h (sve_switcher): Inherit
16431 from aarch64_simd_switcher.
16432 * config/aarch64/aarch64-builtins.cc (aarch64_simd_tuple_modes):
16434 (aarch64_lookup_simd_builtin_type): Use it instead of TYPE_MODE.
16435 (register_tuple_type): Add more asserts. Expect the alignment
16436 of the structure to be subject to flag_pack_struct and
16437 maximum_field_alignment. Set aarch64_simd_tuple_modes.
16438 (aarch64_simd_switcher::aarch64_simd_switcher): New function.
16439 (aarch64_simd_switcher::~aarch64_simd_switcher): Likewise.
16440 (handle_arm_neon_h): Hold an aarch64_simd_switcher throughout.
16441 (aarch64_general_init_builtins): Hold an aarch64_simd_switcher
16442 while calling aarch64_init_simd_builtins.
16443 * config/aarch64/aarch64-sve-builtins.cc (sve_switcher::sve_switcher)
16444 (sve_switcher::~sve_switcher): Remove code now performed by
16445 aarch64_simd_switcher.
16447 2022-04-05 Richard Sandiford <richard.sandiford@arm.com>
16450 * config/aarch64/aarch64-sve-builtins.cc
16451 (function_resolver::infer_vector_or_tuple_type): Use error_n
16452 for "%d vectors" messages.
16454 2022-04-05 Chung-Lin Tang <cltang@codesourcery.com>
16456 * omp-low.cc (lower_omp_target): Use outer context looked-up 'var' as
16457 argument to lang_hooks.decls.omp_array_data, instead of 'ovar' from
16460 2022-04-05 Richard Biener <rguenther@suse.de>
16463 * passes.def (pass_walloca): Move early instance into
16464 pass_build_ssa_passes to make SSA form available.
16466 2022-04-05 liuhongt <hongtao.liu@intel.com>
16469 * config/i386/i386.cc (ix86_split_stlf_stall_load): New
16471 (ix86_reorg): Call ix86_split_stlf_stall_load.
16472 * config/i386/i386.opt (-param=x86-stlf-window-ninsns=): New
16475 2022-04-05 Alexandre Oliva <oliva@adacore.com>
16477 * targhooks.cc (default_zero_call_used_regs): Attempt to group
16478 regs that the target refuses to use in their natural modes.
16479 (zcur_select_mode_rtx): New.
16480 * regs.h (struct target_regs): Add x_hard_regno_max_nregs.
16481 (hard_regno_max_nregs): Define.
16482 * reginfo.cc (init_reg_modes_target): Set hard_regno_max_nregs.
16484 2022-04-04 Alex Coplan <alex.coplan@arm.com>
16486 * doc/match-and-simplify.texi: Fix typos.
16488 2022-04-04 Jakub Jelinek <jakub@redhat.com>
16491 * config/aarch64/t-aarch64 (s-aarch64-tune-md): Do move-if-change
16492 only if configured with --enable-maintainer-mode, otherwise compare
16493 tmp-aarch64-tune.md with $(srcdir)/config/aarch64/aarch64-tune.md and
16494 if they differ, emit a message and fail.
16496 2022-04-04 Jakub Jelinek <jakub@redhat.com>
16499 * config/aarch64/t-aarch64 (s-mddeps): Depend on s-aarch64-tune-md.
16500 * config/aarch64/aarch64-tune.md: Regenerated.
16502 2022-04-04 Richard Biener <rguenther@suse.de>
16504 PR tree-optimization/105132
16505 * tree-vect-stmts.cc (vectorizable_operation): Check that
16506 the input vectors have the same number of elements.
16508 2022-04-04 Richard Biener <rguenther@suse.de>
16510 PR middle-end/105140
16511 * fold-const.cc (fold_convertible_p): Allow a TYPE_P arg.
16513 2022-04-03 Jeff Law <jeffreyalaw@gmail.com>
16516 * config/iq2000/iq2000.md (bbi): New attribute, default to no.
16517 (delay slot descripts): Use different delay slot description when
16518 the insn as the "bbi" attribute.
16519 (bbi, bbin patterns): Set the bbi attribute to yes.
16521 2022-04-03 Jakub Jelinek <jakub@redhat.com>
16524 * config/i386/i386-expand.cc (ix86_expand_vector_init_general): Avoid
16525 using word as target for expand_simple_binop when doing ASHIFT and
16528 2022-04-02 Xi Ruoyao <xry111@mengyan1223.wang>
16530 * config/mips/mips.cc (mips_function_arg): Check if DECL_SIZE is
16531 NULL before dereferencing it.
16533 2022-04-01 Qing Zhao <qing.zhao@oracle.com>
16535 * config/i386/i386.cc (zero_all_st_registers): Return the value of
16537 (ix86_zero_call_used_regs): Update zeroed_hardregs set according to
16538 the return value of zero_all_st_registers.
16539 * doc/tm.texi: Update the documentation of TARGET_ZERO_CALL_USED_REGS.
16540 * function.cc (gen_call_used_regs_seq): Add an assertion.
16541 * target.def: Update the documentation of TARGET_ZERO_CALL_USED_REGS.
16543 2022-04-01 Xi Ruoyao <xry111@mengyan1223.wang>
16546 * config/mips/mips.cc (mips_function_arg): Ignore zero-width
16547 fields, and inform if it causes a psABI change.
16549 2022-04-01 Xi Ruoyao <xry111@mengyan1223.wang>
16552 * config/mips/mips.cc (mips_fpr_return_fields): Detect C++
16553 zero-width bit-fields and set up an indicator.
16554 (mips_return_in_msb): Adapt for mips_fpr_return_fields change.
16555 (mips_function_value_1): Diagnose when the presense of a C++
16556 zero-width bit-field changes function returning in GCC 12.
16558 2022-04-01 Jakub Jelinek <jakub@redhat.com>
16560 PR tree-optimization/104645
16561 * tree-ssa-phiopt.cc (value_replacement): If assign has
16562 CONVERT_EXPR_CODE_P rhs_code, treat it like a preparation
16563 statement with constant evaluation.
16565 2022-04-01 YunQiang Su <yunqiang.su@cipunited.com>
16567 * config/mips/mips.cc (mips_expand_prologue):
16568 IPL is 8bit for MCU ASE.
16570 2022-03-31 Bill Schmidt <wschmidt@linux.ibm.com>
16573 * config/rs6000/rs6000-builtins.def (MFFSL): Mark nosoft.
16574 (MTFSB0): Likewise.
16575 (MTFSB1): Likewise.
16576 (SET_FPSCR_RN): Likewise.
16577 (SET_FPSCR_DRN): Mark nosoft and no32bit.
16579 2022-03-31 Thomas Schwinge <thomas@codesourcery.com>
16581 * doc/options.texi (Option file format): Clarifications around
16582 option definition records' help texts.
16584 2022-03-31 Thomas Schwinge <thomas@codesourcery.com>
16586 * optc-gen.awk <END>: Fix "Multiple different help strings" error
16589 2022-03-31 Andre Vieira <andre.simoesdiasvieira@arm.com>
16591 * config/aarch64/aarch64.cc (aarch64_vector_costs): Define
16592 determine_suggested_unroll_factor and m_has_avg.
16593 (determine_suggested_unroll_factor): New function.
16594 (aarch64_vector_costs::add_stmt_cost): Check for a qualifying pattern
16595 to set m_nosve_pattern.
16596 (aarch64_vector_costs::finish_costs): Use
16597 determine_suggested_unroll_factor.
16598 * config/aarch64/aarch64.opt (aarch64-vect-unroll-limit): New.
16599 * doc/invoke.texi: (aarch64-vect-unroll-limit): Document new option.
16601 2022-03-31 Martin Jambor <mjambor@suse.cz>
16604 * ipa-prop.h (ipa_ancestor_jf_data): New flag keep_null;
16605 (ipa_get_jf_ancestor_keep_null): New function.
16606 * ipa-prop.cc (ipa_set_ancestor_jf): Initialize keep_null field of the
16608 (compute_complex_assign_jump_func): Pass false to keep_null
16609 parameter of ipa_set_ancestor_jf.
16610 (compute_complex_ancestor_jump_func): Pass true to keep_null
16611 parameter of ipa_set_ancestor_jf.
16612 (update_jump_functions_after_inlining): Carry over keep_null from the
16613 original ancestor jump-function or merge them.
16614 (ipa_write_jump_function): Stream keep_null flag.
16615 (ipa_read_jump_function): Likewise.
16616 (ipa_print_node_jump_functions_for_edge): Print the new flag.
16617 * ipa-cp.cc (class ipcp_bits_lattice): Make various getters const. New
16618 member function known_nonzero_p.
16619 (ipcp_bits_lattice::known_nonzero_p): New.
16620 (ipcp_bits_lattice::meet_with_1): New parameter drop_all_ones,
16622 (ipcp_bits_lattice::meet_with): Likewise.
16623 (propagate_bits_across_jump_function): Simplify. Pass true in
16624 drop_all_ones when it is necessary.
16625 (propagate_aggs_across_jump_function): Take care of keep_null
16627 (ipa_get_jf_ancestor_result): Propagate NULL accross keep_null
16630 2022-03-31 Martin Jambor <mjambor@suse.cz>
16633 * ipa-cp.cc (decide_whether_version_node): Skip scalar values
16634 which do not fit the known value_range.
16636 2022-03-31 Martin Jambor <mjambor@suse.cz>
16639 * ipa-prop.cc (propagate_controlled_uses): Add a LOAD reference
16640 always when an ADDR_EXPR constant is known to reach a load because
16641 of inlining, not just when removing an ADDR reference.
16643 2022-03-31 Richard Biener <rguenther@suse.de>
16645 PR tree-optimization/105109
16646 * tree-ssa.cc (execute_update_addresses_taken): Suppress
16647 diagnostics on the load of the other complex component.
16649 2022-03-31 Tom de Vries <tdevries@suse.de>
16651 * config/nvptx/nvptx.h (ASM_SPEC): Use "-m sm_35" for -misa=sm_30.
16653 2022-03-31 Richard Biener <rguenther@suse.de>
16655 PR rtl-optimization/105091
16656 * gimple-expr.cc (mark_addressable): Handle TARGET_MEM_REF
16659 2022-03-31 Richard Biener <rguenther@suse.de>
16662 2021-09-13 Richard Earnshaw <rearnsha@arm.com>
16665 * gimple-fold.c (gimple_fold_builtin_memory_op): Allow folding
16666 memcpy if the size is not more than MOVE_MAX * MOVE_RATIO.
16668 2022-03-31 Sebastian Huber <sebastian.huber@embedded-brains.de>
16670 * gcov-io.cc (gcov_read_string): Reword documentation comment.
16672 2022-03-30 Bill Schmidt <wschmidt@linux.ibm.com>
16674 * config/rs6000/rs6000-builtins.def (NEG_V16QI): Move to [altivec]
16676 (NEG_V4SF): Likewise.
16677 (NEG_V4SI): Likewise.
16678 (NEG_V8HI): Likewise.
16679 (NEG_V2DF): Move to [vsx] stanza.
16680 (NEG_V2DI): Likewise.
16682 2022-03-30 Vladimir N. Makarov <vmakarov@redhat.com>
16684 PR middle-end/105032
16685 * lra-assigns.cc (find_reload_regno_insns): Modify loop condition.
16687 2022-03-30 Tom de Vries <tdevries@suse.de>
16688 Tobias Burnus <tobias@codesourcery.com>
16690 * doc/invoke.texi (march): Document __PTX_SM__.
16691 (mptx): Document __PTX_ISA_VERSION_MAJOR__ and
16692 __PTX_ISA_VERSION_MINOR__.
16694 2022-03-30 Jakub Jelinek <jakub@redhat.com>
16696 PR sanitizer/105093
16697 * ubsan.cc (instrument_object_size): If t is equal to inner and
16698 is a decl other than global var, punt. When emitting call to
16699 UBSAN_OBJECT_SIZE ifn, make sure base is addressable.
16701 2022-03-30 Jakub Jelinek <jakub@redhat.com>
16703 PR tree-optimization/105094
16704 * gimple-ssa-store-merging.cc (mem_valid_for_store_merging): Punt if
16705 bitsize <= 0 rather than just == 0.
16707 2022-03-30 Tom de Vries <tdevries@suse.de>
16709 * doc/invoke.texi (misa, mptx): Update.
16710 (march, march-map): Add.
16712 2022-03-30 Thomas Schwinge <thomas@codesourcery.com>
16714 * opt-functions.awk (n_args): New function.
16715 (lang_enabled_by): Merge function into...
16716 * optc-gen.awk <END>: ... sole user here.
16717 Improve diagnostics.
16719 2022-03-29 Marek Polacek <polacek@redhat.com>
16720 Jakub Jelinek <jakub@redhat.com>
16722 PR middle-end/103597
16723 * gimplify.cc (collect_fallthrough_labels): Don't push UNUSED_LABEL_Ps
16724 into labels. Maybe set prev to the statement preceding UNUSED_LABEL_P.
16725 (gimplify_cond_expr): Set UNUSED_LABEL_P.
16726 * tree.h (UNUSED_LABEL_P): New.
16728 2022-03-29 Michael Meissner <meissner@linux.ibm.com>
16730 * config/rs6000/vsx.md (vsx_extract_<mode>): Allow destination to
16731 be any VSX register.
16733 2022-03-29 Richard Earnshaw <rearnsha@arm.com>
16736 * config/aarch64/aarch64.cc (aapcs_vfp_sub_candidate): Handle
16737 zero-sized bit-fields. Detect cases where a warning may be needed.
16738 (aarch64_vfp_is_call_or_return_candidate): Emit a note if a
16739 zero-sized bit-field has caused parameter passing to change.
16741 2022-03-29 Richard Earnshaw <rearnsha@arm.com>
16744 * config/arm/arm.cc (aapcs_vfp_sub_candidate): Handle zero-sized
16745 bit-fields. Detect cases where a warning may be needed.
16746 (aapcs_vfp_is_call_or_return_candidate): Emit a note if
16747 a zero-sized bit-field has caused parameter passing to change.
16749 2022-03-29 Richard Earnshaw <rearnsha@arm.com>
16752 * config/arm/arm.cc (arm_get_pcs_model): Disable selection of
16753 ARM_PCS_AAPCS_LOCAL.
16755 2022-03-29 Tom de Vries <tdevries@suse.de>
16758 * config/nvptx/nvptx-c.cc (nvptx_cpu_cpp_builtins): Emit
16759 __PTX_ISA_VERSION_MAJOR__ and __PTX_ISA_VERSION_MINOR__.
16760 * config/nvptx/nvptx.cc (ptx_version_to_number): New function.
16761 * config/nvptx/nvptx-protos.h (ptx_version_to_number): Declare.
16763 2022-03-29 Tom de Vries <tdevries@suse.de>
16765 * config/nvptx/nvptx.opt (m64): Update help text to reflect that it
16768 2022-03-29 Tom de Vries <tdevries@suse.de>
16771 * config/nvptx/nvptx.opt (march-map=*): Add aliases.
16773 2022-03-29 Jan Hubicka <hubicka@ucw.cz>
16775 * config/i386/i386-builtins.cc (ix86_vectorize_builtin_gather): Test
16776 TARGET_USE_GATHER_2PARTS and TARGET_USE_GATHER_4PARTS.
16777 * config/i386/i386.h (TARGET_USE_GATHER_2PARTS): New macro.
16778 (TARGET_USE_GATHER_4PARTS): New macro.
16779 * config/i386/x86-tune.def (X86_TUNE_USE_GATHER_2PARTS): New tune
16780 (X86_TUNE_USE_GATHER_4PARTS): New tune
16782 2022-03-29 Tom de Vries <tdevries@suse.de>
16784 * config/nvptx/nvptx.opt (march): Add alias of misa.
16786 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16787 Lulu Cheng <chenglulu@loongson.cn>
16789 * doc/install.texi: Add LoongArch options section.
16790 * doc/invoke.texi: Add LoongArch options section.
16791 * doc/md.texi: Add LoongArch options section.
16793 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16794 Lulu Cheng <chenglulu@loongson.cn>
16796 * config/loongarch/loongarch-c.cc
16798 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16799 Lulu Cheng <chenglulu@loongson.cn>
16801 * config/loongarch/larchintrin.h: New file.
16802 * config/loongarch/loongarch-builtins.cc: New file.
16804 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16805 Lulu Cheng <chenglulu@loongson.cn>
16807 * config/host-linux.cc: Add LoongArch support.
16808 * config/loongarch/loongarch-protos.h: New file.
16809 * config/loongarch/loongarch-tune.h: Likewise.
16810 * config/loongarch/loongarch.cc: Likewise.
16811 * config/loongarch/loongarch.h: Likewise.
16813 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16814 Lulu Cheng <chenglulu@loongson.cn>
16816 * config/loongarch/constraints.md: New file.
16817 * config/loongarch/generic.md: New file.
16818 * config/loongarch/la464.md: New file.
16819 * config/loongarch/loongarch-ftypes.def: New file.
16820 * config/loongarch/loongarch-modes.def: New file.
16821 * config/loongarch/loongarch.md: New file.
16822 * config/loongarch/predicates.md: New file.
16823 * config/loongarch/sync.md: New file.
16825 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16826 Lulu Cheng <chenglulu@loongson.cn>
16828 * configure: Regenerate file.
16830 2022-03-29 Chenghua Xu <xuchenghua@loongson.cn>
16831 Lulu Cheng <chenglulu@loongson.cn>
16833 * common/config/loongarch/loongarch-common.cc: New file.
16834 * config/loongarch/genopts/genstr.sh: New file.
16835 * config/loongarch/genopts/loongarch-strings: New file.
16836 * config/loongarch/genopts/loongarch.opt.in: New file.
16837 * config/loongarch/loongarch-str.h: New file.
16838 * config/loongarch/gnu-user.h: New file.
16839 * config/loongarch/linux.h: New file.
16840 * config/loongarch/loongarch-cpu.cc: New file.
16841 * config/loongarch/loongarch-cpu.h: New file.
16842 * config/loongarch/loongarch-def.c: New file.
16843 * config/loongarch/loongarch-def.h: New file.
16844 * config/loongarch/loongarch-driver.cc: New file.
16845 * config/loongarch/loongarch-driver.h: New file.
16846 * config/loongarch/loongarch-opts.cc: New file.
16847 * config/loongarch/loongarch-opts.h: New file.
16848 * config/loongarch/loongarch.opt: New file.
16849 * config/loongarch/t-linux: New file.
16850 * config/loongarch/t-loongarch: New file.
16851 * config.gcc: Add LoongArch support.
16852 * configure.ac: Add LoongArch support.
16854 2022-03-29 Thomas Schwinge <thomas@codesourcery.com>
16856 * opt-functions.awk (lang_enabled_by): Fix 'enabledby_negargs'
16859 2022-03-29 Richard Biener <rguenther@suse.de>
16861 PR tree-optimization/105080
16862 * tree-ssa-strlen.cc (printf_strlen_execute): Always init
16865 2022-03-28 Indu Bhagat <indu.bhagat@oracle.com>
16867 * ctfout.cc (ctf_preprocess): Use ctfc_get_num_ctf_vars instead.
16868 (output_ctf_vars): Likewise.
16870 2022-03-28 Jason Merrill <jason@redhat.com>
16873 * doc/extend.texi: Refer to __is_trivial instead of __is_pod.
16875 2022-03-28 H.J. Lu <hjl.tools@gmail.com>
16878 * config/i386/sse.md (*ssse3_pshufbv8qi3): Also replace "Yv" with
16881 2022-03-28 Tom de Vries <tdevries@suse.de>
16884 * config/nvptx/gen-opt.sh (ptx_isa): Improve help text.
16885 * config/nvptx/nvptx-gen.opt: Regenerate.
16886 * config/nvptx/nvptx.opt (misa, mptx, ptx_version): Improve help text.
16887 * config/nvptx/t-nvptx (s-nvptx-gen-opt): Add missing dependency on
16890 2022-03-28 David Malcolm <dmalcolm@redhat.com>
16893 * gimple-fold.cc (gimple_fold_builtin_memory_op): When optimizing
16894 to loads then stores, set the location of the new load stmt.
16896 2022-03-28 Richard Biener <rguenther@suse.de>
16898 PR tree-optimization/105070
16899 * tree-switch-conversion.h
16900 (bit_test_cluster::hoist_edge_and_branch_if_true): Add location
16902 * tree-switch-conversion.cc
16903 (bit_test_cluster::hoist_edge_and_branch_if_true): Annotate
16904 cond with location.
16905 (bit_test_cluster::emit): Annotate all generated expressions
16908 2022-03-28 Andre Vieira <andre.simoesdiasvieira@arm.com>
16910 * config/aarch64/aarch64-cores.def: Update Neoverse N2 core entry.
16912 2022-03-28 liuhongt <hongtao.liu@intel.com>
16915 * config/i386/sse.md (vec_set<mode>_0): Change attr "isa" of
16916 alternative 4 from sse4_noavx to noavx.
16918 2022-03-28 Jakub Jelinek <jakub@redhat.com>
16920 PR tree-optimization/105056
16921 * tree-predcom.cc (component::component): Initialize also comp_step.
16923 2022-03-27 H.J. Lu <hjl.tools@gmail.com>
16926 * config/i386/sse.md (*ssse3_pshufbv8qi3): Replace "Yv" with
16929 2022-03-26 Roger Sayle <roger@nextmovesoftware.com>
16931 PR middle-end/104885
16932 * calls.cc (mark_stack_region_used): Check that the region
16933 is within the allocated size of stack_usage_map.
16935 2022-03-26 Jakub Jelinek <jakub@redhat.com>
16937 PR rtl-optimization/103775
16938 * recog.cc (check_invalid_inc_dec): New function.
16939 (insn_invalid_p): Return 1 if REG_INC operand overlaps
16942 2022-03-26 H.J. Lu <hjl.tools@gmail.com>
16945 * config/i386/sse.md (loadiwkey): Replace "v" with "x".
16946 (aes<aesklvariant>u8): Likewise.
16948 2022-03-26 H.J. Lu <hjl.tools@gmail.com>
16951 * config/i386/sse.md (ssse3_ph<plusminus_mnemonic>wv4hi3):
16952 Replace "Yv" with "x".
16953 (ssse3_ph<plusminus_mnemonic>dv2si3): Likewise.
16954 (ssse3_psign<mode>3): Likewise.
16956 2022-03-26 Hans-Peter Nilsson <hp@axis.com>
16958 * reload.cc (find_reloads): Align comment with code where
16959 considering the intersection of register classes then tweaking the
16960 regclass for the current alternative or rejecting it.
16962 2022-03-25 Christophe Lyon <christohe.lyon@arm.com>
16966 2021-06-11 Christophe Lyon <christophe.lyon@linaro.org>
16968 * config/arm/mve.md (mve_vec_unpack<US>_lo_<mode>): Delete.
16969 (mve_vec_unpack<US>_hi_<mode>): Delete.
16970 (@mve_vec_pack_trunc_lo_<mode>): Delete.
16971 (mve_vmovntq_<supf><mode>): Remove '@' prefix.
16972 * config/arm/neon.md (vec_unpack<US>_hi_<mode>): Move back
16973 from vec-common.md.
16974 (vec_unpack<US>_lo_<mode>): Likewise.
16975 (vec_pack_trunc_<mode>): Rename from
16976 neon_quad_vec_pack_trunc_<mode>.
16977 * config/arm/vec-common.md (vec_unpack<US>_hi_<mode>): Delete.
16978 (vec_unpack<US>_lo_<mode>): Delete.
16979 (vec_pack_trunc_<mode>): Delete.
16981 2022-03-25 Vladimir N. Makarov <vmakarov@redhat.com>
16983 PR middle-end/104971
16984 * lra-lives.cc (process_bb_lives): Check hard_regs_live for hard
16985 regs to clear remove_p flag.
16987 2022-03-25 Richard Biener <rguenther@suse.de>
16989 PR tree-optimization/105053
16990 * tree-vect-loop.cc (vect_create_epilog_for_reduction): Pick
16991 the correct live-out stmt for a reduction chain.
16993 2022-03-25 Richard Biener <rguenther@suse.de>
16995 PR middle-end/105049
16996 * tree.cc (uniform_vector_p): Recurse for VECTOR_CST or
16997 CONSTRUCTOR first elements.
16999 2022-03-25 Tobias Burnus <tobias@codesourcery.com>
17002 * doc/invoke.texi (Static Analyzer Options): Move
17003 @ignore block after @gccoptlist's '}' for 'make pdf'.
17005 2022-03-25 David Malcolm <dmalcolm@redhat.com>
17008 * doc/invoke.texi (Static Analyzer Options): Add
17009 -fdump-analyzer-untracked.
17011 2022-03-25 Avinash Sonawane <rootkea@gmail.com>
17014 * doc/invoke.texi: Document that enabling taint analyzer
17015 checker disables some warnings from `-fanalyzer`.
17017 2022-03-24 Alexandre Oliva <oliva@adacore.com>
17020 * gimple-harden-conditionals.cc (detach_value): Keep temps
17023 2022-03-24 Alexandre Oliva <oliva@adacore.com>
17025 PR middle-end/104975
17026 * gimple-harden-conditionals.cc
17027 (pass_harden_compares::execute): Force split in case of
17030 2022-03-24 Jakub Jelinek <jakub@redhat.com>
17033 * fold-const.cc (operand_equal_p) <case COMPONENT_REF>: If either
17034 field0 or field1 is not a FIELD_DECL, return false.
17036 2022-03-24 Richard Biener <rguenther@suse.de>
17038 * tree-predcom.cc (chain::chain): Add CTOR.
17039 (component::component): Likewise.
17040 (pcom_worker::release_chain): Use delete.
17041 (release_components): Likewise.
17042 (pcom_worker::filter_suitable_components): Likewise.
17043 (pcom_worker::split_data_refs_to_components): Use new.
17044 (make_invariant_chain): Likewise.
17045 (make_rooted_chain): Likewise.
17046 (pcom_worker::combine_chains): Likewise.
17047 * tree-vect-loop.cc (vect_create_epilog_for_reduction):
17048 Make sure to release previously constructed scalar_results.
17049 * tree-vect-stmts.cc (vectorizable_load): Use auto_vec
17051 * vr-values.cc (simplify_using_ranges::~simplify_using_ranges):
17052 Release m_flag_set_edges.
17054 2022-03-24 Siddhesh Poyarekar <siddhesh@gotplt.org>
17056 PR tree-optimization/104970
17057 * tree-object-size.cc (parm_object_size): Restrict size
17058 computation scenarios to explicit access attributes.
17060 2022-03-24 Kewen Lin <linkw@linux.ibm.com>
17063 * config/rs6000/rs6000-c.cc (find_instance): Skip instances with null
17066 2022-03-23 Richard Biener <rguenther@suse.de>
17069 * gimple-fold.cc (gimple_fold_builtin_memory_op): Allow the
17070 use of movmisalign when either the source or destination
17071 decl is properly aligned.
17073 2022-03-23 Richard Biener <rguenther@suse.de>
17075 PR rtl-optimization/105028
17076 * ira-color.cc (form_threads_from_copies): Remove unnecessary
17077 copying of the sorted_copies tail.
17079 2022-03-23 Martin Liska <mliska@suse.cz>
17081 * config/rs6000/rs6000-c.cc (altivec_resolve_overloaded_builtin):
17083 * config/rs6000/rs6000.cc (rs6000_option_override_internal):
17084 Reword the error message.
17086 2022-03-23 liuhongt <hongtao.liu@intel.com>
17089 * config/i386/sse.md (ssePSmodelower): New.
17090 (*avx_cmp<mode>3_ltint_not): Force_reg operand before
17091 lowpart_subreg to avoid NULL_RTX.
17092 (<avx512>_fmaddc_<mode>_mask1<round_expand_name>,
17093 <avx512>_fcmaddc_<mode>_mask1<round_expand_name>,
17094 fma_<mode>_fmaddc_bcst, fma_<mode>_fcmaddc_bcst,
17095 <avx512>_<complexopname>_<mode>_mask<round_name>,
17096 avx512fp16_fcmaddcsh_v8hf_mask1<round_expand_name>,
17097 avx512fp16_fcmaddcsh_v8hf_mask3<round_expand_name>,
17098 avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>,
17099 avx512fp16_fmaddcsh_v8hf_mask3<round_expand_name>,
17100 float<floatunssuffix><mode>v4hf2,
17101 float<floatunssuffix>v2div2hf2,
17102 fix<fixunssuffix>_truncv4hf<mode>2,
17103 fix<fixunssuffix>_truncv2hfv2di2, extendv4hf<mode>2,
17105 trunc<mode>v4hf2,truncv2dfv2hf2,
17106 *avx512bw_permvar_truncv16siv16hi_1,
17107 *avx512bw_permvar_truncv16siv16hi_1_hf,
17108 *avx512f_permvar_truncv8siv8hi_1,
17109 *avx512f_permvar_truncv8siv8hi_1_hf,
17110 *avx512f_vpermvar_truncv8div8si_1,
17111 *avx512f_permvar_truncv32hiv32qi_1,
17112 *avx512f_permvar_truncv16hiv16qi_1,
17113 *avx512f_permvar_truncv4div4si_1,
17114 *avx512f_pshufb_truncv8hiv8qi_1,
17115 *avx512f_pshufb_truncv4siv4hi_1,
17116 *avx512f_pshufd_truncv2div2si_1,
17117 sdot_prod<mode>, avx2_pblend<ssemodesuffix>_1,
17118 ashrv2di3,ashrv2di3,usdot_prod<mode>): Ditto.
17120 2022-03-22 Tom de Vries <tdevries@suse.de>
17123 * config/nvptx/nvptx.md (define_insn "nvptx_uniform_warp_check"):
17124 Use % as register prefix.
17126 2022-03-22 Tom de Vries <tdevries@suse.de>
17128 * config/nvptx/nvptx.cc (nvptx_scalar_mode_supported_p)
17129 (nvptx_libgcc_floating_mode_supported_p): Only enable HFmode for
17132 2022-03-22 Tom de Vries <tdevries@suse.de>
17134 * config/nvptx/nvptx.opt (mexperimental): New option.
17136 2022-03-22 Tom de Vries <tdevries@suse.de>
17139 * config/nvptx/nvptx-protos.h (nvptx_asm_output_def_from_decls): Declare.
17140 * config/nvptx/nvptx.cc (write_fn_proto_1): Don't add function marker
17142 (SET_ASM_OP, NVPTX_ASM_OUTPUT_DEF): New macro def.
17143 (nvptx_asm_output_def_from_decls): New function.
17144 * config/nvptx/nvptx.h (ASM_OUTPUT_DEF): New macro def, define to
17145 gcc_unreachable ().
17146 (ASM_OUTPUT_DEF_FROM_DECLS): New macro def, define to
17147 nvptx_asm_output_def_from_decls.
17148 * config/nvptx/nvptx.opt (malias): New opt.
17150 2022-03-22 Tom de Vries <tdevries@suse.de>
17154 * config/nvptx/nvptx.md (define_expand "omp_simt_exit"): Emit warp
17155 sync (or uniform warp check for mptx < 6.0).
17157 2022-03-22 Richard Biener <rguenther@suse.de>
17159 PR tree-optimization/105012
17160 * tree-if-conv.cc (ifcvt_local_dce): Only call
17161 dse_classify_store when we have a VDEF.
17163 2022-03-22 Martin Liska <mliska@suse.cz>
17166 * config/nvptx/nvptx.cc (handle_ptx_version_option):
17167 Fix option wrapping in an error message.
17169 2022-03-22 Martin Liska <mliska@suse.cz>
17172 * config/rs6000/rs6000-c.cc (altivec_resolve_overloaded_builtin):
17173 Wrap const keyword.
17175 2022-03-22 Martin Liska <mliska@suse.cz>
17177 * config/v850/v850-c.cc (pop_data_area): Fix typo in pragma
17180 2022-03-22 Martin Liska <mliska@suse.cz>
17183 * config/rs6000/rs6000.cc (rs6000_option_override_internal):
17184 Use %qs instead of (%qs).
17186 2022-03-22 Martin Liska <mliska@suse.cz>
17189 * config/i386/i386-options.cc (ix86_option_override_internal):
17190 Use '%qs' instead of '(%qs)'.
17192 2022-03-22 Martin Liska <mliska@suse.cz>
17195 * config/aarch64/aarch64.cc (aarch64_handle_attr_arch):
17196 Use 'qs' and remove usage '(%qs)'.
17197 (aarch64_handle_attr_cpu): Likewise.
17198 (aarch64_handle_attr_tune): Likewise.
17199 (aarch64_handle_attr_isa_flags): Likewise.
17201 2022-03-22 Tamar Christina <tamar.christina@arm.com>
17202 Andre Vieira <andre.simoesdiasvieira@arm.com>
17204 * config/aarch64/aarch64.cc (neoversev1_regmove_cost): New tuning
17206 (neoversev1_tunings): Use neoversev1_regmove_cost and update store_int
17208 (neoverse512tvb_tunings): Likewise.
17210 2022-03-22 Tamar Christina <tamar.christina@arm.com>
17211 Andre Vieira <andre.simoesdiasvieira@arm.com>
17213 * config/aarch64/aarch64.cc (demeter_addrcost_table,
17214 demeter_regmove_cost, demeter_advsimd_vector_cost,
17215 demeter_sve_vector_cost, demeter_scalar_issue_info,
17216 demeter_advsimd_issue_info, demeter_sve_issue_info,
17217 demeter_vec_issue_info, demeter_vector_cost,
17218 demeter_tunings): New tuning structs.
17219 (aarch64_ve_op_count::rename_cycles_per_iter): Enable for demeter
17221 * config/aarch64/aarch64-cores.def: Add entry for demeter.
17222 * config/aarch64/aarch64-tune.md (tune): Add demeter to list.
17224 2022-03-22 Tamar Christina <tamar.christina@arm.com>
17225 Andre Vieira <andre.simoesdiasvieira@arm.com>
17227 * config/aarch64/aarch64-protos.h (struct cpu_memmov_cost): New struct.
17228 (struct tune_params): Change type of memmov_cost to use cpu_memmov_cost.
17229 * config/aarch64/aarch64.cc (aarch64_memory_move_cost): Update all
17230 tunings to use cpu_memmov_cost struct.
17232 2022-03-22 Tamar Christina <tamar.christina@arm.com>
17233 Andre Vieira <andre.simoesdiasvieira@arm.com>
17235 * config/aarch64/aarch64.cc (neoversen2_addrcost_table,
17236 neoversen2_regmove_cost, neoversen2_advsimd_vector_cost,
17237 neoversen2_sve_vector_cost, neoversen2_scalar_issue_info,
17238 neoversen2_advsimd_issue_info, neoversen2_sve_issue_info,
17239 neoversen2_vec_issue_info, neoversen2_tunings): New structs.
17240 (neoversen2_tunings): Use new structs and update tuning flags.
17241 (aarch64_vec_op_count::rename_cycles_per_iter): Enable for neoversen2
17244 2022-03-22 Andre Vieira <andre.simoesdiasvieira@arm.com>
17246 * config/aarch64/aarch64.h (AARCH64_FL_FOR_ARCH9): Add FP16 feature
17249 2022-03-22 liuhongt <hongtao.liu@intel.com>
17252 * config/i386/i386.md (*jcc_bt<mode>_mask): Extend the
17253 following splitter to reversed condition.
17255 2022-03-22 Jakub Jelinek <jakub@redhat.com>
17257 PR rtl-optimization/104989
17258 * calls.cc (expand_call): Don't set ECF_NORETURN in flags after
17259 sorry for passing too large argument, instead set sibcall_failure
17260 for pass == 0, or a new normal_failure flag otherwise. If
17261 normal_failure is set, don't assert all stack has been deallocated
17262 at the end and throw away the whole insn sequence.
17264 2022-03-22 Qian Jianhua <qianjh@cn.fujitsu.com>
17266 * print-tree.cc: Change array length
17268 2022-03-22 Hongyu Wang <hongyu.wang@intel.com>
17271 * config/i386/sse.md
17272 (avx512fp16_fmaddcsh_v8hf_mask1<round_expand_name):
17273 Use avx512f_movsf_mask instead of vmovaps or vblend, and
17274 force_reg before lowpart_subreg.
17275 (avx512fp16_fcmaddcsh_v8hf_mask1<round_expand_name): Likewise.
17277 2022-03-21 H.J. Lu <hjl.tools@gmail.com>
17280 * common/config/i386/i386-common.cc
17281 (OPTION_MASK_ISA2_GENERAL_REGS_ONLY_UNSET): Replace
17282 OPTION_MASK_ISA2_AVX512F_UNSET with OPTION_MASK_ISA2_SSE_UNSET.
17284 2022-03-21 H.J. Lu <hjl.tools@gmail.com>
17287 * common/config/i386/cpuinfo.h (get_available_features): Pass
17288 0x19 to __cpuid for bit_AESKLE. Enable FEATURE_AESKLE only if
17291 2022-03-21 Richard Sandiford <richard.sandiford@arm.com>
17293 PR middle-end/104869
17294 * rtl-ssa/accesses.h (clobber_group::prev_clobber): Declare.
17295 (clobber_group::next_clobber): Likewise.
17296 (def_lookup::prev_def): Rename to...
17297 (def_lookup::last_def_of_prev_group): ...this.
17298 (def_lookup::next_def): Rename to...
17299 (def_lookup::first_def_of_next_group): ...this.
17300 (def_lookup::matching_or_prev_def): Rename to...
17301 (def_lookup::matching_set_or_last_def_of_prev_group): ...this.
17302 (def_lookup::matching_or_next_def): Rename to...
17303 (def_lookup::matching_set_or_first_def_of_next_group): ...this.
17304 (def_lookup::prev_def): New function, taking the lookup insn as
17306 (def_lookup::next_def): Likewise.
17307 * rtl-ssa/member-fns.inl (def_lookup::prev_def): Rename to...
17308 (def_lookup::last_def_of_prev_group): ...this.
17309 (def_lookup::next_def): Rename to...
17310 (def_lookup::first_def_of_next_group): ...this.
17311 (def_lookup::matching_or_prev_def): Rename to...
17312 (def_lookup::matching_set_or_last_def_of_prev_group): ...this.
17313 (def_lookup::matching_or_next_def): Rename to...
17314 (def_lookup::matching_set_or_first_def_of_next_group): ...this.
17315 * rtl-ssa/movement.h (restrict_movement_for_dead_range): Update after
17317 * rtl-ssa/accesses.cc (clobber_group::prev_clobber): New function.
17318 (clobber_group::next_clobber): Likewise.
17319 (def_lookup::prev_def): Likewise.
17320 (def_lookup::next_def): Likewise.
17321 (function_info::make_use_available): Pass the lookup insn to
17322 def_lookup::prev_def and def_lookup::next_def.
17324 2022-03-21 Martin Liska <mliska@suse.cz>
17326 * doc/invoke.texi: Document min-pagesize parameter.
17328 2022-03-21 Richard Biener <rguenther@suse.de>
17330 * tree-ssa-loop-niter.cc (estimate_numbers_of_iterations): Dump
17331 we are estimating niter of loop.
17333 2022-03-21 Kito Cheng <kito.cheng@sifive.com>
17335 * common/config/riscv/riscv-common.cc (riscv_ext_flag_table):
17336 Update flag name and mask name.
17337 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins): Define
17338 misc macro for vector extensions.
17339 * config/riscv/riscv-opts.h (MASK_VECTOR_EEW_32): Rename to ...
17340 (MASK_VECTOR_ELEN_32): ... this.
17341 (MASK_VECTOR_EEW_64): Rename to ...
17342 (MASK_VECTOR_ELEN_64): ... this.
17343 (MASK_VECTOR_EEW_FP_32): Rename to ...
17344 (MASK_VECTOR_ELEN_FP_32): ... this.
17345 (MASK_VECTOR_EEW_FP_64): Rename to ...
17346 (MASK_VECTOR_ELEN_FP_64): ... this.
17347 (TARGET_VECTOR_ELEN_32): New.
17348 (TARGET_VECTOR_ELEN_64): Ditto.
17349 (TARGET_VECTOR_ELEN_FP_32): Ditto.
17350 (TARGET_VECTOR_ELEN_FP_64): Ditto.
17351 (TARGET_MIN_VLEN): Ditto.
17352 * config/riscv/riscv.opt (riscv_vector_eew_flags): Rename to ...
17353 (riscv_vector_elen_flags): ... this.
17355 2022-03-21 Hongyu Wang <hongyu.wang@intel.com>
17358 * config/i386/sse.md
17359 (avx512fp16_fma<complexopname>sh_v8hf<mask_scalarcz_name><round_scalarcz_name>):
17360 Correct round operand for intel dialect.
17362 2022-03-19 Arthur Cohen <arthur.cohen@embecosm.com>
17364 * diagnostic.cc (diagnostic_cc_tests): Rename to...
17365 (c_diagnostic_cc_tests): ...this.
17366 * opt-problem.cc (opt_problem_cc_tests): Rename to...
17367 (c_opt_problem_cc_tests): ...this.
17368 * selftest-run-tests.cc (selftest::run_tests): No longer run
17369 opt_problem_cc_tests or diagnostic_cc_tests.
17370 * selftest.h (diagnostic_cc_tests): Remove declaration.
17371 (opt_problem_cc_tests): Likewise.
17373 2022-03-19 Marc Nieper-Wißkirchen <marc@nieper-wisskirchen.de>
17376 * hash-traits.h (struct typed_const_free_remove): New.
17377 (struct free_string_hash): New.
17378 * pass_manager.h: Use free_string_hash.
17379 * passes.cc (pass_manager::register_pass_name): Use free_string_hash.
17380 (pass_manager::~pass_manager): Delete allocated m_name_to_pass_map.
17382 2022-03-19 Jakub Jelinek <jakub@redhat.com>
17384 PR middle-end/104971
17385 * config/i386/i386-expand.cc
17386 (ix86_expand_builtin) <case IX86_BUILTIN_READ_FLAGS>: If ignore,
17387 don't push/pop anything and just return const0_rtx.
17389 2022-03-18 Vladimir N. Makarov <vmakarov@redhat.com>
17391 PR rtl-optimization/104961
17392 * lra-assigns.cc (find_reload_regno_insns): Process reload pseudo clobber.
17394 2022-03-18 Jason Merrill <jason@redhat.com>
17396 * tree.h (IDENTIFIER_LENGTH): Add comment.
17398 2022-03-18 Jakub Jelinek <jakub@redhat.com>
17400 PR middle-end/99578
17401 PR middle-end/100680
17402 PR tree-optimization/100834
17403 * params.opt (--param=min-pagesize=): New parameter.
17405 (compute_objsize_r) <case ARRAY_REF>: Formatting fix.
17406 (compute_objsize_r) <case INTEGER_CST>: Use maximum object size instead
17407 of zero for pointer constants equal or larger than min-pagesize.
17409 2022-03-18 Tom de Vries <tdevries@suse.de>
17411 * gimplify.cc (gimplify_omp_for): Set location using 'input_location'.
17412 Set gfor location only when dealing with a OMP_TASKLOOP.
17414 2022-03-18 Tom de Vries <tdevries@suse.de>
17416 * gimplify.cc (gimplify_omp_for): Set taskloop location.
17418 2022-03-18 Tom de Vries <tdevries@suse.de>
17421 * omp-low.cc (lower_rec_input_clauses): Make sure GOMP_SIMT_XCHG_BFLY
17422 is executed unconditionally.
17424 2022-03-18 liuhongt <hongtao.liu@intel.com>
17427 * config/i386/i386.md (*movhi_internal): Set attr type from HI
17428 to HF for alternative 12 under TARGET_AVX512FP16.
17430 2022-03-18 Cui,Lili <lili.cui@intel.com>
17433 * config/i386/i386.h (PTA_SAPPHIRERAPIDS): change it to base on ICX.
17434 * doc/invoke.texi: Update documents for Intel sapphirerapids.
17436 2022-03-17 Roger Sayle <roger@nextmovesoftware.com>
17439 PR tree-optimization/90356
17440 * config/i386/i386.md (*movtf_internal): Don't guard
17441 standard_sse_constant_p clause by optimize_function_for_size_p.
17442 (*movdf_internal): Likewise.
17443 (*movsf_internal): Likewise.
17445 2022-03-17 Andrew MacLeod <amacleod@redhat.com>
17447 PR tree-optimization/102943
17448 * gimple-range-cache.cc (ranger_cache::range_from_dom): Find range via
17449 dominators and apply intermediary outgoing edge ranges.
17451 2022-03-17 Richard Biener <rguenther@suse.de>
17453 PR tree-optimization/104960
17454 * passes.def: Add pass parameter to pass_sink_code, mark
17455 last one to unsplit edges.
17456 * tree-ssa-sink.cc (pass_sink_code::set_pass_param): New.
17457 (pass_sink_code::execute): Always execute TODO_cleanup_cfg
17458 when we need to unsplit edges.
17460 2022-03-17 Jakub Jelinek <jakub@redhat.com>
17462 PR middle-end/103984
17463 * gimplify.cc (gimplify_target_expr): Gimplify type sizes and
17464 TARGET_EXPR_INITIAL into a temporary sequence, then push clobbers
17465 and asan unpoisioning, then append the temporary sequence and
17466 finally the TARGET_EXPR_CLEANUP clobbers.
17468 2022-03-16 Roger Sayle <roger@nextmovesoftware.com>
17470 * config/i386/sse.md: Delete corrupt character/typo.
17472 2022-03-16 Roger Sayle <roger@nextmovesoftware.com>
17475 * config/i386/sse.md (sse2_movq128): New define_expand to
17476 preserve previous named instruction.
17477 (*sse2_movq128_<mode>): Renamed from sse2_movq128, and
17478 generalized to VI8F_128 (both V2DI and V2DF).
17480 2022-03-16 Siddhesh Poyarekar <siddhesh@gotplt.org>
17482 PR tree-optimization/104941
17483 * tree-object-size.cc (size_for_offset): Make useless conversion
17484 check lighter and assign result of fold_convert to OFFSET.
17486 2022-03-16 H.J. Lu <hjl.tools@gmail.com>
17489 * config/i386/x86gprintrin.h: Also check _SOFT_FLOAT before
17490 pushing target("general-regs-only").
17492 2022-03-16 Kito Cheng <kito.cheng@sifive.com>
17494 * common/config/riscv/riscv-common.cc (riscv_ext_version_table):
17495 Add version info for zk, zks and zkn.
17497 2022-03-16 LiaoShihua <shihua@iscas.ac.cn>
17499 * common/config/riscv/riscv-common.cc
17500 (riscv_combine_info): New.
17501 (riscv_subset_list::handle_combine_ext): Combine back into zk to
17502 maintain the canonical order in isa strings.
17503 (riscv_subset_list::parse): Ditto.
17504 * config/riscv/riscv-subset.h (handle_combine_ext): New.
17506 2022-03-16 Richard Biener <rguenther@suse.de>
17508 PR tree-optimization/102008
17509 * passes.def: Move the added code sinking pass before the
17510 preceeding phiopt pass.
17512 2022-03-16 Patrick Palka <ppalka@redhat.com>
17515 * doc/invoke.texi (C++ Dialect Options): Document
17516 -ffold-simple-inlines.
17518 2022-03-16 Siddhesh Poyarekar <siddhesh@gotplt.org>
17520 PR tree-optimization/104942
17521 * tree-object-size.cc (alloc_object_size): Remove STRIP_NOPS.
17523 2022-03-16 Jakub Jelinek <jakub@redhat.com>
17526 * config/aarch64/aarch64.cc (aarch64_load_symref_appropriately): Copy
17529 2022-03-16 Roger Sayle <roger@nextmovesoftware.com>
17530 Richard Biener <rguenther@suse.de>
17532 * gimple-match-head.cc (single_use): Implement inline using a
17535 2022-03-16 Roger Sayle <roger@nextmovesoftware.com>
17537 * match.pd (X CMP X -> true): Test tree_expr_maybe_nan_p
17538 instead of HONOR_NANS.
17539 (X LTGT X -> false): Enable if X is not tree_expr_maybe_nan_p, as
17540 this can't trap/signal.
17542 2022-03-16 liuhongt <hongtao.liu@intel.com>
17545 * config/i386/i386-builtin.def (BDESC): Add
17546 CODE_FOR_sse4_1_blendvpd for IX86_BUILTIN_BLENDVPD.
17547 * config/i386/i386.cc (ix86_gimple_fold_builtin): Don't fold
17548 __builtin_ia32_blendvpd w/o sse4.2
17550 2022-03-15 Peter Bergner <bergner@linux.ibm.com>
17553 * config/rs6000/predicates.md (mma_disassemble_output_operand): Restrict
17554 acceptable MEM addresses.
17556 2022-03-15 Jakub Jelinek <jakub@redhat.com>
17559 * config/riscv/riscv.cc (riscv_pass_aggregate_in_fpr_pair_p,
17560 riscv_pass_aggregate_in_fpr_and_gpr_p): Pass OPT_Wpsabi instead of 0
17563 2022-03-15 Jakub Jelinek <jakub@redhat.com>
17566 * config/i386/i386.h (LIBGCC2_UNWIND_ATTRIBUTE): Use no-mmx,no-sse
17567 instead of general-regs-only.
17569 2022-03-15 Roger Sayle <roger@nextmovesoftware.com>
17570 Marc Glisse <marc.glisse@inria.fr>
17571 Richard Biener <rguenther@suse.de>
17573 PR tree-optimization/101895
17574 * match.pd (vec_same_elem_p): Handle CONSTRUCTOR_EXPR def.
17575 (plus (vec_perm (mult ...) ...) ...): New reordering simplification.
17577 2022-03-15 Jakub Jelinek <jakub@redhat.com>
17579 PR rtl-optimization/104814
17580 * ifcvt.cc (find_if_case_1, find_if_case_2): Punt if test_bb doesn't
17581 end with onlyjump_p. Assume BB_END (test_bb) is always non-NULL.
17583 2022-03-15 Martin Sebor <msebor@redhat.com>
17585 PR middle-end/104436
17586 * gimple-ssa-warn-access.cc (pass_waccess::check_dangling_stores):
17587 Check for warning suppression. Avoid by-value arguments transformed
17588 into by-transparent-reference.
17590 2022-03-14 Roger Sayle <roger@nextmovesoftware.com>
17591 Uroš Bizjak <ubizjak@gmail.com>
17593 * config/i386/i386.md (peephole2 xorl;movb -> movzbl): Disable
17594 transformation when *zero_extend<mode>si2 is not available.
17596 2022-03-14 Xi Ruoyao <xry111@mengyan1223.wang>
17598 * config/mips/mips.h (SUBTARGET_SHADOW_OFFSET): Define.
17599 * config/mips/mips.cc (mips_option_override): Make
17600 -fsanitize=address imply -fasynchronous-unwind-tables. This is
17601 needed by libasan for stack backtrace on MIPS.
17602 (mips_asan_shadow_offset): Return SUBTARGET_SHADOW_OFFSET.
17604 2022-03-14 Jakub Jelinek <jakub@redhat.com>
17607 * lra.cc (lra_substitute_pseudo): For debug_p mode, simplify
17608 SUBREG, ZERO_EXTEND, SIGN_EXTEND, FLOAT or UNSIGNED_FLOAT if recursive
17609 call simplified the first operand into VOIDmode constant.
17611 2022-03-14 Jakub Jelinek <jakub@redhat.com>
17613 PR tree-optimization/102586
17614 * doc/extend.texi (__builtin_clear_padding): Clearify that for C++
17615 argument type should be pointer to trivially-copyable type unless it
17616 is address of a variable or parameter.
17618 2022-03-14 Jakub Jelinek <jakub@redhat.com>
17621 * config/i386/emmintrin.h (_mm_loadu_si32): Put loaded value into
17622 first rather than last element of the vector, use __m32_u to do
17623 a really unaligned load, use just 0 instead of (int)0.
17624 (_mm_loadu_si16): Put loaded value into first rather than last
17625 element of the vector, use __m16_u to do a really unaligned load,
17626 use just 0 instead of (short)0.
17628 2022-03-14 Jakub Jelinek <jakub@redhat.com>
17631 * config/bfin/bfin.cc (bfin_handle_longcall_attribute): Fix a typo
17632 in diagnostic message - cannott -> cannot. Use %< and %> around
17633 names of attribute. Avoid too long line.
17634 * range-op.cc (operator_logical_and::op1_range): Fix up a typo
17635 in comment - cannott -> cannot. Use 2 spaces after . instead of one.
17637 2022-03-14 liuhongt <hongtao.liu@intel.com>
17640 * config/i386/i386-expand.cc
17641 (ix86_check_builtin_isa_match): New func.
17642 (ix86_expand_builtin): Move code to
17643 ix86_check_builtin_isa_match and call it.
17644 * config/i386/i386-protos.h
17645 (ix86_check_builtin_isa_match): Declare.
17646 * config/i386/i386.cc (ix86_gimple_fold_builtin): Don't fold
17647 builtin into gimple when isa mismatches.
17649 2022-03-13 Tobias Burnus <tobias@codesourcery.com>
17651 * doc/invoke.texi: Fix typos.
17652 * doc/tm.texi.in: Remove duplicated word.
17653 * doc/tm.texi: Regenerate.
17655 2022-03-12 Segher Boessenkool <segher@kernel.crashing.org>
17658 * config/rs6000/rs6000.cc (rs6000_machine_from_flags): Don't output
17659 "ppc" and "ppc64" based on rs6000_cpu.
17661 2022-03-12 Thomas Schwinge <thomas@codesourcery.com>
17663 PR middle-end/100280
17664 PR middle-end/104892
17665 * omp-oacc-kernels-decompose.cc (omp_oacc_kernels_decompose_1):
17666 Remove special handling of 'GOMP_MAP_FORCE_TOFROM'.
17668 2022-03-12 Thomas Schwinge <thomas@codesourcery.com>
17670 PR middle-end/100280
17671 PR middle-end/104086
17672 * omp-oacc-kernels-decompose.cc (omp_oacc_kernels_decompose_1):
17673 Mark variables used in 'present' clauses as addressable.
17674 * omp-low.cc (scan_sharing_clauses) <OMP_CLAUSE_MAP>: Gracefully
17675 handle duplicate 'OMP_CLAUSE_MAP_DECL_MAKE_ADDRESSABLE'.
17677 2022-03-12 Thomas Schwinge <thomas@codesourcery.com>
17680 * tree-core.h (user_omp_claus_code_name): Declare function.
17681 * tree.cc (user_omp_clause_code_name): New function.
17683 2022-03-12 Roger Sayle <roger@nextmovesoftware.com>
17685 PR middle-end/98420
17686 * match.pd (minus @0 @0): Additional checks for -fno-rounding-math
17687 (the defaut) or -fno-signed-zeros.
17689 2022-03-12 Michael Meissner <meissner@linux.ibm.com>
17692 * config/rs6000/vsx.md (extendditi2): Use a 'b' constraint when
17693 moving from a GPR register to an Altivec register.
17695 2022-03-11 Roger Sayle <roger@nextmovesoftware.com>
17697 PR tree-optimization/98335
17698 * config/i386/i386.md (peephole2): Eliminate redundant insv.
17699 Combine movl followed by movb. Transform xorl followed by
17700 a suitable movb or movw into the equivalent movz[bw]l.
17702 2022-03-11 Roger Sayle <roger@nextmovesoftware.com>
17703 Richard Biener <rguenther@suse.de>
17705 PR tree-optimization/98335
17706 * builtins.cc (get_object_alignment_2): Export.
17707 * builtins.h (get_object_alignment_2): Likewise.
17708 * tree-ssa-alias.cc (ao_ref_alignment): New.
17709 * tree-ssa-alias.h (ao_ref_alignment): Declare.
17710 * tree-ssa-dse.cc (compute_trims): Improve logic deciding whether
17711 to align head/tail, writing more bytes but using fewer store insns.
17713 2022-03-11 Richard Biener <rguenther@suse.de>
17715 PR tree-optimization/104880
17716 * tree-ssa.cc (execute_update_address_taken): Remember if we
17717 optimistically made something not addressable and
17718 prepare to undo it.
17720 2022-03-11 Richard Biener <rguenther@suse.de>
17723 * config/i386/i386.cc (ix86_builtin_vectorization_cost): Do not
17724 cost the first lane of SSE pieces as inserts for vec_construct.
17726 2022-03-10 Roger Sayle <roger@nextmovesoftware.com>
17729 * calls.cc (expand_call): Ignore stack adjustments after sorry.
17731 2022-03-10 Vladimir N. Makarov <vmakarov@redhat.com>
17734 * lra-constraints.cc (split_reg): Set up
17735 check_and_force_assignment_correctness_p when splitting hard
17736 register live range.
17738 2022-03-10 Martin Jambor <mjambor@suse.cz>
17741 * ipa-cp.cc (create_specialized_node): Move removal of
17742 self-recursive calls from callers vector before refrence
17745 2022-03-10 Richard Biener <rguenther@suse.de>
17747 PR tree-optimization/102943
17748 * gimple-range-cache.cc (sbr_sparse_bitmap::bitvec):
17749 Make a bitmap_head.
17750 (sbr_sparse_bitmap::sbr_sparse_bitmap): Adjust and switch
17752 (sbr_sparse_bitmap::set_bb_range): Adjust.
17753 (sbr_sparse_bitmap::get_bb_range): Likewise.
17755 2022-03-10 Richard Biener <rguenther@suse.de>
17757 PR tree-optimization/102943
17758 * tree-ssa-dom.cc (back_propagate_equivalences): Only
17759 populate the dominance bitmap if fast queries are not
17760 available. Use a tree view bitmap.
17761 (record_temporary_equivalences): Cache the dominance bitmap
17762 across all equivalences on the edge.
17764 2022-03-10 Tom de Vries <tdevries@suse.de>
17767 * config/nvptx/nvptx.md (define_attr "predicable"): Use no,yes instead
17770 2022-03-10 Tom de Vries <tdevries@suse.de>
17773 * config/nvptx/nvptx.cc (nvptx_init_unisimt_predicate)
17774 (nvptx_output_unisimt_switch): Handle unisimt_outside_simt_predicate.
17775 (nvptx_get_unisimt_outside_simt_predicate): New function.
17776 (predicate_insn): New function, factored out of ...
17777 (nvptx_reorg_uniform_simt): ... here. Predicate all emitted insns.
17778 * config/nvptx/nvptx.h (struct machine_function): Add
17779 unisimt_outside_simt_predicate field.
17780 * config/nvptx/nvptx.md (define_insn "nvptx_warpsync")
17781 (define_insn "nvptx_uniform_warp_check"): Make predicable.
17783 2022-03-10 Tom de Vries <tdevries@suse.de>
17785 * config/nvptx/nvptx.cc (nvptx_unisimt_handle_set): Handle unused
17788 2022-03-10 Tom de Vries <tdevries@suse.de>
17791 * config/nvptx/nvptx.cc (nvptx_print_operand): Handle 'x' operand
17793 * config/nvptx/nvptx.md: Use %x0 destination operand in atom insns.
17795 2022-03-10 Tom de Vries <tdevries@suse.de>
17797 * config/nvptx/nvptx.md (define_insn "atomic_fetch_<logic><mode>"):
17798 Emit atom.and.b64 instead of atom.b64.and.
17800 2022-03-10 Tom de Vries <tdevries@suse.de>
17802 * config/nvptx/t-nvptx (MULTILIB_EXTRA_OPTS): Move mptx=3.1 ...
17803 (MULTILIB_OPTIONS): ... here.
17805 2022-03-10 Tom de Vries <tdevries@suse.de>
17808 * config/nvptx/nvptx.opt (misa): Set default to sm_30.
17809 * config/nvptx/t-nvptx (MULTILIB_EXTRA_OPTS): Remove misa=sm_30.
17811 2022-03-10 Thomas Schwinge <thomas@codesourcery.com>
17813 PR middle-end/90115
17814 PR middle-end/102330
17815 PR middle-end/104774
17816 * omp-low.cc (oacc_privatization_candidate_p)
17817 (oacc_privatization_scan_clause_chain)
17818 (oacc_privatization_scan_decl_chain, lower_oacc_private_marker):
17819 Analyze 'lookup_decl'-translated DECL.
17821 2022-03-10 Jakub Jelinek <jakub@redhat.com>
17824 * config/rs6000/rs6000.h (enum rs6000_builtin_type_index): Remove
17825 RS6000_BTI_ptr_ieee128_float and RS6000_BTI_ptr_ibm128_float.
17826 (ptr_ieee128_float_type_node, ptr_ibm128_float_type_node): Remove.
17827 * config/rs6000/rs6000-builtin.cc (rs6000_type_string): Return
17828 "**NULL**" if type_node is NULL first. Handle
17829 ieee128_float_type_node.
17830 (rs6000_init_builtins): Don't initialize ptr_ieee128_float_type_node
17831 and ptr_ibm128_float_type_node. Set ibm128_float_type_node and
17832 ieee128_float_type_node to NULL rather than long_double_type_node if
17833 they aren't supported. Do support __ibm128 even if
17834 !TARGET_FLOAT128_TYPE when long double is double double.
17835 (rs6000_expand_builtin): Error if bif_is_ibm128 and
17836 !ibm128_float_type_node. Remap RS6000_BIF_{,UN}PACK_IF to
17837 RS6000_BIF_{,UN}PACK_TF much earlier and only use bif_is_ibm128 check
17839 * config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Define
17840 __SIZEOF_FLOAT128__ here and only iff __float128 macro is defined.
17841 (rs6000_cpu_cpp_builtins): Don't define __SIZEOF_FLOAT128__ here.
17842 Define __SIZEOF_IBM128__=16 if ieee128_float_type_node is non-NULL.
17844 * config/rs6000/rs6000-gen-builtins.cc: Document ibm128 attribute.
17845 (struct attrinfo): Add isibm128 member.
17846 (TYPE_MAP_SIZE): Remove.
17847 (type_map): Use [] instead of [TYPE_MAP_SIZE]. For "if" use
17848 ibm128_float_type_node only if it is non-NULL, otherwise fall back
17849 to long_double_type_node. Remove "pif" entry.
17850 (parse_bif_attrs): Handle ibm128 attribute and print it for debugging.
17851 (write_decls): Output bif_ibm128_bit and bif_is_ibm128.
17852 (write_type_node): Use sizeof type_map / sizeof type_map[0]
17853 instead of TYPE_MAP_SIZE.
17854 (write_bif_static_init): Handle isibm128.
17855 * config/rs6000/rs6000-builtins.def: Document ibm128 attribute.
17856 (__builtin_pack_ibm128, __builtin_unpack_ibm128): Add ibm128
17859 2022-03-09 Richard Biener <rguenther@suse.de>
17861 * cfgexpand.cc (expand_gimple_asm): Special-case MEM_REF
17862 with non-decl operand, avoiding a copy.
17864 2022-03-09 Jakub Jelinek <jakub@redhat.com>
17867 * config/i386/i386.h (LIBGCC2_UNWIND_ATTRIBUTE): Define for ia32.
17869 2022-03-09 Richard Biener <rguenther@suse.de>
17871 PR middle-end/104786
17872 * cfgexpand.cc (expand_asm_stmt): Do not generate a copy
17873 for VLAs without an upper size bound.
17875 2022-03-09 Xi Ruoyao <xry111@mengyan1223.wang>
17877 PR tree-optimization/104851
17878 * optabs-query.cc (supports_vec_convert_optab_p): Fix off-by-one
17881 2022-03-09 Jakub Jelinek <jakub@redhat.com>
17884 * doc/invoke.texi (-Wextra): Document that -Wshift-negative-value
17885 is enabled by it only for C++11 to C++17 rather than for C++03 or
17887 (-Wshift-negative-value): Similarly (except here we stated
17888 that it is enabled for C++11 or later).
17890 2022-03-09 Jakub Jelinek <jakub@redhat.com>
17892 PR rtl-optimization/104839
17893 * simplify-rtx.cc (simplify_unary_operation_1) <case SIGN_EXTEND>:
17894 Use SRP_SIGNED instead of incorrect 1 in SUBREG_PROMOTED_SET.
17895 (simplify_unary_operation_1) <case ZERO_EXTEND>: Use SRP_UNSIGNED
17896 instead of incorrect 0 in SUBREG_PROMOTED_SET.
17898 2022-03-09 Xi Ruoyao <xry111@mengyan1223.wang>
17901 * config/mips/mips.h (LUI_OPERAND): Cast the input to an unsigned
17902 value before adding an offset.
17904 2022-03-08 Christophe Lyon <christophe.lyon@arm.com>
17906 * config/arm/arm-builtins.cc
17907 (arm_binop_none_none_unone_qualifiers): Delete.
17908 (BINOP_NONE_NONE_UNONE_QUALIFIERS): Delete.
17910 2022-03-08 Iain Sandoe <iain@sandoe.co.uk>
17912 PR translation/104552
17913 * config/host-darwin.cc (darwin_gt_pch_get_address): Amend
17914 the PCH out of memory error message punctuation and wording.
17916 2022-03-08 Marek Polacek <polacek@redhat.com>
17918 PR rtl-optimization/104777
17919 * rtl.cc (classify_insn): For ASM_OPERANDS, return JUMP_INSN only if
17920 ASM_OPERANDS_LABEL_VEC has at least one element.
17922 2022-03-08 H.J. Lu <hjl.tools@gmail.com>
17925 * config/i386/i386.cc (ix86_expand_epilogue): Sorry if there is
17926 stack realignment or regparm nested function with EH return.
17928 2022-03-08 Andre Vieira <andre.simoesdiasvieira@arm.com>
17931 * config/arm/arm.h (MVE_STN_LDW_MODE): New MACRO.
17932 * config/arm/arm.cc (mve_vector_mem_operand): Relax constraint on base
17933 register for non widening loads or narrowing stores.
17935 2022-03-08 Eric Gallager <egallager@gcc.gnu.org>
17937 PR translation/104552
17938 * params.opt: Fix typo.
17940 2022-03-08 Richard Biener <rguenther@suse.de>
17942 PR tree-optimization/84201
17943 * params.opt (-param=vect-induction-float): Add.
17944 * doc/invoke.texi (vect-induction-float): Document.
17945 * tree-vect-loop.cc (vectorizable_induction): Honor
17946 param_vect_induction_float.
17948 2022-03-08 Martin Jambor <mjambor@suse.cz>
17950 PR translation/104552
17951 * params.opt (ipa-cp-recursive-freq-factor): Remove repeated word
17952 "that" in the description.
17954 2022-03-08 Richard Biener <rguenther@suse.de>
17956 PR tree-optimization/104825
17957 * tree-ssa-sccvn.cc (visit_reference_op_call): Properly
17958 guard modref get_ao_ref on a pointer typed argument.
17960 2022-03-08 liuhongt <hongtao.liu@intel.com>
17962 * config/i386/sse.md (*vec_dupv4si): Disable memory operand
17963 for !TARGET_INTER_UNIT_MOVES_TO_VEC when prefer_for_speed.
17965 2022-03-07 Jonathan Wakely <jwakely@redhat.com>
17967 * doc/invoke.texi (C++ Modules): Remove anachronism.
17969 2022-03-07 Martin Liska <mliska@suse.cz>
17971 PR middle-end/104381
17972 * opts.cc (finish_options): If debug info is disabled
17973 (debug_info_level) and -fvar-tracking is unset, disable it.
17975 2022-03-07 Jakub Jelinek <jakub@redhat.com>
17977 * tree-ssa-propagate.cc: Fix up duplicated word issue in a comment.
17978 * config/riscv/riscv.cc: Likewise.
17979 * config/darwin.h: Likewise.
17980 * config/i386/i386.cc: Likewise.
17981 * config/aarch64/thunderx3t110.md: Likewise.
17982 * config/aarch64/fractional-cost.h: Likewise.
17983 * config/vax/vax.cc: Likewise.
17984 * config/rs6000/pcrel-opt.md: Likewise.
17985 * config/rs6000/predicates.md: Likewise.
17986 * ctfc.h: Likewise.
17987 * tree-ssa-uninit.cc: Likewise.
17988 * value-relation.h: Likewise.
17989 * gimple-range-gori.cc: Likewise.
17990 * ipa-polymorphic-call.cc: Likewise.
17991 * pointer-query.cc: Likewise.
17992 * ipa-sra.cc: Likewise.
17993 * internal-fn.cc: Likewise.
17994 * varasm.cc: Likewise.
17995 * gimple-ssa-warn-access.cc: Likewise.
17997 2022-03-07 Martin Liska <mliska@suse.cz>
18000 * config/arm/arm.cc (arm_option_override_internal): Add missing
18003 2022-03-07 Richard Biener <rguenther@suse.de>
18005 PR tree-optimization/104782
18006 * tree-vect-slp.cc (vectorize_slp_instance_root_stmt):
18007 Re-instantiate r10-5979 fix, add comment.
18009 2022-03-07 Martin Liska <mliska@suse.cz>
18012 * config/msp430/msp430.cc (msp430_expand_delay_cycles): Remove
18013 parenthesis from built-in name.
18015 2022-03-07 Martin Liska <mliska@suse.cz>
18018 * config/arm/arm.cc (arm_option_override_internal): Fix quoting
18019 of options in error messages.
18020 (arm_option_reconfigure_globals): Likewise.
18022 2022-03-07 Martin Liska <mliska@suse.cz>
18025 * config/arm/arm-builtins.cc (arm_expand_builtin): Reuse error
18026 message. Fix ARM_BUILTIN_WRORHI and ARM_BUILTIN_WRORH that can
18027 have only range [0,32].
18029 2022-03-07 Jakub Jelinek <jakub@redhat.com>
18032 * config/s390/s390.md (*cmp_and_trap_unsigned_int<mode>): Use
18033 S constraint instead of T in the last alternative.
18035 2022-03-07 Martin Liska <mliska@suse.cz>
18037 * plugin.cc (default_plugin_dir_name): Remove <dir> from error
18040 2022-03-07 Martin Liska <mliska@suse.cz>
18042 PR translation/90148
18043 * config/rs6000/rs6000.cc (rs6000_linux64_override_options): Put
18044 quote to a proper place.
18045 * plugin.cc (default_plugin_dir_name): Likewise.
18047 2022-03-07 Martin Liska <mliska@suse.cz>
18050 * config/rx/rx.cc (rx_expand_builtin_mvtc): Fix translation
18053 2022-03-07 Jakub Jelinek <jakub@redhat.com>
18056 * config/i386/sse.md (avx512dq_mul<mode>3<mask_name>): New
18057 define_expand pattern. Rename define_insn to ...
18058 (*avx512dq_mul<mode>3<mask_name>): ... this.
18059 (<code><mode>3_mask): New any_logic define_expand pattern.
18060 (<mask_codefor><code><mode>3<mask_name>): Rename to ...
18061 (*<code><mode>3<mask_name>): ... this.
18063 2022-03-05 Jakub Jelinek <jakub@redhat.com>
18065 * gimple-ssa-warn-access.cc (pass_waccess::use_after_inval_p): Remove
18066 visited bitmap and its use. Also punt on EDGE_ABNORMAL edges.
18068 2022-03-05 Roger Sayle <roger@nextmovesoftware.com>
18069 Uroš Bizjak <ubizjak@gmail.com>
18071 PR testsuite/104732
18072 * config/i386/i386.md (SWIM1248x): Renamed from SWIM1248s.
18073 Include DI mode unconditionally.
18074 (*anddi3_doubleword): Remove && TARGET_STV && TARGET_SSE2 condition,
18075 i.e. always split on !TARGET_64BIT.
18076 (*<any_or>di3_doubleword): Likewise.
18077 (*one_cmpldi2_doubleword): Likewise.
18078 (and<mode>3 expander): Update to use SWIM1248x from SWIM1248s.
18079 (<any_or><mode>3 expander): Likewise.
18080 (one_cmpl<mode>2 expander): Likewise.
18082 2022-03-05 Michael Meissner <meissner@linux.ibm.com>
18085 * config/rs6000/vsx.md (UNSPEC_MTVSRD_DITI_W1): Delete.
18086 (mtvsrdd_diti_w1): Delete.
18087 (extendditi2): Convert from define_expand to
18088 define_insn_and_split. Replace with code to deal with both GPR
18089 registers and with altivec registers.
18091 2022-03-04 Segher Boessenkool <segher@kernel.crashing.org>
18093 * config/rs6000/rs6000.cc (rs6000_machine_from_flags): Restructure a
18094 bit. Handle most older CPUs.
18096 2022-03-04 Iain Sandoe <iain@sandoe.co.uk>
18098 * config/darwin.cc (darwin_fold_builtin): Make fcode an int to
18099 avoid a mismatch with DECL_MD_FUNCTION_CODE().
18101 2022-03-04 Iain Sandoe <iain@sandoe.co.uk>
18104 * config/rs6000/darwin.md (@machopic_high_<mode>): New.
18105 (@machopic_low_<mode>): New.
18106 * config/rs6000/predicates.md (macho_pic_address): New.
18107 * config/rs6000/rs6000.cc (rs6000_legitimize_address): Do not
18108 apply the TLS processing to Darwin.
18109 * lra-constraints.cc (process_address_1): Revert the changes
18112 2022-03-04 Peter Bergner <bergner@linux.ibm.com>
18116 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Make the
18117 ISA 2.06 requirement for -mabi=ieeelongdouble conditional on
18119 Move the -mabi=ieeelongdouble and -mabi=ibmlongdouble error checking
18121 * common/config/rs6000/rs6000-common.cc (rs6000_handle_option):
18124 2022-03-04 Jakub Jelinek <jakub@redhat.com>
18126 PR middle-end/104529
18127 * gimplify.cc (gimplify_init_constructor): Clear TREE_READONLY
18128 on automatic objects which will be runtime initialized.
18130 2022-03-04 Thomas Schwinge <thomas@codesourcery.com>
18132 PR middle-end/100280
18133 PR middle-end/104132
18134 PR middle-end/104133
18135 * omp-low.cc (task_shared_vars): Rename to
18136 'make_addressable_vars'. Adjust all users.
18137 (scan_sharing_clauses) <OMP_CLAUSE_MAP> Use it for
18138 'OMP_CLAUSE_MAP_DECL_MAKE_ADDRESSABLE' DECLs, too.
18140 2022-03-04 Thomas Schwinge <thomas@codesourcery.com>
18142 PR middle-end/100280
18143 * tree.h (OMP_CLAUSE_MAP_DECL_MAKE_ADDRESSABLE): New.
18144 * tree-core.h: Document it.
18145 * omp-low.cc (scan_sharing_clauses) <OMP_CLAUSE_MAP>: Handle
18146 'OMP_CLAUSE_MAP_DECL_MAKE_ADDRESSABLE'.
18147 * omp-oacc-kernels-decompose.cc (maybe_build_inner_data_region):
18148 Set 'OMP_CLAUSE_MAP_DECL_MAKE_ADDRESSABLE' instead of
18149 'TREE_ADDRESSABLE'.
18151 2022-03-04 Thomas Schwinge <thomas@codesourcery.com>
18153 PR middle-end/100280
18154 * omp-oacc-kernels-decompose.cc (maybe_build_inner_data_region):
18155 Add diagnostic: "note: OpenACC 'kernels' decomposition: variable
18156 '[...]' declared in block made addressable".
18158 2022-03-04 Thomas Schwinge <thomas@codesourcery.com>
18160 PR middle-end/100400
18161 PR middle-end/103836
18162 PR middle-end/104061
18163 * omp-oacc-kernels-decompose.cc (decompose_kernels_region_body):
18164 Catch 'GIMPLE_DEBUG'.
18166 2022-03-04 Jakub Jelinek <jakub@redhat.com>
18169 * tree.cc (warn_deprecated_use): For types prefer to use node
18170 and only use TYPE_MAIN_VARIANT (node) if TYPE_STUB_DECL (node) is
18173 2022-03-04 H.J. Lu <hjl.tools@gmail.com>
18176 * config/i386/i386.cc (ix86_gen_scratch_sse_rtx): Always return
18179 2022-03-03 Martin Sebor <msebor@redhat.com>
18181 PR middle-end/104761
18182 * gimple-ssa-warn-access.cc (pass_waccess::execute): Call
18183 mark_dfs_back_edges.
18185 2022-03-03 Martin Liska <mliska@suse.cz>
18187 * configure.ac: Use linker plug-in by default.
18188 * configure: Regenerate.
18190 2022-03-03 Martin Liska <mliska@suse.cz>
18192 * configure.ac: Now ld.mold support LTO plugin API, use it.
18193 * configure: Regenerate.
18195 2022-03-03 Tom de Vries <tdevries@suse.de>
18197 * config/nvptx/t-nvptx (MULTILIB_EXTRA_OPTS): Add mptx=3.1.
18199 2022-03-03 Tom de Vries <tdevries@suse.de>
18202 * config/nvptx/t-nvptx (MULTILIB_EXTRA_OPTS): Add misa=sm_30.
18204 2022-03-03 Tom de Vries <tdevries@suse.de>
18206 * config/nvptx/nvptx.h (ASM_SPEC): Add %{misa=sm_30:--no-verify}.
18208 2022-03-03 Jakub Jelinek <jakub@redhat.com>
18210 PR middle-end/104757
18211 * gimplify.cc (gimplify_omp_loop): Call gimplify_expr rather than
18213 (gimplify_expr) <case OMP_SIMD>: Temporarily disable
18214 gimplify_ctxp->into_ssa around call to gimplify_omp_for.
18216 2022-03-03 Jakub Jelinek <jakub@redhat.com>
18218 PR middle-end/104558
18219 * calls.cc (store_one_arg): When not calling emit_push_insn
18220 because size_rtx is const0_rtx, call at least anti_adjust_stack
18221 on arg->locate.alignment_pad if !argblock and the alignment might
18224 2022-03-02 Alexandre Oliva <oliva@adacore.com>
18226 * lra-constraints.cc (undo_optional_reloads): Recognize and
18227 drop insns of multi-word move sequences, tolerate removal
18228 iteration on an already-removed clobber, and refuse to
18229 substitute original pseudos into clobbers.
18231 2022-03-02 Qing Zhao <qing.zhao@oracle.com>
18233 PR middle-end/102276
18234 * common.opt (-Wtrivial-auto-var-init): New option.
18235 * doc/invoke.texi (-Wtrivial-auto-var-init): Document new option.
18236 (-ftrivial-auto-var-init): Update option;
18237 * gimplify.cc (emit_warn_switch_unreachable): New function.
18238 (warn_switch_unreachable_r): Rename to ...
18239 (warn_switch_unreachable_and_auto_init_r): This.
18240 (maybe_warn_switch_unreachable): Rename to ...
18241 (maybe_warn_switch_unreachable_and_auto_init): This.
18242 (gimplify_switch_expr): Update calls to renamed function.
18244 2022-03-02 Richard Biener <rguenther@suse.de>
18246 PR rtl-optimization/104686
18247 * ira-color.cc (object_conflicts_with_allocno_p): New function
18248 using a bitvector test instead of iterating when possible.
18249 (allocnos_conflict_p): Choose the best allocno to iterate over
18251 (update_conflict_hard_regno_costs): Do allocnos_conflict_p test
18254 2022-03-02 Jakub Jelinek <jakub@redhat.com>
18256 * cfg.cc (dump_edge_info): Dump goto_locus if present.
18258 2022-03-02 Jakub Jelinek <jakub@redhat.com>
18260 PR rtl-optimization/104589
18261 * cfgrtl.cc (fixup_reorder_chain): Use loc_equal instead of direct
18262 INSN_LOCATION comparison with goto_locus.
18264 2022-03-02 Jakub Jelinek <jakub@redhat.com>
18266 * tree-ssa-strlen.cc (strlen_pass::handle_assign,
18267 strlen_pass::before_dom_children): Comment spelling fixes.
18269 2022-03-02 Jakub Jelinek <jakub@redhat.com>
18271 * ipa-modref-tree.cc (modref_access_node::contains,
18272 modref_access_node::closer_pair_p, modref_access_node::insert,
18273 modref_access_node::insert_kill): Comment spelling fixes.
18274 * ipa-modref.cc: Likewise.
18275 (modref_summary::finalize, ignore_nondeterminism_p,
18276 class modref_access_analysis,
18277 modref_access_analysis::set_side_effects,
18278 modref_access_analysis::set_nondeterministic,
18279 modref_access_analysis::record_global_memory_load,
18280 modref_access_analysis::propagate, modref_access_analysis::analyze,
18281 struct escape_point, class modref_lattice, modref_lattice::merge,
18282 modref_lattice::merge_deref, class modref_eaf_analysis,
18283 modref_eaf_analysis::merge_call_lhs_flags,
18284 modref_eaf_analysis::analyze_ssa_name, modref_eaf_analysis::propagate,
18285 modref_eaf_analysis::record_escape_points, remap_kills,
18286 update_escape_summary, remove_useless_summaries,
18287 ipa_merge_modref_summary_after_inlining, pass_ipa_modref::execute):
18289 * ipa-modref.h (struct modref_summary, interposable_eaf_flags):
18291 * ipa-modref-tree.h (enum modref_special_parms,
18292 struct modref_access_node): Likewise.
18294 2022-03-01 Jakub Jelinek <jakub@redhat.com>
18296 PR tree-optimization/104715
18297 * gimple-ssa-warn-access.cc (pass_waccess::check_pointer_uses): Don't
18298 unnecessarily test if ptr is a SSA_NAME, it has to be. Only push lhs
18299 of a call if gimple_call_return_arg is equal to ptr, not just when it
18302 2022-03-01 Jakub Jelinek <jakub@redhat.com>
18304 * gimple-ssa-warn-access.cc (warn_string_no_nul,
18305 maybe_warn_nonstring_arg, fndecl_alloc_p, new_delete_mismatch_p,
18306 matching_alloc_calls_p, maybe_warn_alloc_args_overflow,
18307 pass_waccess::check_alloca, pass_waccess::check_strcat,
18308 memmodel_to_uhwi, fntype_argno_type,
18309 pass_waccess::maybe_check_access_sizes,
18310 pass_waccess::check_call_access,
18311 pass_waccess::maybe_check_dealloc_call, pass_waccess::check_stmt):
18312 Comment spelling fixes.
18314 2022-03-01 Richard Biener <rguenther@suse.de>
18316 PR tree-optimization/104716
18317 * tree-loop-distribution.cc (find_seed_stmts_for_distribution):
18318 Check if we can copy the loop.
18320 2022-03-01 H.J. Lu <hjl.tools@gmail.com>
18322 PR middle-end/104721
18323 * cfgexpand.cc (expand_gimple_basic_block): Clear
18324 currently_expanding_gimple_stmt when returning inside the loop.
18326 2022-03-01 Martin Liska <mliska@suse.cz>
18329 * multiple_target.cc (get_attr_len): Move to tree.c.
18330 (expand_target_clones): Remove single value checking.
18331 * tree.cc (get_target_clone_attr_len): New fn.
18332 * tree.h (get_target_clone_attr_len): Likewise.
18334 2022-03-01 Martin Liska <mliska@suse.cz>
18336 PR gcov-profile/104677
18337 * doc/invoke.texi: Document more .gcda file name generation.
18339 2022-03-01 Tom de Vries <tdevries@suse.de>
18342 * config/nvptx/nvptx.cc (nvptx_gen_shuffle): Handle DCmode and CDImode.
18343 * config/nvptx/nvptx.md
18344 (define_predicate "nvptx_register_or_complex_di_df_register_operand"):
18346 (define_expand "omp_simt_xchg_bfly", define_expand "omp_simt_xchg_idx"):
18347 Use nvptx_register_or_complex_di_df_register_operand.
18349 2022-03-01 Tom de Vries <tdevries@suse.de>
18351 * config.gcc (nvptx*-*-*): Add nvptx/nvptx-gen.opt to extra_options.
18352 * config/nvptx/gen-copyright.sh: New file.
18353 * config/nvptx/gen-h.sh: New file.
18354 * config/nvptx/gen-opt.sh: New file.
18355 * config/nvptx/nvptx.h (TARGET_SM35, TARGET_SM53, TARGET_SM70)
18356 (TARGET_SM75, TARGET_SM80): Move ...
18357 * config/nvptx/nvptx-gen.h: ... here. New file, generate.
18358 * config/nvptx/nvptx.opt (Enum ptx_isa): Move ...
18359 * config/nvptx/nvptx-gen.opt: ... here. New file, generate.
18360 * config/nvptx/t-nvptx ($(srcdir)/config/nvptx/nvptx-gen.h)
18361 ($(srcdir)/config/nvptx/nvptx-gen.opt): New make target.
18363 2022-03-01 Tom de Vries <tdevries@suse.de>
18365 * config/nvptx/gen-omp-device-properties.sh: New file.
18366 * config/nvptx/t-omp-device: Use gen-omp-device-properties.sh.
18368 2022-03-01 Tom de Vries <tdevries@suse.de>
18370 * config/nvptx/nvptx-sm.def: New file.
18371 * config/nvptx/nvptx-c.cc (nvptx_cpu_cpp_builtins): Use nvptx-sm.def.
18372 * config/nvptx/nvptx-opts.h (enum ptx_isa): Same.
18373 * config/nvptx/nvptx.cc (sm_version_to_string)
18374 (nvptx_omp_device_kind_arch_isa): Same.
18376 2022-03-01 Robin Dapp <rdapp@linux.ibm.com>
18378 PR rtl-optimization/104154
18379 * config/arc/arc.cc (gen_compare_reg): Return the CC-mode
18380 comparison ifcvt passed us.
18382 2022-03-01 Hongyu Wang <hongyu.wang@intel.com>
18385 * config/i386/i386-expand.cc (ix86_expand_vector_init_duplicate):
18386 Use vec_setv8hf_0 for HF to V8HFmode move instead of subreg.
18388 2022-02-28 Roger Sayle <roger@nextmovesoftware.com>
18390 PR tree-optimization/91384
18391 * config/i386/i386.md (peephole2): Eliminate final testl insn
18392 from the sequence *movsi_internal, *negsi_1, *cmpsi_ccno_1 by
18393 transforming using *negsi_2 for the negation.
18395 2022-02-28 Roger Sayle <roger@nextmovesoftware.com>
18396 Eric Botcazou <ebotcazou@adacore.com>
18398 PR middle-end/80270
18399 * expmed.cc (extract_integral_bit_field): If OP0 is a hard
18400 register, copy it to a pseudo before calling simplify_gen_subreg.
18402 2022-02-28 Vladimir N. Makarov <vmakarov@redhat.com>
18404 PR rtl-optimization/104637
18405 * lra-assigns.cc (lra_split_hard_reg_for): Split hard regs as many
18406 as possible on one subpass.
18408 2022-02-28 Qing Zhao <qing.zhao@oracle.com>
18410 PR middle-end/104550
18411 * gimple-fold.cc (clear_padding_flush): Suppress warnings for new
18414 2022-02-28 Martin Liska <mliska@suse.cz>
18417 * main.cc (main): Use flag_checking instead of CHECKING_P
18418 and run toplev::finalize only if there is not error seen.
18420 2022-02-28 Richard Biener <rguenther@suse.de>
18422 * tree-ssa-pre.cc (compute_avail): Revert part of last change.
18424 2022-02-28 Richard Biener <rguenther@suse.de>
18426 PR tree-optimization/104700
18427 * tree-ssa-pre.cc (get_or_alloc_expr_for): Remove and inline
18429 (find_or_generate_expression): ... here, simplifying code.
18431 2022-02-28 Tom de Vries <tdevries@suse.de>
18433 * config/nvptx/nvptx-opts.h (enum ptx_version): Add
18434 PTX_VERSION_default.
18435 * config/nvptx/nvptx.cc (handle_ptx_version_option): Handle
18436 PTX_VERSION_default.
18437 * config/nvptx/nvptx.opt: Add EnumValue "_" / PTX_VERSION_default.
18439 2022-02-28 Richard Biener <rguenther@suse.de>
18441 PR rtl-optimization/104686
18442 * ira-int.h (minmax_set_iter_cond): Use ctz_hwi to elide loop
18443 skipping bits that are zero.
18444 (ira_object_conflict_iter_cond): Likewise.
18446 2022-02-28 Hongyu Wang <hongyu.wang@intel.com>
18448 * config/i386/avx512fintrin.h (_MM_TERNLOG_ENUM): New enum.
18449 (_mm512_ternarylogic_epi64): Truncate imm to unsigned
18450 char to avoid error when using ~enum as parameter.
18451 (_mm512_mask_ternarylogic_epi64): Likewise.
18452 (_mm512_maskz_ternarylogic_epi64): Likewise.
18453 (_mm512_ternarylogic_epi32): Likewise.
18454 (_mm512_mask_ternarylogic_epi32): Likewise.
18455 (_mm512_maskz_ternarylogic_epi32): Likewise.
18456 * config/i386/avx512vlintrin.h (_mm256_ternarylogic_epi64):
18457 Adjust imm param type to unsigned char.
18458 (_mm256_mask_ternarylogic_epi64): Likewise.
18459 (_mm256_maskz_ternarylogic_epi64): Likewise.
18460 (_mm256_ternarylogic_epi32): Likewise.
18461 (_mm256_mask_ternarylogic_epi32): Likewise.
18462 (_mm256_maskz_ternarylogic_epi32): Likewise.
18463 (_mm_ternarylogic_epi64): Likewise.
18464 (_mm_mask_ternarylogic_epi64): Likewise.
18465 (_mm_maskz_ternarylogic_epi64): Likewise.
18466 (_mm_ternarylogic_epi32): Likewise.
18467 (_mm_mask_ternarylogic_epi32): Likewise.
18468 (_mm_maskz_ternarylogic_epi32): Likewise.
18470 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18471 Marc Glisse <marc.glisse@inria.fr>
18473 PR tree-optimization/104675
18474 * match.pd (t * 2U / 2 -> t & (~0 / 2), t / 2U * 2 -> t & ~1):
18475 Restrict simplifications to INTEGRAL_TYPE_P.
18477 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18480 * config/rs6000/vector.md (movmisalign<mode>): Use rs6000_emit_move.
18482 2022-02-25 Claudiu Zissulescu <claziss@synopsys.com>
18484 * config/arc/arc.cc (gen_compare_reg): Return NULL_RTX if the
18485 comparison is not valid.
18486 * config/arc/arc.md (movsicc): Fail if comparison is not valid.
18487 (movdicc): Likewise.
18488 (movsfcc): Likewise.
18489 (movdfcc): Likewise.
18491 2022-02-25 Richard Biener <rguenther@suse.de>
18493 PR tree-optimization/103037
18494 * tree-ssa-sccvn.h (alloc_vn_nary_op_noinit): Declare.
18495 (vn_nary_length_from_stmt): Likewise.
18496 (init_vn_nary_op_from_stmt): Likewise.
18497 (vn_nary_op_compute_hash): Likewise.
18498 * tree-ssa-sccvn.cc (alloc_vn_nary_op_noinit): Export.
18499 (vn_nary_length_from_stmt): Likewise.
18500 (init_vn_nary_op_from_stmt): Likewise.
18501 (vn_nary_op_compute_hash): Likewise.
18502 * tree-ssa-pre.cc (pre_expr_obstack): New obstack.
18503 (get_or_alloc_expr_for_nary): Pass in the value-id to use,
18504 (re-)compute the hash value and if the expression is not
18505 found allocate it from pre_expr_obstack.
18506 (phi_translate_1): Do not insert the NARY found in the
18507 VN tables but build a PRE expression from the valueized
18508 NARY with the value-id we eventually found.
18509 (find_or_generate_expression): Assert we have an entry
18510 for constant values.
18511 (compute_avail): Insert not valueized expressions into
18512 EXP_GEN using the value-id from the VN tables.
18513 (init_pre): Allocate pre_expr_obstack.
18514 (fini_pre): Free pre_expr_obstack.
18516 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18519 * config/i386/i386.h (enum ix86_stack_slot): Add SLOT_FLOATxFDI_387.
18520 * config/i386/i386.md (splitter to floatdi<mode>2_i387_with_xmm): Use
18521 SLOT_FLOATxFDI_387 rather than SLOT_TEMP.
18523 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18525 * warning-control.cc (get_nowarn_spec): Comment spelling fix.
18527 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18529 PR middle-end/104679
18530 * internal-fn.cc (expand_SPACESHIP): Call do_pending_stack_adjust.
18532 2022-02-25 Jakub Jelinek <jakub@redhat.com>
18534 PR tree-optimization/104675
18535 * match.pd (-A - 1 -> ~A, -1 - A -> ~A): Don't simplify for
18538 2022-02-25 Alexandre Oliva <oliva@adacore.com>
18542 * expr.cc (emit_move_multi_word): Restore clobbers during LRA.
18544 2022-02-25 Alexandre Oliva <oliva@adacore.com>
18546 PR middle-end/104540
18547 * dwarf2cfi.cc (cfi_oprnd_equal_p): Cope with NULL
18550 2022-02-25 Alexandre Oliva <oliva@adacore.com>
18552 PR tree-optimization/103856
18553 * gimple-harden-conditionals.cc (non_eh_succ_edge): Enable the
18554 eh edge to be requested through an extra parameter.
18555 (pass_harden_compares::execute): Copy PHI args in the EH dest
18556 block for the new EH edge added for the inverted compare.
18558 2022-02-24 Palmer Dabbelt <palmer@rivosinc.com>
18560 * doc/invoke.texi (RISC-V -mcmodel=medany): Document the degree
18561 of position independence that -mcmodel=medany affords.
18563 2022-02-24 Jose E. Marchesi <jose.marchesi@oracle.com>
18566 * configure.ac: --disable-gcov if targetting bpf-*.
18567 * configure: Regenerate.
18569 2022-02-24 Richard Biener <rguenther@suse.de>
18571 PR tree-optimization/104676
18572 * tree-loop-distribution.cc (loop_distribution::execute):
18573 Do a full scev_reset.
18575 2022-02-24 Jakub Jelinek <jakub@redhat.com>
18577 PR tree-optimization/104601
18578 * tree-ssa-sccvn.cc (visit_reference_op_call): For calls with
18579 non-SSA_NAME lhs value number vdef to itself instead of e.g. the
18582 2022-02-24 Tom de Vries <tdevries@suse.de>
18583 Tobias Burnus <tobias@codesourcery.com>
18585 * config/nvptx/nvptx.cc (nvptx_omp_device_kind_arch_isa): Handle
18586 sm_70, sm_75 and sm_80.
18587 * config/nvptx/t-omp-device: Add sm_53, sm_70, sm_75 and sm_80.
18589 2022-02-24 Tom de Vries <tdevries@suse.de>
18591 * config/nvptx/nvptx.md (define_insn "rotlsi3", define_insn
18592 "rotrsi3"): New define_insn.
18594 2022-02-24 Tom de Vries <tdevries@suse.de>
18596 * config/nvptx/nvptx.cc (gen_comment): Use
18597 DECL_SOURCE_LOCATION (cfun->decl) instead of cfun->function_start_locus.
18599 2022-02-24 liuhongt <hongtao.liu@intel.com>
18601 * config/i386/sse.md (<code>v1ti3): Add suffix and replace
18602 isa attr of alternative 2 from avx to avx512vl.
18604 2022-02-23 Richard Biener <rguenther@suse.de>
18605 Jakub Jelinek <jakub@redhat.com>
18607 PR tree-optimization/104644
18608 * doc/match-and-simplify.texi: Amend ! documentation.
18609 * genmatch.cc (expr::gen_transform): Code-generate ! support
18611 (parser::parse_expr): Allow ! for GENERIC.
18612 * match.pd (cmp (bswap @0) INTEGER_CST@1): Use ! modifier on
18615 2022-02-23 Richard Biener <rguenther@suse.de>
18617 PR tree-optimization/101636
18618 * tree-vect-slp.cc (vect_print_slp_tree): Dump the
18619 vector type of the node.
18620 (vect_slp_analyze_operations): Make sure the CTOR
18621 is vectorized with an expected type.
18622 (vectorize_slp_instance_root_stmt): Revert r10-5979 fix.
18624 2022-02-23 Jakub Jelinek <jakub@redhat.com>
18627 * gimple-warn-recursion.cc (pass_warn_recursion::find_function_exit):
18628 Don't warn about calls to corresponding builtin from extern inline
18629 gnu_inline wrappers.
18631 2022-02-23 Roger Sayle <roger@nextmovesoftware.com>
18634 * config/nvptx/nvptx.md (*movhf_insn): Add subregs_ok attribute.
18636 2022-02-23 Christophe Lyon <christophe.lyon@arm.com>
18640 * config/arm/vfp.md (thumb2_movhi_vfp, thumb2_movhi_fp16): Fix
18643 2022-02-23 Cui,Lili <lili.cui@intel.com>
18645 * doc/invoke.texi: Update documents for Intel architectures.
18647 2022-02-22 Roger Sayle <roger@nextmovesoftware.com>
18649 * config/i386/i386-expand.cc (ix86_expand_cmpxchg_loop): Restore
18652 2022-02-22 Thomas Schwinge <thomas@codesourcery.com>
18654 * omp-low.cc (omp_build_component_ref): Move function...
18655 * omp-general.cc (omp_build_component_ref): ... here. Remove
18657 * omp-general.h (omp_build_component_ref): Declare function.
18658 * omp-oacc-neuter-broadcast.cc (oacc_build_component_ref): Remove
18660 (build_receiver_ref, build_sender_ref): Call
18661 'omp_build_component_ref' instead.
18663 2022-02-22 Thomas Schwinge <thomas@codesourcery.com>
18665 * omp-oacc-neuter-broadcast.cc (record_field_map_t): Further
18666 simplify. Adjust all users.
18668 2022-02-22 Segher Boessenkool <segher@kernel.crashing.org>
18671 * config/rs6000/rs6000.cc (atomic_hold_decl, atomic_clear_decl,
18672 atomic_update_decl): Add GTY markup.
18674 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18676 * config/arm/arm.h (REG_CLASS_CONTENTS): Add VPR_REG to ALL_REGS.
18678 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18682 * config/arm/arm-builtins.cc (CX_UNARY_UNONE_QUALIFIERS): Use
18684 (CX_BINARY_UNONE_QUALIFIERS): Likewise.
18685 (CX_TERNARY_UNONE_QUALIFIERS): Likewise.
18686 (TERNOP_NONE_NONE_NONE_UNONE_QUALIFIERS): Delete.
18687 (QUADOP_NONE_NONE_NONE_NONE_UNONE_QUALIFIERS): Delete.
18688 (QUADOP_UNONE_UNONE_UNONE_UNONE_UNONE_QUALIFIERS): Delete.
18689 * config/arm/arm_mve_builtins.def: Use predicated qualifiers.
18690 * config/arm/mve.md: Use VxBI instead of HI.
18692 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18696 * config/arm/arm-builtins.cc (STRSBS_P_QUALIFIERS): Use predicate
18698 (STRSBU_P_QUALIFIERS): Likewise.
18699 (LDRGBS_Z_QUALIFIERS): Likewise.
18700 (LDRGBU_Z_QUALIFIERS): Likewise.
18701 (LDRGBWBXU_Z_QUALIFIERS): Likewise.
18702 (LDRGBWBS_Z_QUALIFIERS): Likewise.
18703 (LDRGBWBU_Z_QUALIFIERS): Likewise.
18704 (STRSBWBS_P_QUALIFIERS): Likewise.
18705 (STRSBWBU_P_QUALIFIERS): Likewise.
18706 * config/arm/mve.md: Use VxBI instead of HI.
18708 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18712 * config/arm/arm-builtins.cc (TERNOP_UNONE_UNONE_NONE_UNONE_QUALIFIERS): Change to ...
18713 (TERNOP_UNONE_UNONE_NONE_PRED_QUALIFIERS): ... this.
18714 (TERNOP_UNONE_UNONE_IMM_UNONE_QUALIFIERS): Change to ...
18715 (TERNOP_UNONE_UNONE_IMM_PRED_QUALIFIERS): ... this.
18716 (TERNOP_NONE_NONE_IMM_UNONE_QUALIFIERS): Change to ...
18717 (TERNOP_NONE_NONE_IMM_PRED_QUALIFIERS): ... this.
18718 (TERNOP_NONE_NONE_UNONE_UNONE_QUALIFIERS): Change to ...
18719 (TERNOP_NONE_NONE_UNONE_PRED_QUALIFIERS): ... this.
18720 (QUADOP_UNONE_UNONE_NONE_NONE_UNONE_QUALIFIERS): Change to ...
18721 (QUADOP_UNONE_UNONE_NONE_NONE_PRED_QUALIFIERS): ... this.
18722 (QUADOP_NONE_NONE_NONE_NONE_PRED_QUALIFIERS): New.
18723 (QUADOP_NONE_NONE_NONE_IMM_UNONE_QUALIFIERS): Change to ...
18724 (QUADOP_NONE_NONE_NONE_IMM_PRED_QUALIFIERS): ... this.
18725 (QUADOP_UNONE_UNONE_UNONE_UNONE_PRED_QUALIFIERS): New.
18726 (QUADOP_UNONE_UNONE_NONE_IMM_UNONE_QUALIFIERS): Change to ...
18727 (QUADOP_UNONE_UNONE_NONE_IMM_PRED_QUALIFIERS): ... this.
18728 (QUADOP_NONE_NONE_UNONE_IMM_UNONE_QUALIFIERS): Change to ...
18729 (QUADOP_NONE_NONE_UNONE_IMM_PRED_QUALIFIERS): ... this.
18730 (QUADOP_UNONE_UNONE_UNONE_IMM_UNONE_QUALIFIERS): Change to ...
18731 (QUADOP_UNONE_UNONE_UNONE_IMM_PRED_QUALIFIERS): ... this.
18732 (QUADOP_UNONE_UNONE_UNONE_NONE_UNONE_QUALIFIERS): Change to ...
18733 (QUADOP_UNONE_UNONE_UNONE_NONE_PRED_QUALIFIERS): ... this.
18734 (STRS_P_QUALIFIERS): Use predicate qualifier.
18735 (STRU_P_QUALIFIERS): Likewise.
18736 (STRSU_P_QUALIFIERS): Likewise.
18737 (STRSS_P_QUALIFIERS): Likewise.
18738 (LDRGS_Z_QUALIFIERS): Likewise.
18739 (LDRGU_Z_QUALIFIERS): Likewise.
18740 (LDRS_Z_QUALIFIERS): Likewise.
18741 (LDRU_Z_QUALIFIERS): Likewise.
18742 (QUINOP_UNONE_UNONE_UNONE_UNONE_IMM_UNONE_QUALIFIERS): Change to ...
18743 (QUINOP_UNONE_UNONE_UNONE_UNONE_IMM_PRED_QUALIFIERS): ... this.
18744 (BINOP_NONE_NONE_PRED_QUALIFIERS): New.
18745 (BINOP_UNONE_UNONE_PRED_QUALIFIERS): New.
18746 * config/arm/arm_mve_builtins.def: Use new predicated qualifiers.
18747 * config/arm/mve.md: Use MVE_VPRED instead of HI.
18749 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18753 * config/arm/arm-builtins.cc (BINOP_UNONE_NONE_NONE_QUALIFIERS):
18755 (TERNOP_UNONE_NONE_NONE_UNONE_QUALIFIERS): Change to ...
18756 (TERNOP_PRED_NONE_NONE_PRED_QUALIFIERS): ... this.
18757 (TERNOP_PRED_UNONE_UNONE_PRED_QUALIFIERS): New.
18758 * config/arm/arm_mve_builtins.def (vcmp*q_n_, vcmp*q_m_f): Use new
18759 predicated qualifiers.
18760 * config/arm/mve.md (mve_vcmp<mve_cmp_op>q_n_<mode>)
18761 (mve_vcmp*q_m_f<mode>): Use MVE_VPRED instead of HI.
18763 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18766 * config/arm/arm-protos.h (arm_get_mask_mode): New prototype.
18767 (arm_expand_vector_compare): Update prototype.
18768 * config/arm/arm.cc (TARGET_VECTORIZE_GET_MASK_MODE): New.
18769 (arm_vector_mode_supported_p): Add support for VxBI modes.
18770 (arm_expand_vector_compare): Remove useless generation of vpsel.
18771 (arm_expand_vcond): Fix select operands.
18772 (arm_get_mask_mode): New.
18773 * config/arm/mve.md (vec_cmp<mode><MVE_vpred>): New.
18774 (vec_cmpu<mode><MVE_vpred>): New.
18775 (vcond_mask_<mode><MVE_vpred>): New.
18776 * config/arm/vec-common.md (vec_cmp<mode><v_cmp_result>)
18777 (vec_cmpu<mode><mode, vcond_mask_<mode><v_cmp_result>): Move to ...
18778 * config/arm/neon.md (vec_cmp<mode><v_cmp_result>)
18779 (vec_cmpu<mode><mode, vcond_mask_<mode><v_cmp_result>): ... here
18780 and disable for MVE.
18781 * doc/sourcebuild.texi (arm_mve): Document new effective-target.
18783 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18787 * config/arm/arm-builtins.cc (BINOP_PRED_UNONE_UNONE_QUALIFIERS)
18788 (BINOP_PRED_NONE_NONE_QUALIFIERS)
18789 (TERNOP_NONE_NONE_NONE_PRED_QUALIFIERS)
18790 (TERNOP_UNONE_UNONE_UNONE_PRED_QUALIFIERS): New.
18791 * config/arm/arm-protos.h (mve_bool_vec_to_const): New.
18792 * config/arm/arm.cc (arm_hard_regno_mode_ok): Handle new VxBI
18794 (arm_mode_to_pred_mode): New.
18795 (arm_expand_vector_compare): Use the right VxBI mode instead of
18797 (arm_expand_vcond): Likewise.
18798 (simd_valid_immediate): Handle MODE_VECTOR_BOOL.
18799 (mve_bool_vec_to_const): New.
18800 (neon_make_constant): Call mve_bool_vec_to_const when needed.
18801 * config/arm/arm_mve_builtins.def (vcmpneq_, vcmphiq_, vcmpcsq_)
18802 (vcmpltq_, vcmpleq_, vcmpgtq_, vcmpgeq_, vcmpeqq_, vcmpneq_f)
18803 (vcmpltq_f, vcmpleq_f, vcmpgtq_f, vcmpgeq_f, vcmpeqq_f, vpselq_u)
18804 (vpselq_s, vpselq_f): Use new predicated qualifiers.
18805 * config/arm/constraints.md (DB): New.
18806 * config/arm/iterators.md (MVE_7, MVE_7_HI): New mode iterators.
18807 (MVE_VPRED, MVE_vpred): New attribute iterators.
18808 * config/arm/mve.md (@mve_vcmp<mve_cmp_op>q_<mode>)
18809 (@mve_vcmp<mve_cmp_op>q_f<mode>, @mve_vpselq_<supf><mode>)
18810 (@mve_vpselq_f<mode>): Use MVE_VPRED instead of HI.
18811 (@mve_vpselq_<supf>v2di): Define separately.
18812 (mov<mode>): New expander for VxBI modes.
18813 * config/arm/vfp.md (thumb2_movhi_vfp, thumb2_movhi_fp16): Use
18814 MVE_7_HI iterator and add support for DB constraint.
18816 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18817 Richard Sandiford <richard.sandiford@arm.com>
18821 * config/aarch64/aarch64-modes.def (VNx16BI, VNx8BI, VNx4BI,
18822 VNx2BI): Update definition.
18823 * config/arm/arm-builtins.cc (arm_init_simd_builtin_types): Add new
18825 (arm_init_builtin): Map predicate vectors arguments to HImode.
18826 (arm_expand_builtin_args): Move HImode predicate arguments to VxBI
18827 rtx. Move return value to HImode rtx.
18828 * config/arm/arm-builtins.h (arm_type_qualifiers): Add qualifier_predicate.
18829 * config/arm/arm-modes.def (B2I, B4I, V16BI, V8BI, V4BI): New modes.
18830 * config/arm/arm-simd-builtin-types.def (Pred1x16_t,
18831 Pred2x8_t,Pred4x4_t): New.
18832 * emit-rtl.cc (init_emit_once): Handle all boolean modes.
18833 * genmodes.cc (mode_data): Add boolean field.
18834 (blank_mode): Initialize it.
18835 (make_complex_modes): Fix handling of boolean modes.
18836 (make_vector_modes): Likewise.
18837 (VECTOR_BOOL_MODE): Use new COMPONENT parameter.
18838 (make_vector_bool_mode): Likewise.
18840 (make_bool_mode): New.
18841 (emit_insn_modes_h): Fix generation of boolean modes.
18842 (emit_class_narrowest_mode): Likewise.
18843 * machmode.def: (VECTOR_BOOL_MODE): Document new COMPONENT
18844 parameter. Use new BOOL_MODE instead of FRACTIONAL_INT_MODE to
18846 * rtx-vector-builder.cc (rtx_vector_builder::find_cached_value):
18847 Fix handling of constm1_rtx for VECTOR_BOOL.
18848 * simplify-rtx.cc (native_encode_rtx): Fix support for VECTOR_BOOL.
18849 (native_decode_vector_rtx): Likewise.
18850 (test_vector_ops_duplicate): Skip vec_merge test
18851 with vectors of booleans.
18852 * varasm.cc (output_constant_pool_2): Likewise.
18854 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18856 * config/arm/mve.md (mve_vmvnq_n_<supf><mode>): Use V_elem mode
18859 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18861 * config/arm/arm.cc (arm_class_likely_spilled_p): Handle VPR_REG.
18863 2022-02-22 Christophe Lyon <christophe.lyon@arm.com>
18865 * config/arm/arm.h (reg_class): Add GENERAL_AND_VPR_REGS.
18866 (REG_CLASS_NAMES): Likewise.
18867 (REG_CLASS_CONTENTS): Likewise.
18868 (CLASS_MAX_NREGS): Handle VPR.
18869 * config/arm/arm.cc (arm_hard_regno_nregs): Handle VPR.
18871 2022-02-22 Tobias Burnus <tobias@codesourcery.com>
18872 Tom de Vries <tdevries@suse.de>
18874 * config/nvptx/nvptx-c.cc (nvptx_cpu_cpp_builtins): Handle SM70.
18875 * config/nvptx/nvptx.cc (first_ptx_version_supporting_sm):
18877 * config/nvptx/nvptx.opt (misa): Add sm_70 alias PTX_ISA_SM70.
18879 2022-02-22 Tobias Burnus <tobias@codesourcery.com>
18880 Tom de Vries <tdevries@suse.de>
18882 * config/nvptx/nvptx.opt (mptx): Add 6.0 alias PTX_VERSION_6_0.
18883 * doc/invoke.texi (-mptx): Update for new values and defaults.
18885 2022-02-22 Tom de Vries <tdevries@suse.de>
18887 * config/nvptx/nvptx.cc (gen_comment): New function.
18888 (workaround_uninit_method_1, workaround_uninit_method_2)
18889 (workaround_uninit_method_3): : Use gen_comment.
18890 * config/nvptx/nvptx.opt (mptx-comment): New option.
18892 2022-02-22 Richard Biener <rguenther@suse.de>
18894 * tree-vect-slp.cc (vect_build_slp_tree_2): Dump the def used
18897 2022-02-22 Roger Sayle <roger@nextmovesoftware.com>
18898 Richard Biener <rguenther@suse.de>
18900 * fold-const.cc (ctor_single_nonzero_element): New function to
18901 return the single non-zero element of a (vector) constructor.
18902 * fold-const.h (ctor_single_nonzero_element): Prototype here.
18903 * match.pd (reduc (constructor@0)): Simplify reductions of a
18904 constructor containing a single non-zero element.
18905 (reduc (@0 op VECTOR_CST) -> (reduc @0) op CONST): Simplify
18906 reductions of vector operations of the same operator with
18907 constant vector operands.
18909 2022-02-22 Jakub Jelinek <jakub@redhat.com>
18911 PR tree-optimization/104604
18912 * gimple-range-fold.cc (adjust_imagpart_expr, adjust_realpart_expr):
18913 Only check if gimple_assign_rhs1 is COMPLEX_CST if
18914 gimple_assign_rhs_code is COMPLEX_CST.
18916 2022-02-22 Jakub Jelinek <jakub@redhat.com>
18919 * config/i386/i386-expand.cc (ix86_expand_copysign): Call force_reg
18920 on input operands before calling lowpart_subreg on it. For output
18921 operand, use a vmode pseudo as destination and then move its lowpart
18922 subreg into operands[0] if lowpart_subreg fails on dest.
18923 (ix86_expand_xorsign): Likewise.
18925 2022-02-22 Richard Biener <rguenther@suse.de>
18927 PR tree-optimization/104582
18929 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
18930 Cost GPR to vector register moves for integer vector construction.
18932 2022-02-22 Richard Biener <rguenther@suse.de>
18934 PR tree-optimization/104582
18935 * tree-vectorizer.h (stmt_info_for_cost::node): New field.
18936 (vector_costs::add_stmt_cost): Add SLP node parameter.
18937 (dump_stmt_cost): Likewise.
18938 (add_stmt_cost): Likewise, new overload and adjust.
18939 (add_stmt_costs): Adjust.
18940 (record_stmt_cost): New overload.
18941 * tree-vectorizer.cc (dump_stmt_cost): Dump the SLP node.
18942 (vector_costs::add_stmt_cost): Adjust.
18943 * tree-vect-loop.cc (vect_estimate_min_profitable_iters):
18945 * tree-vect-slp.cc (vect_prologue_cost_for_slp): Record
18946 the SLP node for costing.
18947 (vectorizable_slp_permutation): Likewise.
18948 * tree-vect-stmts.cc (record_stmt_cost): Adjust and add
18950 * config/i386/i386.cc (ix86_vector_costs::add_stmt_cost):
18952 * config/aarch64/aarch64.cc (aarch64_vector_costs::add_stmt_cost):
18954 * config/rs6000/rs6000.cc (rs6000_vector_costs::add_stmt_cost):
18956 (rs6000_cost_data::adjust_vect_cost_per_loop): Likewise.
18958 2022-02-22 Richard Biener <rguenther@suse.de>
18960 PR tree-optimization/104582
18961 * tree-vectorizer.h (add_stmt_cost): New overload.
18962 (record_stmt_cost): Likewise.
18963 * tree-vect-loop.cc (vect_compute_single_scalar_iteration_cost):
18964 Use add_stmt_costs.
18965 (vect_get_known_peeling_cost): Use new overloads.
18966 (vect_estimate_min_profitable_iters): Likewise. Consistently
18967 use scalar_stmt for costing versioning checks.
18968 * tree-vect-stmts.cc (record_stmt_cost): New overload.
18970 2022-02-22 Hongyu Wang <hongyu.wang@intel.com>
18973 * config/i386/i386-expand.cc (ix86_expand_atomic_fetch_op_loop):
18974 Split atomic fetch and loop part.
18975 (ix86_expand_cmpxchg_loop): New expander for cmpxchg loop.
18976 * config/i386/i386-protos.h (ix86_expand_cmpxchg_loop): New
18978 * config/i386/sync.md (atomic_compare_and_swap<mode>): Call new
18979 expander under TARGET_RELAX_CMPXCHG_LOOP.
18980 (atomic_compare_and_swap<mode>): Likewise for doubleword modes.
18982 2022-02-21 Dan Li <ashimida@linux.alibaba.com>
18984 * config/aarch64/aarch64.cc (SLOT_REQUIRED):
18985 Change wb_candidate[12] to wb_push_candidate[12].
18986 (aarch64_layout_frame): Likewise, and
18987 change callee_adjust when scs is enabled.
18988 (aarch64_save_callee_saves):
18989 Change wb_candidate[12] to wb_push_candidate[12].
18990 (aarch64_restore_callee_saves):
18991 Change wb_candidate[12] to wb_pop_candidate[12].
18992 (aarch64_get_separate_components):
18993 Change wb_candidate[12] to wb_push_candidate[12].
18994 (aarch64_expand_prologue): Push x30 onto SCS before it's
18996 (aarch64_expand_epilogue): Pop x30 frome SCS, while
18997 preventing it from being popped from the regular stack again.
18998 (aarch64_override_options_internal): Add SCS compile option check.
18999 (TARGET_HAVE_SHADOW_CALL_STACK): New hook.
19000 * config/aarch64/aarch64.h (struct GTY): Add is_scs_enabled,
19001 wb_pop_candidate[12], and rename wb_candidate[12] to
19002 wb_push_candidate[12].
19003 * config/aarch64/aarch64.md (scs_push): New template.
19004 (scs_pop): Likewise.
19005 * doc/invoke.texi: Document -fsanitize=shadow-call-stack.
19006 * doc/tm.texi: Regenerate.
19007 * doc/tm.texi.in: Add hook have_shadow_call_stack.
19008 * flag-types.h (enum sanitize_code):
19009 Add SANITIZE_SHADOW_CALL_STACK.
19010 * opts.cc (parse_sanitizer_options): Add shadow-call-stack
19011 and exclude SANITIZE_SHADOW_CALL_STACK.
19012 * target.def: New hook.
19013 * toplev.cc (process_options): Add SCS compile option check.
19014 * ubsan.cc (ubsan_expand_null_ifn): Enum type conversion.
19016 2022-02-21 Tom de Vries <tdevries@suse.de>
19019 * config/nvptx/nvptx.cc (workaround_uninit_method_1)
19020 (workaround_uninit_method_2, workaround_uninit_method_3)
19021 (workaround_uninit): New function.
19022 (nvptx_reorg): Use workaround_uninit.
19023 * config/nvptx/nvptx.opt (minit-regs): New option.
19025 2022-02-21 Andre Vieira <andre.simoesdiasvieira@arm.com>
19027 PR rtl-optimization/104498
19028 * alias.cc (compare_base_symbol_refs): Correct distance computation
19029 when swapping x and y.
19031 2022-02-21 Andrew Pinski <apinski@marvell.com>
19034 * tree-ssa.cc (tree_ssa_useless_type_conversion):
19035 Check the inner type before calling useless_type_conversion_p.
19037 2022-02-19 Tom de Vries <tdevries@suse.de>
19039 * config/nvptx/nvptx.cc (nvptx_reorg_uniform_simt): Handle SET insn.
19040 * config/nvptx/nvptx.md
19041 (define_insn "nvptx_atomic_store<mode>"): Rename to ...
19042 (define_insn "nvptx_atomic_store_sm70<mode>"): This.
19043 (define_insn "nvptx_atomic_store<mode>"): New define_insn.
19044 (define_expand "atomic_store<mode>"): Handle rename. Use
19045 nvptx_atomic_store instead of atomic_exchange.
19047 2022-02-19 Tom de Vries <tdevries@suse.de>
19049 * config/nvptx/nvptx.cc (nvptx_reorg_uniform_simt): Handle all
19050 insns with atomic attribute. Assert that all handled insns are
19052 * config/nvptx/nvptx.md (define_insn "nvptx_atomic_store<mode>"):
19053 Set atomic attribute to false.
19055 2022-02-19 Tom de Vries <tdevries@suse.de>
19057 * config/nvptx/nvptx.cc (nvptx_unisimt_handle_set): Change return
19059 (nvptx_reorg_uniform_simt): Insert nvptx_uniform_warp_check or
19060 nvptx_warpsync, if necessary.
19062 2022-02-19 Jakub Jelinek <jakub@redhat.com>
19064 PR sanitizer/102656
19065 * asan.cc (instrument_derefs): If inner is a RESULT_DECL and access is
19066 known to be within bounds, treat it like automatic variables.
19067 If instrumenting access and inner is {VAR,PARM,RESULT}_DECL from
19068 current function and !TREE_STATIC which is not TREE_ADDRESSABLE, mark
19071 2022-02-18 Pat Haugen <pthaugen@linux.ibm.com>
19073 * config/rs6000/rs6000.opt (mpower10-fusion): Mark Undocumented.
19074 (mpower10-fusion-ld-cmpi, mpower10-fusion-2logical,
19075 mpower10-fusion-logical-add, mpower10-fusion-add-logical,
19076 mpower10-fusion-2add, mpower10-fusion-2store): Remove.
19077 * config/rs6000/rs6000-cpus.def (ISA_3_1_MASKS_SERVER,
19078 OTHER_P9_VECTOR_MASKS): Remove Power10 fusion sub-options.
19079 * config/rs6000/rs6000.cc (rs6000_option_override_internal,
19080 power10_sched_reorder): Likewise.
19081 * config/rs6000/genfusion.pl (gen_ld_cmpi_p10, gen_logical_addsubf,
19082 gen_addadd): Likewise
19083 * config/rs6000/fusion.md: Regenerate.
19085 2022-02-18 Jakub Jelinek <jakub@redhat.com>
19089 * config/rs6000/mm_malloc.h (_mm_malloc): Call posix_memalign
19090 rather than __posix_memalign.
19092 2022-02-18 Richard Biener <rguenther@suse.de>
19095 * config/i386/i386.cc (ix86_avx_u128_mode_source): Remove.
19096 (ix86_avx_u128_mode_needed): Return AVX_U128_DIRTY instead
19097 of calling ix86_avx_u128_mode_source which would eventually
19098 have returned AVX_U128_ANY in some very special case.
19100 2022-02-18 Richard Biener <rguenther@suse.de>
19102 PR tree-optimization/96881
19103 * tree-ssa-dce.cc (mark_stmt_if_obviously_necessary): Comment
19105 (control_parents_preserved_p): New function.
19106 (eliminate_unnecessary_stmts): Check that we preserved control
19107 parents before retaining a CLOBBER.
19108 (perform_tree_ssa_dce): Pass down aggressive flag
19109 to eliminate_unnecessary_stmts.
19111 2022-02-17 Jason Merrill <jason@redhat.com>
19113 * tree.cc (warn_deprecated_use): Look for TYPE_STUB_DECL
19114 on TYPE_MAIN_VARIANT.
19116 2022-02-17 Paul A. Clarke <pc@us.ibm.com>
19119 * config/rs6000/bmi2intrin.h: Uglify local variables.
19120 * config/rs6000/emmintrin.h: Likewise.
19121 * config/rs6000/mm_malloc.h: Likewise.
19122 * config/rs6000/mmintrin.h: Likewise.
19123 * config/rs6000/pmmintrin.h: Likewise.
19124 * config/rs6000/smmintrin.h: Likewise.
19125 * config/rs6000/tmmintrin.h: Likewise.
19126 * config/rs6000/xmmintrin.h: Likewise.
19128 2022-02-17 Robin Dapp <rdapp@linux.ibm.com>
19131 * config/rs6000/rs6000.cc (rs6000_emit_int_cmove): Return false
19132 if the expected comparison's first operand is of mode MODE_CC.
19134 2022-02-17 Vladimir N. Makarov <vmakarov@redhat.com>
19136 PR rtl-optimization/104447
19137 * lra-constraints.cc (spill_hard_reg_in_range): Initiate ignore
19138 hard reg set by lra_no_alloc_regs.
19140 2022-02-17 liuhongt <hongtao.liu@intel.com>
19142 PR tree-optimization/104551
19143 PR tree-optimization/103771
19144 * match.pd (cond_expr_convert_p): Add types_match check when
19145 convert is extension.
19146 * tree-vect-patterns.cc
19147 (gimple_cond_expr_convert_p): Adjust comments.
19148 (vect_recog_cond_expr_convert_pattern): Ditto.
19150 2022-02-17 Jakub Jelinek <jakub@redhat.com>
19153 * valtrack.cc (debug_lowpart_subreg): Don't call gen_rtx_raw_SUBREG
19154 if expr has VOIDmode.
19156 2022-02-17 liuhongt <hongtao.liu@intel.com>
19158 * config/i386/cpuid.h (bit_MPX): Removed.
19159 (bit_BNDREGS): Ditto.
19160 (bit_BNDCSR): Ditto.
19162 2022-02-17 Michael Meissner <meissner@the-meissners.org>
19165 * config/rs6000/rs6000-c.cc (rs6000_cpu_cpp_builtins): Define
19166 __SIZEOF_IBM128__ if the IBM 128-bit long double type is created.
19167 Define __SIZEOF_FLOAT128__ if the IEEE 128-bit floating point type
19170 2022-02-16 Andrew MacLeod <amacleod@redhat.com>
19172 * gimple-range-gori.cc (gori_compute::condexpr_adjust): Use
19173 range_compatible_p instead of direct type comparison.
19175 2022-02-16 Jakub Jelinek <jakub@redhat.com>
19177 PR rtl-optimization/104544
19178 * combine.cc (try_combine): When looking for insn whose links
19179 should be updated from i3 to i2, don't stop on debug insns, instead
19182 2022-02-16 Richard Sandiford <richard.sandiford@arm.com>
19185 * config/aarch64/iterators.md (LOGICAL_OR_PLUS): New iterator.
19186 * config/aarch64/aarch64.md: Extend the PR100056 patterns
19187 to handle plus in the same way as ior, if the operands have
19188 no set bits in common.
19190 2022-02-15 Andrew MacLeod <amacleod@redhat.com>
19192 PR tree-optimization/104526
19193 * gimple-range-fold.cc (fold_using_range::range_of_cond_expr): Call
19195 * gimple-range-gori.cc (range_def_chain::get_def_chain): Force a build
19196 of dependency chain if there isn't one.
19197 (gori_compute::condexpr_adjust): New.
19198 * gimple-range-gori.h (class gori_compute): New prototype.
19200 2022-02-15 Richard Sandiford <richard.sandiford@arm.com>
19203 * config/aarch64/aarch64-protos.h (aarch64_maxmin_plus_const):
19205 * config/aarch64/aarch64.cc (aarch64_maxmin_plus_const): New function.
19206 * config/aarch64/aarch64.md (*aarch64_minmax_plus): New pattern.
19208 2022-02-15 Richard Sandiford <richard.sandiford@arm.com>
19210 * tree-vectorizer.h (vect_scalar_ops_slice): New struct.
19211 (vect_scalar_ops_slice_hash): Likewise.
19212 (vect_scalar_ops_slice::op): New function.
19213 * tree-vect-slp.cc (vect_scalar_ops_slice::all_same_p): New function.
19214 (vect_scalar_ops_slice_hash::hash): Likewise.
19215 (vect_scalar_ops_slice_hash::equal): Likewise.
19216 (vect_prologue_cost_for_slp): Check for duplicate vectors.
19217 * config/aarch64/aarch64.cc
19218 (aarch64_vector_costs::m_stp_sequence_cost): New member variable.
19219 (aarch64_aligned_constant_offset_p): New function.
19220 (aarch64_stp_sequence_cost): Likewise.
19221 (aarch64_vector_costs::add_stmt_cost): Handle new STP heuristic.
19222 (aarch64_vector_costs::finish_cost): Likewise.
19224 2022-02-15 Richard Sandiford <richard.sandiford@arm.com>
19226 * tree-vect-slp.cc (vect_bb_vectorization_profitable_p): Fix
19229 2022-02-15 Richard Biener <rguenther@suse.de>
19231 PR tree-optimization/104543
19232 * gimple-loop-jam.cc (unroll_jam_possible_p): Check outer loop exits
19233 come after the inner loop.
19235 2022-02-15 Jakub Jelinek <jakub@redhat.com>
19238 * config/i386/host-cygwin.cc (cygwin_gt_pch_get_address): Use
19239 cannot instead of can%'t in diagnostics. Formatting fixes.
19241 2022-02-15 Jakub Jelinek <jakub@redhat.com>
19243 PR middle-end/104522
19244 * fold-const.h (native_interpret_real): Declare.
19245 * fold-const.cc (native_interpret_real): No longer static. Don't
19246 perform MODE_COMPOSITE_P verification here.
19247 (native_interpret_expr) <case REAL_TYPE>: But perform it here instead
19249 * gimple-fold.cc (clear_padding_type): Call native_interpret_real
19250 instead of native_interpret_expr.
19251 * simplify-rtx.cc (simplify_immed_subreg): Perform the native_encode_rtx
19252 and comparison verification for all FLOAT_MODE_P modes, not just
19255 2022-02-15 Richard Biener <rguenther@suse.de>
19257 PR tree-optimization/104519
19258 * fold-const.cc (multiple_of_p): Remove never true condition.
19259 * tree-ssa-loop-niter.cc (number_of_iterations_ne): Use
19260 the appropriate types for determining whether the difference
19261 of final and base is a multiple of the step.
19263 2022-02-15 Jakub Jelinek <jakub@redhat.com>
19266 * omp-low.cc (task_cpyfns): New variable.
19267 (delete_omp_context): Don't call finalize_task_copyfn from here.
19268 (create_task_copyfn): Push task_stmt into task_cpyfns.
19269 (execute_lower_omp): Call finalize_task_copyfn here on entries from
19270 task_cpyfns vector and release the vector.
19272 2022-02-14 Martin Sebor <msebor@redhat.com>
19274 PR middle-end/104355
19275 * doc/invoke.texi (-Warray-bounds): Update documentation.
19277 2022-02-14 Michael Meissner <meissner@the-meissners.org>
19280 * config/rs6000/rs6000.cc (init_float128_ibm): Update the
19281 conversion functions used to convert IFmode types.
19283 2022-02-14 Andrew Stubbs <ams@codesourcery.com>
19285 * config/gcn/gcn.cc (gcn_expand_reduc_scalar): Use force_reg.
19287 2022-02-14 Richard Biener <rguenther@suse.de>
19289 PR tree-optimization/104528
19290 * tree-ssa.h (find_released_ssa_name): Declare.
19291 * tree-ssa.cc (find_released_ssa_name): Export.
19292 * cfgloop.cc (verify_loop_structure): Look for released
19293 SSA names in loops nb_iterations.
19294 * tree-ssa-dse.cc (pass_dse::execute): Release number of iteration
19297 2022-02-14 Richard Biener <rguenther@suse.de>
19299 PR tree-optimization/104511
19300 * tree-ssa-forwprop.cc (simplify_vector_constructor): Avoid
19301 touching DFP <-> FP conversions.
19303 2022-02-14 Richard Biener <rguenther@suse.de>
19305 PR middle-end/104497
19306 * gimplify.cc (gimplify_compound_lval): Make sure the
19307 base is a non-register if needed and possible.
19309 2022-02-13 liuhongt <hongtao.liu@intel.com>
19312 * match.pd (cond_expr_convert_p): New match.
19313 * tree-vect-patterns.cc (gimple_cond_expr_convert_p): Declare.
19314 (vect_recog_cond_expr_convert_pattern): New.
19316 2022-02-12 Jakub Jelinek <jakub@redhat.com>
19318 PR sanitizer/104449
19319 * asan.cc: Include tree-eh.h.
19320 (handle_builtin_alloca): Handle the case when __builtin_alloca or
19321 __builtin_alloca_with_align can throw.
19323 2022-02-12 Jakub Jelinek <jakub@redhat.com>
19326 * config/i386/i386.md (cvtsd2ss splitter): If operands[1] is xmm16+
19327 and AVX512VL isn't available, move operands[1] to operands[0] first.
19329 2022-02-12 Uroš Bizjak <ubizjak@gmail.com>
19332 * config/i386/i386.cc (type_natural_mode):
19333 Skip decimal float vector modes.
19335 2022-02-11 Iain Sandoe <iain@sandoe.co.uk>
19336 Vladimir Makarov <vmakarov@redhat.com>
19339 * config/rs6000/rs6000.cc (darwin_rs6000_legitimate_lo_sum_const_p):
19340 Check for UNSPEC_MACHOPIC_OFFSET wrappers on symbolic addresses when
19342 (legitimate_lo_sum_address_p): Likewise.
19343 * lra-constraints.cc (process_address_1): Do not attempt to emit a reg
19344 load from an invalid lo_sum address.
19346 2022-02-11 Jakub Jelinek <jakub@redhat.com>
19348 PR tree-optimization/104499
19349 * match.pd ((X & Y) CMP 0 -> X CMP2 ~Y): Use view_convert instead
19352 2022-02-11 Jakub Jelinek <jakub@redhat.com>
19354 * tree.cc (build_common_builtin_nodes): Fix up formatting in
19355 __builtin_clear_padding decl creation.
19356 * gimplify.cc (gimple_add_padding_init_for_auto_var): Encode
19357 for_auto_init in the value of 2nd BUILT_IN_CLEAR_PADDING
19358 argument rather than in 3rd argument.
19359 (gimplify_call_expr): Likewise. Fix up comment formatting.
19360 * gimple-fold.cc (gimple_fold_builtin_clear_padding): Expect
19361 2 arguments instead of 3, take for_auto_init from the value
19364 2022-02-11 Vladimir N. Makarov <vmakarov@redhat.com>
19366 PR rtl-optimization/104400
19367 * lra-constraints.cc (process_alt_operands): Don't make union of
19368 this_alternative_exclude_start_hard_regs when reg class in insn
19369 alternative covers other reg classes in the same alternative.
19371 2022-02-11 Jakub Jelinek <jakub@redhat.com>
19373 PR middle-end/104446
19374 * combine.cc (subst): Don't substitute CONST_INTs into RTX_AUTOINC
19377 2022-02-11 Richard Biener <rguenther@suse.de>
19379 PR middle-end/104496
19380 * internal-fn.cc (vectorized_internal_fn_supported_p):
19381 Bail out for integer mode vector types.
19383 2022-02-11 Jakub Jelinek <jakub@redhat.com>
19385 PR rtl-optimization/104459
19386 * df-scan.cc (df_insn_change_bb): Don't call df_set_bb_dirty when
19387 moving DEBUG_INSNs between bbs.
19389 2022-02-11 liuhongt <hongtao.liu@intel.com>
19391 PR tree-optimization/104479
19392 * match.pd (uncond_op + vec_cond -> cond_op): Add single_use
19393 for the dest of uncond_op.
19395 2022-02-11 Tom de Vries <tdevries@suse.de>
19398 * config/nvptx/nvptx.cc (prevent_branch_around_nothing): Handle asm
19401 2022-02-10 Qing Zhao <qing.zhao@oracle.com>
19403 PR middle-end/100775
19404 * function.cc (gen_call_used_regs_seq): Call
19405 df_update_exit_block_uses when updating df.
19407 2022-02-10 Uroš Bizjak <ubizjak@gmail.com>
19410 * config/i386/sse.md (vec_unpacks_float_lo_v4si):
19411 Change operand 1 constraint to register_operand.
19413 2022-02-10 Richard Biener <rguenther@suse.de>
19415 PR tree-optimization/104373
19416 * tree-ssa-sccvn.h (do_rpo_vn): New export exposing the
19418 * tree-ssa-sccvn.cc (do_rpo_vn): Export, get the default
19419 walk kind as argument.
19420 (run_rpo_vn): Adjust.
19421 (pass_fre::execute): Likewise.
19422 * tree-ssa-uninit.cc (warn_uninitialized_vars): Skip
19423 blocks not reachable.
19424 (execute_late_warn_uninitialized): Mark all edges as
19426 (execute_early_warn_uninitialized): Use VN to compute
19428 (pass_data_early_warn_uninitialized): Enable a dump file,
19429 change dump name to warn_uninit.
19431 2022-02-10 Richard Biener <rguenther@suse.de>
19433 PR middle-end/104467
19434 * match.pd (vector extract simplification): Multiply the
19435 number of CTOR elements with the number of element elements.
19437 2022-02-10 Richard Biener <rguenther@suse.de>
19439 PR tree-optimization/104466
19440 * tree-ssa-alias.cc (refs_may_alias_p_2): Use rbase1/rbase2
19441 for the MR_DEPENDENCE checks as intended.
19443 2022-02-10 Tom de Vries <tdevries@suse.de>
19445 * config/nvptx/nvptx.md (define_insn "nvptx_atomic_store<mode>"): New
19447 (define_expand "atomic_store<mode>"): Use nvptx_atomic_store<mode> for
19449 (define_c_enum "unspecv"): Add UNSPECV_ST.
19451 2022-02-10 Tom de Vries <tdevries@suse.de>
19453 * config/nvptx/nvptx-protos.h (nvptx_mem_maybe_shared_p): Declare.
19454 * config/nvptx/nvptx.cc (nvptx_mem_data_area): New static function.
19455 (nvptx_mem_maybe_shared_p): New function.
19456 * config/nvptx/nvptx.md (define_expand "atomic_store<mode>"): New
19459 2022-02-10 Tom de Vries <tdevries@suse.de>
19462 * config/nvptx/nvptx.md (define_insn "sub<mode>3"): Workaround
19463 driver JIT bug by using sub.s16 instead of sub.u16.
19465 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19467 * config/nvptx/nvptx.md (copysign<mode>3): Allow immediate
19468 floating point constants as operands 1 and/or 2.
19470 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19473 * config/nvptx/nvptx.md (sel_true<mode>): Fix indentation.
19474 (sel_false<mode>): Likewise.
19475 (define_code_iterator eqne): New code iterator for EQ and NE.
19476 (*selp<mode>_neg_<code>): New define_insn_and_split to optimize
19477 the negation of a selp instruction.
19478 (*selp<mode>_not_<code>): New define_insn_and_split to optimize
19479 the bitwise not of a selp instruction.
19480 (*setcc_int<mode>): Use set instruction for neg:SI of a selp.
19482 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19484 * config/nvptx/nvptx.md (any_logic): Move code iterator earlier
19485 in machine description.
19486 (logic): Move code attribute earlier in machine description.
19487 (ilogic): New code attribute, like logic but "ior" for IOR.
19488 (and<mode>3, ior<mode>3, xor<mode>3): Delete. Replace with...
19489 (<ilogic><mode>3): New define_insn for HSDIM logic operations.
19490 (<ilogic>bi3): New define_insn for BI mode logic operations.
19491 (define_split): Lower logic operations from integer modes to
19492 BI mode predicate operations.
19494 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19496 * config/nvptx/nvptx.md (UNSPEC_ISINF): New UNSPEC.
19497 (one_cmplbi2): New define_insn for not.pred.
19498 (mulditi3): New define_expand for signed widening multiply.
19499 (umulditi3): New define_expand for unsigned widening multiply.
19500 (smul<mode>3_highpart): New define_insn for signed highpart mult.
19501 (umul<mode>3_highpart): New define_insn for unsigned highpart mult.
19502 (*smulhi3_highpart_2): Renamed from smulhi3_highpart.
19503 (*smulsi3_highpart_2): Renamed from smulsi3_highpart.
19504 (*umulhi3_highpart_2): Renamed from umulhi3_highpart.
19505 (*umulsi3_highpart_2): Renamed from umulsi3_highpart.
19506 (*setcc<mode>_from_not_bi): New define_insn.
19507 (*setcc_isinf<mode>): New define_insn for testp.infinite.
19508 (isinf<mode>2): New define_expand.
19510 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19512 * config/nvptx/nvptx.md (cmp<mode>): Renamed from *cmp<mode>.
19513 (setcc<mode>_from_bi): Additionally support QImode.
19514 (extendbi<mode>2): Additionally support QImode.
19515 (zero_extendbi<mode>2): Additionally support QImode.
19516 (any_sbinary, any_ubinary, any_sunary, any_uunary): New code
19517 iterators for signed and unsigned, binary and unary operations.
19518 (<sbinary>qi3, <ubinary>qi3, <sunary>qi2, <uunary>qi2): New
19519 expanders to perform QImode operations using SImode instructions.
19520 (cstoreqi4): New define_expand.
19521 (*ext_truncsi2_qi): New define_insn.
19522 (*zext_truncsi2_qi): New define_insn.
19524 2022-02-10 Roger Sayle <roger@nextmovesoftware.com>
19526 * config/nvptx/nvptx.md (*cmpf): New define_insn.
19527 (cstorehf4): New define_expand.
19528 (fmahf4): New define_insn.
19529 (neghf2): New define_insn.
19530 (abshf2): New define_insn.
19532 2022-02-10 Gerald Pfeifer <gerald@pfeifer.com>
19534 * doc/install.texi (Specific): Change the www.bitwizard.nl
19535 reference to use https.
19537 2022-02-10 Marcel Vollweiler <marcel@codesourcery.com>
19539 * gimplify.cc (gimplify_scan_omp_clauses): Added cases for
19540 OMP_CLAUSE_HAS_DEVICE_ADDR
19541 and handle array sections.
19542 (gimplify_adjust_omp_clauses): Added OMP_CLAUSE_HAS_DEVICE_ADDR case.
19543 * omp-low.cc (scan_sharing_clauses): Handle OMP_CLAUSE_HAS_DEVICE_ADDR.
19544 (lower_omp_target): Same.
19545 * tree-core.h (enum omp_clause_code): Same.
19546 * tree-nested.cc (convert_nonlocal_omp_clauses): Same.
19547 (convert_local_omp_clauses): Same.
19548 * tree-pretty-print.cc (dump_omp_clause): Same.
19551 2022-02-10 Eugene Rozenfeld <erozen@microsoft.com>
19553 * auto-profile.cc (afdo_indirect_call): Don't attempt to promote indirect calls
19554 that will result in direct recursive calls.
19556 2022-02-10 Andrew Pinski <apinski@marvell.com>
19559 * config/aarch64/aarch64.cc
19560 (aarch64_sve_expand_vector_init_handle_trailing_constants):
19561 Use CONST0_RTX instead of const0_rtx for the non-constant elements.
19563 2022-02-09 Uroš Bizjak <ubizjak@gmail.com>
19566 * common/config/i386/i386-common.cc (OPTION_MASK_ISA2_XSAVE_UNSET):
19567 Also include OPTION_MASK_ISA2_AVX2_UNSET.
19569 2022-02-09 Uroš Bizjak <ubizjak@gmail.com>
19572 * config/i386/i386-expand.cc (ix86_split_idivmod):
19573 Force operands[2] and operands[3] into a register..
19575 2022-02-09 Jeff Law <jeffreyalaw@gmail.com>
19578 * config/v850/v850.md (*v850_fnmasf4): Renamed from fnmasf4.
19579 (*v850_fnmssf4): Renamed from fnmssf4
19581 2022-02-09 Ian Lance Taylor <iant@golang.org>
19583 * godump.cc (go_force_record_alignment): Really name the alignment
19584 field "_" (complete 2021-12-29 change).
19586 2022-02-09 Bill Schmidt <wschmidt@linux.ibm.com>
19588 * config/rs6000/rs6000-builtins.def (VREPLACE_UN_UV2DI): Change
19589 function prototype.
19590 (VREPLACE_UN_UV4SI): Likewise.
19591 (VREPLACE_UN_V2DF): Likewise.
19592 (VREPLACE_UN_V2DI): Likewise.
19593 (VREPLACE_UN_V4SF): Likewise.
19594 (VREPLACE_UN_V4SI): Likewise.
19595 * config/rs6000/rs6000-overload.def (VEC_REPLACE_UN): Change all
19596 function prototypes.
19597 * config/rs6000/vsx.md (vreplace_un_<mode>): Remove define_expand.
19598 (vreplace_un_<mode>): New define_insn.
19600 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19602 * config/aarch64/iterators.md (VDCSIF): New mode iterator.
19604 (single_wx, single_type, single_dtype, dblq): New mode attributes.
19605 * config/aarch64/aarch64-simd.md (load_pair_lanes<mode>): Extend
19606 from VDC to VDCSIF.
19607 (store_pair_lanes<mode>): Likewise.
19608 (*aarch64_combine_internal<mode>): Likewise.
19609 (*aarch64_combine_internal_be<mode>): Likewise.
19610 (*aarch64_combinez<mode>): Likewise.
19611 (*aarch64_combinez_be<mode>): Likewise.
19612 * config/aarch64/aarch64.cc (aarch64_classify_address): Handle
19613 8-byte modes for ADDR_QUERY_LDP_STP_N.
19614 (aarch64_print_operand): Likewise for %y.
19616 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19618 * config/aarch64/aarch64-simd.md (@aarch64_split_simd_mov<mode>):
19619 Use aarch64_combine instead of move_lo/hi_quad. Tabify.
19620 (move_lo_quad_<mode>, aarch64_simd_move_hi_quad_<mode>): Delete.
19621 (aarch64_simd_move_hi_quad_be_<mode>, move_hi_quad_<mode>): Delete.
19622 (vec_pack_trunc_<mode>): Take general_operand elements and use
19623 aarch64_combine rather than move_lo/hi_quad to combine them.
19624 (vec_pack_trunc_df): Likewise.
19626 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19628 * config/aarch64/aarch64-protos.h (aarch64_split_simd_combine):
19630 * config/aarch64/aarch64-simd.md (@aarch64_combinez<mode>): Rename
19632 (*aarch64_combinez<mode>): ...this.
19633 (@aarch64_combinez_be<mode>): Rename to...
19634 (*aarch64_combinez_be<mode>): ...this.
19635 (@aarch64_vec_concat<mode>): New expander.
19636 (aarch64_combine<mode>): Use it.
19637 (@aarch64_simd_combine<mode>): Delete.
19638 * config/aarch64/aarch64.cc (aarch64_split_simd_combine): Delete.
19639 (aarch64_expand_vector_init): Use aarch64_vec_concat.
19641 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19643 * config/aarch64/predicates.md (aarch64_reg_or_mem_pair_operand):
19645 * config/aarch64/aarch64-simd.md (*aarch64_combine_internal<mode>)
19646 (*aarch64_combine_internal_be<mode>): New patterns.
19648 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19650 * config/aarch64/aarch64-simd.md (move_lo_quad_internal_<mode>)
19651 (move_lo_quad_internal_be_<mode>): Delete.
19652 (move_lo_quad_<mode>): Use aarch64_combine<Vhalf> instead of the above.
19654 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19656 * config/aarch64/aarch64-protos.h (aarch64_mergeable_load_pair_p):
19658 * config/aarch64/aarch64-simd.md (load_pair_lanes<mode>): Use
19659 aarch64_mergeable_load_pair_p instead of inline check.
19660 * config/aarch64/aarch64.cc (aarch64_expand_vector_init): Likewise.
19661 (aarch64_check_consecutive_mems): Allow the reversed parameter
19663 (aarch64_mergeable_load_pair_p): New function.
19665 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19667 * config/aarch64/aarch64-simd.md (vec_set<mode>): Allow the
19668 element to be an aarch64_simd_nonimmediate_operand.
19670 2022-02-09 Richard Sandiford <richard.sandiford@arm.com>
19672 * config/aarch64/aarch64-simd.md (aarch64_simd_vec_set<mode>): Use
19673 aarch64_simd_nonimmediate_operand instead of
19674 aarch64_simd_general_operand.
19675 (@aarch64_combinez<mode>): Use nonimmediate_operand instead of
19677 (@aarch64_combinez_be<mode>): Likewise.
19679 2022-02-09 Richard Biener <rguenther@suse.de>
19681 PR middle-end/104464
19682 * gimple-isel.cc (gimple_expand_vec_cond_expr): Postpone
19683 throwing check to after unproblematic replacement.
19685 2022-02-09 Roger Sayle <roger@nextmovesoftware.com>
19687 PR tree-optimization/104420
19688 * match.pd (mult @0 real_zerop): Tweak conditions for constant
19689 folding X*0.0 (or X*-0.0) to HONOR_SIGNED_ZEROS when appropriate.
19691 2022-02-09 Jakub Jelinek <jakub@redhat.com>
19694 * dwarf2out.cc (mangle_referenced_decls): New function.
19695 (tree_add_const_value_attribute): Don't call rtl_for_decl_init if
19696 early_dwarf. Instead walk the initializer and try to mangle vars or
19697 functions referenced from it.
19699 2022-02-09 Andrew MacLeod <amacleod@redhat.com>
19701 PR tree-optimization/104288
19702 * gimple-range-cache.cc (non_null_ref::set_nonnull): New.
19703 (non_null_ref::adjust_range): Move to header.
19704 (ranger_cache::range_of_def): Don't check non-null.
19705 (ranger_cache::entry_range): Don't check non-null.
19706 (ranger_cache::range_on_edge): Check for nonnull on normal edges.
19707 (ranger_cache::update_to_nonnull): New.
19708 (non_null_loadstore): New.
19709 (ranger_cache::block_apply_nonnull): New.
19710 * gimple-range-cache.h (class non_null_ref): Update prototypes.
19711 (non_null_ref::adjust_range): Move to here and inline.
19712 (class ranger_cache): Update prototypes.
19713 * gimple-range-path.cc (path_range_query::range_defined_in_block): Do
19714 not search dominators.
19715 (path_range_query::adjust_for_non_null_uses): Ditto.
19716 * gimple-range.cc (gimple_ranger::range_of_expr): Check on-entry for
19717 def overrides. Do not check nonnull.
19718 (gimple_ranger::range_on_entry): Check dominators for nonnull.
19719 (gimple_ranger::range_on_edge): Check for nonnull on normal edges..
19720 (gimple_ranger::register_side_effects): New.
19721 * gimple-range.h (gimple_ranger::register_side_effects): New.
19722 * tree-vrp.cc (rvrp_folder::fold_stmt): Call register_side_effects.
19724 2022-02-09 Richard Biener <rguenther@suse.de>
19726 PR tree-optimization/104445
19727 PR tree-optimization/102832
19728 * optabs-query.h (can_vec_extract): New.
19729 * optabs-query.cc (can_vec_extract): Likewise.
19730 * tree-vect-loop.cc (vect_find_reusable_accumulator): Check
19731 we can extract a hi/lo part from the larger vector, rework
19732 check iteration from larger to smaller sizes.
19734 2022-02-09 H.J. Lu <hjl.tools@gmail.com>
19738 * config/i386/gnu-property.cc: Include "i386-protos.h".
19739 (file_end_indicate_exec_stack_and_gnu_property): Generate
19740 a GNU_PROPERTY_1_NEEDED note for -mno-direct-extern-access or
19741 nodirect_extern_access attribute.
19742 * config/i386/i386-options.cc
19743 (handle_nodirect_extern_access_attribute): New function.
19744 (ix86_attribute_table): Add nodirect_extern_access attribute.
19745 * config/i386/i386-protos.h (ix86_force_load_from_GOT_p): Add a
19747 (ix86_has_no_direct_extern_access): New.
19748 * config/i386/i386.cc (ix86_has_no_direct_extern_access): New.
19749 (ix86_force_load_from_GOT_p): Add a bool argument to indicate
19750 call operand. Force non-call load from GOT for
19751 -mno-direct-extern-access or nodirect_extern_access attribute.
19752 (legitimate_pic_address_disp_p): Avoid copy relocation in PIE
19753 for -mno-direct-extern-access or nodirect_extern_access attribute.
19754 (ix86_print_operand): Pass true to ix86_force_load_from_GOT_p
19756 (asm_preferred_eh_data_format): Use PC-relative format for
19757 -mno-direct-extern-access to avoid copy relocation. Check
19758 ptr_mode instead of TARGET_64BIT when selecting DW_EH_PE_sdata4.
19759 (ix86_binds_local_p): Set ix86_has_no_direct_extern_access to
19760 true for -mno-direct-extern-access or nodirect_extern_access
19761 attribute. Don't treat protected data as extern and avoid copy
19762 relocation on common symbol with -mno-direct-extern-access or
19763 nodirect_extern_access attribute.
19764 (ix86_reloc_rw_mask): New to avoid copy relocation for
19765 -mno-direct-extern-access.
19766 (TARGET_ASM_RELOC_RW_MASK): New.
19767 * config/i386/i386.opt: Add -mdirect-extern-access.
19768 * doc/extend.texi: Document nodirect_extern_access attribute.
19769 * doc/invoke.texi: Document -m[no-]direct-extern-access.
19771 2022-02-09 H.J. Lu <hjl.tools@gmail.com>
19774 * config/i386/i386.cc (ix86_avx_u128_mode_source): New function.
19775 (ix86_avx_u128_mode_needed): Return AVX_U128_ANY for debug INSN.
19776 Call ix86_avx_u128_mode_source to check mode for each component
19779 2022-02-09 liuhongt <hongtao.liu@intel.com>
19782 * config/i386/sse.md (<insn><mode>3): lowpart_subreg
19783 operands[2] from SImode to QImode.
19785 2022-02-09 Richard Biener <rguenther@suse.de>
19787 PR middle-end/104450
19788 * gimple-isel.cc: Pass cfun around.
19789 (+gimple_expand_vec_cond_expr): Do not combine a throwing
19790 comparison with the select.
19792 2022-02-09 Richard Biener <rguenther@suse.de>
19795 * config/i386/i386.cc (ix86_gimple_fold_builtin): Guard shift
19796 folding for NULL LHS.
19798 2022-02-08 Robin Dapp <rdapp@linux.ibm.com>
19800 PR rtl-optimization/104198
19801 PR rtl-optimization/104153
19802 * ifcvt.cc (noce_convert_multiple_sets_1): Copy rtx instead of
19803 using it directly. Rework comparison handling and always
19804 perform a second pass.
19806 2022-02-08 Jakub Jelinek <jakub@redhat.com>
19809 * config/rs6000/rs6000.cc (vspltis_shifted): Return false also if
19810 split1 pass has finished already.
19812 2022-02-08 Bill Schmidt <wschmidt@linux.ibm.com>
19814 * config/rs6000/rs6000-builtins.def (VMSUMCUD): New.
19815 * config/rs6000/rs6000-overload.def (VEC_MSUMC): New.
19816 * config/rs6000/vsx.md (UNSPEC_VMSUMCUD): New constant.
19817 (vmsumcud): New define_insn.
19819 2022-02-08 Tom de Vries <tdevries@suse.de>
19821 * config/nvptx/nvptx-opts.h (enum ptx_isa): Add PTX_ISA_SM70.
19822 * config/nvptx/nvptx.h (TARGET_SM70): Define.
19824 2022-02-08 Robin Dapp <rdapp@linux.ibm.com>
19826 * config/s390/s390.cc (s390_rtx_costs): Increase costs for load
19828 * config/s390/s390.md: Use paradoxical subreg.
19830 2022-02-08 Robin Dapp <rdapp@linux.ibm.com>
19832 * combine.cc (reg_subword_p): Check for paradoxical subreg.
19834 2022-02-08 Tom de Vries <tdevries@suse.de>
19837 * config/nvptx/nvptx-opts.h (enum ptx_version): Add PTX_VERSION_3_0
19838 and PTX_VERSION_4_2.
19839 * config/nvptx/nvptx.cc (first_ptx_version_supporting_sm)
19840 (default_ptx_version_option, ptx_version_to_string)
19841 (sm_version_to_string, handle_ptx_version_option): New function.
19842 (nvptx_option_override): Call handle_ptx_version_option.
19843 (nvptx_file_start): Use ptx_version_to_string and sm_version_to_string.
19844 * config/nvptx/nvptx.md (define_insn "nvptx_shuffle<mode>")
19845 (define_insn "nvptx_vote_ballot"): Use TARGET_PTX_6_0.
19846 * config/nvptx/nvptx.opt (mptx): Remove 'Init'.
19848 2022-02-08 Maciej W. Rozycki <macro@embecosm.com>
19850 * doc/install.texi (Configuration): Document `--with-isa-spec='
19852 * doc/invoke.texi (Option Summary): List `-misa-spec=' RISC-V
19854 (RISC-V Options): Document it.
19856 2022-02-08 Maciej W. Rozycki <macro@embecosm.com>
19858 * config/riscv/t-riscv (riscv-sr.o): Add $(TM_H) dependency.
19860 2022-02-08 Tom de Vries <tdevries@suse.de>
19862 * config/nvptx/nvptx.cc (write_fn_proto_1): Handle 'main (int)'.
19864 2022-02-08 Tom de Vries <tdevries@suse.de>
19867 * config/nvptx/nvptx-protos.h (nvptx_mem_local_p): Declare.
19868 * config/nvptx/nvptx.cc (nvptx_reorg_uniform_simt): Assert that
19869 change is validated.
19870 (nvptx_mem_local_p): New function.
19871 * config/nvptx/nvptx.md: Use nvptx_mem_local_p.
19872 (define_c_enum "unspecv"): Add UNSPECV_CAS_LOCAL.
19873 (define_insn "atomic_compare_and_swap<mode>_1_local"): New
19874 non-atomic, non-predicable define_insn, factored out of ...
19875 (define_insn "atomic_compare_and_swap<mode>_1"): ... here.
19876 Make predicable again.
19877 (define_expand "atomic_compare_and_swap<mode>"): Use
19878 atomic_compare_and_swap<mode>_1_local.
19880 2022-02-08 liuhongt <hongtao.liu@intel.com>
19882 PR rtl-optimization/104059
19883 * regcprop.cc (copyprop_hardreg_forward_1): Don't propagate
19884 for a more expensive reg-reg move.
19886 2022-02-07 Tamar Christina <tamar.christina@arm.com>
19888 * config/arm/arm_neon.h (vusdotq_s32, vusdot_laneq_s32,
19889 vusdotq_laneq_s32, vsudot_laneq_s32, vsudotq_laneq_s32): New
19890 * config/arm/arm_neon_builtins.def (usdot): Add V16QI.
19891 (usdot_laneq, sudot_laneq): New.
19892 * config/arm/neon.md (neon_<sup>dot_laneq<vsi2qi>): New.
19893 (neon_<sup>dot_lane<vsi2qi>): Remote unneeded code.
19895 2022-02-07 Tamar Christina <tamar.christina@arm.com>
19897 * config/arm/arm_neon.h (vdot_laneq_u32, vdotq_laneq_u32,
19898 vdot_laneq_s32, vdotq_laneq_s32): New.
19899 * config/arm/arm_neon_builtins.def (sdot_laneq, udot_laneq): New.
19900 * config/arm/neon.md (neon_<sup>dot<vsi2qi>): New.
19901 (<sup>dot_prod<vsi2qi>): Re-order rtl.
19902 (neon_<sup>dot_lane<vsi2qi>): Fix rtl order and endiannes.
19903 (neon_<sup>dot_laneq<vsi2qi>): New.
19905 2022-02-07 Andreas Krebbel <krebbel@linux.ibm.com>
19908 * config/s390/s390.cc (s390_can_inline_p): Accept a few more flags
19909 if always_inline is set. Don't inline when tune differs without
19912 2022-02-07 Richard Biener <rguenther@suse.de>
19914 PR middle-end/104402
19915 * gimple-expr.cc (is_gimple_condexpr): _Complex typed
19916 compares are not valid.
19917 * tree-cfg.cc (verify_gimple_assign_ternary): For COND_EXPR
19918 check is_gimple_condexpr.
19920 2022-02-07 Kewen Lin <linkw@linux.ibm.com>
19923 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Move the
19924 hunk affecting VSX and ALTIVEC to appropriate place.
19926 2022-02-07 Kewen Lin <linkw@linux.ibm.com>
19929 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Disable
19930 MMA if !TARGET_VSX.
19932 2022-02-06 Jakub Jelinek <jakub@redhat.com>
19936 * fold-const.h (folding_initializer): Adjust comment.
19937 (folding_cxx_constexpr): Declare.
19938 * fold-const.cc (folding_initializer): Adjust comment.
19939 (folding_cxx_constexpr): New variable.
19940 (address_compare): Restrict the decl vs. STRING_CST
19941 or vice versa or STRING_CST vs. STRING_CST or
19942 is_global_var != is_global_var optimizations to !folding_cxx_constexpr.
19943 Punt for FUNCTION_DECLs with non-zero offsets. If folding_initializer,
19944 assume non-aliased functions have non-zero size and have different
19945 addresses. For folding_cxx_constexpr, punt on comparisons of start
19946 of some object and end of another one, regardless whether it is a decl
19947 or string literal. Also punt for folding_cxx_constexpr on
19948 STRING_CST vs. STRING_CST comparisons if the two literals could be
19951 2022-02-05 Jakub Jelinek <jakub@redhat.com>
19953 PR tree-optimization/104389
19954 * match.pd (x * 0 -> 0): Punt if x maybe infinite and NaNs are
19957 2022-02-05 Kito Cheng <kito.cheng@sifive.com>
19959 * configure.ac: Fix detection for zifencei support.
19960 * configure: Regenerate.
19962 2022-02-05 Kito Cheng <kito.cheng@sifive.com>
19965 * config.gcc (riscv*-*-*): Normalize the with_isa_spec value.
19966 (all_defaults): Add isa_spec.
19967 * config/riscv/riscv.h (OPTION_DEFAULT_SPECS): Add isa_spec.
19969 2022-02-04 Bill Schmidt <wschmidt@linux.ibm.com>
19971 * config/rs6000/rs6000-c.cc (resolve_vec_mul): Accept args and types
19972 parameters instead of arglist and nargs. Simplify accordingly. Remove
19973 unnecessary test for argument count mismatch.
19974 (resolve_vec_cmpne): Likewise.
19975 (resolve_vec_adde_sube): Likewise.
19976 (resolve_vec_addec_subec): Likewise.
19977 (altivec_resolve_overloaded_builtin): Move overload special handling
19978 after the gathering of arguments into args[] and types[] and the test
19979 for correct number of arguments. Don't perform the test for correct
19980 number of arguments for certain special cases. Call the other special
19981 cases with args and types instead of arglist and nargs.
19983 2022-02-04 Bill Schmidt <wschmidt@linux.ibm.com>
19986 * doc/extend.texi (Basic PowerPC Built-in Functions Available on ISA
19987 3.1): Provide consistent type names. Remove unnecessary semicolons.
19988 Fix bad line breaks.
19990 2022-02-04 Jakub Jelinek <jakub@redhat.com>
19993 * config/rs6000/rs6000.cc (rs6000_mangle_decl_assembler_name): Also
19994 adjust mangling of __builtin*printf_chk.
19996 2022-02-04 Jonathan Wakely <jwakely@redhat.com>
19998 * doc/cpp.texi (Variadic Macros): Replace C++2a with C++20.
20000 2022-02-04 Richard Biener <rguenther@suse.de>
20001 Bin Cheng <bin.cheng@linux.alibaba.com>
20003 PR tree-optimization/100499
20004 * fold-const.h (multiple_of_p): Add nowrap parameter, defaulted
20006 * fold-const.cc (multiple_of_p): Likewise. Honor it for
20007 MULT_EXPR, PLUS_EXPR and MINUS_EXPR and pass it along,
20008 switching to false for conversions.
20009 * tree-ssa-loop-niter.cc (number_of_iterations_ne): Do not
20010 claim the outermost expression does not wrap when calling
20011 multiple_of_p. Refactor the check done to check the
20012 original IV, avoiding a bias that might wrap.
20014 2022-02-04 Richard Biener <rguenther@suse.de>
20016 * fold-const.cc (multiple_of_p): Re-write and move LSHIFT_EXPR
20019 2022-02-04 Eric Botcazou <ebotcazou@adacore.com>
20022 * dwarf2out.cc (dwarf2out_finish): Empty base_types.
20023 (dwarf2out_early_finish): Likewise.
20025 2022-02-04 Eric Botcazou <ebotcazou@adacore.com>
20027 PR tree-optimization/104356
20028 * match.pd (X / bool_range_Y is X): Add guard.
20029 (X / X is one): Likewise.
20030 (X / abs (X) is X < 0 ? -1 : 1): Likewise.
20031 (X / -X is -1): Likewise.
20032 (1 / X -> X == 1): Likewise.
20034 2022-02-04 Richard Biener <rguenther@suse.de>
20036 PR tree-optimization/103641
20037 * tree-vect-patterns.cc (vect_synth_mult_by_constant):
20038 Pass the vector mode to choose_mult_variant.
20040 2022-02-04 Roger Sayle <roger@nextmovesoftware.com>
20042 PR rtl-optimization/101885
20043 * combine.cc (try_combine): When splitting a parallel into two
20044 sequential sets, check not only that the first doesn't clobber
20045 the second but also that the second doesn't clobber the first.
20047 2022-02-04 Richard Biener <rguenther@suse.de>
20049 PR middle-end/90348
20050 PR middle-end/104092
20051 * tree-core.h (clobber_kind): New enum.
20052 (tree_base::u::bits::address_space): Document use in CONSTRUCTORs.
20053 * tree.h (CLOBBER_KIND): Add.
20054 (build_clobber): Add clobber kind argument, defaulted to
20056 * tree.cc (build_clobber): Likewise.
20057 * gimple.h (gimple_clobber_p): New overload with specified kind.
20058 * tree-streamer-in.cc (streamer_read_tree_bitfields): Stream
20060 * tree-streamer-out.cc (streamer_write_tree_bitfields):
20062 * tree-pretty-print.cc (dump_generic_node): Mark EOL CLOBBERs.
20063 * gimplify.cc (gimplify_bind_expr): Build storage end-of-life clobbers
20065 (gimplify_target_expr): Likewise.
20066 * tree-inline.cc (expand_call_inline): Likewise.
20067 * tree-ssa-ccp.cc (insert_clobber_before_stack_restore): Likewise.
20068 * gimple-ssa-warn-access.cc (pass_waccess::check_stmt): Only treat
20069 CLOBBER_EOL clobbers as ending lifetime of storage.
20071 2022-02-04 Martin Sebor <msebor@redhat.com>
20073 * pointer-query.h (pointer_query::cache_type): Use auto_vec for auto
20076 2022-02-03 Martin Sebor <msebor@redhat.com>
20078 PR middle-end/104260
20079 * passes.def (pass_warn_access): Adjust pass placement.
20081 2022-02-03 Uroš Bizjak <ubizjak@gmail.com>
20084 * config/i386/i386.cc (find_drap_reg): For 32bit targets
20085 return DI_REG if function uses __builtin_eh_return.
20087 2022-02-03 Martin Sebor <msebor@redhat.com>
20089 * gimple-ssa-warn-restrict.cc (class pass_wrestrict): Outline ctor.
20090 (pass_wrestrict::m_ptr_qry): New member.
20091 (wrestrict_walk): Rename...
20092 (pass_wrestrict::check_block): ...to this.
20093 (pass_wrestrict::execute): Set up and tear down pointer_query and
20095 (builtin_memref::builtin_memref): Change ctor argument. Simplify.
20096 (builtin_access::builtin_access): Same.
20097 (builtin_access::m_ptr_qry): New member.
20098 (check_call): Rename...
20099 (pass_wrestrict::check_call): ...to this.
20100 (check_bounds_or_overlap): Change argument.
20101 * gimple-ssa-warn-restrict.h (check_bounds_or_overlap): Same.
20103 2022-02-03 Martin Sebor <msebor@redhat.com>
20105 * gimple-array-bounds.cc (array_bounds_checker::array_bounds_checker):
20107 (array_bounds_checker::get_value_range): Use new member.
20108 (array_bounds_checker::check_mem_ref): Same.
20109 * gimple-array-bounds.h (array_bounds_checker::array_bounds_checker):
20111 (array_bounds_checker::m_ptr_query): New member.
20113 2022-02-03 Martin Sebor <msebor@redhat.com>
20115 * gimple-ssa-warn-access.cc (pass_waccess::pass_waccess): Remove
20116 pointer_query cache.
20117 * pointer-query.cc (pointer_query::pointer_query): Remove cache
20118 argument. Zero-initialize new cache member.
20119 (pointer_query::get_ref): Replace cache pointer with direct access.
20120 (pointer_query::put_ref): Same.
20121 (pointer_query::flush_cache): Same.
20122 (pointer_query::dump): Same.
20123 * pointer-query.h (class pointer_query): Remove cache argument from
20124 ctor. Change cache pointer to cache subobject member.
20125 * tree-ssa-strlen.cc: Remove pointer_query cache.
20127 2022-02-03 Martin Sebor <msebor@redhat.com>
20129 PR tree-optimization/104119
20130 * gimple-ssa-sprintf.cc (struct directive): Change argument type.
20131 (format_none): Same.
20132 (format_percent): Same.
20133 (format_integer): Same.
20134 (format_floating): Same.
20135 (get_string_length): Same.
20136 (format_character): Same.
20137 (format_string): Same.
20138 (format_plain): Same.
20139 (format_directive): Same.
20140 (compute_format_length): Same.
20141 (handle_printf_call): Same.
20142 * tree-ssa-strlen.cc (get_range_strlen_dynamic): Same. Call
20144 (get_range_strlen_phi): Same.
20145 (get_maxbound): New function.
20146 (strlen_pass::get_len_or_size): Adjust to parameter change.
20147 * tree-ssa-strlen.h (get_range_strlen_dynamic): Change argument type.
20149 2022-02-03 Bill Schmidt <wschmidt@linux.ibm.com>
20152 * config/rs6000/rs6000-builtin.cc (rs6000_gimple_fold_builtin): Remove
20153 test for !rs6000_fold_gimple.
20154 * config/rs6000/rs6000.cc (rs6000_option_override_internal): Likewise.
20155 * config/rs6000/rs6000.opt (mfold-gimple): Remove.
20157 2022-02-03 Bill Schmidt <wschmidt@linux.ibm.com>
20160 * config/rs6000/rs6000-builtin.cc (rs6000_expand_builtin): Handle
20161 endianness for vclzlsbb and vctzlsbb.
20162 * config/rs6000/rs6000-builtins.def (VCLZLSBB_V16QI): Change
20163 default pattern and indicate a different pattern will be used for
20165 (VCLZLSBB_V4SI): Likewise.
20166 (VCLZLSBB_V8HI): Likewise.
20167 (VCTZLSBB_V16QI): Likewise.
20168 (VCTZLSBB_V4SI): Likewise.
20169 (VCTZLSBB_V8HI): Likewise.
20171 2022-02-03 Bill Schmidt <wschmidt@linux.ibm.com>
20173 * config.gcc (powerpc*-*-*): Add rs6000-builtin.o to extra_objs.
20174 * config/rs6000/rs6000-builtin.cc: New file, containing code moved
20176 * config/rs6000/rs6000-call.cc (cpu_is_info): Move to
20178 (cpu_supports_info): Likewise.
20179 (rs6000_type_string): Likewise.
20180 (altivec_expand_predicate_builtin): Likewise.
20181 (rs6000_htm_spr_icode): Likewise.
20182 (altivec_expand_vec_init_builtin): Likewise.
20183 (get_element_number): Likewise.
20184 (altivec_expand_vec_set_builtin): Likewise.
20185 (altivec_expand_vec_ext_builtin): Likewise.
20186 (rs6000_invalid_builtin): Likewise.
20187 (rs6000_fold_builtin): Likewise.
20188 (fold_build_vec_cmp): Likewise.
20189 (fold_compare_helper): Likewise.
20190 (map_to_integral_tree_type): Likewise.
20191 (fold_mergehl_helper): Likewise.
20192 (fold_mergeeo_helper): Likewise.
20193 (rs6000_builtin_valid_without_lhs): Likewise.
20194 (rs6000_builtin_is_supported): Likewise.
20195 (rs6000_gimple_fold_mma_builtin): Likewise.
20196 (rs6000_gimple_fold_builtin): Likewise.
20197 (rs6000_expand_ldst_mask): Likewise.
20198 (cpu_expand_builtin): Likewise.
20199 (elemrev_icode): Likewise.
20200 (ldv_expand_builtin): Likewise.
20201 (lxvrse_expand_builtin): Likewise.
20202 (lxvrze_expand_builtin): Likewise.
20203 (stv_expand_builtin): Likewise.
20204 (mma_expand_builtin): Likewise.
20205 (htm_spr_num): Likewise.
20206 (htm_expand_builtin): Likewise.
20207 (rs6000_expand_builtin): Likewise.
20208 (rs6000_vector_type): Likewise.
20209 (rs6000_init_builtins): Likewise. Remove initialization of
20210 builtin_mode_to_type entries.
20211 (rs6000_builtin_decl): Move to rs6000-builtin.cc.
20212 * config/rs6000/rs6000.cc (rs6000_builtin_mask_for_load): New
20213 external declaration.
20214 (rs6000_builtin_md_vectorized_function): Likewise.
20215 (rs6000_builtin_reciprocal): Likewise.
20216 (altivec_builtin_mask_for_load): Move to rs6000-builtin.cc.
20217 (rs6000_builtin_types): Likewise.
20218 (builtin_mode_to_type): Remove.
20219 (rs6000_builtin_mask_for_load): Move to rs6000-builtin.cc. Remove
20221 (rs6000_builtin_md_vectorized_function): Likewise.
20222 (rs6000_builtin_reciprocal): Likewise.
20223 * config/rs6000/rs6000.h (builtin_mode_to_type): Remove.
20224 * config/rs6000/t-rs6000 (rs6000-builtin.o): New target.
20226 2022-02-03 Richard Biener <rguenther@suse.de>
20229 * tree-nrv.cc (pass_nrv::execute): Remove tieing result and found
20230 together via DECL_ABSTRACT_ORIGIN.
20232 2022-02-03 Bill Schmidt <wschmidt@linux.ibm.com>
20234 * config/rs6000/rs6000-call.cc (rs6000_expand_builtin): Revise error
20235 message for RES_BITS case.
20237 2022-02-03 Aldy Hernandez <aldyh@redhat.com>
20239 * gimple-range-fold.cc (fur_list::fur_list): Set m_local[1] correctly.
20241 2022-02-03 Jakub Jelinek <jakub@redhat.com>
20243 * config/arm/arm.opt (mfix-cortex-a57-aes-1742098,
20244 mfix-cortex-a72-aes-1655431): Ensure description ends with full stop.
20246 2022-02-03 Aldy Hernandez <aldyh@redhat.com>
20248 * cfganal.cc (verify_marked_backedges): New.
20249 * cfganal.h (verify_marked_backedges): New.
20250 * gimple-range-path.cc (path_range_query::path_range_query):
20251 Verify freshness of back edges.
20252 * tree-ssa-loop-ch.cc (ch_base::copy_headers): Call
20253 mark_dfs_back_edges.
20254 * tree-ssa-threadbackward.cc (back_threader::back_threader): Move
20255 path_range_query construction after backedges have been
20258 2022-02-03 Richard Sandiford <richard.sandiford@arm.com>
20260 * config/aarch64/aarch64-simd.md (movmisalign<mode>): Extend from
20263 2022-02-03 Richard Sandiford <richard.sandiford@arm.com>
20265 * config/aarch64/iterators.md (VALL_F16MOV): Delete.
20266 * config/aarch64/aarch64-simd.md (mov<mode>): Use VALL_F16 instead
20269 2022-02-03 Martin Liska <mliska@suse.cz>
20271 * config/i386/i386-options.cc (ix86_valid_target_attribute_inner_p):
20272 Change subject and object in the error message.
20273 * config/s390/s390.cc (s390_valid_target_attribute_inner_p):
20276 2022-02-03 Martin Liska <mliska@suse.cz>
20278 * config/s390/s390.cc (s390_valid_target_attribute_inner_p):
20279 Use the error message for i386 target.
20281 2022-02-03 Jakub Jelinek <jakub@redhat.com>
20283 PR tree-optimization/104334
20284 * range-op.cc (range_operator::wi_fold_in_parts): Change lh_range
20285 and rh_range type to widest_int and subtract in widest_int. Remove
20286 ov_rh, ov_lh and sign vars, always perform comparisons as signed
20287 and use >, < and == operators for it.
20289 2022-02-03 Martin Sebor <msebor@redhat.com>
20291 * common.opt (-Wuse-after-free): Correct typos.
20293 2022-02-02 David Malcolm <dmalcolm@redhat.com>
20296 * doc/invoke.texi (-ftrivial-auto-var-init=): Add reference to
20297 -Wanalyzer-use-of-uninitialized-value to paragraph documenting that
20298 -ftrivial-auto-var-init= doesn't suppress warnings.
20300 2022-02-02 Martin Liska <mliska@suse.cz>
20302 * dwarf2out.cc (TEXT_SECTION_NAME): Remove unused macro.
20304 2022-02-02 Bernd Kuhls <bernd.kuhls@t-online.de>
20307 * config/or1k/linux.h (CPP_SPEC): Define.
20309 2022-02-02 Tamar Christina <tamar.christina@arm.com>
20311 PR tree-optimization/102819
20312 PR tree-optimization/103169
20313 * config/arm/vec-common.md (cml<fcmac1><conj_op><mode>4): Use
20316 2022-02-02 Tamar Christina <tamar.christina@arm.com>
20318 PR tree-optimization/102819
20319 PR tree-optimization/103169
20320 * config/aarch64/aarch64-simd.md (cml<fcmac1><conj_op><mode>4): Use
20322 * config/aarch64/aarch64-sve.md (cml<fcmac1><conj_op><mode>4): Likewise.
20324 2022-02-02 Tamar Christina <tamar.christina@arm.com>
20326 PR tree-optimization/102819
20327 PR tree-optimization/103169
20328 * doc/md.texi: Update docs for cfms, cfma.
20329 * tree-data-ref.h (same_data_refs): Accept optional offset.
20330 * tree-vect-slp-patterns.cc (is_linear_load_p): Fix issue with repeating
20332 (vect_normalize_conj_loc): Remove.
20333 (is_eq_or_top): Change to take two nodes.
20334 (enum _conj_status, compatible_complex_nodes_p,
20335 vect_validate_multiplication): New.
20336 (class complex_add_pattern, complex_add_pattern::matches,
20337 complex_add_pattern::recognize, class complex_mul_pattern,
20338 complex_mul_pattern::recognize, class complex_fms_pattern,
20339 complex_fms_pattern::recognize, class complex_operations_pattern,
20340 complex_operations_pattern::recognize, addsub_pattern::recognize): Pass
20342 (complex_fms_pattern::matches, complex_mul_pattern::matches): Pass new
20343 cache and use new validation code.
20344 * tree-vect-slp.cc (vect_match_slp_patterns_2, vect_match_slp_patterns,
20345 vect_analyze_slp): Pass along cache.
20346 (compatible_calls_p): Expose.
20347 * tree-vectorizer.h (compatible_calls_p, slp_node_hash,
20348 slp_compat_nodes_map_t): New.
20349 (class vect_pattern): Update signatures include new cache.
20351 2022-02-02 Hans-Peter Nilsson <hp@axis.com>
20353 * config/cris/cris.cc (cris_preferred_reload_class): Reject
20354 "eliminated" registers and small-enough constants unless
20355 reloaded into a class that is a subset of GENERAL_REGS.
20356 * config/cris/cris.md (attribute "cpu_variant"): New.
20357 (attribute "enabled"): Conditionalize on a matching attribute
20358 cpu_variant, if specified.
20359 ("*movsi_internal<setcc><setnz><setnzvc>"): For moves to and from
20360 memory, add cpu-variant-enabled variants for "r" alternatives on
20361 the far side of the "x" alternatives, preferring the "x" ones
20362 only for variants where MOF is present (in addition to SRP).
20364 2022-02-02 Hans-Peter Nilsson <hp@axis.com>
20366 * config/cris/cris.cc (cris_register_move_cost): Remove special pre-ira
20367 extra cost for ALL_REGS.
20369 2022-02-02 Hans-Peter Nilsson <hp@axis.com>
20371 * config/cris/constraints.md (define_register_constraint "b"): Now
20373 * config/cris/cris.md (CRIS_ACR_REGNUM): Remove.
20374 * config/cris/cris.h: (reg_class, REG_CLASS_NAMES)
20375 (REG_CLASS_CONTENTS): Remove ACR_REGS, SPEC_ACR_REGS, GENNONACR_REGS,
20376 and SPEC_GENNONACR_REGS.
20377 * config/cris/cris.cc (cris_preferred_reload_class): Don't mention
20378 ACR_REGS and return GENERAL_REGS instead of GENNONACR_REGS.
20380 2022-02-02 Hans-Peter Nilsson <hp@axis.com>
20382 * config/cris/cris.md ("*movsi_internal<setcc><setnz><setnzvc>"):
20383 Conditionalize on (sub-)register operands or operand 1 being 0.
20385 2022-02-02 Hans-Peter Nilsson <hp@axis.com>
20387 * config/cris/cris.h (TARGET_DEFAULT): Don't include MASK_MUL_BUG.
20388 (MUL_BUG_ASM_DEFAULT): New macro.
20389 (MAYBE_AS_NO_MUL_BUG_ABORT): Define in terms of MUL_BUG_ASM_DEFAULT.
20390 * doc/invoke.texi (CRIS Options, -mmul-bug-workaround): Adjust
20393 2022-02-01 Eugene Rozenfeld <erozen@microsoft.com>
20395 * opts.cc (common_handle_option): Don't set param_early_inliner_max_iterations
20398 2022-02-01 Eugene Rozenfeld <erozen@microsoft.com>
20400 * auto-profile.cc (auto_profile): Hard-code the number of iterations (10).
20402 2022-02-01 Andrew Pinski <apinski@marvell.com>
20404 * doc/install.texi:
20406 2022-02-01 Ilya Leoshkevich <iii@linux.ibm.com>
20408 * config/s390/s390.cc (s390_code_end): Do not switch back to
20411 2022-02-01 Jakub Jelinek <jakub@redhat.com>
20414 * config/rs6000/t-rs6000 (EXTRA_GTYPE_DEPS): Append rs6000-builtins.h
20415 rather than $(srcdir)/config/rs6000/rs6000-builtins.def.
20416 * config/rs6000/rs6000-gen-builtins.cc (write_decls): Don't use
20417 GTY((user)) for struct bifdata and struct ovlddata. Instead add
20418 GTY((skip(""))) to members with pointer and enum types that don't need
20419 to be tracked. Add GTY(()) to rs6000_builtin_info and rs6000_instance_info
20420 declarations. Don't emit gt_ggc_mx and gt_pch_nx declarations.
20421 (write_extern_fntype, write_fntype): Remove.
20422 (write_fntype_init): Emit the fntype vars as automatic vars instead
20423 of file scope ones.
20424 (write_header_file): Don't iterate with write_extern_fntype.
20425 (write_init_file): Don't iterate with write_fntype. Don't emit
20426 gt_ggc_mx and gt_pch_nx definitions.
20428 2022-02-01 Jason Merrill <jason@redhat.com>
20430 * tree.h (struct tree_vec_map_cache_hasher): Move from...
20431 * tree.cc (struct tree_vec_map_cache_hasher): ...here.
20433 2022-02-01 Tom de Vries <tdevries@suse.de>
20435 * config/nvptx/nvptx.cc (nvptx_single): Use nvptx_uniform_warp_check.
20436 * config/nvptx/nvptx.md (define_c_enum "unspecv"): Add
20437 UNSPECV_UNIFORM_WARP_CHECK.
20438 (define_insn "nvptx_uniform_warp_check"): New define_insn.
20440 2022-02-01 Tom de Vries <tdevries@suse.de>
20442 * config/nvptx/nvptx.cc (nvptx_single): Use nvptx_warpsync.
20443 * config/nvptx/nvptx.md (define_c_enum "unspecv"): Add
20445 (define_insn "nvptx_warpsync"): New define_insn.
20447 2022-02-01 Tom de Vries <tdevries@suse.de>
20449 * config/nvptx/nvptx.opt (mptx): Set to PTX_VERSION_6_3 by default.
20451 2022-02-01 Tom de Vries <tdevries@suse.de>
20453 * config/nvptx/nvptx-opts.h (enum ptx_version): Add PTX_VERSION_6_0.
20454 * config/nvptx/nvptx.h (TARGET_PTX_6_0): New macro.
20455 * config/nvptx/nvptx.md (define_insn "nvptx_barsync"): Use barrier
20456 insn for TARGET_PTX_6_0.
20458 2022-02-01 Tom de Vries <tdevries@suse.de>
20461 * config/nvptx/nvptx.cc (prevent_branch_around_nothing): Handle nop
20464 2022-02-01 Tom de Vries <tdevries@suse.de>
20466 * config/nvptx/nvptx.md (define_insn "atomic_compare_and_swap<mode>_1")
20467 (define_insn "atomic_exchange<mode>")
20468 (define_insn "atomic_fetch_add<mode>")
20469 (define_insn "atomic_fetch_addsf")
20470 (define_insn "atomic_fetch_<logic><mode>"): Output non-atomic version
20471 if memory operands is frame-relative.
20473 2022-02-01 Tom de Vries <tdevries@suse.de>
20475 * config/nvptx/nvptx.cc (enum nvptx_builtins): Add
20476 NVPTX_BUILTIN_MEMBAR_GL and NVPTX_BUILTIN_MEMBAR_CTA.
20478 (nvptx_init_builtins): Add MEMBAR_GL and MEMBAR_CTA.
20479 (nvptx_expand_builtin): Handle NVPTX_BUILTIN_MEMBAR_GL and
20480 NVPTX_BUILTIN_MEMBAR_CTA.
20481 (nvptx_lockfull_update): Add level parameter. Emit barriers.
20482 (nvptx_reduction_update, nvptx_goacc_reduction_fini): Update call to
20483 nvptx_lockfull_update.
20484 * config/nvptx/nvptx.md (define_c_enum "unspecv"): Add
20486 (define_expand "nvptx_membar_gl"): New expand.
20487 (define_insn "*nvptx_membar_gl"): New insn.
20489 2022-02-01 Martin Liska <mliska@suse.cz>
20491 * doc/install.texi: Remove option for GCC < 4.8.
20493 2022-02-01 Jakub Jelinek <jakub@redhat.com>
20495 PR middle-end/104307
20496 * tree-vect-generic.cc (expand_vector_comparison): Don't push debug
20497 stmts to uses vector, just set vec_cond_expr_only to false for
20498 non-VEC_COND_EXPRs instead of pushing them into uses. Treat
20499 VEC_COND_EXPRs that use lhs not just in rhs1, but rhs2 or rhs3 too
20500 like non-VEC_COND_EXPRs.
20502 2022-02-01 Bill Schmidt <wschmidt@linux.ibm.com>
20504 * config/rs6000/rs6000-overload.def (VEC_ABSD): Remove #ifdef token.
20505 (VEC_BLENDV): Likewise.
20506 (VEC_BPERM): Likewise.
20507 (VEC_CFUGE): Likewise.
20508 (VEC_CIPHER_BE): Likewise.
20509 (VEC_CIPHERLAST_BE): Likewise.
20510 (VEC_CLRL): Likewise.
20511 (VEC_CLRR): Likewise.
20512 (VEC_CMPNEZ): Likewise.
20513 (VEC_CNTLZ): Likewise.
20514 (VEC_CNTLZM): Likewise.
20515 (VEC_CNTTZM): Likewise.
20516 (VEC_CNTLZ_LSBB): Likewise.
20517 (VEC_CNTM): Likewise.
20518 (VEC_CNTTZ): Likewise.
20519 (VEC_CNTTZ_LSBB): Likewise.
20520 (VEC_CONVERT_4F32_8F16): Likewise.
20521 (VEC_DIV): Likewise.
20522 (VEC_DIVE): Likewise.
20523 (VEC_EQV): Likewise.
20524 (VEC_EXPANDM): Likewise.
20525 (VEC_EXTRACT_FP_FROM_SHORTH): Likewise.
20526 (VEC_EXTRACT_FP_FROM_SHORTL): Likewise.
20527 (VEC_EXTRACTH): Likewise.
20528 (VEC_EXTRACTL): Likewise.
20529 (VEC_EXTRACTM): Likewise.
20530 (VEC_EXTRACT4B): Likewise.
20531 (VEC_EXTULX): Likewise.
20532 (VEC_EXTURX): Likewise.
20533 (VEC_FIRSTMATCHINDEX): Likewise.
20534 (VEC_FIRSTMACHOREOSINDEX): Likewise.
20535 (VEC_FIRSTMISMATCHINDEX): Likewise.
20536 (VEC_FIRSTMISMATCHOREOSINDEX): Likewise.
20537 (VEC_GB): Likewise.
20538 (VEC_GENBM): Likewise.
20539 (VEC_GENHM): Likewise.
20540 (VEC_GENWM): Likewise.
20541 (VEC_GENDM): Likewise.
20542 (VEC_GENQM): Likewise.
20543 (VEC_GENPCVM): Likewise.
20544 (VEC_GNB): Likewise.
20545 (VEC_INSERTH): Likewise.
20546 (VEC_INSERTL): Likewise.
20547 (VEC_INSERT4B): Likewise.
20548 (VEC_LXVL): Likewise.
20549 (VEC_MERGEE): Likewise.
20550 (VEC_MERGEO): Likewise.
20551 (VEC_MOD): Likewise.
20552 (VEC_MSUB): Likewise.
20553 (VEC_MULH): Likewise.
20554 (VEC_NAND): Likewise.
20555 (VEC_NCIPHER_BE): Likewise.
20556 (VEC_NCIPHERLAST_BE): Likewise.
20557 (VEC_NEARBYINT): Likewise.
20558 (VEC_NMADD): Likewise.
20559 (VEC_ORC): Likewise.
20560 (VEC_PDEP): Likewise.
20561 (VEC_PERMX): Likewise.
20562 (VEC_PEXT): Likewise.
20563 (VEC_POPCNT): Likewise.
20564 (VEC_PARITY_LSBB): Likewise.
20565 (VEC_REPLACE_ELT): Likewise.
20566 (VEC_REPLACE_UN): Likewise.
20567 (VEC_REVB): Likewise.
20568 (VEC_RINT): Likewise.
20569 (VEC_RLMI): Likewise.
20570 (VEC_RLNM): Likewise.
20571 (VEC_SBOX_BE): Likewise.
20572 (VEC_SIGNEXTI): Likewise.
20573 (VEC_SIGNEXTLL): Likewise.
20574 (VEC_SIGNEXTQ): Likewise.
20575 (VEC_SLDB): Likewise.
20576 (VEC_SLV): Likewise.
20577 (VEC_SPLATI): Likewise.
20578 (VEC_SPLATID): Likewise.
20579 (VEC_SPLATI_INS): Likewise.
20580 (VEC_SQRT): Likewise.
20581 (VEC_SRDB): Likewise.
20582 (VEC_SRV): Likewise.
20583 (VEC_STRIL): Likewise.
20584 (VEC_STRIL_P): Likewise.
20585 (VEC_STRIR): Likewise.
20586 (VEC_STRIR_P): Likewise.
20587 (VEC_STXVL): Likewise.
20588 (VEC_TERNARYLOGIC): Likewise.
20589 (VEC_TEST_LSBB_ALL_ONES): Likewise.
20590 (VEC_TEST_LSBB_ALL_ZEROS): Likewise.
20591 (VEC_VEE): Likewise.
20592 (VEC_VES): Likewise.
20593 (VEC_VIE): Likewise.
20594 (VEC_VPRTYB): Likewise.
20595 (VEC_VSCEEQ): Likewise.
20596 (VEC_VSCEGT): Likewise.
20597 (VEC_VSCELT): Likewise.
20598 (VEC_VSCEUO): Likewise.
20599 (VEC_VSEE): Likewise.
20600 (VEC_VSES): Likewise.
20601 (VEC_VSIE): Likewise.
20602 (VEC_VSTDC): Likewise.
20603 (VEC_VSTDCN): Likewise.
20604 (VEC_VTDC): Likewise.
20605 (VEC_XL): Likewise.
20606 (VEC_XL_BE): Likewise.
20607 (VEC_XL_LEN_R): Likewise.
20608 (VEC_XL_SEXT): Likewise.
20609 (VEC_XL_ZEXT): Likewise.
20610 (VEC_XST): Likewise.
20611 (VEC_XST_BE): Likewise.
20612 (VEC_XST_LEN_R): Likewise.
20613 (VEC_XST_TRUNC): Likewise.
20614 (VEC_XXPERMDI): Likewise.
20615 (VEC_XXSLDWI): Likewise.
20616 (VEC_TSTSFI_EQ_DD): Likewise.
20617 (VEC_TSTSFI_EQ_TD): Likewise.
20618 (VEC_TSTSFI_GT_DD): Likewise.
20619 (VEC_TSTSFI_GT_TD): Likewise.
20620 (VEC_TSTSFI_LT_DD): Likewise.
20621 (VEC_TSTSFI_LT_TD): Likewise.
20622 (VEC_TSTSFI_OV_DD): Likewise.
20623 (VEC_TSTSFI_OV_TD): Likewise.
20624 (VEC_VADDCUQ): Likewise.
20625 (VEC_VADDECUQ): Likewise.
20626 (VEC_VADDEUQM): Likewise.
20627 (VEC_VADDUDM): Likewise.
20628 (VEC_VADDUQM): Likewise.
20629 (VEC_VBPERMQ): Likewise.
20630 (VEC_VCLZB): Likewise.
20631 (VEC_VCLZD): Likewise.
20632 (VEC_VCLZH): Likewise.
20633 (VEC_VCLZW): Likewise.
20634 (VEC_VCTZB): Likewise.
20635 (VEC_VCTZD): Likewise.
20636 (VEC_VCTZH): Likewise.
20637 (VEC_VCTZW): Likewise.
20638 (VEC_VEEDP): Likewise.
20639 (VEC_VEESP): Likewise.
20640 (VEC_VESDP): Likewise.
20641 (VEC_VESSP): Likewise.
20642 (VEC_VIEDP): Likewise.
20643 (VEC_VIESP): Likewise.
20644 (VEC_VPKSDSS): Likewise.
20645 (VEC_VPKSDUS): Likewise.
20646 (VEC_VPKUDUM): Likewise.
20647 (VEC_VPKUDUS): Likewise.
20648 (VEC_VPOPCNT): Likewise.
20649 (VEC_VPOPCNTB): Likewise.
20650 (VEC_VPOPCNTD): Likewise.
20651 (VEC_VPOPCNTH): Likewise.
20652 (VEC_VPOPCNTW): Likewise.
20653 (VEC_VPRTYBD): Likewise.
20654 (VEC_VPRTYBQ): Likewise.
20655 (VEC_VPRTYBW): Likewise.
20656 (VEC_VRLD): Likewise.
20657 (VEC_VSLD): Likewise.
20658 (VEC_VSRAD): Likewise.
20659 (VEC_VSRD): Likewise.
20660 (VEC_VSTDCDP): Likewise.
20661 (VEC_VSTDCNDP): Likewise.
20662 (VEC_VSTDCNQP): Likewise.
20663 (VEC_VSTDCNSP): Likewise.
20664 (VEC_VSTDCQP): Likewise.
20665 (VEC_VSTDCSP): Likewise.
20666 (VEC_VSUBECUQ): Likewise.
20667 (VEC_VSUBEUQM): Likewise.
20668 (VEC_VSUBUDM): Likewise.
20669 (VEC_VSUBUQM): Likewise.
20670 (VEC_VTDCDP): Likewise.
20671 (VEC_VTDCSP): Likewise.
20672 (VEC_VUPKHSW): Likewise.
20673 (VEC_VUPKLSW): Likewise.
20675 2022-02-01 Andreas Krebbel <krebbel@linux.ibm.com>
20677 PR rtl-optimization/101260
20678 * regcprop.cc (maybe_mode_change): Invoke mode_change_ok also for
20681 2022-02-01 Xi Ruoyao <xry111@mengyan1223.wang>
20683 PR middle-end/95115
20684 * fold-const.cc (const_binop): Do not fold NaN result from
20687 2022-02-01 Tom de Vries <tdevries@suse.de>
20689 * tree-loop-distribution.cc (generate_reduction_builtin_1): Check for
20690 -ftree-loop-distribute-patterns.
20691 (loop_distribution::execute): Don't call transform_reduction_loop for
20692 -fno-tree-loop-distribute-patterns.
20694 2022-01-31 Andrew Pinski <apinski@marvell.com>
20696 * fold-const.h (operand_compare::operand_equal_p):
20697 Fix comment about OEP_* flags.
20699 2022-01-31 Jakub Jelinek <jakub@redhat.com>
20702 * config/rs6000/aix.h (OPTION_GLIBC): Remove.
20703 * config/rs6000/darwin.h (OPTION_GLIBC): Likewise.
20704 * config/rs6000/option-defaults.h (OPTION_GLIBC): Define to 0
20705 if not already defined.
20707 2022-01-31 Martin Sebor <msebor@redhat.com>
20709 PR middle-end/104232
20710 * gimple-ssa-warn-access.cc (pointers_related_p): Add argument.
20711 Handle PHIs. Add a synonymous overload.
20712 (pass_waccess::check_pointer_uses): Call pointers_related_p.
20714 2022-01-31 Richard Biener <rguenther@suse.de>
20716 PR tree-optimization/100499
20717 * fold-const.cc (multiple_of_p): Pass the correct type of
20718 the expression to the recursive invocation of multiple_of_p
20719 for conversions and use CASE_CONVERT.
20721 2022-01-31 Eric Botcazou <ebotcazou@adacore.com>
20724 * config/sparc/linux64.h (TARGET_DEFAULT): Add MASK_V8PLUS.
20726 2022-01-31 Richard Biener <rguenther@suse.de>
20728 PR tree-optimization/100499
20729 * tree-cfg.cc (verify_gimple_assign_ternary): Use multiple_p
20730 on poly-ints instead of multiple_of_p.
20731 * tree-ssa.cc (maybe_rewrite_mem_ref_base): Likewise.
20732 (non_rewritable_mem_ref_base): Likewise.
20733 (non_rewritable_lvalue_p): Likewise.
20734 (execute_update_addresses_taken): Likewise.
20736 2022-01-29 Jakub Jelinek <jakub@redhat.com>
20737 Andrew Pinski <apinski@marvell.com>
20739 PR tree-optimization/104279
20740 PR tree-optimization/104280
20741 PR tree-optimization/104281
20742 * match.pd (1 / X -> X == 1 for unsigned X): Build eq with
20743 boolean_type_node and convert to type. Formatting fixes.
20745 2022-01-28 Yoshinori Sato <yo-satoh@sios.com>
20747 * config/sh/t-linux (MULTILIB_EXCEPTIONS): Add m1, mb/m1 and m2a.
20749 2022-01-28 Navid Rahimi <navidrahimi@microsoft.com>
20751 PR tree-optimization/103514
20752 * match.pd (a & b) ^ (a == b) -> !(a | b): New optimization.
20753 (a & b) == (a ^ b) -> !(a | b): New optimization.
20755 2022-01-28 Marek Polacek <polacek@redhat.com>
20757 * doc/invoke.texi: Update -Wbidi-chars documentation.
20759 2022-01-28 Iain Sandoe <iain@sandoe.co.uk>
20761 * config/rs6000/darwin.h (OPTION_GLIBC): Define to 0.
20763 2022-01-28 Zhao Wei Liew <zhaoweiliew@gmail.com>
20765 PR tree-optimization/95424
20766 * match.pd: Simplify 1 / X where X is an integer.
20768 2022-01-28 Jakub Jelinek <jakub@redhat.com>
20770 PR tree-optimization/104263
20771 * gimple-ssa-store-merging.cc (get_status_for_store_merging): For
20772 cfun->can_throw_non_call_exceptions && cfun->eh test whether
20773 last non-debug stmt in the bb is store_valid_for_store_merging_p
20774 rather than last stmt.
20776 2022-01-28 Martin Liska <mliska@suse.cz>
20778 * diagnostic.cc (diagnostic_action_after_output): Remove extra
20781 2022-01-28 Martin Liska <mliska@suse.cz>
20783 * config/rs6000/host-darwin.cc (segv_crash_handler):
20784 Do not use leading capital letter.
20785 (segv_handler): Likewise.
20786 * ipa-sra.cc (verify_splitting_accesses): Likewise.
20787 * varasm.cc (get_section): Likewise.
20789 2022-01-28 Richard Biener <rguenther@suse.de>
20791 PR tree-optimization/104267
20792 * tree-vect-stmts.cc (vectorizable_call): Properly use the
20793 per-argument determined vector type for externals and
20796 2022-01-28 Richard Biener <rguenther@suse.de>
20798 PR tree-optimization/104263
20799 * tree-cfg.cc (gimple_purge_dead_abnormal_call_edges):
20800 Purge edges also when !cfun->has_nonlocal_label
20801 and !cfun->calls_setjmp.
20803 2022-01-28 Maciej W. Rozycki <macro@embecosm.com>
20805 * config/riscv/riscv.md: Document `auipc' and `bitmanip' `type'
20808 2022-01-28 Jakub Jelinek <jakub@redhat.com>
20811 * cfgrtl.cc (loc_equal): New function.
20812 (unique_locus_on_edge_between_p): Use it.
20814 2022-01-28 Richard Biener <rguenther@suse.de>
20816 * cfganal.h (mark_dfs_back_edges): Provide API with struct
20818 * cfganal.cc (mark_dfs_back_edges): Take a struct function
20819 to work on, add a wrapper passing cfun.
20820 * graph.cc (draw_cfg_nodes_no_loops): Replace stray cfun
20821 uses with fun which is already passed.
20822 (draw_cfg_edges): Likewise.
20823 (draw_cfg_nodes_for_loop): Do not use draw_cfg_nodes_for_loop
20826 2022-01-27 Patrick Palka <ppalka@redhat.com>
20829 * tree.cc (build_call_vec): Add const to second parameter.
20830 * tree.h (build_call_vec): Likewise.
20832 2022-01-27 Martin Liska <mliska@suse.cz>
20835 * diagnostic.cc (diagnostic_initialize):
20836 Initialize report_bug flag.
20837 (diagnostic_action_after_output):
20838 Explain that -freport-bug option can be used for pre-processed
20839 file creation. Make the message shorter.
20840 (error_recursion): Rename Internal to internal.
20841 * diagnostic.h (struct diagnostic_context): New field.
20842 * opts.cc (common_handle_option): Init the field here.
20844 2022-01-27 Kewen Lin <linkw@linux.ibm.com>
20847 * config/rs6000/rs6000.cc
20848 (rs6000_cost_data::update_target_cost_per_stmt): Fix one wrong
20849 assertion with early return.
20851 2022-01-27 Chung-Lin Tang <cltang@codesourcery.com>
20853 PR middle-end/103642
20854 * gimplify.cc (gimplify_scan_omp_clauses): Do not do indir_p handling
20855 for non-pointer or non-reference-to-pointer cases.
20857 2022-01-27 Jakub Jelinek <jakub@redhat.com>
20859 PR tree-optimization/104196
20860 * gimple-fold.h (rewrite_to_defined_overflow): Add IN_PLACE argument.
20861 * gimple-fold.cc (rewrite_to_defined_overflow): Likewise. If true,
20862 return NULL and emit needed stmts before and after stmt.
20863 * tree-ssa-reassoc.cc (update_range_test): For inter-bb range opt
20864 pick as operand_entry that will hold the merged test the one feeding
20865 earliest condition, ensure that by swapping range->idx with some
20866 other range's idx if needed. If seq is non-NULL, don't actually swap
20867 it but instead rewrite stmts with undefined overflow in between
20869 (maybe_optimize_range_tests): Set ops[]->id to bb->index with the
20870 corresponding condition even if they have non-NULL ops[]->op.
20873 2022-01-26 Jakub Jelinek <jakub@redhat.com>
20876 * config/rs6000/emmintrin.h (_mm_sad_epu8): Use __asm__ instead of
20878 * config/rs6000/smmintrin.h (_mm_minpos_epu16): Declare iterator
20879 before for loop instead of for init clause.
20880 * config/rs6000/bmi2intrin.h (_pext_u64): Likewise.
20882 2022-01-26 Jakub Jelinek <jakub@redhat.com>
20885 * config/rs6000/bmiintrin.h: Test _X86GPRINTRIN_H_INCLUDED instead of
20886 _X86INTRIN_H_INCLUDED and adjust #error wording.
20887 * config/rs6000/bmi2intrin.h: Likewise.
20889 2022-01-26 Jakub Jelinek <jakub@redhat.com>
20892 * dwarf2out.cc (long_double_as_float128): New function.
20893 (modified_type_die): For powerpc64le IEEE 754 quad long double
20894 and complex long double emit those as DW_TAG_typedef to
20895 _Float128 or complex _Float128 base type.
20897 2022-01-26 Marek Polacek <polacek@redhat.com>
20900 * gimple-ssa-warn-access.cc (pass_waccess::warn_invalid_pointer): Don't
20901 warn when the SSA_NAME_VAR of REF has supressed -Wuse-after-free.
20903 2022-01-26 Martin Liska <mliska@suse.cz>
20905 * ipa-modref-tree.cc (modref_access_node::update):
20906 Remove "--param param=foo" with "--param foo".
20907 (modref_access_node::insert): Likewise.
20908 (modref_access_node::insert_kill): Likewise.
20909 * ipa-modref-tree.h (struct modref_ref_node): Likewise.
20910 (struct modref_base_node): Likewise.
20911 (struct modref_tree): Likewise.
20913 2022-01-26 Raoni Fassina Firmino <raoni@linux.ibm.com>
20916 * builtins.cc (expand_builtin_feclear_feraise_except): Add op0
20919 2022-01-25 Martin Sebor <msebor@redhat.com>
20921 PR tree-optimization/104203
20922 * gimple-ssa-warn-access.cc (pass_data pass_data_waccess): Use
20924 * pointer-query.cc (access_ref::merge_ref): Change return type.
20925 Convert failure to a conservative success.
20926 (access_ref::get_ref): Adjust to the change above. Short-circuit
20927 PHI evaluation after first failure turned into conservative success.
20928 * pointer-query.h (access_ref::merge_ref): Change return type.
20929 * timevar.def (TV_WARN_ACCESS): New timer variable.
20931 2022-01-25 David Edelsohn <dje.gcc@gmail.com>
20933 * config/rs6000/aix.h (OPTION_GLIBC): Define as 0.
20935 2022-01-25 Richard Biener <rguenther@suse.de>
20937 PR tree-optimization/104214
20938 * tree-ssa-loop-niter.cc (number_of_iterations_cond): Use
20939 stronger guarantees for relational pointer compares when
20940 rewriting BASE0 + STEP0 cmp BASE1 + STEP1 as
20941 BASE0 + STEP0 - STEP1 cmp BASE1.
20943 2022-01-25 Jakub Jelinek <jakub@redhat.com>
20946 * config/rs6000/rs6000-internal.h (rs6000_passes_ieee128): Don't
20948 * config/rs6000/rs6000.cc (rs6000_passes_ieee128,
20949 ieee128_mangling_gcc_8_1): Remove.
20950 (TARGET_ASM_GLOBALIZE_DECL_NAME): Don't redefine.
20951 (rs6000_mangle_type): Return "u9__ieee128" instead of
20952 ieee128_mangling_gcc_8_1 ? "U10__float128" : "u9__ieee128".
20953 (rs6000_globalize_decl_name): Remove.
20954 * config/rs6000/rs6000-call.cc (init_cumulative_args,
20955 rs6000_function_arg_advance_1): Don't set rs6000_passes_ieee128.
20957 2022-01-24 Martin Sebor <msebor@redhat.com>
20959 * pointer-query.cc (pointer_query::dump): Remove duplicate
20962 2022-01-24 Marek Polacek <polacek@redhat.com>
20964 PR preprocessor/104030
20965 * doc/invoke.texi: Update documentation for -Wbidi-chars.
20967 2022-01-24 Raoni Fassina Firmino <raoni@linux.ibm.com>
20970 * builtins.cc (expand_builtin_fegetround): New function.
20971 (expand_builtin_feclear_feraise_except): New function.
20972 (expand_builtin): Add cases for BUILT_IN_FEGETROUND,
20973 BUILT_IN_FECLEAREXCEPT and BUILT_IN_FERAISEEXCEPT.
20974 * config/rs6000/rs6000.md (fegetroundsi): New pattern.
20975 (feclearexceptsi): New Pattern.
20976 (feraiseexceptsi): New Pattern.
20977 * doc/extend.texi: Add a new introductory paragraph about the
20979 * doc/md.texi: (fegetround@var{m}): Document new optab.
20980 (feclearexcept@var{m}): Document new optab.
20981 (feraiseexcept@var{m}): Document new optab.
20982 * optabs.def (fegetround_optab): New optab.
20983 (feclearexcept_optab): New optab.
20984 (feraiseexcept_optab): New optab.
20986 2022-01-24 Richard Biener <rguenther@suse.de>
20987 Jiufu Guo <guojiufu@linux.ibm.com>
20989 PR tree-optimization/100740
20990 PR tree-optimization/101508
20991 PR tree-optimization/101972
20992 PR tree-optimization/102131
20993 * tree-ssa-loop-niter.cc (number_of_iterations_cond): Properly
20994 constrain BASE0 + STEP0 cmp BASE1 + STEP1 to
20995 BASE0 + STEP0 - STEP1 cmp BASE1 transform.
20997 2022-01-24 Jakub Jelinek <jakub@redhat.com>
20999 PR sanitizer/104158
21000 * opt-functions.awk (var_set): Handle EnumBitSet property.
21001 * optc-gen.awk: Don't disallow RejectNegative if EnumBitSet is
21003 * opts.h (enum cl_enum_var_value): New type.
21004 * opts-common.cc (decode_cmdline_option): Use CLEV_* values.
21005 Handle CLEV_BITSET.
21006 (cmdline_handle_error): Handle CLEV_BITSET.
21007 * opts.cc (test_enum_sets): Also test EnumBitSet requirements.
21008 * doc/options.texi (EnumBitSet): Document.
21009 * common.opt (fsanitize-coverage=): Use EnumBitSet instead of
21011 (trace-pc, trace-cmp): Drop Set properties.
21013 2022-01-24 Jakub Jelinek <jakub@redhat.com>
21015 PR sanitizer/104158
21016 * common.opt (flag_sanitize_coverage): Remove Variable entry.
21017 (fsanitize-coverage=): Remove RejectNegative property, add
21018 Var(flag_sanitize_coverage) and EnumSet properties.
21019 (trace-pc): Add Set(1) property.
21020 (trace-cmp): Add Set(2) property.
21021 * opts.cc (common_handle_option): Don't handle
21022 OPT_fsanitize_coverage_.
21024 2022-01-24 Jakub Jelinek <jakub@redhat.com>
21026 PR sanitizer/104158
21027 * opt-functions.awk (var_set): Handle EnumSet property.
21028 * optc-gen.awk: Don't disallow RejectNegative if EnumSet is
21030 * opt-read.awk: Handle Set property.
21031 * opts.h (CL_ENUM_SET_SHIFT, CL_ERR_ENUM_SET_ARG): Define.
21032 (struct cl_decoded_option): Mention enum in value description.
21034 (set_option): Add mask argument defaulted to 0.
21035 * opts.cc (test_enum_sets): New function.
21036 (opts_cc_tests): Call it.
21037 * opts-common.cc (enum_arg_to_value): Change return argument
21038 from bool to int, on success return index into the cl_enum_arg
21039 array, on failure -1. Add len argument, if non-0, use strncmp
21041 (opt_enum_arg_to_value): Adjust caller.
21042 (decode_cmdline_option): Handle EnumSet represented as
21043 CLVC_ENUM with non-zero var_value. Initialize decoded->mask.
21044 (decode_cmdline_options_to_array): CLear opt_array[0].mask.
21045 (handle_option): Pass decoded->mask to set_options last argument.
21046 (generate_option): Clear decoded->mask.
21047 (generate_option_input_file): Likewise.
21048 (cmdline_handle_error): Handle CL_ERR_ENUM_SET_ARG.
21049 (set_option): Add mask argument, use it for CLVC_ENUM.
21050 (control_warning_option): Adjust enum_arg_to_value caller.
21051 * doc/options.texi: Document Set and EnumSet properties.
21053 2022-01-24 Jakub Jelinek <jakub@redhat.com>
21055 PR bootstrap/104170
21056 * config/linux.h (OPTION_GLIBC_P, OPTION_UCLIBC_P,
21057 OPTION_BIONIC_P, OPTION_MUSL_P): Define.
21058 (OPTION_GLIBC, OPTION_UCLIBC, OPTION_BIONIC, OPTION_MUSL): Redefine
21059 using OPTION_*_P macros.
21060 * config/alpha/linux.h (OPTION_GLIBC_P, OPTION_UCLIBC_P,
21061 OPTION_BIONIC_P, OPTION_MUSL_P): Define.
21062 (OPTION_GLIBC, OPTION_UCLIBC, OPTION_BIONIC, OPTION_MUSL): Redefine
21063 using OPTION_*_P macros.
21064 * config/rs6000/linux.h (OPTION_GLIBC_P, OPTION_UCLIBC_P,
21065 OPTION_BIONIC_P, OPTION_MUSL_P): Define.
21066 (OPTION_GLIBC, OPTION_UCLIBC, OPTION_BIONIC, OPTION_MUSL): Redefine
21067 using OPTION_*_P macros.
21068 * config/rs6000/linux64.h (OPTION_GLIBC_P, OPTION_UCLIBC_P,
21069 OPTION_BIONIC_P, OPTION_MUSL_P): Define.
21070 (OPTION_GLIBC, OPTION_UCLIBC, OPTION_BIONIC, OPTION_MUSL): Redefine
21071 using OPTION_*_P macros.
21072 * config/fuchsia.h (OPTION_MUSL_P): Redefine.
21073 * config/glibc-stdint.h (OPTION_MUSL_P): Define if not defined.
21074 * common/config/s390/s390-common.cc (s390_supports_split_stack): Re-add
21075 ATTRIBUTE_UNUSED to opts parameter. If OPTION_GLIBC_P is defined, use
21076 OPTION_GLIBC_P (opts) as condition, otherwise assume if (false).
21077 * common/config/i386/i386-common.cc (ix86_supports_split_stack): If
21078 OPTION_GLIBC_P is defined use !OPTION_GLIBC_P (opts) as condition,
21079 otherwise assume if (true).
21081 2022-01-24 Kito Cheng <kito.cheng@sifive.com>
21083 * common/config/riscv/riscv-common.cc (riscv_subset_list::to_string):
21084 Skip zicsr and zifencei if I-ext is 2.0.
21086 2022-01-24 Jia-Wei Chen <jiawei@iscas.ac.cn>
21088 * config.gcc: Modify default isa_spec version.
21090 2022-01-24 Jiufu Guo <guojiufu@linux.ibm.com>
21092 PR tree-optimization/102087
21093 * tree-ssa-loop-niter.cc (number_of_iterations_until_wrap):
21094 Correct PLUS result type.
21096 2022-01-24 H.J. Lu <hjl.tools@gmail.com>
21099 * config/i386/predicates.md (bcst_mem_operand): Also check mode
21100 of memory broadcast.
21102 2022-01-23 Andrew Pinski <apinski@marvell.com>
21105 * config/aarch64/aarch64-builtins.cc
21106 (aarch64_general_gimple_fold_builtin): Handle
21107 __builtin_aarch64_sqrt* and simplify into SQRT internal
21110 2022-01-22 Jakub Jelinek <jakub@redhat.com>
21113 * opts-global.cc (handle_common_deferred_options): Quote
21114 --enable-plugin in diagnostics to avoid -Werror=format-diag.
21116 2022-01-21 Michael Meissner <meissner@the-meissners.org>
21119 * config/rs6000/rs6000-protos.h (prefixed_xxsplti_p): Delete.
21120 * config/rs6000/rs6000.cc (prefixed_xxsplti_p): Delete.
21121 * config/rs6000/rs6000.md (prefixed attribute): Delete section
21122 that sets the prefixed attribute for xxspltiw, xxspltidp, and
21123 xxsplti32dx instructions.
21124 (movsf_hardfloat): Explicitly set the prefixed attribute
21125 when xxspltiw and xxspltidp instructions are generated.
21126 (mov<mode>_hardfloat32): Likewise.
21127 (mov<mode>_hardfloat64): Likewise.
21128 * config/rs6000/vsx.md (vsx_mov<mode>_64bit): Explicitly set the
21129 prefixed attribute for xxspltiw and xxspltidp instructions.
21130 (vsx_mov<mode>_32bit): Likewise.
21132 2022-01-21 H.J. Lu <hjl.tools@gmail.com>
21134 PR bootstrap/104170
21135 * common/config/i386/i386-common.cc (ix86_supports_split_stack):
21136 Return true only on glibc.
21137 * config/i386/gnu-user-common.h (STACK_CHECK_STATIC_BUILTIN):
21138 Revert commit c163647ffbc.
21139 * config/i386/gnu.h (TARGET_LIBC_PROVIDES_SSP): Likewise.
21141 2022-01-21 Sören Tempel <soeren@soeren-tempel.net>
21143 * common/config/s390/s390-common.cc (s390_supports_split_stack):
21144 Only support split-stack on glibc targets.
21145 * config/i386/gnu-user-common.h (STACK_CHECK_STATIC_BUILTIN): Ditto.
21146 * config/i386/gnu.h (defined): Ditto.
21148 2022-01-21 Bill Schmidt <wschmidt@linux.ibm.com>
21150 * config/rs6000/rs6000-overload.def (VEC_SLDW): Add instances for
21151 vector float and vector double.
21153 2022-01-21 Bill Seurer <seurer@gcc.gnu.org>
21155 * config/rs6000/rs6000.cc (rs6000_get_function_versions_dispatcher):
21156 Fix mention of ifunc in string.
21158 2022-01-21 Roger Sayle <roger@nextmovesoftware.com>
21160 PR middle-end/104140
21161 * tree-ssa-math-opts.cc (convert_mult_to_highpart): Check that the
21162 operands of the widening multiplication are either both signed or
21163 both unsigned, and abort the conversion if mismatched.
21164 * doc/generic.texi (WIDEN_MULT_EXPR): Describe expression node.
21165 (MULT_HIGHPART_EXPR): Clarify that operands must have the same
21167 * tree.def (MULT_HIGHPART_EXPR): Document both operands must have
21168 integer types with the same precision and signedness.
21169 (WIDEN_MULT_EXPR): Document that operands must have integer types
21170 with the same precision, but possibly differing signedness.
21171 * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins): Defend against
21172 riscv_current_subset_list returning a NULL pointer (empty list).
21174 2022-01-21 Vladimir N. Makarov <vmakarov@redhat.com>
21177 * ira.h (struct target_ira): Add member
21178 x_ira_exclude_class_mode_regs.
21179 (ira_exclude_class_mode_regs): New macro.
21180 * lra.h (lra_create_new_reg): Add arg exclude_start_hard_regs and
21182 * lra-int.h: ... to here.
21183 (lra_create_new_reg_with_unique_value): Add arg
21184 exclude_start_hard_regs.
21185 (class lra_reg): Add member exclude_start_hard_regs.
21186 * lra-assigns.cc (find_hard_regno_for_1): Setup
21187 impossible_start_hard_regs from exclude_start_hard_regs.
21188 * lra-constraints.cc (get_reload_reg): Add arg exclude_start_hard_regs and pass
21189 it lra_create_new_reg[_with_unique_value].
21190 (match_reload): Ditto.
21191 (check_and_process_move): Pass NULL
21192 exclude_start_hard_regs to lra_create_new_reg_with_unique_value.
21193 (goal_alt_exclude_start_hard_regs): New static variable.
21194 (process_addr_reg, simplify_operand_subreg): Pass NULL
21195 exclude_start_hard_regs to lra_create_new_reg_with_unique_value
21196 and get_reload_reg.
21197 (process_alt_operands): Setup goal_alt_exclude_start_hard_regs.
21198 Use this_alternative_exclude_start_hard_regs additionally to find
21199 winning operand alternative.
21200 (base_to_reg, base_plus_disp_to_reg, index_part_to_reg): Pass NULL
21201 exclude_start_hard_regs to lra_create_new_reg.
21202 (process_address_1, emit_inc): Ditto.
21203 (curr_insn_transform): Pass exclude_start_hard_regs value to
21204 lra_create_new_reg, get_reload_reg, match_reload.
21205 (inherit_reload_reg, split_reg): Pass NULL exclude_start_hard_regs
21206 to lra_create_new_reg.
21207 (process_invariant_for_inheritance): Ditto.
21208 * lra-remat.cc (update_scratch_ops): Ditto.
21209 * lra.cc (lra_create_new_reg_with_unique_value): Add arg
21210 exclude_start_hard_regs. Setup the corresponding member of
21212 (lra_create_new_reg): Add arg exclude_start_hard_regs and pass it
21213 to lra_create_new_reg_with_unique_value.
21214 (initialize_lra_reg_info_element): Initialize member
21215 exclude_start_hard_regs.
21216 (get_scratch_reg): Pass NULL to lra_create_new_reg.
21217 * ira.cc (setup_prohibited_class_mode_regs): Rename to
21218 setup_prohibited_and_exclude_class_mode_regs and calculate
21219 ira_exclude_class_mode_regs.
21221 2022-01-21 Martin Liska <mliska@suse.cz>
21223 * configure.ac: Detect ld_is_mold and use it for
21224 comdat_group=yes and gcc_cv_ld_hidden=yes.
21225 * configure: Regenerate.
21227 2022-01-21 Richard Biener <rguenther@suse.de>
21229 PR tree-optimization/100089
21230 * tree-vect-slp.cc (vect_slp_region): Reject BB vectorization
21231 of if-converted loops with unvectorized COND_EXPRs for
21232 all but the unlimited cost models.
21234 2022-01-21 Ard Biesheuvel <ardb@kernel.org>
21236 * config/arm/arm-opts.h (enum stack_protector_guard): New.
21237 * config/arm/arm-protos.h (arm_stack_protect_tls_canary_mem):
21239 * config/arm/arm.cc (TARGET_STACK_PROTECT_GUARD): Define.
21240 (arm_option_override_internal): Handle and put in error checks.
21241 for stack protector guard options.
21242 (arm_option_reconfigure_globals): Likewise.
21243 (arm_stack_protect_tls_canary_mem): New.
21244 (arm_stack_protect_guard): New.
21245 * config/arm/arm.md (stack_protect_set): New.
21246 (stack_protect_set_tls): Likewise.
21247 (stack_protect_test): Likewise.
21248 (stack_protect_test_tls): Likewise.
21249 (reload_tp_hard): Likewise.
21250 * config/arm/arm.opt (-mstack-protector-guard): New
21251 (-mstack-protector-guard-offset): New.
21252 * doc/invoke.texi: Document new options.
21254 2022-01-21 Richard Biener <rguenther@suse.de>
21256 PR tree-optimization/104156
21257 * tree-ssa-loop-unswitch.cc (tree_unswitch_outer_loop):
21258 Collect and reset debug stmts with out-of-loop uses when
21260 (find_loop_guard): Adjust.
21261 (empty_bb_without_guard_p): Likewise. Ignore debug stmts.
21262 (used_outside_loop_p): Push debug uses to a vector of
21263 debug stmts to reset.
21264 (hoist_guard): Adjust -fopt-info category.
21266 2022-01-21 Richard Biener <rguenther@suse.de>
21268 PR tree-optimization/104152
21269 * tree-vect-slp.cc (vect_build_slp_tree_2): Add missing
21270 can_duplicate_and_interleave_p check.
21272 2022-01-21 Jakub Jelinek <jakub@redhat.com>
21274 * gimple-ssa-warn-access.cc (pass_waccess::warn_invalid_pointer):
21275 Avoid passing var to warning_at when the format string doesn't
21278 2022-01-21 Aldy Hernandez <aldyh@redhat.com>
21280 PR tree-optimization/103721
21281 * gimple-range-path.cc
21282 (path_range_query::relations_may_be_invalidated): New.
21283 (path_range_query::compute_ranges_in_block): Reset relations if
21284 they may be invalidated.
21285 (path_range_query::maybe_register_phi_relation): Exit if relations
21286 may be invalidated on incoming edge.
21287 (path_range_query::compute_phi_relations): Pass incoming PHI edge
21288 to maybe_register_phi_relation.
21289 * gimple-range-path.h (relations_may_be_invalidated): New.
21290 (maybe_register_phi_relation): Pass edge instead of tree.
21291 * tree-ssa-threadbackward.cc (back_threader::back_threader):
21293 * value-relation.cc (path_oracle::path_oracle): Call
21294 mark_dfs_back_edges.
21295 (path_oracle::register_relation): Add SSA names to m_registered
21297 (path_oracle::reset_path): Clear m_registered bitmap.
21298 * value-relation.h (path_oracle::set_root_oracle): New.
21300 2022-01-21 Jakub Jelinek <jakub@redhat.com>
21302 PR rtl-optimization/102478
21303 * optabs.cc (prepare_cmp_insn): If !can_create_pseudo_p (), don't
21304 force_reg constants and for -fnon-call-exceptions fail if copy_to_reg
21307 2022-01-20 Richard Biener <rguenther@suse.de>
21309 PR middle-end/100786
21310 * gimple-fold.cc (get_symbol_constant_value): Only return
21311 values of compatible type to the symbol.
21313 2022-01-20 Andrew MacLeod <amacleod@redhat.com>
21315 * value-relation.cc (relation_oracle::valid_equivs): Query and add
21316 if valid members of a set.
21317 (equiv_oracle::register_equiv): Call valid_equivs rather than
21318 bitmap direct operations.
21319 (path_oracle::register_equiv): Ditto.
21320 * value-relation.h (relation_oracle::valid_equivs): New prototype.
21322 2022-01-20 Richard Biener <rguenther@suse.de>
21325 * config/i386/i386.cc (ix86_gimple_fold_builtin): Check for
21326 LHS before folding __builtin_ia32_shufpd and friends.
21328 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21330 * config/arm/crypto.md (aes_op_protect): Allow moves from core
21331 registers and from memory.
21332 (aes_op_protect_misalign_load): New pattern.
21333 (aes_op_protect_neon_vld1v16qi): New pattern.
21335 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21337 * config/arm/crypto.md (crypto_<CRYPTO_AESMC:crypto_pattern>_protected):
21339 (aarch32_crypto_aese_fused_protected): Likewise.
21340 (aarch32_crypto_aesd_fused_protected): Likewise.
21342 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21344 * config/arm/crypto.md (crypto_<CRYPTO_AES:crypto_pattern>): Convert
21345 to define_expand. Add mitigation for the Cortex-A AES erratum
21347 (*crypto_<CRYPTO_AES:crypto_pattern>_insn): New pattern, based
21348 on original crypto_<CRYPTO_AES:crypto_pattern> insn.
21349 (aes_op_protect): New pattern.
21350 * config/arm/unspecs.md (unspec): Add UNSPEC_AES_PROTECT.
21352 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21354 * config/arm/arm-cpus.in (quirk_aes_1742098): New quirk feature
21355 (ALL_QUIRKS): Add it.
21356 (cortex-a57, cortex-a72): Enable it.
21357 (cortex-a57.cortex-a53, cortex-a72.cortex-a53): Likewise.
21358 * config/arm/arm.opt (mfix-cortex-a57-aes-1742098): New command-line
21360 (mfix-cortex-a72-aes-1655431): New option alias.
21361 * config/arm/arm.cc (arm_option_override): Handle default settings
21362 for AES erratum switch.
21363 * doc/invoke.texi (Arm Options): Document new options.
21365 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21367 * config/arm/crypto.md (crypto_<CYRPTO_AES:crypto_pattern>): Use
21368 <crypto_mode> rather than hard-coding the mode.
21369 (crypto_<CRYPTO_AESMC:crypto_pattern>): Fix white space.
21370 (crypto_<CRYPTO_AES:crypto_pattern>): Likewise.
21371 (*aarch32_crypto_aese_fused): Likewise.
21372 (*aarch32_crypto_aesd_fused): Likewise.
21373 (crypto_<CRYPTO_BINARY:crypto_pattern>): Likewise.
21374 (crypto_<CRYPTO_TERNARY:crypto_pattern>): Likewise.
21375 (crypto_sha1h_lb): Likewise.
21376 (crypto_vmullp64): Likewise.
21377 (crypto_<CRYPTO_SELECTING:crypto_pattern>): Likewise.
21378 (crypto_<CRYPTO_SELECTING:crypto_pattern>_lb): Likewise.
21380 2022-01-20 Richard Earnshaw <rearnsha@arm.com>
21382 * config/arm/crypto.md (crypto_<CRYPTO_AESMC:crypto_pattern>): Add
21383 iterator to pattern name to disambiguate.
21384 (crypto_<CRYPTO_AES:crypto_pattern>): Likewise.
21385 (crypto_<CRYPTO_BINARY:crypto_pattern>): Likewise.
21386 (crypto_<CRYPTO_TERNARY:crypto_pattern>): Likewise.
21387 (crypto_<CRYPTO_SELECTING:crypto_pattern>): Likewise.
21388 (crypto_<CRYPTO_SELECTING:crypto_pattern>_lb): Likewise.
21390 2022-01-20 Martin Liska <mliska@suse.cz>
21392 PR bootstrap/104135
21393 * emit-rtl.cc (make_insn_raw): Fix -Wformat-diag warnings.
21394 * rtl.cc: Partially disable -Wformat-diag for RTL checking
21397 2022-01-20 Jakub Jelinek <jakub@redhat.com>
21400 * dwarf2out.cc (index_rnglists): For !HAVE_AS_LEB128 and
21401 block_num > 0, index entry even if !have_multiple_function_sections.
21403 2022-01-20 liuhongt <hongtao.liu@intel.com>
21406 * tree-vect-stmts.cc (supportable_narrowing_operation): Enhance
21407 integral mode mask pack by multi steps which takes
21408 vec_pack_sbool_trunc_optab as start when elements number is
21409 less than BITS_PER_UNITS.
21411 2022-01-20 Richard Biener <rguenther@suse.de>
21413 PR tree-optimization/104114
21414 * tree-vect-generic.cc (expand_vector_piecewise): Do not diagnose
21415 single element vector decomposition.
21417 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21419 * ifcvt.cc (noce_convert_multiple_sets_1): New function.
21420 (noce_convert_multiple_sets): Call function a second time if we can
21421 improve the first try.
21423 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21425 * ifcvt.cc (cond_exec_get_condition): New parameter to allow getting the
21426 reversed comparison.
21427 (try_emit_cmove_seq): New function to facilitate creating a cmov
21429 (noce_convert_multiple_sets): Create two sequences and use the less
21432 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21434 * rtl.h (struct rtx_comparison): New struct that holds an rtx
21436 * config/rs6000/rs6000.cc (rs6000_emit_minmax): Use struct instead of
21438 (rs6000_emit_swsqrt): Likewise.
21439 * expmed.cc (expand_sdiv_pow2): Likewise.
21440 (emit_store_flag): Likewise.
21441 * expr.cc (expand_cond_expr_using_cmove): Likewise.
21442 (expand_expr_real_2): Likewise.
21443 * ifcvt.cc (noce_emit_cmove): Add compare and reversed compare
21445 * optabs.cc (emit_conditional_move_1): New function.
21446 (expand_doubleword_shift_condmove): Use struct.
21447 (emit_conditional_move): Use struct and allow to call directly
21448 without going through preparation steps.
21449 * optabs.h (emit_conditional_move): Use struct.
21451 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21453 * ifcvt.cc (bb_ok_for_noce_convert_multiple_sets): Estimate insns costs.
21454 (noce_process_if_block): Use potential costs.
21456 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21458 * ifcvt.cc (noce_convert_multiple_sets): Allow constants.
21459 (bb_ok_for_noce_convert_multiple_sets): Likewise.
21461 2022-01-19 Robin Dapp <rdapp@linux.ibm.com>
21463 * ifcvt.cc (need_cmov_or_rewire): New function.
21464 (noce_convert_multiple_sets): Call it.
21466 2022-01-19 David Malcolm <dmalcolm@redhat.com>
21468 * attribs.cc (attribute_c_tests): Rename to...
21469 (attribs_cc_tests): ...this.
21470 * bitmap.cc (bitmap_c_tests): Rename to...
21471 (bitmap_cc_tests): ...this.
21472 * cgraph.cc (cgraph_c_finalize): Rename to...
21473 (cgraph_cc_finalize): ...this.
21474 (cgraph_c_tests): Rename to...
21475 (cgraph_cc_tests): ...this.
21476 * cgraph.h (cgraph_c_finalize): Rename to...
21477 (cgraph_cc_finalize): ...this.
21478 (cgraphunit_c_finalize): Rename to...
21479 (cgraphunit_cc_finalize): ...this.
21480 * cgraphunit.cc (cgraphunit_c_finalize): Rename to...
21481 (cgraphunit_cc_finalize): ...this.
21482 * convert.cc (convert_c_tests): Rename to...
21483 (convert_cc_tests): ...this.
21484 * dbgcnt.cc (dbgcnt_c_tests): Rename to...
21485 (dbgcnt_cc_tests): ...this.
21486 * diagnostic-show-locus.cc (diagnostic_show_locus_c_tests): Rename to...
21487 (diagnostic_show_locus_cc_tests): ...this.
21488 * diagnostic.cc (diagnostic_c_tests): Rename to...
21489 (diagnostic_cc_tests): ...this.
21490 * dumpfile.cc (dumpfile_c_tests): Rename to...
21491 (dumpfile_cc_tests): ...this.
21492 * dwarf2out.cc (dwarf2out_c_finalize): Rename to...
21493 (dwarf2out_cc_finalize): ...this.
21494 * dwarf2out.h (dwarf2out_c_finalize): Rename to...
21495 (dwarf2out_cc_finalize): ...this.
21496 * edit-context.cc (edit_context_c_tests): Rename to...
21497 (edit_context_cc_tests): ...this.
21498 * et-forest.cc (et_forest_c_tests): Rename to...
21499 (et_forest_cc_tests): ...this.
21500 * fibonacci_heap.cc (fibonacci_heap_c_tests): Rename to...
21501 (fibonacci_heap_cc_tests): ...this.
21502 * fold-const.cc (fold_const_c_tests): Rename to...
21503 (fold_const_cc_tests): ...this.
21504 * function-tests.cc (function_tests_c_tests): Rename to...
21505 (function_tests_cc_tests): ...this.
21506 * gcse.cc (gcse_c_finalize): Rename to...
21507 (gcse_cc_finalize): ...this.
21508 * gcse.h (gcse_c_finalize): Rename to...
21509 (gcse_cc_finalize): ...this.
21510 * ggc-tests.cc (ggc_tests_c_tests): Rename to...
21511 (ggc_tests_cc_tests): ...this.
21512 * gimple-ssa-store-merging.cc (store_merging_c_tests): Rename to...
21513 (store_merging_cc_tests): ...this.
21514 * gimple.cc (gimple_c_tests): Rename to...
21515 (gimple_cc_tests): ...this.
21516 * hash-map-tests.cc (hash_map_tests_c_tests): Rename to...
21517 (hash_map_tests_cc_tests): ...this.
21518 * hash-set-tests.cc (hash_set_tests_c_tests): Rename to...
21519 (hash_set_tests_cc_tests): ...this.
21520 * input.cc (input_c_tests): Rename to...
21521 (input_cc_tests): ...this.
21522 * ipa-cp.cc (ipa_cp_c_finalize): Rename to...
21523 (ipa_cp_cc_finalize): ...this.
21524 * ipa-fnsummary.cc (ipa_fnsummary_c_finalize): Rename to...
21525 (ipa_fnsummary_cc_finalize): ...this.
21526 * ipa-fnsummary.h (ipa_fnsummary_c_finalize): Rename to...
21527 (ipa_fnsummary_cc_finalize): ...this.
21528 * ipa-modref-tree.cc (ipa_modref_tree_c_tests): Rename to...
21529 (ipa_modref_tree_cc_tests): ...this.
21530 * ipa-modref-tree.h (modref_c_tests): Delete bogus decl.
21531 * ipa-modref.cc (ipa_modref_c_finalize): Rename to...
21532 (ipa_modref_cc_finalize): ...this.
21533 * ipa-modref.h (ipa_modref_c_finalize): Rename to...
21534 (ipa_modref_cc_finalize): ...this.
21535 * ipa-prop.h (ipa_cp_c_finalize): Rename to...
21536 (ipa_cp_cc_finalize): ...this.
21537 * ipa-reference.cc (ipa_reference_c_finalize): Rename to...
21538 (ipa_reference_cc_finalize): ...this.
21539 * ipa-reference.h (ipa_reference_c_finalize): Rename to...
21540 (ipa_reference_cc_finalize): ...this.
21541 * ira-costs.cc (ira_costs_c_finalize): Rename to...
21542 (ira_costs_cc_finalize): ...this.
21543 * ira.h (ira_costs_c_finalize): Rename to...
21544 (ira_costs_cc_finalize): ...this.
21545 * opt-suggestions.cc (opt_proposer_c_tests): Rename to...
21546 (opt_suggestions_cc_tests): ...this.
21547 * opts.cc (opts_c_tests): Rename to...
21548 (opts_cc_tests): ...this.
21549 * predict.cc (predict_c_tests): Rename to...
21550 (predict_cc_tests): ...this.
21551 * pretty-print.cc (pretty_print_c_tests): Rename to...
21552 (pretty_print_cc_tests): ...this.
21553 * read-rtl-function.cc (read_rtl_function_c_tests): Rename to...
21554 (read_rtl_function_cc_tests): ...this.
21555 * rtl-tests.cc (rtl_tests_c_tests): Rename to...
21556 (rtl_tests_cc_tests): ...this.
21557 * sbitmap.cc (sbitmap_c_tests): Rename to...
21558 (sbitmap_cc_tests): ...this.
21559 * selftest-run-tests.cc (selftest::run_tests): Update calls for
21560 _c_ to _cc_ function renamings; fix names of attribs and
21561 opt-suggestions tests.
21562 * selftest.cc (selftest_c_tests): Rename to...
21563 (selftest_cc_tests): ...this.
21564 * selftest.h (attribute_c_tests): Rename to...
21565 (attribs_cc_tests): ...this.
21566 (bitmap_c_tests): Rename to...
21567 (bitmap_cc_tests): ...this.
21568 (cgraph_c_tests): Rename to...
21569 (cgraph_cc_tests): ...this.
21570 (convert_c_tests): Rename to...
21571 (convert_cc_tests): ...this.
21572 (diagnostic_c_tests): Rename to...
21573 (diagnostic_cc_tests): ...this.
21574 (diagnostic_show_locus_c_tests): Rename to...
21575 (diagnostic_show_locus_cc_tests): ...this.
21576 (dumpfile_c_tests): Rename to...
21577 (dumpfile_cc_tests): ...this.
21578 (edit_context_c_tests): Rename to...
21579 (edit_context_cc_tests): ...this.
21580 (et_forest_c_tests): Rename to...
21581 (et_forest_cc_tests): ...this.
21582 (fibonacci_heap_c_tests): Rename to...
21583 (fibonacci_heap_cc_tests): ...this.
21584 (fold_const_c_tests): Rename to...
21585 (fold_const_cc_tests): ...this.
21586 (function_tests_c_tests): Rename to...
21587 (function_tests_cc_tests): ...this.
21588 (ggc_tests_c_tests): Rename to...
21589 (ggc_tests_cc_tests): ...this.
21590 (gimple_c_tests): Rename to...
21591 (gimple_cc_tests): ...this.
21592 (hash_map_tests_c_tests): Rename to...
21593 (hash_map_tests_cc_tests): ...this.
21594 (hash_set_tests_c_tests): Rename to...
21595 (hash_set_tests_cc_tests): ...this.
21596 (input_c_tests): Rename to...
21597 (input_cc_tests): ...this.
21598 (opts_c_tests): Rename to...
21599 (opts_cc_tests): ...this.
21600 (predict_c_tests): Rename to...
21601 (predict_cc_tests): ...this.
21602 (pretty_print_c_tests): Rename to...
21603 (pretty_print_cc_tests): ...this.
21604 (read_rtl_function_c_tests): Rename to...
21605 (read_rtl_function_cc_tests): ...this.
21606 (rtl_tests_c_tests): Rename to...
21607 (rtl_tests_cc_tests): ...this.
21608 (sbitmap_c_tests): Rename to...
21609 (sbitmap_cc_tests): ...this.
21610 (selftest_c_tests): Rename to...
21611 (selftest_cc_tests): ...this.
21612 (simplify_rtx_c_tests): Rename to...
21613 (simplify_rtx_cc_tests): ...this.
21614 (spellcheck_c_tests): Rename to...
21615 (spellcheck_cc_tests): ...this.
21616 (spellcheck_tree_c_tests): Rename to...
21617 (spellcheck_tree_cc_tests): ...this.
21618 (sreal_c_tests): Rename to...
21619 (sreal_cc_tests): ...this.
21620 (store_merging_c_tests): Rename to...
21621 (store_merging_cc_tests): ...this.
21622 (tree_c_tests): Rename to...
21623 (tree_cc_tests): ...this.
21624 (tree_cfg_c_tests): Rename to...
21625 (tree_cfg_cc_tests): ...this.
21626 (typed_splay_tree_c_tests): Rename to...
21627 (typed_splay_tree_cc_tests): ...this.
21628 (vec_c_tests): Rename to...
21629 (vec_cc_tests): ...this.
21630 (vec_perm_indices_c_tests): Rename to...
21631 (vec_perm_indices_cc_tests): ..this.
21632 (opt_proposer_c_tests): Rename to...
21633 (opt_suggestions_cc_tests): ...this.
21634 (dbgcnt_c_tests): Rename to...
21635 (dbgcnt_cc_tests): ...this.
21636 (ipa_modref_tree_c_tests): Rename to...
21637 (ipa_modref_tree_cc_tests): ...this.
21638 * simplify-rtx.cc (simplify_rtx_c_tests): Rename to...
21639 (simplify_rtx_cc_tests): ...this.
21640 * spellcheck-tree.cc (spellcheck_tree_c_tests): Rename to...
21641 (spellcheck_tree_cc_tests): ...this.
21642 * spellcheck.cc (spellcheck_c_tests): Rename to...
21643 (spellcheck_cc_tests): ...this.
21644 * sreal.cc (sreal_c_tests): Rename to...
21645 (sreal_cc_tests): ...this.
21646 * toplev.cc (toplev::finalize): Update calls for _c_ to _cc_
21647 function renamings.
21648 * tree-cfg.cc (tree_cfg_c_tests): Rename to...
21649 (tree_cfg_cc_tests): ...this.
21650 * tree.cc (tree_c_tests): Rename to...
21651 (tree_cc_tests): ...this.
21652 * typed-splay-tree.cc (typed_splay_tree_c_tests): Rename to...
21653 (typed_splay_tree_cc_tests): ...this.
21654 * vec-perm-indices.cc (vec_perm_indices_c_tests): Rename to...
21655 (vec_perm_indices_cc_tests): ...this.
21656 * vec.cc (vec_c_tests): Rename to...
21657 (vec_cc_tests): ...this.
21659 2022-01-19 Andre Vieira <andre.simoesdiasvieira@arm.com>
21661 PR tree-optimization/103997
21662 * tree-vect-loop.cc (vect_analyze_loop): Fix mode skipping for epilogue
21665 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21667 PR middle-end/102860
21668 * match.pd (x %[fl] y -> x % y): New simplification for
21669 unsigned integral types.
21670 * optabs-tree.cc (optab_for_tree_code): Return unknown_optab
21671 for {CEIL,FLOOR,ROUND}_{DIV,MOD}_EXPR with VECTOR_TYPE.
21673 2022-01-19 Richard Biener <rguenther@suse.de>
21675 PR tree-optimization/104112
21676 * tree-vect-loop.cc (vect_find_reusable_accumulator): Check
21677 for required intermediate vector types.
21679 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21681 * config/rs6000/rs6000.cc (rs6000_machine_from_flags): Add default:.
21683 2022-01-19 Martin Liska <mliska@suse.cz>
21685 * configure.ac: Remove -Wno-error=format-diag.
21686 * configure: Regenerate.
21688 2022-01-19 Martin Liska <mliska@suse.cz>
21690 * config/riscv/riscv.cc (riscv_handle_type_attribute):
21691 Update one -Wformat-diag string in warning message.
21693 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21695 PR middle-end/104103
21696 * gimple-ssa-warn-access.cc (pass_waccess::check_call): Don't check
21699 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21702 * fold-const.cc (address_compare): Consider different STRING_CSTs
21703 with the same lengths that memcmp the same as equal, not different.
21705 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21707 * config/i386/sse.md (*aes<aeswideklvariant>u*): Use %0 instead of
21710 2022-01-19 Martin Liska <mliska@suse.cz>
21711 Thomas Schwinge <thomas@codesourcery.com>
21713 * config/nvptx/nvptx.cc (nvptx_goacc_validate_dims_1): Update
21716 2022-01-19 Sebastian Huber <sebastian.huber@embedded-brains.de>
21719 * config/rs6000/rs6000.cc (rs6000_machine_from_flags): Use also
21722 2022-01-19 Jakub Jelinek <jakub@redhat.com>
21725 * config/i386/sse.md
21726 (<avx512>_<complexopname>_<mode><maskc_name><round_name>,
21727 avx512fp16_<complexopname>sh_v8hf<mask_scalarc_name><round_scalarcz_name>,
21728 avx512dq_mul<mode>3<mask_name>, <avx2_avx512>_permvar<mode><mask_name>,
21729 avx2_perm<mode>_1<mask_name>, avx512f_perm<mode>_1<mask_name>,
21730 avx512dq_rangep<mode><mask_name><round_saeonly_name>,
21731 avx512dq_ranges<mode><mask_scalar_name><round_saeonly_scalar_name>,
21732 <avx512>_getmant<mode><mask_name><round_saeonly_name>,
21733 avx512f_vgetmant<mode><mask_scalar_name><round_saeonly_scalar_name>):
21734 Use vxorps\t%x0, %x0, %x0 instead of vxorps\t{%x0, %x0, %x0}.
21736 2022-01-19 Martin Sebor <msebor@redhat.com>
21738 PR middle-end/104069
21739 * gimple-ssa-warn-access.cc (pointers_related_p): Return false for
21740 an unknown result as documented.
21742 2022-01-18 Andrew Pinski <apinski@marvell.com>
21744 * ipa-split.cc (visit_bb): Fix comment before the
21745 warning/error attribute checking code.
21747 2022-01-18 David Faust <david.faust@oracle.com>
21749 * config/bpf/coreout.cc (bpf_core_reloc_add): Do not account
21750 for base strtab offset yet as it may change.
21751 (output_asm_btfext_core_reloc): Do so here instead.
21752 (output_btfext_core_sections): Likewise.
21754 2022-01-18 David Faust <david.faust@oracle.com>
21756 * config/bpf/coreout.cc (output_btfext_header): Account for
21757 4-byte record size in core_relo_len.
21758 (output_btfext_core_sections): Only write record size once.
21759 * config/bpf/coreout.h (btf_ext_section_header): Delete unused
21762 2022-01-18 Maciej W. Rozycki <macro@embecosm.com>
21764 * common/config/riscv/riscv-common.cc
21765 (riscv_subset_list::parse_multiletter_ext): Move pointer
21766 arithmetic ahead of `free'.
21768 2022-01-18 Jason Merrill <jason@redhat.com>
21771 * gimplify.cc (gimple_push_cleanup): Handle eh_only in conditional
21774 2022-01-18 Sandra Loosemore <sandra@codesourcery.com>
21776 PR middle-end/103163
21777 * emit-rtl.cc (init_emit_regs): Initialize stack_limit_rtx here...
21778 (init_emit_once): ...not here.
21780 2022-01-18 Martin Liska <mliska@suse.cz>
21782 * collect2.cc (scan_libraries): Fix -Wformat-diag issues.
21783 * config/aarch64/aarch64-builtins.cc (aarch64_simd_expand_builtin): Likewise.
21784 * config/arc/arc.md: Likewise.
21785 * config/avr/avr.cc (avr_section_type_flags): Likewise.
21786 * config/bfin/bfin.cc (bfin_option_override): Likewise.
21787 (bfin_handle_longcall_attribute): Likewise.
21788 * config/cris/cris.h (FUNCTION_PROFILER): Likewise.
21789 * config/frv/frv.cc (frv_expand_builtin): Likewise.
21790 * config/ia64/ia64-c.cc (ia64_hpux_handle_builtin_pragma): Likewise.
21791 * config/iq2000/iq2000.cc (save_restore_insns): Likewise.
21792 (iq2000_print_operand_address): Likewise.
21793 (iq2000_print_operand): Likewise.
21794 * config/m32c/m32c-pragma.cc (m32c_pragma_memregs): Likewise.
21795 (m32c_pragma_address): Likewise.
21796 * config/m68k/m68k.cc (m68k_handle_fndecl_attribute): Likewise.
21797 * config/mips/mips.cc (mips_handle_interrupt_attr): Likewise.
21798 (mips_set_compression_mode): Likewise.
21799 * config/mmix/mmix.cc (mmix_function_profiler): Likewise.
21800 (mmix_print_operand): Likewise.
21801 (mmix_output_shiftvalue_op_from_str): Likewise.
21802 (mmix_output_shifted_value): Likewise.
21803 * config/msp430/driver-msp430.cc (msp430_select_hwmult_lib): Likewise.
21804 * config/msp430/msp430.cc (msp430_option_override): Likewise.
21805 (msp430_attr): Likewise.
21806 (msp430_expand_delay_cycles): Likewise.
21807 (msp430_expand_builtin): Likewise.
21808 * config/rs6000/aix73.h: Likewise.
21809 * config/rs6000/rtems.h (INVALID_64BIT): Likewise.
21810 * config/rx/rx.cc (rx_expand_builtin_mvtc): Likewise.
21811 (valid_psw_flag): Likewise.
21812 * config/sh/sh.cc (parse_validate_atomic_model_option): Likewise.
21813 * config/stormy16/stormy16.cc (xstormy16_function_profiler): Likewise.
21814 (xstormy16_expand_builtin_va_start): Likewise.
21815 (xstormy16_handle_below100_attribute): Likewise.
21817 2022-01-18 Martin Liska <mliska@suse.cz>
21819 * config/vms/vms-c.cc (vms_pragma_nostandard): Fix -Wformat-diag
21821 (vms_pragma_standard): Likewise.
21822 (vms_pragma_extern_prefix): Likewise.
21824 2022-01-18 Martin Liska <mliska@suse.cz>
21826 * config/xtensa/xtensa.cc (print_operand): Fix warnings.
21827 (print_operand_address): Likewise.
21828 (xtensa_multibss_section_type_flags): Likewise.
21830 2022-01-18 Martin Liska <mliska@suse.cz>
21832 * config/rs6000/rs6000-call.cc (rs6000_invalid_builtin): Change
21833 wording of an error message.
21835 2022-01-18 Martin Liska <mliska@suse.cz>
21837 * config/v850/v850-c.cc (pop_data_area): Fix -Wformat-diag
21839 (ghs_pragma_section): Likewise.
21840 (ghs_pragma_interrupt): Likewise.
21841 (ghs_pragma_starttda): Likewise.
21842 (ghs_pragma_startsda): Likewise.
21843 (ghs_pragma_startzda): Likewise.
21844 (ghs_pragma_endtda): Likewise.
21845 (ghs_pragma_endsda): Likewise.
21846 (ghs_pragma_endzda): Likewise.
21848 2022-01-18 Martin Liska <mliska@suse.cz>
21850 * config/nds32/nds32-intrinsic.cc (nds32_expand_builtin_impl):
21852 * config/nds32/nds32-intrinsic.md: Likewise.
21853 * config/nds32/nds32-isr.cc (nds32_check_isr_attrs_conflict): Likewise.
21854 * config/nds32/nds32.cc (nds32_print_operand): Likewise.
21855 (nds32_insert_attributes): Likewise.
21857 2022-01-18 Martin Liska <mliska@suse.cz>
21859 * config/nvptx/nvptx.cc (nvptx_goacc_validate_dims_1): Wrap
21861 * config/nvptx/nvptx.md: Remove trailing dot.
21863 2022-01-18 Martin Liska <mliska@suse.cz>
21865 * common/config/riscv/riscv-common.cc (riscv_subset_list::add):
21866 Wrap keywords with quotes and remove trailing dots.
21867 (riscv_subset_list::parsing_subset_version): Likewise.
21868 (riscv_subset_list::parse_std_ext): Likewise.
21869 (riscv_subset_list::parse_multiletter_ext): Likewise.
21870 * config/riscv/riscv.cc (riscv_handle_type_attribute): Likewise.
21872 2022-01-18 Andre Vieira <andre.simoesdiasvieira@arm.com>
21874 * tree-vect-loop.cc (vect_estimate_min_profitable_iters): Pass new
21875 argument suggested_unroll_factor.
21876 (vect_analyze_loop_costing): Likewise.
21877 (_loop_vec_info::_loop_vec_info): Initialize new member
21878 suggested_unroll_factor.
21879 (vect_determine_partial_vectors_and_peeling): Make epilogue of unrolled
21880 main loop use partial vectors.
21881 (vect_analyze_loop_2): Pass and use new argument
21882 suggested_unroll_factor.
21883 (vect_analyze_loop_1): Change to intialize local
21884 suggested_unroll_factor and use it.
21885 (vectorizable_reduction): Don't use single_defuse_cycle when unrolling.
21886 * tree-vectorizer.h (_loop_vec_info::_loop_vec_info): Add new member
21887 suggested_unroll_factor.
21888 (vector_costs::vector_costs): Add new member m_suggested_unroll_factor.
21889 (vector_costs::suggested_unroll_factor): New getter function.
21890 (finish_cost): Set return argument suggested_unroll_factor.
21892 2022-01-18 Andrew MacLeod <amacleod@redhat.com>
21894 PR tree-optimization/104038
21895 * doc/invoke.texi (relation-block-limit): New.
21896 * params.opt (relation-block-limit): New.
21897 * value-relation.cc (dom_oracle::register_relation): Check for NULL
21898 record before invoking transitive registery.
21899 (dom_oracle::set_one_relation): Check limit before creating record.
21900 (dom_oracle::register_transitives): Stop when no record created.
21901 * value-relation.h (relation_chain_head::m_num_relations): New.
21903 2022-01-18 Richard Biener <rguenther@suse.de>
21906 * ipa-inline.cc (inline_small_functions): Do not enqueue call
21907 edges originating in functions compiled with -Og.
21909 2022-01-18 Richard Biener <rguenther@suse.de>
21912 * passes.def (pass_all_optimizations_g): Remove pass_modref
21913 and pass_local_pure_const.
21915 2022-01-18 Martin Liska <mliska@suse.cz>
21917 * config/s390/s390.cc: Fix -Wformat-diag warnings.
21919 2022-01-18 Martin Liska <mliska@suse.cz>
21921 * config/s390/s390-c.cc (s390_expand_overloaded_builtin): Wrap
21923 (s390_resolve_overloaded_builtin): Remove trailing dot.
21924 * config/s390/s390.cc (s390_const_operand_ok): Use - for range.
21925 (s390_expand_builtin): Remove trailing dot.
21926 (s390_emit_prologue): Likewise, use semicolon.
21927 (s390_option_override_internal): Update keyword.
21928 * varasm.cc (do_assemble_alias): Wrap keyword in quotes.
21930 2022-01-18 Martin Liska <mliska@suse.cz>
21932 * config/rs6000/rs6000-call.cc (rs6000_expand_builtin): Wrap
21933 keywords and use %qs instead of %<%s%>.
21935 2022-01-18 Richard Biener <rguenther@suse.de>
21937 PR tree-optimization/103987
21938 * tree-ssa-dse.cc (dse_optimize_call): Properly guard modref
21939 query with a pointer check.
21941 2022-01-18 Richard Sandiford <richard.sandiford@arm.com>
21944 * config/aarch64/aarch64.cc (aarch64_check_consecutive_mems):
21945 When using MEM_EXPR, require the base to be a decl.
21947 2022-01-18 Richard Biener <rguenther@suse.de>
21949 * cgraph.h (struct cgraph_simd_clone_arg): Re-arrange fields to
21951 * function.h (struct function): Likewise.
21953 2022-01-18 Arnaud Charlet <charlet@adacore.com>
21955 * doc/install.texi: Update prerequisites for GNAT
21957 2022-01-18 Andrew Pinski <apinski@marvell.com>
21959 PR tree-optimization/101941
21960 * ipa-split.cc (visit_bb): Disallow function calls where
21961 the function has either error or warning attribute.
21963 2022-01-18 Richard Biener <rguenther@suse.de>
21965 PR tree-optimization/104064
21966 * tree-vect-data-refs.cc (vect_analyze_data_ref_accesses): Check
21967 DR_INIT fits in a signed HWI, represent the difference from the
21968 first DR in unsigned.
21970 2022-01-17 Martin Liska <mliska@suse.cz>
21972 * Makefile.in: Rename .c names to .cc.
21973 * config.gcc: Likewise.
21974 * configure: Regenerate. Likewise.
21975 * configure.ac: Likewise.
21976 * gengtype.cc (set_gc_used): Likewise.
21977 (source_dot_c_frul): Likewise.
21978 (source_dot_cc_frul): Likewise.
21979 (struct file_rule_st): Likewise.
21980 (close_output_files): Likewise.
21981 * config/avr/t-avr: Use CXXFLAGS_* and CXX_FOR_BUILD.
21983 2022-01-17 Martin Liska <mliska@suse.cz>
21985 * Makefile.in: Rename .c names to .cc.
21986 * alias.h: Likewise.
21987 * asan.cc: Likewise.
21988 * auto-profile.h: Likewise.
21989 * basic-block.h (struct basic_block_d): Likewise.
21990 * btfout.cc: Likewise.
21991 * builtins.cc (expand_builtin_longjmp): Likewise.
21992 (validate_arg): Likewise.
21993 (access_ref::offset_bounded): Likewise.
21994 * caller-save.cc (reg_restore_code): Likewise.
21995 (setup_save_areas): Likewise.
21996 * calls.cc (initialize_argument_information): Likewise.
21997 (expand_call): Likewise.
21998 (emit_library_call_value_1): Likewise.
21999 * cfg-flags.def (RTL): Likewise.
22000 (SIBCALL): Likewise.
22001 (CAN_FALLTHRU): Likewise.
22002 * cfganal.cc (post_order_compute): Likewise.
22003 * cfgcleanup.cc (try_simplify_condjump): Likewise.
22004 (merge_blocks_move_predecessor_nojumps): Likewise.
22005 (merge_blocks_move_successor_nojumps): Likewise.
22006 (merge_blocks_move): Likewise.
22007 (old_insns_match_p): Likewise.
22008 (try_crossjump_bb): Likewise.
22009 * cfgexpand.cc (expand_gimple_stmt): Likewise.
22010 * cfghooks.cc (split_block_before_cond_jump): Likewise.
22011 (profile_record_check_consistency): Likewise.
22012 * cfghooks.h: Likewise.
22013 * cfgrtl.cc (pass_free_cfg::execute): Likewise.
22014 (rtl_can_merge_blocks): Likewise.
22015 (try_redirect_by_replacing_jump): Likewise.
22016 (make_pass_outof_cfg_layout_mode): Likewise.
22017 (cfg_layout_can_merge_blocks_p): Likewise.
22018 * cgraph.cc (release_function_body): Likewise.
22019 (cgraph_node::get_fun): Likewise.
22020 * cgraph.h (struct cgraph_node): Likewise.
22021 (asmname_hasher::equal): Likewise.
22022 (cgraph_inline_failed_type): Likewise.
22023 (thunk_adjust): Likewise.
22024 (dump_callgraph_transformation): Likewise.
22025 (record_references_in_initializer): Likewise.
22026 (ipa_discover_variable_flags): Likewise.
22027 * cgraphclones.cc (GTY): Likewise.
22028 * cgraphunit.cc (symbol_table::finalize_compilation_unit): Likewise.
22029 * collect-utils.h (GCC_COLLECT_UTILS_H): Likewise.
22030 * collect2-aix.h (GCC_COLLECT2_AIX_H): Likewise.
22031 * collect2.cc (maybe_run_lto_and_relink): Likewise.
22032 * combine-stack-adj.cc: Likewise.
22033 * combine.cc (setup_incoming_promotions): Likewise.
22034 (combine_simplify_rtx): Likewise.
22035 (count_rtxs): Likewise.
22036 * common.opt: Likewise.
22037 * common/config/aarch64/aarch64-common.cc: Likewise.
22038 * common/config/arm/arm-common.cc (arm_asm_auto_mfpu): Likewise.
22039 * common/config/avr/avr-common.cc: Likewise.
22040 * common/config/i386/i386-isas.h (struct _isa_names_table): Likewise.
22041 * conditions.h: Likewise.
22042 * config.gcc: Likewise.
22043 * config/aarch64/aarch64-builtins.cc (aarch64_resolve_overloaded_memtag): Likewise.
22044 * config/aarch64/aarch64-protos.h (aarch64_classify_address): Likewise.
22045 (aarch64_get_extension_string_for_isa_flags): Likewise.
22046 * config/aarch64/aarch64-sve-builtins.cc (function_builder::add_function): Likewise.
22047 * config/aarch64/aarch64.cc (aarch64_regmode_natural_size): Likewise.
22048 (aarch64_sched_first_cycle_multipass_dfa_lookahead): Likewise.
22049 (aarch64_option_valid_attribute_p): Likewise.
22050 (aarch64_short_vector_p): Likewise.
22051 (aarch64_float_const_representable_p): Likewise.
22052 * config/aarch64/aarch64.h (DBX_REGISTER_NUMBER): Likewise.
22053 (ASM_OUTPUT_POOL_EPILOGUE): Likewise.
22055 * config/aarch64/cortex-a57-fma-steering.cc: Likewise.
22056 * config/aarch64/driver-aarch64.cc (contains_core_p): Likewise.
22057 * config/aarch64/t-aarch64: Likewise.
22058 * config/aarch64/x-aarch64: Likewise.
22059 * config/aarch64/x-darwin: Likewise.
22060 * config/alpha/alpha-protos.h: Likewise.
22061 * config/alpha/alpha.cc (alpha_scalar_mode_supported_p): Likewise.
22062 * config/alpha/alpha.h (LONG_DOUBLE_TYPE_SIZE): Likewise.
22063 (enum reg_class): Likewise.
22064 * config/alpha/alpha.md: Likewise.
22065 * config/alpha/driver-alpha.cc (AMASK_LOCKPFTCHOK): Likewise.
22066 * config/alpha/x-alpha: Likewise.
22067 * config/arc/arc-protos.h (arc_eh_uses): Likewise.
22068 * config/arc/arc.cc (ARC_OPT): Likewise.
22069 (arc_ccfsm_advance): Likewise.
22070 (arc_arg_partial_bytes): Likewise.
22071 (conditionalize_nonjump): Likewise.
22072 * config/arc/arc.md: Likewise.
22073 * config/arc/builtins.def: Likewise.
22074 * config/arc/t-arc: Likewise.
22075 * config/arm/arm-c.cc (arm_resolve_overloaded_builtin): Likewise.
22076 (arm_pragma_target_parse): Likewise.
22077 * config/arm/arm-protos.h (save_restore_target_globals): Likewise.
22078 (arm_cpu_cpp_builtins): Likewise.
22079 * config/arm/arm.cc (vfp3_const_double_index): Likewise.
22080 (shift_op): Likewise.
22081 (thumb2_final_prescan_insn): Likewise.
22082 (arm_final_prescan_insn): Likewise.
22083 (arm_asm_output_labelref): Likewise.
22084 (arm_small_register_classes_for_mode_p): Likewise.
22085 * config/arm/arm.h: Likewise.
22086 * config/arm/arm.md: Likewise.
22087 * config/arm/driver-arm.cc: Likewise.
22088 * config/arm/symbian.h: Likewise.
22089 * config/arm/t-arm: Likewise.
22090 * config/arm/thumb1.md: Likewise.
22091 * config/arm/x-arm: Likewise.
22092 * config/avr/avr-c.cc (avr_register_target_pragmas): Likewise.
22093 * config/avr/avr-fixed.md: Likewise.
22094 * config/avr/avr-log.cc (avr_log_vadump): Likewise.
22095 * config/avr/avr-mcus.def: Likewise.
22096 * config/avr/avr-modes.def (FRACTIONAL_INT_MODE): Likewise.
22097 * config/avr/avr-passes.def (INSERT_PASS_BEFORE): Likewise.
22098 * config/avr/avr-protos.h (make_avr_pass_casesi): Likewise.
22099 * config/avr/avr.cc (avr_option_override): Likewise.
22100 (avr_build_builtin_va_list): Likewise.
22101 (avr_mode_dependent_address_p): Likewise.
22102 (avr_function_arg_advance): Likewise.
22103 (avr_asm_output_aligned_decl_common): Likewise.
22104 * config/avr/avr.h (RETURN_ADDR_RTX): Likewise.
22105 (SUPPORTS_INIT_PRIORITY): Likewise.
22106 * config/avr/avr.md: Likewise.
22107 * config/avr/builtins.def: Likewise.
22108 * config/avr/gen-avr-mmcu-specs.cc (IN_GEN_AVR_MMCU_TEXI): Likewise.
22109 * config/avr/gen-avr-mmcu-texi.cc (IN_GEN_AVR_MMCU_TEXI): Likewise.
22111 * config/avr/t-avr: Likewise.
22112 * config/bfin/bfin.cc (frame_related_constant_load): Likewise.
22113 * config/bpf/bpf-protos.h (GCC_BPF_PROTOS_H): Likewise.
22114 * config/bpf/bpf.h (enum reg_class): Likewise.
22115 * config/bpf/t-bpf: Likewise.
22116 * config/c6x/c6x-protos.h (GCC_C6X_PROTOS_H): Likewise.
22117 * config/cr16/cr16-protos.h: Likewise.
22118 * config/cris/cris.cc (cris_address_cost): Likewise.
22119 (cris_side_effect_mode_ok): Likewise.
22120 (cris_init_machine_status): Likewise.
22121 (cris_emit_movem_store): Likewise.
22122 * config/cris/cris.h (INDEX_REG_CLASS): Likewise.
22123 (enum reg_class): Likewise.
22124 (struct cum_args): Likewise.
22125 * config/cris/cris.opt: Likewise.
22126 * config/cris/sync.md: Likewise.
22127 * config/csky/csky.cc (csky_expand_prologue): Likewise.
22128 * config/darwin-c.cc: Likewise.
22129 * config/darwin-f.cc: Likewise.
22130 * config/darwin-sections.def (zobj_const_section): Likewise.
22131 * config/darwin.cc (output_objc_section_asm_op): Likewise.
22132 (fprintf): Likewise.
22133 * config/darwin.h (GTY): Likewise.
22134 * config/elfos.h: Likewise.
22135 * config/epiphany/epiphany-sched.md: Likewise.
22136 * config/epiphany/epiphany.cc (epiphany_function_value): Likewise.
22137 * config/epiphany/epiphany.h (GTY): Likewise.
22138 (NO_FUNCTION_CSE): Likewise.
22139 * config/epiphany/mode-switch-use.cc: Likewise.
22140 * config/epiphany/predicates.md: Likewise.
22141 * config/epiphany/t-epiphany: Likewise.
22142 * config/fr30/fr30-protos.h: Likewise.
22143 * config/frv/frv-protos.h: Likewise.
22144 * config/frv/frv.cc (TLS_BIAS): Likewise.
22145 * config/frv/frv.h (ASM_OUTPUT_ALIGNED_LOCAL): Likewise.
22146 * config/ft32/ft32-protos.h: Likewise.
22147 * config/gcn/gcn-hsa.h (ASM_APP_OFF): Likewise.
22148 * config/gcn/gcn.cc (gcn_init_libfuncs): Likewise.
22149 * config/gcn/mkoffload.cc (copy_early_debug_info): Likewise.
22150 * config/gcn/t-gcn-hsa: Likewise.
22151 * config/gcn/t-omp-device: Likewise.
22152 * config/h8300/h8300-protos.h (GCC_H8300_PROTOS_H): Likewise.
22153 (same_cmp_following_p): Likewise.
22154 * config/h8300/h8300.cc (F): Likewise.
22155 * config/h8300/h8300.h (struct cum_arg): Likewise.
22156 (BRANCH_COST): Likewise.
22157 * config/i386/cygming.h (DEFAULT_PCC_STRUCT_RETURN): Likewise.
22158 * config/i386/djgpp.h (TARGET_ASM_LTO_END): Likewise.
22159 * config/i386/dragonfly.h (NO_PROFILE_COUNTERS): Likewise.
22160 * config/i386/driver-i386.cc (detect_caches_intel): Likewise.
22161 * config/i386/freebsd.h (NO_PROFILE_COUNTERS): Likewise.
22162 * config/i386/i386-c.cc (ix86_target_macros): Likewise.
22163 * config/i386/i386-expand.cc (get_mode_wider_vector): Likewise.
22164 * config/i386/i386-options.cc (ix86_set_func_type): Likewise.
22165 * config/i386/i386-protos.h (ix86_extract_perm_from_pool_constant): Likewise.
22166 (ix86_register_pragmas): Likewise.
22167 (ix86_d_has_stdcall_convention): Likewise.
22168 (i386_pe_seh_init_sections): Likewise.
22169 * config/i386/i386.cc (ix86_function_arg_regno_p): Likewise.
22170 (ix86_function_value_regno_p): Likewise.
22171 (ix86_compute_frame_layout): Likewise.
22172 (legitimize_pe_coff_symbol): Likewise.
22173 (output_pic_addr_const): Likewise.
22174 * config/i386/i386.h (defined): Likewise.
22175 (host_detect_local_cpu): Likewise.
22176 (CONSTANT_ADDRESS_P): Likewise.
22177 (DEFAULT_LARGE_SECTION_THRESHOLD): Likewise.
22178 (struct machine_frame_state): Likewise.
22179 * config/i386/i386.md: Likewise.
22180 * config/i386/lynx.h (ASM_OUTPUT_ALIGN): Likewise.
22181 * config/i386/mmx.md: Likewise.
22182 * config/i386/sse.md: Likewise.
22183 * config/i386/t-cygming: Likewise.
22184 * config/i386/t-djgpp: Likewise.
22185 * config/i386/t-gnu-property: Likewise.
22186 * config/i386/t-i386: Likewise.
22187 * config/i386/t-intelmic: Likewise.
22188 * config/i386/t-omp-device: Likewise.
22189 * config/i386/winnt-cxx.cc (i386_pe_type_dllimport_p): Likewise.
22190 (i386_pe_adjust_class_at_definition): Likewise.
22191 * config/i386/winnt.cc (gen_stdcall_or_fastcall_suffix): Likewise.
22192 (i386_pe_mangle_decl_assembler_name): Likewise.
22193 (i386_pe_encode_section_info): Likewise.
22194 * config/i386/x-cygwin: Likewise.
22195 * config/i386/x-darwin: Likewise.
22196 * config/i386/x-i386: Likewise.
22197 * config/i386/x-mingw32: Likewise.
22198 * config/i386/x86-tune-sched-core.cc: Likewise.
22199 * config/i386/x86-tune.def: Likewise.
22200 * config/i386/xm-djgpp.h (STANDARD_STARTFILE_PREFIX_1): Likewise.
22201 * config/ia64/freebsd.h: Likewise.
22202 * config/ia64/hpux.h (REGISTER_TARGET_PRAGMAS): Likewise.
22203 * config/ia64/ia64-protos.h (ia64_except_unwind_info): Likewise.
22204 * config/ia64/ia64.cc (ia64_function_value_regno_p): Likewise.
22205 (ia64_secondary_reload_class): Likewise.
22206 (bundling): Likewise.
22207 * config/ia64/ia64.h: Likewise.
22208 * config/ia64/ia64.md: Likewise.
22209 * config/ia64/predicates.md: Likewise.
22210 * config/ia64/sysv4.h: Likewise.
22211 * config/ia64/t-ia64: Likewise.
22212 * config/iq2000/iq2000.h (FUNCTION_MODE): Likewise.
22213 * config/iq2000/iq2000.md: Likewise.
22214 * config/linux.h (TARGET_HAS_BIONIC): Likewise.
22216 * config/m32c/m32c.cc (m32c_function_needs_enter): Likewise.
22217 * config/m32c/m32c.h (MAX_REGS_PER_ADDRESS): Likewise.
22218 * config/m32c/t-m32c: Likewise.
22219 * config/m32r/m32r-protos.h: Likewise.
22220 * config/m32r/m32r.cc (m32r_print_operand): Likewise.
22221 * config/m32r/m32r.h: Likewise.
22222 * config/m32r/m32r.md: Likewise.
22223 * config/m68k/m68k-isas.def: Likewise.
22224 * config/m68k/m68k-microarchs.def: Likewise.
22225 * config/m68k/m68k-protos.h (strict_low_part_peephole_ok): Likewise.
22226 (m68k_epilogue_uses): Likewise.
22227 * config/m68k/m68k.cc (m68k_call_tls_get_addr): Likewise.
22228 (m68k_sched_adjust_cost): Likewise.
22229 (m68k_sched_md_init): Likewise.
22230 * config/m68k/m68k.h (__transfer_from_trampoline): Likewise.
22231 (enum m68k_function_kind): Likewise.
22232 * config/m68k/m68k.md: Likewise.
22233 * config/m68k/m68kemb.h: Likewise.
22234 * config/m68k/uclinux.h (ENDFILE_SPEC): Likewise.
22235 * config/mcore/mcore-protos.h: Likewise.
22236 * config/mcore/mcore.cc (mcore_expand_insv): Likewise.
22237 (mcore_expand_prolog): Likewise.
22238 * config/mcore/mcore.h (TARGET_MCORE): Likewise.
22239 * config/mcore/mcore.md: Likewise.
22240 * config/microblaze/microblaze-protos.h: Likewise.
22241 * config/microblaze/microblaze.cc (microblaze_legitimate_pic_operand): Likewise.
22242 (microblaze_function_prologue): Likewise.
22243 (microblaze_function_epilogue): Likewise.
22244 (microblaze_select_section): Likewise.
22245 (microblaze_asm_output_mi_thunk): Likewise.
22246 (microblaze_eh_return): Likewise.
22247 * config/microblaze/microblaze.h: Likewise.
22248 * config/microblaze/microblaze.md: Likewise.
22249 * config/microblaze/t-microblaze: Likewise.
22250 * config/mips/driver-native.cc: Likewise.
22251 * config/mips/loongson2ef.md: Likewise.
22252 * config/mips/mips-protos.h (mips_expand_vec_cmp_expr): Likewise.
22253 * config/mips/mips.cc (mips_rtx_costs): Likewise.
22254 (mips_output_filename): Likewise.
22255 (mips_output_function_prologue): Likewise.
22256 (mips_output_function_epilogue): Likewise.
22257 (mips_output_mi_thunk): Likewise.
22258 * config/mips/mips.h: Likewise.
22259 * config/mips/mips.md: Likewise.
22260 * config/mips/t-mips: Likewise.
22261 * config/mips/x-native: Likewise.
22262 * config/mmix/mmix-protos.h: Likewise.
22263 * config/mmix/mmix.cc (mmix_option_override): Likewise.
22264 (mmix_dbx_register_number): Likewise.
22265 (mmix_expand_prologue): Likewise.
22266 * config/mmix/mmix.h: Likewise.
22267 * config/mmix/mmix.md: Likewise.
22268 * config/mmix/predicates.md: Likewise.
22269 * config/mn10300/mn10300.cc (mn10300_symbolic_operand): Likewise.
22270 (mn10300_legitimate_pic_operand_p): Likewise.
22271 * config/mn10300/mn10300.h (enum reg_class): Likewise.
22272 (NO_FUNCTION_CSE): Likewise.
22273 * config/moxie/moxie-protos.h: Likewise.
22274 * config/moxie/uclinux.h (TARGET_LIBC_HAS_FUNCTION): Likewise.
22275 * config/msp430/msp430-devices.cc (extract_devices_dir_from_exec_prefix): Likewise.
22276 * config/msp430/msp430.cc (msp430_gimplify_va_arg_expr): Likewise.
22277 (msp430_incoming_return_addr_rtx): Likewise.
22278 * config/msp430/msp430.h (msp430_get_linker_devices_include_path): Likewise.
22279 * config/msp430/t-msp430: Likewise.
22280 * config/nds32/nds32-cost.cc (nds32_rtx_costs_speed_prefer): Likewise.
22281 (nds32_rtx_costs_size_prefer): Likewise.
22282 (nds32_init_rtx_costs): Likewise.
22283 * config/nds32/nds32-doubleword.md: Likewise.
22284 * config/nds32/nds32.cc (nds32_memory_move_cost): Likewise.
22285 (nds32_builtin_decl): Likewise.
22286 * config/nds32/nds32.h (enum nds32_16bit_address_type): Likewise.
22287 (enum nds32_isr_nested_type): Likewise.
22288 (enum reg_class): Likewise.
22289 * config/nds32/predicates.md: Likewise.
22290 * config/nds32/t-nds32: Likewise.
22291 * config/nios2/nios2.cc (nios2_pragma_target_parse): Likewise.
22292 * config/nvptx/nvptx-protos.h: Likewise.
22293 * config/nvptx/nvptx.cc (nvptx_goacc_expand_var_decl): Likewise.
22294 * config/nvptx/nvptx.h (TARGET_CPU_CPP_BUILTINS): Likewise.
22295 * config/nvptx/t-nvptx: Likewise.
22296 * config/nvptx/t-omp-device: Likewise.
22297 * config/pa/elf.h: Likewise.
22298 * config/pa/pa-linux.h (GLOBAL_ASM_OP): Likewise.
22299 * config/pa/pa-netbsd.h (GLOBAL_ASM_OP): Likewise.
22300 * config/pa/pa-openbsd.h (TARGET_ASM_GLOBALIZE_LABEL): Likewise.
22301 * config/pa/pa-protos.h (pa_eh_return_handler_rtx): Likewise.
22302 (pa_legitimize_reload_address): Likewise.
22303 (pa_can_use_return_insn): Likewise.
22304 * config/pa/pa.cc (mem_shadd_or_shadd_rtx_p): Likewise.
22305 (som_output_text_section_asm_op): Likewise.
22306 * config/pa/pa.h (PROFILE_BEFORE_PROLOGUE): Likewise.
22307 * config/pa/pa.md: Likewise.
22308 * config/pa/som.h: Likewise.
22309 * config/pa/t-pa: Likewise.
22310 * config/pdp11/pdp11.cc (decode_pdp11_d): Likewise.
22311 * config/pdp11/pdp11.h: Likewise.
22312 * config/pdp11/pdp11.md: Likewise.
22313 * config/pdp11/t-pdp11: Likewise.
22314 * config/pru/pru.md: Likewise.
22315 * config/pru/t-pru: Likewise.
22316 * config/riscv/riscv-protos.h (NUM_SYMBOL_TYPES): Likewise.
22317 (riscv_gpr_save_operation_p): Likewise.
22318 (riscv_d_register_target_info): Likewise.
22319 (riscv_init_builtins): Likewise.
22320 * config/riscv/riscv.cc (riscv_output_mi_thunk): Likewise.
22321 * config/riscv/riscv.h (CSW_MAX_OFFSET): Likewise.
22322 * config/riscv/t-riscv: Likewise.
22323 * config/rl78/rl78.cc (rl78_asm_ctor_dtor): Likewise.
22324 * config/rl78/t-rl78: Likewise.
22325 * config/rs6000/aix.h: Likewise.
22326 * config/rs6000/aix71.h (ASM_SPEC_COMMON): Likewise.
22327 * config/rs6000/aix72.h (ASM_SPEC_COMMON): Likewise.
22328 * config/rs6000/aix73.h (ASM_SPEC_COMMON): Likewise.
22329 * config/rs6000/darwin.h (TARGET_ASM_GLOBALIZE_LABEL): Likewise.
22330 * config/rs6000/driver-rs6000.cc: Likewise.
22331 * config/rs6000/freebsd.h: Likewise.
22332 * config/rs6000/freebsd64.h: Likewise.
22333 * config/rs6000/lynx.h (ASM_OUTPUT_ALIGN): Likewise.
22334 * config/rs6000/rbtree.cc: Likewise.
22335 * config/rs6000/rbtree.h: Likewise.
22336 * config/rs6000/rs6000-c.cc (rs6000_target_modify_macros): Likewise.
22337 * config/rs6000/rs6000-call.cc (rs6000_invalid_builtin): Likewise.
22338 (rs6000_expand_builtin): Likewise.
22339 (rs6000_init_builtins): Likewise.
22340 * config/rs6000/rs6000-cpus.def: Likewise.
22341 * config/rs6000/rs6000-gen-builtins.cc (write_init_ovld_table): Likewise.
22342 * config/rs6000/rs6000-internal.h (ALTIVEC_REG_BIT): Likewise.
22343 (quad_address_offset_p): Likewise.
22344 * config/rs6000/rs6000-logue.cc (interesting_frame_related_regno): Likewise.
22345 (rs6000_emit_epilogue): Likewise.
22346 * config/rs6000/rs6000-overload.def: Likewise.
22347 * config/rs6000/rs6000-p8swap.cc: Likewise.
22348 * config/rs6000/rs6000-protos.h (GCC_RS6000_PROTOS_H): Likewise.
22349 (rs6000_const_f32_to_i32): Likewise.
22350 * config/rs6000/rs6000.cc (legitimate_lo_sum_address_p): Likewise.
22351 (rs6000_debug_legitimize_address): Likewise.
22352 (rs6000_mode_dependent_address): Likewise.
22353 (rs6000_adjust_priority): Likewise.
22354 (rs6000_c_mode_for_suffix): Likewise.
22355 * config/rs6000/rs6000.h (defined): Likewise.
22356 (LONG_DOUBLE_TYPE_SIZE): Likewise.
22357 * config/rs6000/rs6000.md: Likewise.
22358 * config/rs6000/sysv4.h: Likewise.
22359 * config/rs6000/t-linux: Likewise.
22360 * config/rs6000/t-linux64: Likewise.
22361 * config/rs6000/t-rs6000: Likewise.
22362 * config/rs6000/x-darwin: Likewise.
22363 * config/rs6000/x-darwin64: Likewise.
22364 * config/rs6000/x-rs6000: Likewise.
22365 * config/rs6000/xcoff.h (ASM_OUTPUT_LABELREF): Likewise.
22366 * config/rx/rx.cc (rx_expand_builtin): Likewise.
22367 * config/s390/constraints.md: Likewise.
22368 * config/s390/driver-native.cc: Likewise.
22369 * config/s390/htmxlintrin.h: Likewise.
22370 * config/s390/s390-builtins.def (B_DEF): Likewise.
22371 (OB_DEF_VAR): Likewise.
22372 * config/s390/s390-builtins.h: Likewise.
22373 * config/s390/s390-c.cc: Likewise.
22374 * config/s390/s390-opts.h: Likewise.
22375 * config/s390/s390-protos.h (s390_check_symref_alignment): Likewise.
22376 (s390_register_target_pragmas): Likewise.
22377 * config/s390/s390.cc (s390_init_builtins): Likewise.
22378 (s390_expand_plus_operand): Likewise.
22379 (s390_expand_atomic): Likewise.
22380 (s390_valid_target_attribute_inner_p): Likewise.
22381 * config/s390/s390.h (LONG_DOUBLE_TYPE_SIZE): Likewise.
22382 * config/s390/s390.md: Likewise.
22383 * config/s390/t-s390: Likewise.
22384 * config/s390/vx-builtins.md: Likewise.
22385 * config/s390/x-native: Likewise.
22386 * config/sh/divtab-sh4-300.cc (main): Likewise.
22387 * config/sh/divtab-sh4.cc (main): Likewise.
22388 * config/sh/divtab.cc (main): Likewise.
22389 * config/sh/elf.h: Likewise.
22390 * config/sh/sh-protos.h (sh_fsca_int2sf): Likewise.
22391 * config/sh/sh.cc (SYMBOL_FLAG_FUNCVEC_FUNCTION): Likewise.
22392 (sh_struct_value_rtx): Likewise.
22393 (sh_remove_reg_dead_or_unused_notes): Likewise.
22394 * config/sh/sh.h (MIN_UNITS_PER_WORD): Likewise.
22395 * config/sh/t-sh: Likewise.
22396 * config/sol2-protos.h (solaris_override_options): Likewise.
22397 * config/sol2.h: Likewise.
22398 * config/sparc/driver-sparc.cc: Likewise.
22399 * config/sparc/freebsd.h: Likewise.
22400 * config/sparc/sparc-protos.h (make_pass_work_around_errata): Likewise.
22401 * config/sparc/sparc.cc (sparc_output_mi_thunk): Likewise.
22402 (sparc_asan_shadow_offset): Likewise.
22403 * config/sparc/sparc.h: Likewise.
22404 * config/sparc/sparc.md: Likewise.
22405 * config/sparc/t-sparc: Likewise.
22406 * config/sparc/x-sparc: Likewise.
22407 * config/stormy16/stormy16.cc (xstormy16_mode_dependent_address_p): Likewise.
22408 * config/t-darwin: Likewise.
22409 * config/t-dragonfly: Likewise.
22410 * config/t-freebsd: Likewise.
22411 * config/t-glibc: Likewise.
22412 * config/t-linux: Likewise.
22413 * config/t-netbsd: Likewise.
22414 * config/t-openbsd: Likewise.
22415 * config/t-pnt16-warn: Likewise.
22416 * config/t-sol2: Likewise.
22417 * config/t-vxworks: Likewise.
22418 * config/t-winnt: Likewise.
22419 * config/tilegx/t-tilegx: Likewise.
22420 * config/tilegx/tilegx-c.cc: Likewise.
22421 * config/tilegx/tilegx-protos.h (tilegx_function_profiler): Likewise.
22422 * config/tilegx/tilegx.md: Likewise.
22423 * config/tilepro/t-tilepro: Likewise.
22424 * config/tilepro/tilepro-c.cc: Likewise.
22425 * config/v850/t-v850: Likewise.
22426 * config/v850/v850-protos.h: Likewise.
22427 * config/v850/v850.cc (F): Likewise.
22428 * config/v850/v850.h (enum reg_class): Likewise.
22429 (SLOW_BYTE_ACCESS): Likewise.
22430 * config/vax/vax.cc (vax_mode_dependent_address_p): Likewise.
22431 * config/vax/vax.h (enum reg_class): Likewise.
22432 * config/vax/vax.md: Likewise.
22433 * config/visium/visium.cc (visium_legitimate_address_p): Likewise.
22434 * config/visium/visium.h: Likewise.
22435 * config/vms/t-vms: Likewise.
22436 * config/vms/vms-crtlmap.map: Likewise.
22437 * config/vms/vms-protos.h (vms_c_get_vms_ver): Likewise.
22438 * config/vx-common.h: Likewise.
22439 * config/x-darwin: Likewise.
22440 * config/x-hpux: Likewise.
22441 * config/x-linux: Likewise.
22442 * config/x-netbsd: Likewise.
22443 * config/x-openbsd: Likewise.
22444 * config/x-solaris: Likewise.
22445 * config/xtensa/xtensa-protos.h (xtensa_mem_offset): Likewise.
22446 * config/xtensa/xtensa.cc (xtensa_option_override): Likewise.
22447 * config/xtensa/xtensa.h: Likewise.
22448 * configure.ac: Likewise.
22449 * context.cc: Likewise.
22450 * convert.h: Likewise.
22451 * coretypes.h: Likewise.
22452 * coverage.cc: Likewise.
22453 * coverage.h: Likewise.
22454 * cppdefault.h (struct default_include): Likewise.
22455 * cprop.cc (local_cprop_pass): Likewise.
22456 (one_cprop_pass): Likewise.
22457 * cse.cc (hash_rtx_cb): Likewise.
22458 (fold_rtx): Likewise.
22459 * ctfc.h (ctfc_get_num_vlen_bytes): Likewise.
22460 * data-streamer.h (bp_unpack_var_len_int): Likewise.
22461 (streamer_write_widest_int): Likewise.
22462 * dbgcnt.def: Likewise.
22463 * dbxout.cc (dbxout_early_global_decl): Likewise.
22464 (dbxout_common_check): Likewise.
22465 * dbxout.h: Likewise.
22466 * debug.h (struct gcc_debug_hooks): Likewise.
22467 (dump_go_spec_init): Likewise.
22468 * df-core.cc: Likewise.
22469 * df-scan.cc (df_insn_info_delete): Likewise.
22470 (df_insn_delete): Likewise.
22471 * df.h (debug_df_chain): Likewise.
22472 (can_move_insns_across): Likewise.
22473 * dfp.cc (decimal_from_binary): Likewise.
22474 * diagnostic-color.cc: Likewise.
22475 * diagnostic-event-id.h: Likewise.
22476 * diagnostic-show-locus.cc (test_one_liner_labels): Likewise.
22477 * diagnostic.cc (bt_callback): Likewise.
22478 (num_digits): Likewise.
22479 * doc/avr-mmcu.texi: Likewise.
22480 * doc/cfg.texi: Likewise.
22481 * doc/contrib.texi: Likewise.
22482 * doc/cppinternals.texi: Likewise.
22483 * doc/extend.texi: Likewise.
22484 * doc/generic.texi: Likewise.
22485 * doc/gimple.texi: Likewise.
22486 * doc/gty.texi: Likewise.
22487 * doc/invoke.texi: Likewise.
22488 * doc/loop.texi: Likewise.
22489 * doc/lto.texi: Likewise.
22490 * doc/match-and-simplify.texi: Likewise.
22491 * doc/md.texi: Likewise.
22492 * doc/optinfo.texi: Likewise.
22493 * doc/options.texi: Likewise.
22494 * doc/passes.texi: Likewise.
22495 * doc/plugins.texi: Likewise.
22496 * doc/rtl.texi: Likewise.
22497 * doc/sourcebuild.texi: Likewise.
22498 * doc/tm.texi: Likewise.
22499 * doc/tm.texi.in: Likewise.
22500 * doc/tree-ssa.texi: Likewise.
22501 * dojump.cc (do_jump): Likewise.
22502 * dojump.h: Likewise.
22503 * dumpfile.cc (test_impl_location): Likewise.
22504 (test_capture_of_dump_calls): Likewise.
22505 * dumpfile.h (enum dump_kind): Likewise.
22506 (class dump_location_t): Likewise.
22507 (dump_enabled_p): Likewise.
22508 (enable_rtl_dump_file): Likewise.
22509 (dump_combine_total_stats): Likewise.
22510 * dwarf2asm.cc (dw2_asm_output_delta_uleb128): Likewise.
22511 * dwarf2ctf.h (ctf_debug_finish): Likewise.
22512 * dwarf2out.cc (dwarf2out_begin_prologue): Likewise.
22513 (struct loc_descr_context): Likewise.
22514 (rtl_for_decl_location): Likewise.
22515 (gen_subprogram_die): Likewise.
22516 (gen_label_die): Likewise.
22517 (is_trivial_indirect_ref): Likewise.
22518 (dwarf2out_late_global_decl): Likewise.
22519 (dwarf_file_hasher::hash): Likewise.
22520 (dwarf2out_end_source_file): Likewise.
22521 (dwarf2out_define): Likewise.
22522 (dwarf2out_early_finish): Likewise.
22523 * dwarf2out.h (struct dw_fde_node): Likewise.
22524 (struct dw_discr_list_node): Likewise.
22525 (output_loc_sequence_raw): Likewise.
22526 * emit-rtl.cc (gen_raw_REG): Likewise.
22527 (maybe_set_max_label_num): Likewise.
22528 * emit-rtl.h (struct rtl_data): Likewise.
22529 * errors.cc (internal_error): Likewise.
22530 (trim_filename): Likewise.
22531 * et-forest.cc: Likewise.
22532 * except.cc (init_eh_for_function): Likewise.
22533 * explow.cc (promote_ssa_mode): Likewise.
22534 (get_dynamic_stack_size): Likewise.
22535 * explow.h: Likewise.
22536 * expmed.h: Likewise.
22537 * expr.cc (safe_from_p): Likewise.
22538 (expand_expr_real_2): Likewise.
22539 (expand_expr_real_1): Likewise.
22540 * file-prefix-map.cc (remap_filename): Likewise.
22541 * final.cc (app_enable): Likewise.
22542 (make_pass_compute_alignments): Likewise.
22543 (final_scan_insn_1): Likewise.
22544 (final_scan_insn): Likewise.
22545 * fixed-value.h (fixed_from_string): Likewise.
22546 * flag-types.h (NO_DEBUG): Likewise.
22547 (DWARF2_DEBUG): Likewise.
22548 (VMS_DEBUG): Likewise.
22549 (BTF_DEBUG): Likewise.
22550 (enum ctf_debug_info_levels): Likewise.
22551 * fold-const.cc (const_binop): Likewise.
22552 (fold_binary_loc): Likewise.
22553 (fold_checksum_tree): Likewise.
22554 * fp-test.cc: Likewise.
22555 * function.cc (expand_function_end): Likewise.
22556 * function.h (struct function): Likewise.
22557 * fwprop.cc (should_replace_address): Likewise.
22558 * gcc-main.cc: Likewise.
22559 * gcc-rich-location.h (class gcc_rich_location): Likewise.
22560 * gcc-symtab.h: Likewise.
22561 * gcc.cc (MIN_FATAL_STATUS): Likewise.
22562 (driver_handle_option): Likewise.
22563 (quote_spec_arg): Likewise.
22564 (driver::finalize): Likewise.
22565 * gcc.h (set_input): Likewise.
22566 * gcov-dump.cc: Likewise.
22567 * gcov.cc (solve_flow_graph): Likewise.
22568 * gcse-common.cc: Likewise.
22569 * gcse.cc (make_pass_rtl_hoist): Likewise.
22570 * genattr-common.cc: Likewise.
22571 * genattrtab.cc (min_fn): Likewise.
22572 (write_const_num_delay_slots): Likewise.
22573 * genautomata.cc: Likewise.
22574 * genconditions.cc (write_one_condition): Likewise.
22575 * genconstants.cc: Likewise.
22576 * genemit.cc (gen_exp): Likewise.
22577 * generic-match-head.cc: Likewise.
22578 * genextract.cc: Likewise.
22579 * gengenrtl.cc (always_void_p): Likewise.
22580 * gengtype-parse.cc (gtymarker_opt): Likewise.
22581 * gengtype-state.cc (state_writer::state_writer): Likewise.
22582 (write_state_trailer): Likewise.
22583 (equals_type_number): Likewise.
22584 (read_state): Likewise.
22585 * gengtype.cc (open_base_files): Likewise.
22586 (struct file_rule_st): Likewise.
22587 (header_dot_h_frul): Likewise.
22588 * gengtype.h: Likewise.
22589 * genmatch.cc (main): Likewise.
22590 * genmddeps.cc: Likewise.
22591 * genmodes.cc (emit_mode_inner): Likewise.
22592 (emit_mode_unit_size): Likewise.
22593 * genpeep.cc (gen_peephole): Likewise.
22594 * genpreds.cc (write_tm_preds_h): Likewise.
22595 * genrecog.cc (validate_pattern): Likewise.
22596 (write_header): Likewise.
22598 * gensupport.cc (change_subst_attribute): Likewise.
22599 (traverse_c_tests): Likewise.
22600 (add_predicate): Likewise.
22601 (init_predicate_table): Likewise.
22602 * gensupport.h (struct optab_pattern): Likewise.
22603 (get_num_insn_codes): Likewise.
22604 (maybe_eval_c_test): Likewise.
22605 (struct pred_data): Likewise.
22606 * ggc-internal.h: Likewise.
22607 * gimple-fold.cc (maybe_fold_reference): Likewise.
22608 (get_range_strlen_tree): Likewise.
22609 * gimple-fold.h (gimple_stmt_integer_valued_real_p): Likewise.
22610 * gimple-low.cc: Likewise.
22611 * gimple-match-head.cc (directly_supported_p): Likewise.
22612 * gimple-pretty-print.h: Likewise.
22613 * gimple-ssa-sprintf.cc (format_percent): Likewise.
22614 (adjust_range_for_overflow): Likewise.
22615 * gimple-streamer.h: Likewise.
22616 * gimple.h (struct GTY): Likewise.
22617 (is_gimple_resx): Likewise.
22618 * gimplify.cc (gimplify_expr): Likewise.
22619 (gimplify_init_constructor): Likewise.
22620 (omp_construct_selector_matches): Likewise.
22621 (gimplify_omp_target_update): Likewise.
22622 (gimplify_omp_ordered): Likewise.
22623 (gimplify_va_arg_expr): Likewise.
22624 * graphite-isl-ast-to-gimple.cc (should_copy_to_new_region): Likewise.
22625 * haifa-sched.cc (increase_insn_priority): Likewise.
22626 (try_ready): Likewise.
22627 (sched_create_recovery_edges): Likewise.
22628 * ifcvt.cc (find_if_case_1): Likewise.
22629 (find_if_case_2): Likewise.
22630 * inchash.h: Likewise.
22631 * incpath.cc (add_env_var_paths): Likewise.
22632 * input.cc (dump_location_info): Likewise.
22633 (assert_loceq): Likewise.
22634 (test_lexer_string_locations_concatenation_1): Likewise.
22635 (test_lexer_string_locations_concatenation_2): Likewise.
22636 (test_lexer_string_locations_concatenation_3): Likewise.
22637 * input.h (BUILTINS_LOCATION): Likewise.
22638 (class string_concat_db): Likewise.
22639 * internal-fn.cc (expand_MUL_OVERFLOW): Likewise.
22640 (expand_LOOP_VECTORIZED): Likewise.
22641 * ipa-cp.cc (make_pass_ipa_cp): Likewise.
22642 * ipa-fnsummary.cc (remap_freqcounting_preds_after_dup): Likewise.
22643 (ipa_fn_summary_t::duplicate): Likewise.
22644 (make_pass_ipa_fn_summary): Likewise.
22645 * ipa-fnsummary.h (enum ipa_hints_vals): Likewise.
22646 * ipa-free-lang-data.cc (fld_simplified_type): Likewise.
22647 (free_lang_data_in_decl): Likewise.
22648 * ipa-inline.cc (compute_inlined_call_time): Likewise.
22649 (inline_always_inline_functions): Likewise.
22650 * ipa-inline.h (free_growth_caches): Likewise.
22651 (inline_account_function_p): Likewise.
22652 * ipa-modref.cc (modref_access_analysis::analyze_stmt): Likewise.
22653 (modref_eaf_analysis::analyze_ssa_name): Likewise.
22654 * ipa-param-manipulation.cc (ipa_param_body_adjustments::mark_dead_statements): Likewise.
22655 (ipa_param_body_adjustments::remap_with_debug_expressions): Likewise.
22656 * ipa-prop.cc (ipa_set_node_agg_value_chain): Likewise.
22657 * ipa-prop.h (IPA_UNDESCRIBED_USE): Likewise.
22658 (unadjusted_ptr_and_unit_offset): Likewise.
22659 * ipa-reference.cc (make_pass_ipa_reference): Likewise.
22660 * ipa-reference.h (GCC_IPA_REFERENCE_H): Likewise.
22661 * ipa-split.cc (consider_split): Likewise.
22662 * ipa-sra.cc (isra_read_node_info): Likewise.
22663 * ipa-utils.h (struct ipa_dfs_info): Likewise.
22664 (recursive_call_p): Likewise.
22665 (ipa_make_function_pure): Likewise.
22666 * ira-build.cc (ira_create_allocno): Likewise.
22667 (ira_flattening): Likewise.
22668 * ira-color.cc (do_coloring): Likewise.
22669 (update_curr_costs): Likewise.
22670 * ira-conflicts.cc (process_regs_for_copy): Likewise.
22671 * ira-int.h (struct ira_emit_data): Likewise.
22672 (ira_prohibited_mode_move_regs): Likewise.
22673 (ira_get_dup_out_num): Likewise.
22674 (ira_destroy): Likewise.
22675 (ira_tune_allocno_costs): Likewise.
22676 (ira_implicitly_set_insn_hard_regs): Likewise.
22677 (ira_build_conflicts): Likewise.
22678 (ira_color): Likewise.
22679 * ira-lives.cc (process_bb_node_lives): Likewise.
22680 * ira.cc (class ira_spilled_reg_stack_slot): Likewise.
22681 (setup_uniform_class_p): Likewise.
22682 (def_dominates_uses): Likewise.
22683 * ira.h (ira_nullify_asm_goto): Likewise.
22684 * langhooks.cc (lhd_post_options): Likewise.
22685 * langhooks.h (class substring_loc): Likewise.
22686 (struct lang_hooks_for_tree_inlining): Likewise.
22687 (struct lang_hooks_for_types): Likewise.
22688 (struct lang_hooks): Likewise.
22689 * libfuncs.h (synchronize_libfunc): Likewise.
22690 * loop-doloop.cc (doloop_condition_get): Likewise.
22691 * loop-init.cc (fix_loop_structure): Likewise.
22692 * loop-invariant.cc: Likewise.
22693 * lower-subreg.h: Likewise.
22694 * lra-constraints.cc (curr_insn_transform): Likewise.
22695 * lra-int.h (struct lra_insn_reg): Likewise.
22696 (lra_undo_inheritance): Likewise.
22697 (lra_setup_reload_pseudo_preferenced_hard_reg): Likewise.
22698 (lra_split_hard_reg_for): Likewise.
22699 (lra_coalesce): Likewise.
22700 (lra_final_code_change): Likewise.
22701 * lra-spills.cc (lra_final_code_change): Likewise.
22702 * lra.cc (lra_process_new_insns): Likewise.
22703 * lto-compress.h (struct lto_compression_stream): Likewise.
22704 * lto-streamer-out.cc (DFS::DFS_write_tree_body): Likewise.
22705 (write_symbol): Likewise.
22706 * lto-streamer.h (enum LTO_tags): Likewise.
22707 (lto_value_range_error): Likewise.
22708 (lto_append_block): Likewise.
22709 (lto_streamer_hooks_init): Likewise.
22710 (stream_read_tree_ref): Likewise.
22711 (lto_prepare_function_for_streaming): Likewise.
22712 (select_what_to_stream): Likewise.
22713 (omp_lto_input_declare_variant_alt): Likewise.
22714 (cl_optimization_stream_in): Likewise.
22715 * lto-wrapper.cc (append_compiler_options): Likewise.
22716 * machmode.def: Likewise.
22717 * machmode.h (struct int_n_data_t): Likewise.
22718 * main.cc (main): Likewise.
22719 * match.pd: Likewise.
22720 * omp-builtins.def (BUILT_IN_GOMP_CRITICAL_NAME_END): Likewise.
22721 (BUILT_IN_GOMP_LOOP_ULL_ORDERED_RUNTIME_NEXT): Likewise.
22722 * omp-expand.cc (expand_omp_atomic_fetch_op): Likewise.
22723 (make_pass_expand_omp_ssa): Likewise.
22724 * omp-low.cc (struct omp_context): Likewise.
22725 (struct omp_taskcopy_context): Likewise.
22726 (lower_omp): Likewise.
22727 * omp-oacc-neuter-broadcast.cc (omp_sese_active_worker_call): Likewise.
22728 (mask_name): Likewise.
22729 (omp_sese_dump_pars): Likewise.
22730 (worker_single_simple): Likewise.
22731 * omp-offload.cc (omp_finish_file): Likewise.
22732 (execute_oacc_loop_designation): Likewise.
22733 * optabs-query.cc (lshift_cheap_p): Likewise.
22734 * optc-gen.awk: Likewise.
22735 * optc-save-gen.awk: Likewise.
22736 * optinfo-emit-json.cc (optrecord_json_writer::optrecord_json_writer): Likewise.
22737 * opts-common.cc: Likewise.
22738 * output.h (app_enable): Likewise.
22739 (output_operand_lossage): Likewise.
22740 (insn_current_reference_address): Likewise.
22741 (get_insn_template): Likewise.
22742 (output_quoted_string): Likewise.
22743 * pass_manager.h (struct register_pass_info): Likewise.
22744 * plugin.cc: Likewise.
22745 * plugin.def (PLUGIN_ANALYZER_INIT): Likewise.
22746 * plugin.h (invoke_plugin_callbacks): Likewise.
22747 * pointer-query.cc (handle_mem_ref): Likewise.
22748 * postreload-gcse.cc (alloc_mem): Likewise.
22749 * predict.h (enum prediction): Likewise.
22750 (add_reg_br_prob_note): Likewise.
22751 * prefix.h: Likewise.
22752 * profile.h (get_working_sets): Likewise.
22753 * read-md.cc: Likewise.
22754 * read-md.h (struct mapping): Likewise.
22755 (class md_reader): Likewise.
22756 (class noop_reader): Likewise.
22757 * read-rtl-function.cc (function_reader::create_function): Likewise.
22758 (function_reader::extra_parsing_for_operand_code_0): Likewise.
22759 * read-rtl.cc (initialize_iterators): Likewise.
22760 * real.cc: Likewise.
22761 * real.h (struct real_value): Likewise.
22762 (format_helper::format_helper): Likewise.
22763 (real_hash): Likewise.
22764 (real_can_shorten_arithmetic): Likewise.
22765 * recog.cc (struct target_recog): Likewise.
22766 (offsettable_nonstrict_memref_p): Likewise.
22767 (constrain_operands): Likewise.
22768 * recog.h (MAX_RECOG_ALTERNATIVES): Likewise.
22769 (which_op_alt): Likewise.
22770 (struct insn_gen_fn): Likewise.
22771 * reg-notes.def (REG_NOTE): Likewise.
22772 * reg-stack.cc: Likewise.
22773 * regs.h (reg_is_parm_p): Likewise.
22774 * regset.h: Likewise.
22775 * reload.cc (push_reload): Likewise.
22776 (find_reloads): Likewise.
22777 (find_reloads_address_1): Likewise.
22778 (find_replacement): Likewise.
22779 (refers_to_regno_for_reload_p): Likewise.
22780 (refers_to_mem_for_reload_p): Likewise.
22781 * reload.h (push_reload): Likewise.
22782 (deallocate_reload_reg): Likewise.
22783 * reload1.cc (emit_input_reload_insns): Likewise.
22784 * reorg.cc (relax_delay_slots): Likewise.
22785 * rtl.def (UNKNOWN): Likewise.
22786 (SEQUENCE): Likewise.
22787 (BARRIER): Likewise.
22788 (ASM_OPERANDS): Likewise.
22789 (EQ_ATTR_ALT): Likewise.
22790 * rtl.h (struct GTY): Likewise.
22791 (LABEL_NAME): Likewise.
22792 (LABEL_ALT_ENTRY_P): Likewise.
22793 (SUBREG_BYTE): Likewise.
22794 (get_stack_check_protect): Likewise.
22795 (dump_rtx_statistics): Likewise.
22796 (unwrap_const_vec_duplicate): Likewise.
22797 (subreg_promoted_mode): Likewise.
22798 (gen_lowpart_common): Likewise.
22799 (operand_subword): Likewise.
22800 (immed_wide_int_const): Likewise.
22801 (decide_function_section): Likewise.
22802 (active_insn_p): Likewise.
22803 (delete_related_insns): Likewise.
22804 (try_split): Likewise.
22805 (val_signbit_known_clear_p): Likewise.
22806 (simplifiable_subregs): Likewise.
22807 (set_insn_deleted): Likewise.
22808 (subreg_get_info): Likewise.
22809 (remove_free_EXPR_LIST_node): Likewise.
22810 (finish_subregs_of_mode): Likewise.
22811 (get_mem_attrs): Likewise.
22812 (lookup_constant_def): Likewise.
22813 (rtx_to_tree_code): Likewise.
22814 (hash_rtx): Likewise.
22815 (condjump_in_parallel_p): Likewise.
22816 (validate_subreg): Likewise.
22817 (make_compound_operation): Likewise.
22818 (schedule_ebbs): Likewise.
22819 (print_inline_rtx): Likewise.
22820 (fixup_args_size_notes): Likewise.
22821 (expand_dec): Likewise.
22822 (prepare_copy_insn): Likewise.
22823 (mark_elimination): Likewise.
22824 (valid_mode_changes_for_regno): Likewise.
22825 (make_debug_expr_from_rtl): Likewise.
22826 (delete_vta_debug_insns): Likewise.
22827 (simplify_using_condition): Likewise.
22828 (set_insn_locations): Likewise.
22829 (fatal_insn_not_found): Likewise.
22830 (word_register_operation_p): Likewise.
22831 * rtlanal.cc (get_call_fndecl): Likewise.
22832 (side_effects_p): Likewise.
22833 (subreg_nregs): Likewise.
22834 (rtx_cost): Likewise.
22835 (canonicalize_condition): Likewise.
22836 * rtlanal.h (rtx_properties::try_to_add_note): Likewise.
22837 * run-rtl-passes.cc (run_rtl_passes): Likewise.
22838 * sanitizer.def (BUILT_IN_ASAN_VERSION_MISMATCH_CHECK): Likewise.
22839 * sched-deps.cc (add_dependence_1): Likewise.
22840 * sched-ebb.cc (begin_move_insn): Likewise.
22841 (add_deps_for_risky_insns): Likewise.
22842 (advance_target_bb): Likewise.
22843 * sched-int.h (reemit_notes): Likewise.
22844 (struct _haifa_insn_data): Likewise.
22846 (DEP_CANCELLED): Likewise.
22847 (debug_ds): Likewise.
22848 (number_in_ready): Likewise.
22849 (schedule_ebbs_finish): Likewise.
22850 (find_modifiable_mems): Likewise.
22851 * sched-rgn.cc (debug_rgn_dependencies): Likewise.
22852 * sel-sched-dump.cc (dump_lv_set): Likewise.
22853 * sel-sched-dump.h: Likewise.
22854 * sel-sched-ir.cc (sel_insn_rtx_cost): Likewise.
22855 (setup_id_reg_sets): Likewise.
22856 (has_dependence_p): Likewise.
22857 (sel_num_cfg_preds_gt_1): Likewise.
22858 (bb_ends_ebb_p): Likewise.
22859 * sel-sched-ir.h (struct _list_node): Likewise.
22860 (struct idata_def): Likewise.
22861 (bb_next_bb): Likewise.
22862 * sel-sched.cc (vinsn_writes_one_of_regs_p): Likewise.
22863 (choose_best_pseudo_reg): Likewise.
22864 (verify_target_availability): Likewise.
22865 (can_speculate_dep_p): Likewise.
22866 (sel_rank_for_schedule): Likewise.
22867 * selftest-run-tests.cc (selftest::run_tests): Likewise.
22868 * selftest.h (class auto_fix_quotes): Likewise.
22869 * shrink-wrap.cc (handle_simple_exit): Likewise.
22870 * shrink-wrap.h: Likewise.
22871 * simplify-rtx.cc (simplify_context::simplify_associative_operation): Likewise.
22872 (simplify_context::simplify_gen_vec_select): Likewise.
22873 * spellcheck-tree.h: Likewise.
22874 * spellcheck.h: Likewise.
22875 * statistics.h (struct function): Likewise.
22876 * stmt.cc (conditional_probability): Likewise.
22877 * stmt.h: Likewise.
22878 * stor-layout.h: Likewise.
22879 * streamer-hooks.h: Likewise.
22880 * stringpool.h: Likewise.
22881 * symtab.cc (symbol_table::change_decl_assembler_name): Likewise.
22882 * target.def (HOOK_VECTOR_END): Likewise.
22884 * target.h (union cumulative_args_t): Likewise.
22885 (by_pieces_ninsns): Likewise.
22886 (class predefined_function_abi): Likewise.
22887 * targhooks.cc (default_translate_mode_attribute): Likewise.
22888 * timevar.def: Likewise.
22889 * timevar.h (class timer): Likewise.
22890 * toplev.h (enable_rtl_dump_file): Likewise.
22891 * trans-mem.cc (collect_bb2reg): Likewise.
22892 * tree-call-cdce.cc (gen_conditions_for_pow): Likewise.
22893 * tree-cfg.cc (remove_bb): Likewise.
22894 (verify_gimple_debug): Likewise.
22895 (remove_edge_and_dominated_blocks): Likewise.
22896 (push_fndecl): Likewise.
22897 * tree-cfgcleanup.h (GCC_TREE_CFGCLEANUP_H): Likewise.
22898 * tree-complex.cc (expand_complex_multiplication): Likewise.
22899 (expand_complex_div_straight): Likewise.
22900 * tree-core.h (enum tree_index): Likewise.
22901 (enum operand_equal_flag): Likewise.
22902 * tree-eh.cc (honor_protect_cleanup_actions): Likewise.
22903 * tree-if-conv.cc (if_convertible_gimple_assign_stmt_p): Likewise.
22904 * tree-inline.cc (initialize_inlined_parameters): Likewise.
22905 * tree-inline.h (force_value_to_type): Likewise.
22906 * tree-nested.cc (get_chain_decl): Likewise.
22907 (walk_all_functions): Likewise.
22908 * tree-object-size.h: Likewise.
22909 * tree-outof-ssa.cc: Likewise.
22910 * tree-parloops.cc (create_parallel_loop): Likewise.
22911 * tree-pretty-print.cc (print_generic_expr_to_str): Likewise.
22912 (dump_generic_node): Likewise.
22913 * tree-profile.cc (tree_profiling): Likewise.
22914 * tree-sra.cc (maybe_add_sra_candidate): Likewise.
22915 * tree-ssa-address.cc: Likewise.
22916 * tree-ssa-alias.cc: Likewise.
22917 * tree-ssa-alias.h (ao_ref::max_size_known_p): Likewise.
22918 (dump_alias_stats): Likewise.
22919 * tree-ssa-ccp.cc: Likewise.
22920 * tree-ssa-coalesce.h: Likewise.
22921 * tree-ssa-live.cc (remove_unused_scope_block_p): Likewise.
22922 * tree-ssa-loop-manip.cc (copy_phi_node_args): Likewise.
22923 * tree-ssa-loop-unswitch.cc: Likewise.
22924 * tree-ssa-math-opts.cc: Likewise.
22925 * tree-ssa-operands.cc (class operands_scanner): Likewise.
22926 * tree-ssa-pre.cc: Likewise.
22927 * tree-ssa-reassoc.cc (optimize_ops_list): Likewise.
22928 (debug_range_entry): Likewise.
22929 * tree-ssa-sccvn.cc (eliminate_dom_walker::eliminate_stmt): Likewise.
22930 * tree-ssa-sccvn.h (TREE_SSA_SCCVN_H): Likewise.
22931 * tree-ssa-scopedtables.cc (add_expr_commutative): Likewise.
22932 (equal_mem_array_ref_p): Likewise.
22933 * tree-ssa-strlen.cc (is_strlen_related_p): Likewise.
22934 * tree-ssa-strlen.h (get_range_strlen_dynamic): Likewise.
22935 * tree-ssa-tail-merge.cc (stmt_local_def): Likewise.
22936 * tree-ssa-ter.h: Likewise.
22937 * tree-ssa-threadupdate.h (enum bb_dom_status): Likewise.
22938 * tree-streamer-in.cc (lto_input_ts_block_tree_pointers): Likewise.
22939 * tree-streamer-out.cc (pack_ts_block_value_fields): Likewise.
22940 (write_ts_block_tree_pointers): Likewise.
22941 * tree-streamer.h (struct streamer_tree_cache_d): Likewise.
22942 (streamer_read_tree_bitfields): Likewise.
22943 (streamer_write_integer_cst): Likewise.
22944 * tree-vect-patterns.cc (apply_binop_and_append_stmt): Likewise.
22945 (vect_synth_mult_by_constant): Likewise.
22946 * tree-vect-stmts.cc (vectorizable_operation): Likewise.
22947 * tree-vectorizer.cc: Likewise.
22948 * tree-vectorizer.h (class auto_purge_vect_location): Likewise.
22949 (vect_update_inits_of_drs): Likewise.
22950 (vect_get_mask_type_for_stmt): Likewise.
22951 (vect_rgroup_iv_might_wrap_p): Likewise.
22952 (cse_and_gimplify_to_preheader): Likewise.
22953 (vect_free_slp_tree): Likewise.
22954 (vect_pattern_recog): Likewise.
22955 (vect_stmt_dominates_stmt_p): Likewise.
22956 * tree.cc (initialize_tree_contains_struct): Likewise.
22957 (need_assembler_name_p): Likewise.
22958 (type_with_interoperable_signedness): Likewise.
22959 * tree.def (SWITCH_EXPR): Likewise.
22960 * tree.h (TYPE_SYMTAB_ADDRESS): Likewise.
22961 (poly_int_tree_p): Likewise.
22962 (inlined_function_outer_scope_p): Likewise.
22963 (tree_code_for_canonical_type_merging): Likewise.
22964 * value-prof.cc: Likewise.
22965 * value-prof.h (get_nth_most_common_value): Likewise.
22966 (find_func_by_profile_id): Likewise.
22967 * value-range.cc (vrp_operand_equal_p): Likewise.
22968 * value-range.h: Likewise.
22969 * var-tracking.cc: Likewise.
22970 * varasm.cc (default_function_section): Likewise.
22971 (function_section_1): Likewise.
22972 (assemble_variable): Likewise.
22973 (handle_vtv_comdat_section): Likewise.
22974 * vec.h (struct vec_prefix): Likewise.
22975 * vmsdbgout.cc (full_name): Likewise.
22976 * vtable-verify.cc: Likewise.
22977 * vtable-verify.h (struct vtv_graph_node): Likewise.
22978 * xcoffout.cc: Likewise.
22979 * xcoffout.h (DEBUG_SYMS_TEXT): Likewise.
22981 2022-01-17 Martin Liska <mliska@suse.cz>
22983 * adjust-alignment.c: Moved to...
22984 * adjust-alignment.cc: ...here.
22985 * alias.c: Moved to...
22986 * alias.cc: ...here.
22987 * alloc-pool.c: Moved to...
22988 * alloc-pool.cc: ...here.
22989 * asan.c: Moved to...
22990 * asan.cc: ...here.
22991 * attribs.c: Moved to...
22992 * attribs.cc: ...here.
22993 * auto-inc-dec.c: Moved to...
22994 * auto-inc-dec.cc: ...here.
22995 * auto-profile.c: Moved to...
22996 * auto-profile.cc: ...here.
22997 * bb-reorder.c: Moved to...
22998 * bb-reorder.cc: ...here.
22999 * bitmap.c: Moved to...
23000 * bitmap.cc: ...here.
23001 * btfout.c: Moved to...
23002 * btfout.cc: ...here.
23003 * builtins.c: Moved to...
23004 * builtins.cc: ...here.
23005 * caller-save.c: Moved to...
23006 * caller-save.cc: ...here.
23007 * calls.c: Moved to...
23008 * calls.cc: ...here.
23009 * ccmp.c: Moved to...
23010 * ccmp.cc: ...here.
23011 * cfg.c: Moved to...
23013 * cfganal.c: Moved to...
23014 * cfganal.cc: ...here.
23015 * cfgbuild.c: Moved to...
23016 * cfgbuild.cc: ...here.
23017 * cfgcleanup.c: Moved to...
23018 * cfgcleanup.cc: ...here.
23019 * cfgexpand.c: Moved to...
23020 * cfgexpand.cc: ...here.
23021 * cfghooks.c: Moved to...
23022 * cfghooks.cc: ...here.
23023 * cfgloop.c: Moved to...
23024 * cfgloop.cc: ...here.
23025 * cfgloopanal.c: Moved to...
23026 * cfgloopanal.cc: ...here.
23027 * cfgloopmanip.c: Moved to...
23028 * cfgloopmanip.cc: ...here.
23029 * cfgrtl.c: Moved to...
23030 * cfgrtl.cc: ...here.
23031 * cgraph.c: Moved to...
23032 * cgraph.cc: ...here.
23033 * cgraphbuild.c: Moved to...
23034 * cgraphbuild.cc: ...here.
23035 * cgraphclones.c: Moved to...
23036 * cgraphclones.cc: ...here.
23037 * cgraphunit.c: Moved to...
23038 * cgraphunit.cc: ...here.
23039 * collect-utils.c: Moved to...
23040 * collect-utils.cc: ...here.
23041 * collect2-aix.c: Moved to...
23042 * collect2-aix.cc: ...here.
23043 * collect2.c: Moved to...
23044 * collect2.cc: ...here.
23045 * combine-stack-adj.c: Moved to...
23046 * combine-stack-adj.cc: ...here.
23047 * combine.c: Moved to...
23048 * combine.cc: ...here.
23049 * common/common-targhooks.c: Moved to...
23050 * common/common-targhooks.cc: ...here.
23051 * common/config/aarch64/aarch64-common.c: Moved to...
23052 * common/config/aarch64/aarch64-common.cc: ...here.
23053 * common/config/alpha/alpha-common.c: Moved to...
23054 * common/config/alpha/alpha-common.cc: ...here.
23055 * common/config/arc/arc-common.c: Moved to...
23056 * common/config/arc/arc-common.cc: ...here.
23057 * common/config/arm/arm-common.c: Moved to...
23058 * common/config/arm/arm-common.cc: ...here.
23059 * common/config/avr/avr-common.c: Moved to...
23060 * common/config/avr/avr-common.cc: ...here.
23061 * common/config/bfin/bfin-common.c: Moved to...
23062 * common/config/bfin/bfin-common.cc: ...here.
23063 * common/config/bpf/bpf-common.c: Moved to...
23064 * common/config/bpf/bpf-common.cc: ...here.
23065 * common/config/c6x/c6x-common.c: Moved to...
23066 * common/config/c6x/c6x-common.cc: ...here.
23067 * common/config/cr16/cr16-common.c: Moved to...
23068 * common/config/cr16/cr16-common.cc: ...here.
23069 * common/config/cris/cris-common.c: Moved to...
23070 * common/config/cris/cris-common.cc: ...here.
23071 * common/config/csky/csky-common.c: Moved to...
23072 * common/config/csky/csky-common.cc: ...here.
23073 * common/config/default-common.c: Moved to...
23074 * common/config/default-common.cc: ...here.
23075 * common/config/epiphany/epiphany-common.c: Moved to...
23076 * common/config/epiphany/epiphany-common.cc: ...here.
23077 * common/config/fr30/fr30-common.c: Moved to...
23078 * common/config/fr30/fr30-common.cc: ...here.
23079 * common/config/frv/frv-common.c: Moved to...
23080 * common/config/frv/frv-common.cc: ...here.
23081 * common/config/gcn/gcn-common.c: Moved to...
23082 * common/config/gcn/gcn-common.cc: ...here.
23083 * common/config/h8300/h8300-common.c: Moved to...
23084 * common/config/h8300/h8300-common.cc: ...here.
23085 * common/config/i386/i386-common.c: Moved to...
23086 * common/config/i386/i386-common.cc: ...here.
23087 * common/config/ia64/ia64-common.c: Moved to...
23088 * common/config/ia64/ia64-common.cc: ...here.
23089 * common/config/iq2000/iq2000-common.c: Moved to...
23090 * common/config/iq2000/iq2000-common.cc: ...here.
23091 * common/config/lm32/lm32-common.c: Moved to...
23092 * common/config/lm32/lm32-common.cc: ...here.
23093 * common/config/m32r/m32r-common.c: Moved to...
23094 * common/config/m32r/m32r-common.cc: ...here.
23095 * common/config/m68k/m68k-common.c: Moved to...
23096 * common/config/m68k/m68k-common.cc: ...here.
23097 * common/config/mcore/mcore-common.c: Moved to...
23098 * common/config/mcore/mcore-common.cc: ...here.
23099 * common/config/microblaze/microblaze-common.c: Moved to...
23100 * common/config/microblaze/microblaze-common.cc: ...here.
23101 * common/config/mips/mips-common.c: Moved to...
23102 * common/config/mips/mips-common.cc: ...here.
23103 * common/config/mmix/mmix-common.c: Moved to...
23104 * common/config/mmix/mmix-common.cc: ...here.
23105 * common/config/mn10300/mn10300-common.c: Moved to...
23106 * common/config/mn10300/mn10300-common.cc: ...here.
23107 * common/config/msp430/msp430-common.c: Moved to...
23108 * common/config/msp430/msp430-common.cc: ...here.
23109 * common/config/nds32/nds32-common.c: Moved to...
23110 * common/config/nds32/nds32-common.cc: ...here.
23111 * common/config/nios2/nios2-common.c: Moved to...
23112 * common/config/nios2/nios2-common.cc: ...here.
23113 * common/config/nvptx/nvptx-common.c: Moved to...
23114 * common/config/nvptx/nvptx-common.cc: ...here.
23115 * common/config/or1k/or1k-common.c: Moved to...
23116 * common/config/or1k/or1k-common.cc: ...here.
23117 * common/config/pa/pa-common.c: Moved to...
23118 * common/config/pa/pa-common.cc: ...here.
23119 * common/config/pdp11/pdp11-common.c: Moved to...
23120 * common/config/pdp11/pdp11-common.cc: ...here.
23121 * common/config/pru/pru-common.c: Moved to...
23122 * common/config/pru/pru-common.cc: ...here.
23123 * common/config/riscv/riscv-common.c: Moved to...
23124 * common/config/riscv/riscv-common.cc: ...here.
23125 * common/config/rs6000/rs6000-common.c: Moved to...
23126 * common/config/rs6000/rs6000-common.cc: ...here.
23127 * common/config/rx/rx-common.c: Moved to...
23128 * common/config/rx/rx-common.cc: ...here.
23129 * common/config/s390/s390-common.c: Moved to...
23130 * common/config/s390/s390-common.cc: ...here.
23131 * common/config/sh/sh-common.c: Moved to...
23132 * common/config/sh/sh-common.cc: ...here.
23133 * common/config/sparc/sparc-common.c: Moved to...
23134 * common/config/sparc/sparc-common.cc: ...here.
23135 * common/config/tilegx/tilegx-common.c: Moved to...
23136 * common/config/tilegx/tilegx-common.cc: ...here.
23137 * common/config/tilepro/tilepro-common.c: Moved to...
23138 * common/config/tilepro/tilepro-common.cc: ...here.
23139 * common/config/v850/v850-common.c: Moved to...
23140 * common/config/v850/v850-common.cc: ...here.
23141 * common/config/vax/vax-common.c: Moved to...
23142 * common/config/vax/vax-common.cc: ...here.
23143 * common/config/visium/visium-common.c: Moved to...
23144 * common/config/visium/visium-common.cc: ...here.
23145 * common/config/xstormy16/xstormy16-common.c: Moved to...
23146 * common/config/xstormy16/xstormy16-common.cc: ...here.
23147 * common/config/xtensa/xtensa-common.c: Moved to...
23148 * common/config/xtensa/xtensa-common.cc: ...here.
23149 * compare-elim.c: Moved to...
23150 * compare-elim.cc: ...here.
23151 * config/aarch64/aarch64-bti-insert.c: Moved to...
23152 * config/aarch64/aarch64-bti-insert.cc: ...here.
23153 * config/aarch64/aarch64-builtins.c: Moved to...
23154 * config/aarch64/aarch64-builtins.cc: ...here.
23155 * config/aarch64/aarch64-c.c: Moved to...
23156 * config/aarch64/aarch64-c.cc: ...here.
23157 * config/aarch64/aarch64-d.c: Moved to...
23158 * config/aarch64/aarch64-d.cc: ...here.
23159 * config/aarch64/aarch64.c: Moved to...
23160 * config/aarch64/aarch64.cc: ...here.
23161 * config/aarch64/cortex-a57-fma-steering.c: Moved to...
23162 * config/aarch64/cortex-a57-fma-steering.cc: ...here.
23163 * config/aarch64/driver-aarch64.c: Moved to...
23164 * config/aarch64/driver-aarch64.cc: ...here.
23165 * config/aarch64/falkor-tag-collision-avoidance.c: Moved to...
23166 * config/aarch64/falkor-tag-collision-avoidance.cc: ...here.
23167 * config/aarch64/host-aarch64-darwin.c: Moved to...
23168 * config/aarch64/host-aarch64-darwin.cc: ...here.
23169 * config/alpha/alpha.c: Moved to...
23170 * config/alpha/alpha.cc: ...here.
23171 * config/alpha/driver-alpha.c: Moved to...
23172 * config/alpha/driver-alpha.cc: ...here.
23173 * config/arc/arc-c.c: Moved to...
23174 * config/arc/arc-c.cc: ...here.
23175 * config/arc/arc.c: Moved to...
23176 * config/arc/arc.cc: ...here.
23177 * config/arc/driver-arc.c: Moved to...
23178 * config/arc/driver-arc.cc: ...here.
23179 * config/arm/aarch-common.c: Moved to...
23180 * config/arm/aarch-common.cc: ...here.
23181 * config/arm/arm-builtins.c: Moved to...
23182 * config/arm/arm-builtins.cc: ...here.
23183 * config/arm/arm-c.c: Moved to...
23184 * config/arm/arm-c.cc: ...here.
23185 * config/arm/arm-d.c: Moved to...
23186 * config/arm/arm-d.cc: ...here.
23187 * config/arm/arm.c: Moved to...
23188 * config/arm/arm.cc: ...here.
23189 * config/arm/driver-arm.c: Moved to...
23190 * config/arm/driver-arm.cc: ...here.
23191 * config/avr/avr-c.c: Moved to...
23192 * config/avr/avr-c.cc: ...here.
23193 * config/avr/avr-devices.c: Moved to...
23194 * config/avr/avr-devices.cc: ...here.
23195 * config/avr/avr-log.c: Moved to...
23196 * config/avr/avr-log.cc: ...here.
23197 * config/avr/avr.c: Moved to...
23198 * config/avr/avr.cc: ...here.
23199 * config/avr/driver-avr.c: Moved to...
23200 * config/avr/driver-avr.cc: ...here.
23201 * config/avr/gen-avr-mmcu-specs.c: Moved to...
23202 * config/avr/gen-avr-mmcu-specs.cc: ...here.
23203 * config/avr/gen-avr-mmcu-texi.c: Moved to...
23204 * config/avr/gen-avr-mmcu-texi.cc: ...here.
23205 * config/bfin/bfin.c: Moved to...
23206 * config/bfin/bfin.cc: ...here.
23207 * config/bpf/bpf.c: Moved to...
23208 * config/bpf/bpf.cc: ...here.
23209 * config/bpf/coreout.c: Moved to...
23210 * config/bpf/coreout.cc: ...here.
23211 * config/c6x/c6x.c: Moved to...
23212 * config/c6x/c6x.cc: ...here.
23213 * config/cr16/cr16.c: Moved to...
23214 * config/cr16/cr16.cc: ...here.
23215 * config/cris/cris.c: Moved to...
23216 * config/cris/cris.cc: ...here.
23217 * config/csky/csky.c: Moved to...
23218 * config/csky/csky.cc: ...here.
23219 * config/darwin-c.c: Moved to...
23220 * config/darwin-c.cc: ...here.
23221 * config/darwin-d.c: Moved to...
23222 * config/darwin-d.cc: ...here.
23223 * config/darwin-driver.c: Moved to...
23224 * config/darwin-driver.cc: ...here.
23225 * config/darwin-f.c: Moved to...
23226 * config/darwin-f.cc: ...here.
23227 * config/darwin.c: Moved to...
23228 * config/darwin.cc: ...here.
23229 * config/default-c.c: Moved to...
23230 * config/default-c.cc: ...here.
23231 * config/default-d.c: Moved to...
23232 * config/default-d.cc: ...here.
23233 * config/dragonfly-d.c: Moved to...
23234 * config/dragonfly-d.cc: ...here.
23235 * config/epiphany/epiphany.c: Moved to...
23236 * config/epiphany/epiphany.cc: ...here.
23237 * config/epiphany/mode-switch-use.c: Moved to...
23238 * config/epiphany/mode-switch-use.cc: ...here.
23239 * config/epiphany/resolve-sw-modes.c: Moved to...
23240 * config/epiphany/resolve-sw-modes.cc: ...here.
23241 * config/fr30/fr30.c: Moved to...
23242 * config/fr30/fr30.cc: ...here.
23243 * config/freebsd-d.c: Moved to...
23244 * config/freebsd-d.cc: ...here.
23245 * config/frv/frv.c: Moved to...
23246 * config/frv/frv.cc: ...here.
23247 * config/ft32/ft32.c: Moved to...
23248 * config/ft32/ft32.cc: ...here.
23249 * config/gcn/driver-gcn.c: Moved to...
23250 * config/gcn/driver-gcn.cc: ...here.
23251 * config/gcn/gcn-run.c: Moved to...
23252 * config/gcn/gcn-run.cc: ...here.
23253 * config/gcn/gcn-tree.c: Moved to...
23254 * config/gcn/gcn-tree.cc: ...here.
23255 * config/gcn/gcn.c: Moved to...
23256 * config/gcn/gcn.cc: ...here.
23257 * config/gcn/mkoffload.c: Moved to...
23258 * config/gcn/mkoffload.cc: ...here.
23259 * config/glibc-c.c: Moved to...
23260 * config/glibc-c.cc: ...here.
23261 * config/glibc-d.c: Moved to...
23262 * config/glibc-d.cc: ...here.
23263 * config/h8300/h8300.c: Moved to...
23264 * config/h8300/h8300.cc: ...here.
23265 * config/host-darwin.c: Moved to...
23266 * config/host-darwin.cc: ...here.
23267 * config/host-hpux.c: Moved to...
23268 * config/host-hpux.cc: ...here.
23269 * config/host-linux.c: Moved to...
23270 * config/host-linux.cc: ...here.
23271 * config/host-netbsd.c: Moved to...
23272 * config/host-netbsd.cc: ...here.
23273 * config/host-openbsd.c: Moved to...
23274 * config/host-openbsd.cc: ...here.
23275 * config/host-solaris.c: Moved to...
23276 * config/host-solaris.cc: ...here.
23277 * config/i386/djgpp.c: Moved to...
23278 * config/i386/djgpp.cc: ...here.
23279 * config/i386/driver-i386.c: Moved to...
23280 * config/i386/driver-i386.cc: ...here.
23281 * config/i386/driver-mingw32.c: Moved to...
23282 * config/i386/driver-mingw32.cc: ...here.
23283 * config/i386/gnu-property.c: Moved to...
23284 * config/i386/gnu-property.cc: ...here.
23285 * config/i386/host-cygwin.c: Moved to...
23286 * config/i386/host-cygwin.cc: ...here.
23287 * config/i386/host-i386-darwin.c: Moved to...
23288 * config/i386/host-i386-darwin.cc: ...here.
23289 * config/i386/host-mingw32.c: Moved to...
23290 * config/i386/host-mingw32.cc: ...here.
23291 * config/i386/i386-builtins.c: Moved to...
23292 * config/i386/i386-builtins.cc: ...here.
23293 * config/i386/i386-c.c: Moved to...
23294 * config/i386/i386-c.cc: ...here.
23295 * config/i386/i386-d.c: Moved to...
23296 * config/i386/i386-d.cc: ...here.
23297 * config/i386/i386-expand.c: Moved to...
23298 * config/i386/i386-expand.cc: ...here.
23299 * config/i386/i386-features.c: Moved to...
23300 * config/i386/i386-features.cc: ...here.
23301 * config/i386/i386-options.c: Moved to...
23302 * config/i386/i386-options.cc: ...here.
23303 * config/i386/i386.c: Moved to...
23304 * config/i386/i386.cc: ...here.
23305 * config/i386/intelmic-mkoffload.c: Moved to...
23306 * config/i386/intelmic-mkoffload.cc: ...here.
23307 * config/i386/msformat-c.c: Moved to...
23308 * config/i386/msformat-c.cc: ...here.
23309 * config/i386/winnt-cxx.c: Moved to...
23310 * config/i386/winnt-cxx.cc: ...here.
23311 * config/i386/winnt-d.c: Moved to...
23312 * config/i386/winnt-d.cc: ...here.
23313 * config/i386/winnt-stubs.c: Moved to...
23314 * config/i386/winnt-stubs.cc: ...here.
23315 * config/i386/winnt.c: Moved to...
23316 * config/i386/winnt.cc: ...here.
23317 * config/i386/x86-tune-sched-atom.c: Moved to...
23318 * config/i386/x86-tune-sched-atom.cc: ...here.
23319 * config/i386/x86-tune-sched-bd.c: Moved to...
23320 * config/i386/x86-tune-sched-bd.cc: ...here.
23321 * config/i386/x86-tune-sched-core.c: Moved to...
23322 * config/i386/x86-tune-sched-core.cc: ...here.
23323 * config/i386/x86-tune-sched.c: Moved to...
23324 * config/i386/x86-tune-sched.cc: ...here.
23325 * config/ia64/ia64-c.c: Moved to...
23326 * config/ia64/ia64-c.cc: ...here.
23327 * config/ia64/ia64.c: Moved to...
23328 * config/ia64/ia64.cc: ...here.
23329 * config/iq2000/iq2000.c: Moved to...
23330 * config/iq2000/iq2000.cc: ...here.
23331 * config/linux.c: Moved to...
23332 * config/linux.cc: ...here.
23333 * config/lm32/lm32.c: Moved to...
23334 * config/lm32/lm32.cc: ...here.
23335 * config/m32c/m32c-pragma.c: Moved to...
23336 * config/m32c/m32c-pragma.cc: ...here.
23337 * config/m32c/m32c.c: Moved to...
23338 * config/m32c/m32c.cc: ...here.
23339 * config/m32r/m32r.c: Moved to...
23340 * config/m32r/m32r.cc: ...here.
23341 * config/m68k/m68k.c: Moved to...
23342 * config/m68k/m68k.cc: ...here.
23343 * config/mcore/mcore.c: Moved to...
23344 * config/mcore/mcore.cc: ...here.
23345 * config/microblaze/microblaze-c.c: Moved to...
23346 * config/microblaze/microblaze-c.cc: ...here.
23347 * config/microblaze/microblaze.c: Moved to...
23348 * config/microblaze/microblaze.cc: ...here.
23349 * config/mips/driver-native.c: Moved to...
23350 * config/mips/driver-native.cc: ...here.
23351 * config/mips/frame-header-opt.c: Moved to...
23352 * config/mips/frame-header-opt.cc: ...here.
23353 * config/mips/mips-d.c: Moved to...
23354 * config/mips/mips-d.cc: ...here.
23355 * config/mips/mips.c: Moved to...
23356 * config/mips/mips.cc: ...here.
23357 * config/mmix/mmix.c: Moved to...
23358 * config/mmix/mmix.cc: ...here.
23359 * config/mn10300/mn10300.c: Moved to...
23360 * config/mn10300/mn10300.cc: ...here.
23361 * config/moxie/moxie.c: Moved to...
23362 * config/moxie/moxie.cc: ...here.
23363 * config/msp430/driver-msp430.c: Moved to...
23364 * config/msp430/driver-msp430.cc: ...here.
23365 * config/msp430/msp430-c.c: Moved to...
23366 * config/msp430/msp430-c.cc: ...here.
23367 * config/msp430/msp430-devices.c: Moved to...
23368 * config/msp430/msp430-devices.cc: ...here.
23369 * config/msp430/msp430.c: Moved to...
23370 * config/msp430/msp430.cc: ...here.
23371 * config/nds32/nds32-cost.c: Moved to...
23372 * config/nds32/nds32-cost.cc: ...here.
23373 * config/nds32/nds32-fp-as-gp.c: Moved to...
23374 * config/nds32/nds32-fp-as-gp.cc: ...here.
23375 * config/nds32/nds32-intrinsic.c: Moved to...
23376 * config/nds32/nds32-intrinsic.cc: ...here.
23377 * config/nds32/nds32-isr.c: Moved to...
23378 * config/nds32/nds32-isr.cc: ...here.
23379 * config/nds32/nds32-md-auxiliary.c: Moved to...
23380 * config/nds32/nds32-md-auxiliary.cc: ...here.
23381 * config/nds32/nds32-memory-manipulation.c: Moved to...
23382 * config/nds32/nds32-memory-manipulation.cc: ...here.
23383 * config/nds32/nds32-pipelines-auxiliary.c: Moved to...
23384 * config/nds32/nds32-pipelines-auxiliary.cc: ...here.
23385 * config/nds32/nds32-predicates.c: Moved to...
23386 * config/nds32/nds32-predicates.cc: ...here.
23387 * config/nds32/nds32-relax-opt.c: Moved to...
23388 * config/nds32/nds32-relax-opt.cc: ...here.
23389 * config/nds32/nds32-utils.c: Moved to...
23390 * config/nds32/nds32-utils.cc: ...here.
23391 * config/nds32/nds32.c: Moved to...
23392 * config/nds32/nds32.cc: ...here.
23393 * config/netbsd-d.c: Moved to...
23394 * config/netbsd-d.cc: ...here.
23395 * config/netbsd.c: Moved to...
23396 * config/netbsd.cc: ...here.
23397 * config/nios2/nios2.c: Moved to...
23398 * config/nios2/nios2.cc: ...here.
23399 * config/nvptx/mkoffload.c: Moved to...
23400 * config/nvptx/mkoffload.cc: ...here.
23401 * config/nvptx/nvptx-c.c: Moved to...
23402 * config/nvptx/nvptx-c.cc: ...here.
23403 * config/nvptx/nvptx.c: Moved to...
23404 * config/nvptx/nvptx.cc: ...here.
23405 * config/openbsd-d.c: Moved to...
23406 * config/openbsd-d.cc: ...here.
23407 * config/or1k/or1k.c: Moved to...
23408 * config/or1k/or1k.cc: ...here.
23409 * config/pa/pa-d.c: Moved to...
23410 * config/pa/pa-d.cc: ...here.
23411 * config/pa/pa.c: Moved to...
23412 * config/pa/pa.cc: ...here.
23413 * config/pdp11/pdp11.c: Moved to...
23414 * config/pdp11/pdp11.cc: ...here.
23415 * config/pru/pru-passes.c: Moved to...
23416 * config/pru/pru-passes.cc: ...here.
23417 * config/pru/pru-pragma.c: Moved to...
23418 * config/pru/pru-pragma.cc: ...here.
23419 * config/pru/pru.c: Moved to...
23420 * config/pru/pru.cc: ...here.
23421 * config/riscv/riscv-builtins.c: Moved to...
23422 * config/riscv/riscv-builtins.cc: ...here.
23423 * config/riscv/riscv-c.c: Moved to...
23424 * config/riscv/riscv-c.cc: ...here.
23425 * config/riscv/riscv-d.c: Moved to...
23426 * config/riscv/riscv-d.cc: ...here.
23427 * config/riscv/riscv-shorten-memrefs.c: Moved to...
23428 * config/riscv/riscv-shorten-memrefs.cc: ...here.
23429 * config/riscv/riscv-sr.c: Moved to...
23430 * config/riscv/riscv-sr.cc: ...here.
23431 * config/riscv/riscv.c: Moved to...
23432 * config/riscv/riscv.cc: ...here.
23433 * config/rl78/rl78-c.c: Moved to...
23434 * config/rl78/rl78-c.cc: ...here.
23435 * config/rl78/rl78.c: Moved to...
23436 * config/rl78/rl78.cc: ...here.
23437 * config/rs6000/driver-rs6000.c: Moved to...
23438 * config/rs6000/driver-rs6000.cc: ...here.
23439 * config/rs6000/host-darwin.c: Moved to...
23440 * config/rs6000/host-darwin.cc: ...here.
23441 * config/rs6000/host-ppc64-darwin.c: Moved to...
23442 * config/rs6000/host-ppc64-darwin.cc: ...here.
23443 * config/rs6000/rbtree.c: Moved to...
23444 * config/rs6000/rbtree.cc: ...here.
23445 * config/rs6000/rs6000-c.c: Moved to...
23446 * config/rs6000/rs6000-c.cc: ...here.
23447 * config/rs6000/rs6000-call.c: Moved to...
23448 * config/rs6000/rs6000-call.cc: ...here.
23449 * config/rs6000/rs6000-d.c: Moved to...
23450 * config/rs6000/rs6000-d.cc: ...here.
23451 * config/rs6000/rs6000-gen-builtins.c: Moved to...
23452 * config/rs6000/rs6000-gen-builtins.cc: ...here.
23453 * config/rs6000/rs6000-linux.c: Moved to...
23454 * config/rs6000/rs6000-linux.cc: ...here.
23455 * config/rs6000/rs6000-logue.c: Moved to...
23456 * config/rs6000/rs6000-logue.cc: ...here.
23457 * config/rs6000/rs6000-p8swap.c: Moved to...
23458 * config/rs6000/rs6000-p8swap.cc: ...here.
23459 * config/rs6000/rs6000-pcrel-opt.c: Moved to...
23460 * config/rs6000/rs6000-pcrel-opt.cc: ...here.
23461 * config/rs6000/rs6000-string.c: Moved to...
23462 * config/rs6000/rs6000-string.cc: ...here.
23463 * config/rs6000/rs6000.c: Moved to...
23464 * config/rs6000/rs6000.cc: ...here.
23465 * config/rx/rx.c: Moved to...
23466 * config/rx/rx.cc: ...here.
23467 * config/s390/driver-native.c: Moved to...
23468 * config/s390/driver-native.cc: ...here.
23469 * config/s390/s390-c.c: Moved to...
23470 * config/s390/s390-c.cc: ...here.
23471 * config/s390/s390-d.c: Moved to...
23472 * config/s390/s390-d.cc: ...here.
23473 * config/s390/s390.c: Moved to...
23474 * config/s390/s390.cc: ...here.
23475 * config/sh/divtab-sh4-300.c: Moved to...
23476 * config/sh/divtab-sh4-300.cc: ...here.
23477 * config/sh/divtab-sh4.c: Moved to...
23478 * config/sh/divtab-sh4.cc: ...here.
23479 * config/sh/divtab.c: Moved to...
23480 * config/sh/divtab.cc: ...here.
23481 * config/sh/sh-c.c: Moved to...
23482 * config/sh/sh-c.cc: ...here.
23483 * config/sh/sh.c: Moved to...
23484 * config/sh/sh.cc: ...here.
23485 * config/sol2-c.c: Moved to...
23486 * config/sol2-c.cc: ...here.
23487 * config/sol2-cxx.c: Moved to...
23488 * config/sol2-cxx.cc: ...here.
23489 * config/sol2-d.c: Moved to...
23490 * config/sol2-d.cc: ...here.
23491 * config/sol2-stubs.c: Moved to...
23492 * config/sol2-stubs.cc: ...here.
23493 * config/sol2.c: Moved to...
23494 * config/sol2.cc: ...here.
23495 * config/sparc/driver-sparc.c: Moved to...
23496 * config/sparc/driver-sparc.cc: ...here.
23497 * config/sparc/sparc-c.c: Moved to...
23498 * config/sparc/sparc-c.cc: ...here.
23499 * config/sparc/sparc-d.c: Moved to...
23500 * config/sparc/sparc-d.cc: ...here.
23501 * config/sparc/sparc.c: Moved to...
23502 * config/sparc/sparc.cc: ...here.
23503 * config/stormy16/stormy16.c: Moved to...
23504 * config/stormy16/stormy16.cc: ...here.
23505 * config/tilegx/mul-tables.c: Moved to...
23506 * config/tilegx/mul-tables.cc: ...here.
23507 * config/tilegx/tilegx-c.c: Moved to...
23508 * config/tilegx/tilegx-c.cc: ...here.
23509 * config/tilegx/tilegx.c: Moved to...
23510 * config/tilegx/tilegx.cc: ...here.
23511 * config/tilepro/mul-tables.c: Moved to...
23512 * config/tilepro/mul-tables.cc: ...here.
23513 * config/tilepro/tilepro-c.c: Moved to...
23514 * config/tilepro/tilepro-c.cc: ...here.
23515 * config/tilepro/tilepro.c: Moved to...
23516 * config/tilepro/tilepro.cc: ...here.
23517 * config/v850/v850-c.c: Moved to...
23518 * config/v850/v850-c.cc: ...here.
23519 * config/v850/v850.c: Moved to...
23520 * config/v850/v850.cc: ...here.
23521 * config/vax/vax.c: Moved to...
23522 * config/vax/vax.cc: ...here.
23523 * config/visium/visium.c: Moved to...
23524 * config/visium/visium.cc: ...here.
23525 * config/vms/vms-c.c: Moved to...
23526 * config/vms/vms-c.cc: ...here.
23527 * config/vms/vms-f.c: Moved to...
23528 * config/vms/vms-f.cc: ...here.
23529 * config/vms/vms.c: Moved to...
23530 * config/vms/vms.cc: ...here.
23531 * config/vxworks-c.c: Moved to...
23532 * config/vxworks-c.cc: ...here.
23533 * config/vxworks.c: Moved to...
23534 * config/vxworks.cc: ...here.
23535 * config/winnt-c.c: Moved to...
23536 * config/winnt-c.cc: ...here.
23537 * config/xtensa/xtensa.c: Moved to...
23538 * config/xtensa/xtensa.cc: ...here.
23539 * context.c: Moved to...
23540 * context.cc: ...here.
23541 * convert.c: Moved to...
23542 * convert.cc: ...here.
23543 * coverage.c: Moved to...
23544 * coverage.cc: ...here.
23545 * cppbuiltin.c: Moved to...
23546 * cppbuiltin.cc: ...here.
23547 * cppdefault.c: Moved to...
23548 * cppdefault.cc: ...here.
23549 * cprop.c: Moved to...
23550 * cprop.cc: ...here.
23551 * cse.c: Moved to...
23553 * cselib.c: Moved to...
23554 * cselib.cc: ...here.
23555 * ctfc.c: Moved to...
23556 * ctfc.cc: ...here.
23557 * ctfout.c: Moved to...
23558 * ctfout.cc: ...here.
23559 * data-streamer-in.c: Moved to...
23560 * data-streamer-in.cc: ...here.
23561 * data-streamer-out.c: Moved to...
23562 * data-streamer-out.cc: ...here.
23563 * data-streamer.c: Moved to...
23564 * data-streamer.cc: ...here.
23565 * dbgcnt.c: Moved to...
23566 * dbgcnt.cc: ...here.
23567 * dbxout.c: Moved to...
23568 * dbxout.cc: ...here.
23569 * dce.c: Moved to...
23571 * ddg.c: Moved to...
23573 * debug.c: Moved to...
23574 * debug.cc: ...here.
23575 * df-core.c: Moved to...
23576 * df-core.cc: ...here.
23577 * df-problems.c: Moved to...
23578 * df-problems.cc: ...here.
23579 * df-scan.c: Moved to...
23580 * df-scan.cc: ...here.
23581 * dfp.c: Moved to...
23583 * diagnostic-color.c: Moved to...
23584 * diagnostic-color.cc: ...here.
23585 * diagnostic-show-locus.c: Moved to...
23586 * diagnostic-show-locus.cc: ...here.
23587 * diagnostic-spec.c: Moved to...
23588 * diagnostic-spec.cc: ...here.
23589 * diagnostic.c: Moved to...
23590 * diagnostic.cc: ...here.
23591 * dojump.c: Moved to...
23592 * dojump.cc: ...here.
23593 * dominance.c: Moved to...
23594 * dominance.cc: ...here.
23595 * domwalk.c: Moved to...
23596 * domwalk.cc: ...here.
23597 * double-int.c: Moved to...
23598 * double-int.cc: ...here.
23599 * dse.c: Moved to...
23601 * dumpfile.c: Moved to...
23602 * dumpfile.cc: ...here.
23603 * dwarf2asm.c: Moved to...
23604 * dwarf2asm.cc: ...here.
23605 * dwarf2cfi.c: Moved to...
23606 * dwarf2cfi.cc: ...here.
23607 * dwarf2ctf.c: Moved to...
23608 * dwarf2ctf.cc: ...here.
23609 * dwarf2out.c: Moved to...
23610 * dwarf2out.cc: ...here.
23611 * early-remat.c: Moved to...
23612 * early-remat.cc: ...here.
23613 * edit-context.c: Moved to...
23614 * edit-context.cc: ...here.
23615 * emit-rtl.c: Moved to...
23616 * emit-rtl.cc: ...here.
23617 * errors.c: Moved to...
23618 * errors.cc: ...here.
23619 * et-forest.c: Moved to...
23620 * et-forest.cc: ...here.
23621 * except.c: Moved to...
23622 * except.cc: ...here.
23623 * explow.c: Moved to...
23624 * explow.cc: ...here.
23625 * expmed.c: Moved to...
23626 * expmed.cc: ...here.
23627 * expr.c: Moved to...
23628 * expr.cc: ...here.
23629 * fibonacci_heap.c: Moved to...
23630 * fibonacci_heap.cc: ...here.
23631 * file-find.c: Moved to...
23632 * file-find.cc: ...here.
23633 * file-prefix-map.c: Moved to...
23634 * file-prefix-map.cc: ...here.
23635 * final.c: Moved to...
23636 * final.cc: ...here.
23637 * fixed-value.c: Moved to...
23638 * fixed-value.cc: ...here.
23639 * fold-const-call.c: Moved to...
23640 * fold-const-call.cc: ...here.
23641 * fold-const.c: Moved to...
23642 * fold-const.cc: ...here.
23643 * fp-test.c: Moved to...
23644 * fp-test.cc: ...here.
23645 * function-tests.c: Moved to...
23646 * function-tests.cc: ...here.
23647 * function.c: Moved to...
23648 * function.cc: ...here.
23649 * fwprop.c: Moved to...
23650 * fwprop.cc: ...here.
23651 * gcc-ar.c: Moved to...
23652 * gcc-ar.cc: ...here.
23653 * gcc-main.c: Moved to...
23654 * gcc-main.cc: ...here.
23655 * gcc-rich-location.c: Moved to...
23656 * gcc-rich-location.cc: ...here.
23657 * gcc.c: Moved to...
23659 * gcov-dump.c: Moved to...
23660 * gcov-dump.cc: ...here.
23661 * gcov-io.c: Moved to...
23662 * gcov-io.cc: ...here.
23663 * gcov-tool.c: Moved to...
23664 * gcov-tool.cc: ...here.
23665 * gcov.c: Moved to...
23666 * gcov.cc: ...here.
23667 * gcse-common.c: Moved to...
23668 * gcse-common.cc: ...here.
23669 * gcse.c: Moved to...
23670 * gcse.cc: ...here.
23671 * genattr-common.c: Moved to...
23672 * genattr-common.cc: ...here.
23673 * genattr.c: Moved to...
23674 * genattr.cc: ...here.
23675 * genattrtab.c: Moved to...
23676 * genattrtab.cc: ...here.
23677 * genautomata.c: Moved to...
23678 * genautomata.cc: ...here.
23679 * gencfn-macros.c: Moved to...
23680 * gencfn-macros.cc: ...here.
23681 * gencheck.c: Moved to...
23682 * gencheck.cc: ...here.
23683 * genchecksum.c: Moved to...
23684 * genchecksum.cc: ...here.
23685 * gencodes.c: Moved to...
23686 * gencodes.cc: ...here.
23687 * genconditions.c: Moved to...
23688 * genconditions.cc: ...here.
23689 * genconfig.c: Moved to...
23690 * genconfig.cc: ...here.
23691 * genconstants.c: Moved to...
23692 * genconstants.cc: ...here.
23693 * genemit.c: Moved to...
23694 * genemit.cc: ...here.
23695 * genenums.c: Moved to...
23696 * genenums.cc: ...here.
23697 * generic-match-head.c: Moved to...
23698 * generic-match-head.cc: ...here.
23699 * genextract.c: Moved to...
23700 * genextract.cc: ...here.
23701 * genflags.c: Moved to...
23702 * genflags.cc: ...here.
23703 * gengenrtl.c: Moved to...
23704 * gengenrtl.cc: ...here.
23705 * gengtype-parse.c: Moved to...
23706 * gengtype-parse.cc: ...here.
23707 * gengtype-state.c: Moved to...
23708 * gengtype-state.cc: ...here.
23709 * gengtype.c: Moved to...
23710 * gengtype.cc: ...here.
23711 * genhooks.c: Moved to...
23712 * genhooks.cc: ...here.
23713 * genmatch.c: Moved to...
23714 * genmatch.cc: ...here.
23715 * genmddeps.c: Moved to...
23716 * genmddeps.cc: ...here.
23717 * genmddump.c: Moved to...
23718 * genmddump.cc: ...here.
23719 * genmodes.c: Moved to...
23720 * genmodes.cc: ...here.
23721 * genopinit.c: Moved to...
23722 * genopinit.cc: ...here.
23723 * genoutput.c: Moved to...
23724 * genoutput.cc: ...here.
23725 * genpeep.c: Moved to...
23726 * genpeep.cc: ...here.
23727 * genpreds.c: Moved to...
23728 * genpreds.cc: ...here.
23729 * genrecog.c: Moved to...
23730 * genrecog.cc: ...here.
23731 * gensupport.c: Moved to...
23732 * gensupport.cc: ...here.
23733 * gentarget-def.c: Moved to...
23734 * gentarget-def.cc: ...here.
23735 * genversion.c: Moved to...
23736 * genversion.cc: ...here.
23737 * ggc-common.c: Moved to...
23738 * ggc-common.cc: ...here.
23739 * ggc-none.c: Moved to...
23740 * ggc-none.cc: ...here.
23741 * ggc-page.c: Moved to...
23742 * ggc-page.cc: ...here.
23743 * ggc-tests.c: Moved to...
23744 * ggc-tests.cc: ...here.
23745 * gimple-builder.c: Moved to...
23746 * gimple-builder.cc: ...here.
23747 * gimple-expr.c: Moved to...
23748 * gimple-expr.cc: ...here.
23749 * gimple-fold.c: Moved to...
23750 * gimple-fold.cc: ...here.
23751 * gimple-iterator.c: Moved to...
23752 * gimple-iterator.cc: ...here.
23753 * gimple-laddress.c: Moved to...
23754 * gimple-laddress.cc: ...here.
23755 * gimple-loop-jam.c: Moved to...
23756 * gimple-loop-jam.cc: ...here.
23757 * gimple-low.c: Moved to...
23758 * gimple-low.cc: ...here.
23759 * gimple-match-head.c: Moved to...
23760 * gimple-match-head.cc: ...here.
23761 * gimple-pretty-print.c: Moved to...
23762 * gimple-pretty-print.cc: ...here.
23763 * gimple-ssa-backprop.c: Moved to...
23764 * gimple-ssa-backprop.cc: ...here.
23765 * gimple-ssa-evrp-analyze.c: Moved to...
23766 * gimple-ssa-evrp-analyze.cc: ...here.
23767 * gimple-ssa-evrp.c: Moved to...
23768 * gimple-ssa-evrp.cc: ...here.
23769 * gimple-ssa-isolate-paths.c: Moved to...
23770 * gimple-ssa-isolate-paths.cc: ...here.
23771 * gimple-ssa-nonnull-compare.c: Moved to...
23772 * gimple-ssa-nonnull-compare.cc: ...here.
23773 * gimple-ssa-split-paths.c: Moved to...
23774 * gimple-ssa-split-paths.cc: ...here.
23775 * gimple-ssa-sprintf.c: Moved to...
23776 * gimple-ssa-sprintf.cc: ...here.
23777 * gimple-ssa-store-merging.c: Moved to...
23778 * gimple-ssa-store-merging.cc: ...here.
23779 * gimple-ssa-strength-reduction.c: Moved to...
23780 * gimple-ssa-strength-reduction.cc: ...here.
23781 * gimple-ssa-warn-alloca.c: Moved to...
23782 * gimple-ssa-warn-alloca.cc: ...here.
23783 * gimple-ssa-warn-restrict.c: Moved to...
23784 * gimple-ssa-warn-restrict.cc: ...here.
23785 * gimple-streamer-in.c: Moved to...
23786 * gimple-streamer-in.cc: ...here.
23787 * gimple-streamer-out.c: Moved to...
23788 * gimple-streamer-out.cc: ...here.
23789 * gimple-walk.c: Moved to...
23790 * gimple-walk.cc: ...here.
23791 * gimple-warn-recursion.c: Moved to...
23792 * gimple-warn-recursion.cc: ...here.
23793 * gimple.c: Moved to...
23794 * gimple.cc: ...here.
23795 * gimplify-me.c: Moved to...
23796 * gimplify-me.cc: ...here.
23797 * gimplify.c: Moved to...
23798 * gimplify.cc: ...here.
23799 * godump.c: Moved to...
23800 * godump.cc: ...here.
23801 * graph.c: Moved to...
23802 * graph.cc: ...here.
23803 * graphds.c: Moved to...
23804 * graphds.cc: ...here.
23805 * graphite-dependences.c: Moved to...
23806 * graphite-dependences.cc: ...here.
23807 * graphite-isl-ast-to-gimple.c: Moved to...
23808 * graphite-isl-ast-to-gimple.cc: ...here.
23809 * graphite-optimize-isl.c: Moved to...
23810 * graphite-optimize-isl.cc: ...here.
23811 * graphite-poly.c: Moved to...
23812 * graphite-poly.cc: ...here.
23813 * graphite-scop-detection.c: Moved to...
23814 * graphite-scop-detection.cc: ...here.
23815 * graphite-sese-to-poly.c: Moved to...
23816 * graphite-sese-to-poly.cc: ...here.
23817 * graphite.c: Moved to...
23818 * graphite.cc: ...here.
23819 * haifa-sched.c: Moved to...
23820 * haifa-sched.cc: ...here.
23821 * hash-map-tests.c: Moved to...
23822 * hash-map-tests.cc: ...here.
23823 * hash-set-tests.c: Moved to...
23824 * hash-set-tests.cc: ...here.
23825 * hash-table.c: Moved to...
23826 * hash-table.cc: ...here.
23827 * hooks.c: Moved to...
23828 * hooks.cc: ...here.
23829 * host-default.c: Moved to...
23830 * host-default.cc: ...here.
23831 * hw-doloop.c: Moved to...
23832 * hw-doloop.cc: ...here.
23833 * hwint.c: Moved to...
23834 * hwint.cc: ...here.
23835 * ifcvt.c: Moved to...
23836 * ifcvt.cc: ...here.
23837 * inchash.c: Moved to...
23838 * inchash.cc: ...here.
23839 * incpath.c: Moved to...
23840 * incpath.cc: ...here.
23841 * init-regs.c: Moved to...
23842 * init-regs.cc: ...here.
23843 * input.c: Moved to...
23844 * input.cc: ...here.
23845 * internal-fn.c: Moved to...
23846 * internal-fn.cc: ...here.
23847 * intl.c: Moved to...
23848 * intl.cc: ...here.
23849 * ipa-comdats.c: Moved to...
23850 * ipa-comdats.cc: ...here.
23851 * ipa-cp.c: Moved to...
23852 * ipa-cp.cc: ...here.
23853 * ipa-devirt.c: Moved to...
23854 * ipa-devirt.cc: ...here.
23855 * ipa-fnsummary.c: Moved to...
23856 * ipa-fnsummary.cc: ...here.
23857 * ipa-icf-gimple.c: Moved to...
23858 * ipa-icf-gimple.cc: ...here.
23859 * ipa-icf.c: Moved to...
23860 * ipa-icf.cc: ...here.
23861 * ipa-inline-analysis.c: Moved to...
23862 * ipa-inline-analysis.cc: ...here.
23863 * ipa-inline-transform.c: Moved to...
23864 * ipa-inline-transform.cc: ...here.
23865 * ipa-inline.c: Moved to...
23866 * ipa-inline.cc: ...here.
23867 * ipa-modref-tree.c: Moved to...
23868 * ipa-modref-tree.cc: ...here.
23869 * ipa-modref.c: Moved to...
23870 * ipa-modref.cc: ...here.
23871 * ipa-param-manipulation.c: Moved to...
23872 * ipa-param-manipulation.cc: ...here.
23873 * ipa-polymorphic-call.c: Moved to...
23874 * ipa-polymorphic-call.cc: ...here.
23875 * ipa-predicate.c: Moved to...
23876 * ipa-predicate.cc: ...here.
23877 * ipa-profile.c: Moved to...
23878 * ipa-profile.cc: ...here.
23879 * ipa-prop.c: Moved to...
23880 * ipa-prop.cc: ...here.
23881 * ipa-pure-const.c: Moved to...
23882 * ipa-pure-const.cc: ...here.
23883 * ipa-ref.c: Moved to...
23884 * ipa-ref.cc: ...here.
23885 * ipa-reference.c: Moved to...
23886 * ipa-reference.cc: ...here.
23887 * ipa-split.c: Moved to...
23888 * ipa-split.cc: ...here.
23889 * ipa-sra.c: Moved to...
23890 * ipa-sra.cc: ...here.
23891 * ipa-utils.c: Moved to...
23892 * ipa-utils.cc: ...here.
23893 * ipa-visibility.c: Moved to...
23894 * ipa-visibility.cc: ...here.
23895 * ipa.c: Moved to...
23897 * ira-build.c: Moved to...
23898 * ira-build.cc: ...here.
23899 * ira-color.c: Moved to...
23900 * ira-color.cc: ...here.
23901 * ira-conflicts.c: Moved to...
23902 * ira-conflicts.cc: ...here.
23903 * ira-costs.c: Moved to...
23904 * ira-costs.cc: ...here.
23905 * ira-emit.c: Moved to...
23906 * ira-emit.cc: ...here.
23907 * ira-lives.c: Moved to...
23908 * ira-lives.cc: ...here.
23909 * ira.c: Moved to...
23911 * jump.c: Moved to...
23912 * jump.cc: ...here.
23913 * langhooks.c: Moved to...
23914 * langhooks.cc: ...here.
23915 * lcm.c: Moved to...
23917 * lists.c: Moved to...
23918 * lists.cc: ...here.
23919 * loop-doloop.c: Moved to...
23920 * loop-doloop.cc: ...here.
23921 * loop-init.c: Moved to...
23922 * loop-init.cc: ...here.
23923 * loop-invariant.c: Moved to...
23924 * loop-invariant.cc: ...here.
23925 * loop-iv.c: Moved to...
23926 * loop-iv.cc: ...here.
23927 * loop-unroll.c: Moved to...
23928 * loop-unroll.cc: ...here.
23929 * lower-subreg.c: Moved to...
23930 * lower-subreg.cc: ...here.
23931 * lra-assigns.c: Moved to...
23932 * lra-assigns.cc: ...here.
23933 * lra-coalesce.c: Moved to...
23934 * lra-coalesce.cc: ...here.
23935 * lra-constraints.c: Moved to...
23936 * lra-constraints.cc: ...here.
23937 * lra-eliminations.c: Moved to...
23938 * lra-eliminations.cc: ...here.
23939 * lra-lives.c: Moved to...
23940 * lra-lives.cc: ...here.
23941 * lra-remat.c: Moved to...
23942 * lra-remat.cc: ...here.
23943 * lra-spills.c: Moved to...
23944 * lra-spills.cc: ...here.
23945 * lra.c: Moved to...
23947 * lto-cgraph.c: Moved to...
23948 * lto-cgraph.cc: ...here.
23949 * lto-compress.c: Moved to...
23950 * lto-compress.cc: ...here.
23951 * lto-opts.c: Moved to...
23952 * lto-opts.cc: ...here.
23953 * lto-section-in.c: Moved to...
23954 * lto-section-in.cc: ...here.
23955 * lto-section-out.c: Moved to...
23956 * lto-section-out.cc: ...here.
23957 * lto-streamer-in.c: Moved to...
23958 * lto-streamer-in.cc: ...here.
23959 * lto-streamer-out.c: Moved to...
23960 * lto-streamer-out.cc: ...here.
23961 * lto-streamer.c: Moved to...
23962 * lto-streamer.cc: ...here.
23963 * lto-wrapper.c: Moved to...
23964 * lto-wrapper.cc: ...here.
23965 * main.c: Moved to...
23966 * main.cc: ...here.
23967 * mcf.c: Moved to...
23969 * mode-switching.c: Moved to...
23970 * mode-switching.cc: ...here.
23971 * modulo-sched.c: Moved to...
23972 * modulo-sched.cc: ...here.
23973 * multiple_target.c: Moved to...
23974 * multiple_target.cc: ...here.
23975 * omp-expand.c: Moved to...
23976 * omp-expand.cc: ...here.
23977 * omp-general.c: Moved to...
23978 * omp-general.cc: ...here.
23979 * omp-low.c: Moved to...
23980 * omp-low.cc: ...here.
23981 * omp-offload.c: Moved to...
23982 * omp-offload.cc: ...here.
23983 * omp-simd-clone.c: Moved to...
23984 * omp-simd-clone.cc: ...here.
23985 * opt-suggestions.c: Moved to...
23986 * opt-suggestions.cc: ...here.
23987 * optabs-libfuncs.c: Moved to...
23988 * optabs-libfuncs.cc: ...here.
23989 * optabs-query.c: Moved to...
23990 * optabs-query.cc: ...here.
23991 * optabs-tree.c: Moved to...
23992 * optabs-tree.cc: ...here.
23993 * optabs.c: Moved to...
23994 * optabs.cc: ...here.
23995 * opts-common.c: Moved to...
23996 * opts-common.cc: ...here.
23997 * opts-global.c: Moved to...
23998 * opts-global.cc: ...here.
23999 * opts.c: Moved to...
24000 * opts.cc: ...here.
24001 * passes.c: Moved to...
24002 * passes.cc: ...here.
24003 * plugin.c: Moved to...
24004 * plugin.cc: ...here.
24005 * postreload-gcse.c: Moved to...
24006 * postreload-gcse.cc: ...here.
24007 * postreload.c: Moved to...
24008 * postreload.cc: ...here.
24009 * predict.c: Moved to...
24010 * predict.cc: ...here.
24011 * prefix.c: Moved to...
24012 * prefix.cc: ...here.
24013 * pretty-print.c: Moved to...
24014 * pretty-print.cc: ...here.
24015 * print-rtl-function.c: Moved to...
24016 * print-rtl-function.cc: ...here.
24017 * print-rtl.c: Moved to...
24018 * print-rtl.cc: ...here.
24019 * print-tree.c: Moved to...
24020 * print-tree.cc: ...here.
24021 * profile-count.c: Moved to...
24022 * profile-count.cc: ...here.
24023 * profile.c: Moved to...
24024 * profile.cc: ...here.
24025 * read-md.c: Moved to...
24026 * read-md.cc: ...here.
24027 * read-rtl-function.c: Moved to...
24028 * read-rtl-function.cc: ...here.
24029 * read-rtl.c: Moved to...
24030 * read-rtl.cc: ...here.
24031 * real.c: Moved to...
24032 * real.cc: ...here.
24033 * realmpfr.c: Moved to...
24034 * realmpfr.cc: ...here.
24035 * recog.c: Moved to...
24036 * recog.cc: ...here.
24037 * ree.c: Moved to...
24039 * reg-stack.c: Moved to...
24040 * reg-stack.cc: ...here.
24041 * regcprop.c: Moved to...
24042 * regcprop.cc: ...here.
24043 * reginfo.c: Moved to...
24044 * reginfo.cc: ...here.
24045 * regrename.c: Moved to...
24046 * regrename.cc: ...here.
24047 * regstat.c: Moved to...
24048 * regstat.cc: ...here.
24049 * reload.c: Moved to...
24050 * reload.cc: ...here.
24051 * reload1.c: Moved to...
24052 * reload1.cc: ...here.
24053 * reorg.c: Moved to...
24054 * reorg.cc: ...here.
24055 * resource.c: Moved to...
24056 * resource.cc: ...here.
24057 * rtl-error.c: Moved to...
24058 * rtl-error.cc: ...here.
24059 * rtl-tests.c: Moved to...
24060 * rtl-tests.cc: ...here.
24061 * rtl.c: Moved to...
24063 * rtlanal.c: Moved to...
24064 * rtlanal.cc: ...here.
24065 * rtlhash.c: Moved to...
24066 * rtlhash.cc: ...here.
24067 * rtlhooks.c: Moved to...
24068 * rtlhooks.cc: ...here.
24069 * rtx-vector-builder.c: Moved to...
24070 * rtx-vector-builder.cc: ...here.
24071 * run-rtl-passes.c: Moved to...
24072 * run-rtl-passes.cc: ...here.
24073 * sancov.c: Moved to...
24074 * sancov.cc: ...here.
24075 * sanopt.c: Moved to...
24076 * sanopt.cc: ...here.
24077 * sbitmap.c: Moved to...
24078 * sbitmap.cc: ...here.
24079 * sched-deps.c: Moved to...
24080 * sched-deps.cc: ...here.
24081 * sched-ebb.c: Moved to...
24082 * sched-ebb.cc: ...here.
24083 * sched-rgn.c: Moved to...
24084 * sched-rgn.cc: ...here.
24085 * sel-sched-dump.c: Moved to...
24086 * sel-sched-dump.cc: ...here.
24087 * sel-sched-ir.c: Moved to...
24088 * sel-sched-ir.cc: ...here.
24089 * sel-sched.c: Moved to...
24090 * sel-sched.cc: ...here.
24091 * selftest-diagnostic.c: Moved to...
24092 * selftest-diagnostic.cc: ...here.
24093 * selftest-rtl.c: Moved to...
24094 * selftest-rtl.cc: ...here.
24095 * selftest-run-tests.c: Moved to...
24096 * selftest-run-tests.cc: ...here.
24097 * selftest.c: Moved to...
24098 * selftest.cc: ...here.
24099 * sese.c: Moved to...
24100 * sese.cc: ...here.
24101 * shrink-wrap.c: Moved to...
24102 * shrink-wrap.cc: ...here.
24103 * simplify-rtx.c: Moved to...
24104 * simplify-rtx.cc: ...here.
24105 * sparseset.c: Moved to...
24106 * sparseset.cc: ...here.
24107 * spellcheck-tree.c: Moved to...
24108 * spellcheck-tree.cc: ...here.
24109 * spellcheck.c: Moved to...
24110 * spellcheck.cc: ...here.
24111 * sreal.c: Moved to...
24112 * sreal.cc: ...here.
24113 * stack-ptr-mod.c: Moved to...
24114 * stack-ptr-mod.cc: ...here.
24115 * statistics.c: Moved to...
24116 * statistics.cc: ...here.
24117 * stmt.c: Moved to...
24118 * stmt.cc: ...here.
24119 * stor-layout.c: Moved to...
24120 * stor-layout.cc: ...here.
24121 * store-motion.c: Moved to...
24122 * store-motion.cc: ...here.
24123 * streamer-hooks.c: Moved to...
24124 * streamer-hooks.cc: ...here.
24125 * stringpool.c: Moved to...
24126 * stringpool.cc: ...here.
24127 * substring-locations.c: Moved to...
24128 * substring-locations.cc: ...here.
24129 * symtab.c: Moved to...
24130 * symtab.cc: ...here.
24131 * target-globals.c: Moved to...
24132 * target-globals.cc: ...here.
24133 * targhooks.c: Moved to...
24134 * targhooks.cc: ...here.
24135 * timevar.c: Moved to...
24136 * timevar.cc: ...here.
24137 * toplev.c: Moved to...
24138 * toplev.cc: ...here.
24139 * tracer.c: Moved to...
24140 * tracer.cc: ...here.
24141 * trans-mem.c: Moved to...
24142 * trans-mem.cc: ...here.
24143 * tree-affine.c: Moved to...
24144 * tree-affine.cc: ...here.
24145 * tree-call-cdce.c: Moved to...
24146 * tree-call-cdce.cc: ...here.
24147 * tree-cfg.c: Moved to...
24148 * tree-cfg.cc: ...here.
24149 * tree-cfgcleanup.c: Moved to...
24150 * tree-cfgcleanup.cc: ...here.
24151 * tree-chrec.c: Moved to...
24152 * tree-chrec.cc: ...here.
24153 * tree-complex.c: Moved to...
24154 * tree-complex.cc: ...here.
24155 * tree-data-ref.c: Moved to...
24156 * tree-data-ref.cc: ...here.
24157 * tree-dfa.c: Moved to...
24158 * tree-dfa.cc: ...here.
24159 * tree-diagnostic.c: Moved to...
24160 * tree-diagnostic.cc: ...here.
24161 * tree-dump.c: Moved to...
24162 * tree-dump.cc: ...here.
24163 * tree-eh.c: Moved to...
24164 * tree-eh.cc: ...here.
24165 * tree-emutls.c: Moved to...
24166 * tree-emutls.cc: ...here.
24167 * tree-if-conv.c: Moved to...
24168 * tree-if-conv.cc: ...here.
24169 * tree-inline.c: Moved to...
24170 * tree-inline.cc: ...here.
24171 * tree-into-ssa.c: Moved to...
24172 * tree-into-ssa.cc: ...here.
24173 * tree-iterator.c: Moved to...
24174 * tree-iterator.cc: ...here.
24175 * tree-loop-distribution.c: Moved to...
24176 * tree-loop-distribution.cc: ...here.
24177 * tree-nested.c: Moved to...
24178 * tree-nested.cc: ...here.
24179 * tree-nrv.c: Moved to...
24180 * tree-nrv.cc: ...here.
24181 * tree-object-size.c: Moved to...
24182 * tree-object-size.cc: ...here.
24183 * tree-outof-ssa.c: Moved to...
24184 * tree-outof-ssa.cc: ...here.
24185 * tree-parloops.c: Moved to...
24186 * tree-parloops.cc: ...here.
24187 * tree-phinodes.c: Moved to...
24188 * tree-phinodes.cc: ...here.
24189 * tree-predcom.c: Moved to...
24190 * tree-predcom.cc: ...here.
24191 * tree-pretty-print.c: Moved to...
24192 * tree-pretty-print.cc: ...here.
24193 * tree-profile.c: Moved to...
24194 * tree-profile.cc: ...here.
24195 * tree-scalar-evolution.c: Moved to...
24196 * tree-scalar-evolution.cc: ...here.
24197 * tree-sra.c: Moved to...
24198 * tree-sra.cc: ...here.
24199 * tree-ssa-address.c: Moved to...
24200 * tree-ssa-address.cc: ...here.
24201 * tree-ssa-alias.c: Moved to...
24202 * tree-ssa-alias.cc: ...here.
24203 * tree-ssa-ccp.c: Moved to...
24204 * tree-ssa-ccp.cc: ...here.
24205 * tree-ssa-coalesce.c: Moved to...
24206 * tree-ssa-coalesce.cc: ...here.
24207 * tree-ssa-copy.c: Moved to...
24208 * tree-ssa-copy.cc: ...here.
24209 * tree-ssa-dce.c: Moved to...
24210 * tree-ssa-dce.cc: ...here.
24211 * tree-ssa-dom.c: Moved to...
24212 * tree-ssa-dom.cc: ...here.
24213 * tree-ssa-dse.c: Moved to...
24214 * tree-ssa-dse.cc: ...here.
24215 * tree-ssa-forwprop.c: Moved to...
24216 * tree-ssa-forwprop.cc: ...here.
24217 * tree-ssa-ifcombine.c: Moved to...
24218 * tree-ssa-ifcombine.cc: ...here.
24219 * tree-ssa-live.c: Moved to...
24220 * tree-ssa-live.cc: ...here.
24221 * tree-ssa-loop-ch.c: Moved to...
24222 * tree-ssa-loop-ch.cc: ...here.
24223 * tree-ssa-loop-im.c: Moved to...
24224 * tree-ssa-loop-im.cc: ...here.
24225 * tree-ssa-loop-ivcanon.c: Moved to...
24226 * tree-ssa-loop-ivcanon.cc: ...here.
24227 * tree-ssa-loop-ivopts.c: Moved to...
24228 * tree-ssa-loop-ivopts.cc: ...here.
24229 * tree-ssa-loop-manip.c: Moved to...
24230 * tree-ssa-loop-manip.cc: ...here.
24231 * tree-ssa-loop-niter.c: Moved to...
24232 * tree-ssa-loop-niter.cc: ...here.
24233 * tree-ssa-loop-prefetch.c: Moved to...
24234 * tree-ssa-loop-prefetch.cc: ...here.
24235 * tree-ssa-loop-split.c: Moved to...
24236 * tree-ssa-loop-split.cc: ...here.
24237 * tree-ssa-loop-unswitch.c: Moved to...
24238 * tree-ssa-loop-unswitch.cc: ...here.
24239 * tree-ssa-loop.c: Moved to...
24240 * tree-ssa-loop.cc: ...here.
24241 * tree-ssa-math-opts.c: Moved to...
24242 * tree-ssa-math-opts.cc: ...here.
24243 * tree-ssa-operands.c: Moved to...
24244 * tree-ssa-operands.cc: ...here.
24245 * tree-ssa-phiopt.c: Moved to...
24246 * tree-ssa-phiopt.cc: ...here.
24247 * tree-ssa-phiprop.c: Moved to...
24248 * tree-ssa-phiprop.cc: ...here.
24249 * tree-ssa-pre.c: Moved to...
24250 * tree-ssa-pre.cc: ...here.
24251 * tree-ssa-propagate.c: Moved to...
24252 * tree-ssa-propagate.cc: ...here.
24253 * tree-ssa-reassoc.c: Moved to...
24254 * tree-ssa-reassoc.cc: ...here.
24255 * tree-ssa-sccvn.c: Moved to...
24256 * tree-ssa-sccvn.cc: ...here.
24257 * tree-ssa-scopedtables.c: Moved to...
24258 * tree-ssa-scopedtables.cc: ...here.
24259 * tree-ssa-sink.c: Moved to...
24260 * tree-ssa-sink.cc: ...here.
24261 * tree-ssa-strlen.c: Moved to...
24262 * tree-ssa-strlen.cc: ...here.
24263 * tree-ssa-structalias.c: Moved to...
24264 * tree-ssa-structalias.cc: ...here.
24265 * tree-ssa-tail-merge.c: Moved to...
24266 * tree-ssa-tail-merge.cc: ...here.
24267 * tree-ssa-ter.c: Moved to...
24268 * tree-ssa-ter.cc: ...here.
24269 * tree-ssa-threadbackward.c: Moved to...
24270 * tree-ssa-threadbackward.cc: ...here.
24271 * tree-ssa-threadedge.c: Moved to...
24272 * tree-ssa-threadedge.cc: ...here.
24273 * tree-ssa-threadupdate.c: Moved to...
24274 * tree-ssa-threadupdate.cc: ...here.
24275 * tree-ssa-uncprop.c: Moved to...
24276 * tree-ssa-uncprop.cc: ...here.
24277 * tree-ssa-uninit.c: Moved to...
24278 * tree-ssa-uninit.cc: ...here.
24279 * tree-ssa.c: Moved to...
24280 * tree-ssa.cc: ...here.
24281 * tree-ssanames.c: Moved to...
24282 * tree-ssanames.cc: ...here.
24283 * tree-stdarg.c: Moved to...
24284 * tree-stdarg.cc: ...here.
24285 * tree-streamer-in.c: Moved to...
24286 * tree-streamer-in.cc: ...here.
24287 * tree-streamer-out.c: Moved to...
24288 * tree-streamer-out.cc: ...here.
24289 * tree-streamer.c: Moved to...
24290 * tree-streamer.cc: ...here.
24291 * tree-switch-conversion.c: Moved to...
24292 * tree-switch-conversion.cc: ...here.
24293 * tree-tailcall.c: Moved to...
24294 * tree-tailcall.cc: ...here.
24295 * tree-vect-data-refs.c: Moved to...
24296 * tree-vect-data-refs.cc: ...here.
24297 * tree-vect-generic.c: Moved to...
24298 * tree-vect-generic.cc: ...here.
24299 * tree-vect-loop-manip.c: Moved to...
24300 * tree-vect-loop-manip.cc: ...here.
24301 * tree-vect-loop.c: Moved to...
24302 * tree-vect-loop.cc: ...here.
24303 * tree-vect-patterns.c: Moved to...
24304 * tree-vect-patterns.cc: ...here.
24305 * tree-vect-slp-patterns.c: Moved to...
24306 * tree-vect-slp-patterns.cc: ...here.
24307 * tree-vect-slp.c: Moved to...
24308 * tree-vect-slp.cc: ...here.
24309 * tree-vect-stmts.c: Moved to...
24310 * tree-vect-stmts.cc: ...here.
24311 * tree-vector-builder.c: Moved to...
24312 * tree-vector-builder.cc: ...here.
24313 * tree-vectorizer.c: Moved to...
24314 * tree-vectorizer.cc: ...here.
24315 * tree-vrp.c: Moved to...
24316 * tree-vrp.cc: ...here.
24317 * tree.c: Moved to...
24318 * tree.cc: ...here.
24319 * tsan.c: Moved to...
24320 * tsan.cc: ...here.
24321 * typed-splay-tree.c: Moved to...
24322 * typed-splay-tree.cc: ...here.
24323 * ubsan.c: Moved to...
24324 * ubsan.cc: ...here.
24325 * valtrack.c: Moved to...
24326 * valtrack.cc: ...here.
24327 * value-prof.c: Moved to...
24328 * value-prof.cc: ...here.
24329 * var-tracking.c: Moved to...
24330 * var-tracking.cc: ...here.
24331 * varasm.c: Moved to...
24332 * varasm.cc: ...here.
24333 * varpool.c: Moved to...
24334 * varpool.cc: ...here.
24335 * vec-perm-indices.c: Moved to...
24336 * vec-perm-indices.cc: ...here.
24337 * vec.c: Moved to...
24339 * vmsdbgout.c: Moved to...
24340 * vmsdbgout.cc: ...here.
24341 * vr-values.c: Moved to...
24342 * vr-values.cc: ...here.
24343 * vtable-verify.c: Moved to...
24344 * vtable-verify.cc: ...here.
24345 * web.c: Moved to...
24347 * xcoffout.c: Moved to...
24348 * xcoffout.cc: ...here.
24350 2022-01-17 qing zhao <qing.zhao@oracle.com>
24352 * tree-ssa-uninit.c (warn_uninit): Delete the 4th parameter. Handle
24353 .DEFERRED_INIT call with an anonymous SSA_NAME specially.
24354 (check_defs): Handle .DEFERRED_INIT call with an anonymous SSA_NAME
24356 (warn_uninit_phi_uses): Delete the 4th actual when call warn_uninit.
24357 (warn_uninitialized_vars): Likewise.
24358 (warn_uninitialized_phi): Likewise.
24360 2022-01-17 Jason Merrill <jason@redhat.com>
24362 * diagnostic.h (struct diagnostic_context): Add includes_seen.
24363 * diagnostic.c (diagnostic_initialize): Initialize it.
24364 (diagnostic_finish): Clean it up.
24365 (includes_seen): New function.
24366 (diagnostic_report_current_module): Use it.
24368 2022-01-17 Richard Biener <rguenther@suse.de>
24370 PR middle-end/101292
24371 * diagnostic-spec.c (copy_warning): Make sure to not
24372 reference old hashtable content on possible resize.
24373 * warning-control.cc (copy_warning): Likewise.
24375 2022-01-17 Jakub Jelinek <jakub@redhat.com>
24378 * tree-cfg.h (cond_only_block_p): Declare.
24379 * tree-ssa-phiopt.c (cond_only_block_p): Move function to ...
24380 * tree-cfg.c (cond_only_block_p): ... here. No longer static.
24381 * optabs.def (spaceship_optab): New optab.
24382 * internal-fn.def (SPACESHIP): New internal function.
24383 * internal-fn.h (expand_SPACESHIP): Declare.
24384 * internal-fn.c (expand_PHI): Formatting fix.
24385 (expand_SPACESHIP): New function.
24386 * tree-ssa-math-opts.c (optimize_spaceship): New function.
24387 (math_opts_dom_walker::after_dom_children): Use it.
24388 * config/i386/i386.md (spaceship<mode>3): New define_expand.
24389 * config/i386/i386-protos.h (ix86_expand_fp_spaceship): Declare.
24390 * config/i386/i386-expand.c (ix86_expand_fp_spaceship): New function.
24391 * doc/md.texi (spaceship@var{m}3): Document.
24393 2022-01-17 Kewen Lin <linkw@linux.ibm.com>
24395 * config/rs6000/altivec.md (altivec_vreveti2): Remove.
24396 * config/rs6000/vsx.md (*vsx_extract_si, *vsx_extract_si_<uns>float_df,
24397 *vsx_extract_si_<uns>float_<mode>, *vsx_insert_extract_v4sf_p9): Use
24398 known constant values to simplify code.
24400 2022-01-17 Haochen Gui <guihaoc@gcc.gnu.org>
24403 * config/rs6000/vsx.md (split pattern for TI to V1TI move): Defined.
24405 2022-01-16 wwwhhhyyy <hongyu.wang@intel.com>
24407 * config/i386/i386.h (TARGET_DEST_FALSE_DEP_FOR_GLC): New macro.
24408 * config/i386/sse.md (<avx512>_<complexopname>_<mode><maskc_name><round_name>):
24409 Insert zero-idiom in output template when attr enabled, set new attribute to
24410 true for non-mask/maskz insn.
24411 (avx512fp16_<complexopname>sh_v8hf<mask_scalarc_name><round_scalarcz_name>):
24413 (avx512dq_mul<mode>3<mask_name>): Likewise.
24414 (<avx2_avx512>_permvar<mode><mask_name>): Likewise.
24415 (avx2_perm<mode>_1<mask_name>): Likewise.
24416 (avx512f_perm<mode>_1<mask_name>): Likewise.
24417 (avx512dq_rangep<mode><mask_name><round_saeonly_name>): Likewise.
24418 (avx512dq_ranges<mode><mask_scalar_name><round_saeonly_scalar_name>):
24420 (<avx512>_getmant<mode><mask_name><round_saeonly_name>): Likewise.
24421 (avx512f_vgetmant<mode><mask_scalar_name><round_saeonly_scalar_name>):
24423 * config/i386/subst.md (mask3_dest_false_dep_for_glc_cond): New
24425 (mask4_dest_false_dep_for_glc_cond): Likewise.
24426 (mask6_dest_false_dep_for_glc_cond): Likewise.
24427 (mask10_dest_false_dep_for_glc_cond): Likewise.
24428 (maskc_dest_false_dep_for_glc_cond): Likewise.
24429 (mask_scalar4_dest_false_dep_for_glc_cond): Likewise.
24430 (mask_scalarc_dest_false_dep_for_glc_cond): Likewise.
24431 * config/i386/x86-tune.def (X86_TUNE_DEST_FALSE_DEP_FOR_GLC): New
24432 DEF_TUNE enabled for m_SAPPHIRERAPIDS and m_ALDERLAKE
24434 2022-01-15 Martin Sebor <msebor@redhat.com>
24437 * diagnostic-spec.c (nowarn_spec_t::nowarn_spec_t): Handle
24438 -Wdangling-pointer.
24439 * doc/invoke.texi (-Wdangling-pointer): Document new option.
24440 * gimple-ssa-warn-access.cc (pass_waccess::clone): Set new member.
24441 (pass_waccess::check_pointer_uses): New function.
24442 (pass_waccess::gimple_call_return_arg): New function.
24443 (pass_waccess::gimple_call_return_arg_ref): New function.
24444 (pass_waccess::check_call_dangling): New function.
24445 (pass_waccess::check_dangling_uses): New function overloads.
24446 (pass_waccess::check_dangling_stores): New function.
24447 (pass_waccess::check_dangling_stores): New function.
24448 (pass_waccess::m_clobbers): New data member.
24449 (pass_waccess::m_func): New data member.
24450 (pass_waccess::m_run_number): New data member.
24451 (pass_waccess::m_check_dangling_p): New data member.
24452 (pass_waccess::check_alloca): Check m_early_checks_p.
24453 (pass_waccess::check_alloc_size_call): Same.
24454 (pass_waccess::check_strcat): Same.
24455 (pass_waccess::check_strncat): Same.
24456 (pass_waccess::check_stxcpy): Same.
24457 (pass_waccess::check_stxncpy): Same.
24458 (pass_waccess::check_strncmp): Same.
24459 (pass_waccess::check_memop_access): Same.
24460 (pass_waccess::check_read_access): Same.
24461 (pass_waccess::check_builtin): Call check_pointer_uses.
24462 (pass_waccess::warn_invalid_pointer): Add arguments.
24463 (is_auto_decl): New function.
24464 (pass_waccess::check_stmt): New function.
24465 (pass_waccess::check_block): Call check_stmt.
24466 (pass_waccess::execute): Call check_dangling_uses,
24467 check_dangling_stores. Empty m_clobbers.
24468 * passes.def (pass_warn_access): Invoke pass two more times.
24470 2022-01-15 Martin Sebor <msebor@redhat.com>
24472 PR tree-optimization/80532
24473 * common.opt (-Wuse-after-free): New options.
24474 * diagnostic-spec.c (nowarn_spec_t::nowarn_spec_t): Handle
24475 OPT_Wreturn_local_addr and OPT_Wuse_after_free_.
24476 * diagnostic-spec.h (NW_DANGLING): New enumerator.
24477 * doc/invoke.texi (-Wuse-after-free): Document new option.
24478 * gimple-ssa-warn-access.cc (pass_waccess::check_call): Rename...
24479 (pass_waccess::check_call_access): ...to this.
24480 (pass_waccess::check): Rename...
24481 (pass_waccess::check_block): ...to this.
24482 (pass_waccess::check_pointer_uses): New function.
24483 (pass_waccess::gimple_call_return_arg): New function.
24484 (pass_waccess::warn_invalid_pointer): New function.
24485 (pass_waccess::check_builtin): Handle free and realloc.
24486 (gimple_use_after_inval_p): New function.
24487 (get_realloc_lhs): New function.
24488 (maybe_warn_mismatched_realloc): New function.
24489 (pointers_related_p): New function.
24490 (pass_waccess::check_call): Call check_pointer_uses.
24491 (pass_waccess::execute): Compute and free dominance info.
24493 2022-01-15 Uroš Bizjak <ubizjak@gmail.com>
24495 * config/i386/i386-expand.c (ix86_expand_sse_movcc): Use
24496 expand_simple_unop and expand_simple_binop instead of manually
24497 constructing NOT, AND and IOR RTXes. Use vector_all_ones_operand
24498 consistently. Eliminate common subexpressions and simplify code.
24499 * config/i386/sse.md (<any_logic:code><MODEF:mode>3): New expander.
24500 (<any_logic:code><MODEF:mode>3): Make public.
24502 2022-01-14 Eric Botcazou <ebotcazou@adacore.com>
24504 * ipa-param-manipulation.c (ipa_dump_adjusted_parameters): Dump
24505 reverse flag as "reverse" for the sake of consistency.
24506 * ipa-sra.c: Fix copyright year.
24507 (ipa_sra_function_summaries::duplicate): Copy the reverse flag.
24508 (dump_isra_access): Tweak dump line.
24509 (isra_write_node_summary): Write the reverse flag.
24510 (isra_read_node_info): Read it.
24511 (pull_accesses_from_callee): Test its consistency and copy it.
24513 2022-01-14 Richard Sandiford <richard.sandiford@arm.com>
24515 PR middle-end/104026
24516 * tree-vect-loop.c (_loop_vec_info::_loop_vec_info): Initialize
24517 partial_load_store_bias.
24519 2022-01-14 Martin Sebor <msebor@redhat.com>
24521 PR middle-end/101475
24522 * pointer-query.cc (handle_component_ref): Use the size of
24523 the enclosing object if it's smaller than the member.
24525 2022-01-14 Martin Liska <mliska@suse.cz>
24527 * configure: Regenerate.
24529 2022-01-14 Uroš Bizjak <ubizjak@gmail.com>
24531 * config/i386/i386.md (*add<mode>_1_slp"):
24532 Mark alternative 1 output operand earlyclobbered.
24533 (*sub<mode>_1_slp): Ditto.
24534 (*and<mode>_1_slp): Ditto.
24535 (*<code><mode>_1_slp): Ditto.
24536 (*neg<mode>_1_slp): Ditto.
24537 (*one_cmpl<mode>_1_slp): Ditto.
24538 (*ashl<mode>3_1_slp): Ditto.
24539 (*<insn><mode>3_1_slp): Ditto.
24540 (*<insn><mode>3_1_slp): Ditto.
24542 2022-01-14 Kewen Lin <linkw@linux.ibm.com>
24544 PR tree-optimization/104015
24545 * tree-vect-loop.c (vect_analyze_loop): Check
24546 param_vect_partial_vector_usage for supports_partial_vectors.
24548 2022-01-14 Jakub Jelinek <jakub@redhat.com>
24551 * fold-const.c (address_compare): Punt on comparison of address of
24552 one object with address of end of another object if
24553 folding_initializer.
24555 2022-01-14 Jakub Jelinek <jakub@redhat.com>
24558 * tree-ssa-forwprop.c (simplify_builtin_call): Canonicalize
24559 __atomic_fetch_op (p, x, y) op x into __atomic_op_fetch (p, x, y)
24560 and __atomic_op_fetch (p, x, y) iop x into
24561 __atomic_fetch_op (p, x, y).
24563 2022-01-14 Claudiu Zissulescu <claziss@synopsys.com>
24565 * config/arc/arc.h (DWARF_FRAME_REGNUM): Update definition.
24566 (DWARF_FRAME_RETURN_COLUMN): Use RETURN_ADDR_REGNUM macro.
24567 (INCOMING_RETURN_ADDR_RTX): Likewise.
24568 (DWARF_ALT_FRAME_RETURN_COLUMN): Define.
24570 2022-01-14 Claudiu Zissulescu <claziss@synopsys.com>
24572 * config/arc/arc.c (arc_compute_frame_size): Remove condition when
24573 computin checking accumulator regs.
24574 (arc_expand_prologue): Update comments.
24575 (arc_expand_epilogue): Likewise.
24577 2022-01-14 Roger Sayle <roger@nextmovesoftware.com>
24578 Uroš Bizjak <ubizjak@gmail.com>
24580 * config/i386/i386-expand.c (ix86_expand_v1ti_to_ti): Use force_reg.
24581 (ix86_expand_ti_to_v1ti): Use force_reg.
24582 (ix86_expand_v1ti_shift): Use force_reg.
24583 (ix86_expand_v1ti_rotate): Use force_reg.
24584 (ix86_expand_v1ti_ashiftrt): Provide new three operation
24585 implementations for shifts by 111..126 bits. Use force_reg.
24587 2022-01-14 Martin Liska <mliska@suse.cz>
24589 * common/config/arm/arm-common.c (arm_target_mode): Fix
24590 warning: unterminated quoting directive [-Wformat=].
24592 2022-01-14 Siddhesh Poyarekar <siddhesh@gotplt.org>
24594 PR tree-optimization/104009
24595 * tree-object-size.c (compute_builtin_object_size): Bail out on
24597 (plus_stmt_object_size): Return maximum of wholesize and minimum
24598 of 0 for negative offset.
24600 2022-01-14 liuhongt <hongtao.liu@intel.com>
24605 * config/i386/i386.md (*xor2andn): Refine predicate of
24606 operands[0] from nonimmediate_operand to
24607 register_operand, remove TARGET_AVX512BW from condition.
24609 2022-01-14 David Malcolm <dmalcolm@redhat.com>
24611 * doc/extend.texi (Function Attributes): Note that "tainted_args" can
24612 be used on field decls.
24613 (Common Function Attributes): Add entry on "tainted_args" attribute.
24615 2022-01-13 Anthony Sharp <anthonysharp15@gmail.com>
24616 Jason Merrill <jason@redhat.com>
24619 * doc/invoke.texi: Documentation for Wmissing-template-keyword.
24621 2022-01-13 Uroš Bizjak <ubizjak@gmail.com>
24624 * config/i386/i386.md (*ashlqi_ext<mode>_2): New insn pattern.
24625 (*<any_shiftrt:insn>qi_ext<mode>_2): Ditto.
24626 * config/i386/mmx.md (<any_shift:insn>v2qi):
24627 New insn_and_split pattern.
24629 2022-01-13 Robin Dapp <rdapp@linux.ibm.com>
24631 * internal-fn.c (expand_partial_load_optab_fn): Add bias.
24632 (expand_partial_store_optab_fn): Likewise.
24633 (internal_len_load_store_bias): New function.
24634 * internal-fn.h (VECT_PARTIAL_BIAS_UNSUPPORTED): New define.
24635 (internal_len_load_store_bias): New function.
24636 * tree-vect-loop-manip.c (vect_set_loop_controls_directly): Set bias.
24637 (vect_set_loop_condition_partial_vectors): Add header_seq parameter.
24638 * tree-vect-loop.c (vect_verify_loop_lens): Verify bias.
24639 (vect_estimate_min_profitable_iters): Account for bias.
24640 (vect_get_loop_len): Add bias-adjusted length.
24641 * tree-vect-stmts.c (vectorizable_store): Use.
24642 (vectorizable_load): Use.
24643 * tree-vectorizer.h (struct rgroup_controls): Add bias-adjusted length.
24644 (LOOP_VINFO_PARTIAL_LOAD_STORE_BIAS): New macro.
24645 * config/rs6000/vsx.md: Use const0 bias predicate.
24646 * doc/md.texi: Document bias value.
24648 2022-01-13 Andrew MacLeod <amacleod@redhat.com>
24650 PR tree-optimization/83072
24651 PR tree-optimization/83073
24652 PR tree-optimization/97909
24653 * fold-const.c (expr_not_equal_to): Use a multi-range class.
24655 2022-01-13 Andrew MacLeod <amacleod@redhat.com>
24657 PR tree-optimization/96707
24658 * range-op.cc (operator_rshift::lhs_op1_relation): New.
24660 2022-01-13 Uroš Bizjak <ubizjak@gmail.com>
24662 * config/i386/mmx.md (negv2qi): Disparage GPR alternative a bit.
24663 Disable for TARGET_PARTIAL_REG_STALL unless optimizing for size.
24664 (negv2qi splitters): Use lowpart_subreg instead of
24665 gen_lowpart to create subreg.
24666 (<plusminus:insn>v2qi3): Disparage GPR alternative a bit.
24667 Disable for TARGET_PARTIAL_REG_STALL unless optimizing for size.
24668 (<plusminus:insn>v2qi3 splitters): Use lowpart_subreg instead of
24669 gen_lowpart to create subreg.
24670 * config/i386/i386.md (*subqi_ext<mode>_2): Move.
24672 2022-01-13 Uroš Bizjak <ubizjak@gmail.com>
24675 * config/i386/mmx.md (*xop_pcmov_<mode>): Use VI_16_32 mode iterator.
24677 2022-01-13 Martin Liska <mliska@suse.cz>
24679 * common/config/arm/arm-common.c (arm_target_mode): Wrap
24680 keywords with %<, %> and remove trailing punctuation char.
24681 (arm_canon_arch_option_1): Likewise.
24682 (arm_asm_auto_mfpu): Likewise.
24683 * config/arm/arm-builtins.c (arm_expand_builtin): Likewise.
24684 * config/arm/arm.c (arm_options_perform_arch_sanity_checks): Likewise.
24685 (use_vfp_abi): Likewise.
24686 (aapcs_vfp_is_call_or_return_candidate): Likewise.
24687 (arm_handle_cmse_nonsecure_entry): Likewise.
24688 (arm_handle_cmse_nonsecure_call): Likewise.
24689 (thumb1_md_asm_adjust): Likewise.
24691 2022-01-13 Paul A. Clarke <pc@us.ibm.com>
24693 * config/rs6000/smmintrin.h (_mm_round_pd, _mm_round_ps,
24694 _mm_round_sd, _mm_round_ss, _MM_FROUND_TO_NEAREST_INT,
24695 _MM_FROUND_TO_ZERO, _MM_FROUND_TO_POS_INF, _MM_FROUND_TO_NEG_INF,
24696 _MM_FROUND_CUR_DIRECTION, _MM_FROUND_RAISE_EXC, _MM_FROUND_NO_EXC,
24697 _MM_FROUND_NINT, _MM_FROUND_FLOOR, _MM_FROUND_CEIL, _MM_FROUND_TRUNC,
24698 _MM_FROUND_RINT, _MM_FROUND_NEARBYINT): New.
24699 (_mm_ceil_pd, _mm_ceil_ps, _mm_ceil_sd, _mm_ceil_ss, _mm_floor_pd,
24700 _mm_floor_ps, _mm_floor_sd, _mm_floor_ss): Convert from function to
24703 2022-01-13 Jakub Jelinek <jakub@redhat.com>
24705 PR tree-optimization/103989
24706 * tree-inline.c (setup_one_parameter): Don't copy parms with
24709 2022-01-13 Thomas Schwinge <thomas@codesourcery.com>
24711 * tree-streamer-out.c (pack_ts_base_value_fields): Don't pack
24712 'TYPE_ADDR_SPACE' for offloading.
24713 * tree-streamer-in.c (unpack_ts_base_value_fields): Don't unpack
24714 'TYPE_ADDR_SPACE' for offloading.
24716 2022-01-13 Julian Brown <julian@codesourcery.com>
24717 Thomas Schwinge <thomas@codesourcery.com>
24719 * omp-oacc-kernels-decompose.cc (add_wait): New function, split out
24721 (add_async_clauses_and_wait): ...here. Call new outlined function.
24722 (decompose_kernels_region_body): Add wait at the end of
24723 explicitly-asynchronous kernels regions.
24725 2022-01-13 Thomas Schwinge <thomas@codesourcery.com>
24727 PR middle-end/100280
24728 * omp-oacc-kernels-decompose.cc (maybe_build_inner_data_region):
24729 Mark variables used in synthesized data clauses as addressable.
24731 2022-01-13 Martin Liska <mliska@suse.cz>
24733 * config/epiphany/epiphany.c (epiphany_mode_priority):
24734 Use gcc_unreachable for not handled cases.
24736 2022-01-13 Martin Liska <mliska@suse.cz>
24738 * config/epiphany/epiphany.c (epiphany_handle_interrupt_attribute):
24739 Use %qs format specifier.
24740 (epiphany_override_options): Wrap keyword in %<, %>.
24742 2022-01-13 Haochen Jiang <haochen.jiang@intel.com>
24745 * config/i386/i386.md (*xor2andn): New define_insn_and_split.
24747 2022-01-13 Xionghu Luo <luoxhu@linux.ibm.com>
24749 * config/rs6000/altivec.md (sldoi_to_mov<mode>): New.
24751 2022-01-12 Uroš Bizjak <ubizjak@gmail.com>
24755 * config/i386/i386-expand.c (ix86_emit_vec_binop): New static function.
24756 (ix86_expand_sse_movcc): Use ix86_emit_vec_binop instead of gen_rtx_X
24757 when constructing vector logic RTXes.
24758 (expand_vec_perm_pshufb2): Ditto.
24759 * config/i386/mmx.md (negv2qi): Disparage GPR alternative a bit.
24760 (<plusminus:insn>v2qi3): Ditto.
24761 (vcond<mode><mode>): Re-enable for TARGET_SSE2.
24762 (vcondu<mode><mode>): Ditto.
24763 (vcond_mask_<mode><mode>): Ditto.
24764 (one_cmpl<VI_32:mode>2): Remove expander.
24765 (one_cmpl<VI_16_32:mode>2): Rename from one_cmplv2qi.
24766 Use VI_16_32 mode iterator.
24767 (one_cmpl<VI_16_32:mode>2 splitters): Use VI_16_32 mode iterator.
24768 Use lowpart_subreg instead of gen_lowpart to create subreg.
24769 (*andnot<VI_16_32:mode>3): Merge from "*andnot<VI_32:mode>" and
24770 "*andnotv2qi3" insn patterns using VI_16_32 mode iterator.
24771 Disparage GPR alternative a bit. Add CC clobber.
24772 (*andnot<VI_16_32:mode>3 splitters): Use VI_16_32 mode iterator.
24773 Use lowpart_subreg instead of gen_lowpart to create subreg.
24774 (*<any_logic:code><VI_16_32:mode>3): Merge from
24775 "*<any_logic:code><VI_32:mode>" and "*<any_logic:code>v2qi3" insn patterns
24776 using VI_16_32 mode iterator. Disparage GPR alternative a bit.
24778 (*<any_logic:code><VI_16_32:mode>3 splitters):Use VI_16_32 mode
24779 iterator. Use lowpart_subreg instead of gen_lowpart to create subreg.
24781 2022-01-12 Clément Chigot <clement.chigot@atos.net>
24783 * configure.ac: Check sizeof ino_t and dev_t.
24784 (HOST_STAT_FOR_64BIT_INODES): New AC_DEFINE to provide stat
24785 syscall being able to handle 64bit inodes.
24786 * config.in: Regenerate.
24787 * configure: Regenerate.
24788 * incpath.c (HOST_STAT_FOR_64BIT_INODES): New define.
24789 (remove_duplicates): Use it.
24791 2022-01-12 Andrew MacLeod <amacleod@redhat.com>
24793 PR tree-optimization/103551
24794 * tree-vrp.c (execute_ranger_vrp): Always set EDGE_EXECUTABLE.
24796 2022-01-12 Richard Biener <rguenther@suse.de>
24798 PR tree-optimization/103990
24799 * tree-pass.h (tail_merge_optimize): Drop unused argument.
24800 * tree-ssa-tail-merge.c (tail_merge_optimize): Likewise.
24801 * tree-ssa-pre.c (pass_pre::execute): Retain TODO_cleanup_cfg
24802 and adjust call to tail_merge_optimize.
24804 2022-01-12 Andre Vieira <andre.simoesdiasvieira@arm.com>
24806 * tree-vect-loop.c (vect-analyze-loop): Handle scenario where target
24807 does not add autovectorize_vector_modes.
24809 2022-01-12 Martin Liska <mliska@suse.cz>
24811 * config/aarch64/aarch64.c (aarch64_parse_boolean_options): Use
24812 %qs where possible.
24813 (aarch64_parse_sve_width_string): Likewise.
24814 (aarch64_override_options_internal): Likewise.
24815 (aarch64_print_hint_for_extensions): Likewise.
24816 (aarch64_validate_sls_mitigation): Likewise.
24817 (aarch64_handle_attr_arch): Likewise.
24818 (aarch64_handle_attr_cpu): Likewise.
24819 (aarch64_handle_attr_tune): Likewise.
24820 (aarch64_handle_attr_isa_flags): Likewise.
24822 2022-01-12 Martin Liska <mliska@suse.cz>
24824 * config.gcc: Include elfos.h before ${tm_file}.
24826 2022-01-12 Hans-Peter Nilsson <hp@axis.com>
24828 * config/cris/cris.c: Quote identifiers in parameters to error
24829 and internal_error, and remove extraneous spaces with punctuation.
24830 * config/cris/cris.h (CRIS_ASSERT): When passing on stringified
24831 expression to internal_error, pass it as a parameter instead of
24832 appending it to the format part.
24834 2022-01-12 Hans-Peter Nilsson <hp@axis.com>
24836 * config/cris/cris.c (cris_postdbr_cmpelim): Parenthesize
24839 2022-01-11 qing zhao <qing.zhao@oracle.com>
24841 * gimplify.c (gimple_add_init_for_auto_var): Delete the 3rd argument.
24842 Change the 3rd argument of function .DEFERRED_INIT to the name of the
24844 (gimplify_decl_expr): Delete the 3rd argument when call
24845 gimple_add_init_for_auto_var.
24846 * internal-fn.c (expand_DEFERRED_INIT): Update comments to reflect
24847 the 3rd argument change of function .DEFERRED_INIT.
24848 * tree-cfg.c (verify_gimple_call): Update comments and verification
24849 to reflect the 3rd argument change of function .DEFERRED_INIT.
24850 * tree-sra.c (generate_subtree_deferred_init): Delete the 3rd argument.
24851 (sra_modify_deferred_init): Change the 3rd argument of function
24852 .DEFERRED_INIT to the name of the decl.
24854 2022-01-11 Thomas Koenig <tkoenig@gcc.gnu.org>
24856 * flag-types.h (enum gfc_convert): Add flags for
24859 2022-01-11 Michael Meissner <meissner@the-meissners.org>
24861 * config/rs6000/rs6000.c (rs6000_option_override_internal): Remove
24862 checks for only C/C++ front ends before allowing the long double
24863 format to change without a warning.
24865 2022-01-11 Richard Sandiford <richard.sandiford@arm.com>
24867 PR rtl-optimization/103974
24868 * ira-int.h (ira_subloop_allocnos_can_differ_p): Take an
24869 extra argument, default true, that says whether old-reload
24870 targets should be excluded.
24871 * ira-color.c (color_pass): Pass false.
24873 2022-01-11 Uroš Bizjak <ubizjak@gmail.com>
24876 * config/i386/mmx.md (vcond<mode><mode>):
24877 Use VI_16_32 mode iterator. Enable for TARGET_SSE4_1.
24878 (vcondu<mode><mode>): Ditto.
24879 (vcond_mask_<mode><mode>): Ditto.
24880 (mmx_pblendvb_v8qi): Rename from mmx_pblendvb64.
24881 (mmx_pblendvb_<mode>): Rename from mmx_pblendvb32.
24882 Use VI_16_32 mode iterator.
24883 * config/i386/i386-expand.c (ix86_expand_sse_movcc):
24884 Update for rename. Handle V2QImode.
24885 (expand_vec_perm_blend): Update for rename.
24887 2022-01-11 Jakub Jelinek <jakub@redhat.com>
24890 * tree.def (OBJ_TYPE_REF): Document type of OBJ_TYPE_REF_TOKEN.
24892 2022-01-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
24894 PR middle-end/70090
24895 * tree-object-size.c (size_valid_p): New function.
24896 (size_for_offset): Remove OFFSET constness assertion.
24897 (addr_object_size): Build dynamic expressions for object
24898 sizes and use size_valid_p to decide if it is valid for the
24899 given OBJECT_SIZE_TYPE.
24900 (compute_builtin_object_size): Allow dynamic offsets when
24901 computing size at O0.
24902 (call_object_size): Call size_valid_p.
24903 (plus_stmt_object_size): Allow non-constant offset and use
24904 size_valid_p to decide if it is valid for the given
24907 2022-01-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
24909 PR middle-end/70090
24910 * tree-object-size.c (alloc_object_size): Make and return
24911 non-constant size expression.
24912 (call_object_size): Return expression or unknown based on
24913 whether dynamic object size is requested.
24915 2022-01-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
24917 PR middle-end/70090
24918 * tree-object-size.c: Include tree-dfa.h.
24919 (parm_object_size): New function.
24920 (collect_object_sizes_for): Call it.
24922 2022-01-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
24924 PR middle-end/70090
24925 * builtins.c (fold_builtin_object_size): Adjust for dynamic size
24927 * tree-object-size.c: Include gimplify-me.h.
24928 (struct object_size_info): New member UNKNOWNS.
24929 (size_initval_p, size_usable_p, object_sizes_get_raw): New
24931 (object_sizes_get): Return suitable gimple variable for
24933 (bundle_sizes): New function.
24934 (object_sizes_set): Use it and handle dynamic object size
24936 (object_sizes_set_temp): New function.
24937 (size_for_offset): Adjust for dynamic size expressions.
24938 (emit_phi_nodes, propagate_unknowns, gimplify_size_expressions):
24940 (compute_builtin_object_size): Call gimplify_size_expressions
24942 (dynamic_object_size): New function.
24943 (cond_expr_object_size): Use it.
24944 (phi_dynamic_object_size): New function.
24945 (collect_object_sizes_for): Call it for OST_DYNAMIC. Adjust to
24946 accommodate dynamic object sizes.
24948 2022-01-11 Siddhesh Poyarekar <siddhesh@gotplt.org>
24949 Jakub Jelinek <jakub@redhat.com>
24951 PR tree-optimization/103961
24952 * tree-object-size.c (plus_stmt_object_size): Always avoid
24953 computing offset for -1 size.
24955 2022-01-11 Andrew MacLeod <amacleod@redhat.com>
24957 PR tree-optimization/103821
24958 * range-op.cc (range_operator::fold_range): Only do precise ranges
24959 when there are not too many subranges.
24961 2022-01-11 Richard Sandiford <richard.sandiford@arm.com>
24963 * ira-costs.c (ira_tune_allocno_costs): Fix missing rclass
24964 definition in IRA_HARD_REGNO_ADD_COST_MULTIPLIER code.
24966 2022-01-11 Roger Sayle <roger@nextmovesoftware.com>
24967 Richard Biener <rguenther@suse.de>
24969 * tree-ssa-math-opts.c (struct widen_mul_stats): Add a
24970 highpart_mults_inserted field.
24971 (convert_mult_to_highpart): New function to convert right shift
24972 of a widening multiply into a MULT_HIGHPART_EXPR.
24973 (math_opts_dom_walker::after_dom_children) [RSHIFT_EXPR]:
24974 Call new convert_mult_to_highpart function.
24975 (pass_optimize_widening_mul::execute): Add a statistics counter
24976 for tracking "highpart multiplications inserted" events.
24978 2022-01-11 Xionghu Luo <luoxhu@linux.ibm.com>
24981 * config/rs6000/rs6000-protos.h (rs6000_is_valid_rotate_dot_mask): New
24983 * config/rs6000/rs6000.c (rs6000_is_valid_rotate_dot_mask): New
24985 * config/rs6000/rs6000.md (*branch_anddi3_dot): New.
24987 2022-01-11 Olivier Hainque <hainque@adacore.com>
24989 * gcc.c (driver_handle_option): State --sysroot as
24992 2022-01-11 Kewen Lin <linkw@linux.ibm.com>
24994 * config/rs6000/rs6000.c (rs6000_disable_incompatible_switches): Remove
24995 useless related to option -mno-power10.
24997 2022-01-11 Haochen Jiang <haochen.jiang@intel.com>
25000 * config/i386/sse.md (*andnot<mode>3): Extend predicate of
25001 operands[1] from register_operand to vector_operand.
25003 2022-01-10 Uroš Bizjak <ubizjak@gmail.com>
25006 * config/i386/i386-expand.c (ix86_expand_int_sse_cmp):
25008 * config/i386/mmx.md (<sat_plusminus:insn><mode>3):
25009 Use VI1_16_32 mode iterator.
25010 (*eq<mode>3): Ditto.
25011 (*gt<mode>3): Ditto.
25012 (*xop_maskcmp<mode>3): Ditto.
25013 (*xop_maskcmp_uns<mode>3): Ditto.
25014 (vec_cmp<mode><mode>): Ditto.
25015 (vec_cmpu<mode><mode>): Ditto.
25017 2022-01-10 Richard Biener <rguenther@suse.de>
25019 PR tree-optimization/103948
25020 * tree-vect-generic.c (expand_vector_condition): Return true if
25021 all ones vector is returned for true, all zeros vector for false
25022 and the target defines corresponding vec_cmp{,u}MN named RTX pattern.
25024 2022-01-10 Paul A. Clarke <pc@us.ibm.com>
25026 * config/rs6000/smmintrin.h (_mm_blendv_epi8): Use vec_blendv
25027 when _ARCH_PWR10. Use signed types.
25028 (_mm_blendv_ps): Use vec_blendv when _ARCH_PWR10.
25029 (_mm_blendv_pd): Likewise.
25031 2022-01-10 Andre Vieira <andre.simoesdiasvieira@arm.com>
25033 * tree-vectorizer.c (better_epilogue_loop_than_p): Round factors up for
25035 * tree-vect-loop.c (vect_analyze_loop): Re-analyze all modes for
25036 epilogues, unless we are guaranteed that we can't have partial vectors.
25037 * genopinit.c: (partial_vectors_supported): Generate new function.
25039 2022-01-10 Jakub Jelinek <jakub@redhat.com>
25042 * config/i386/i386.c (classify_argument): Add zero_width_bitfields
25043 argument, when seeing DECL_FIELD_CXX_ZERO_WIDTH_BIT_FIELD bitfields,
25044 always ignore them, when seeing other zero sized bitfields, either
25045 set zero_width_bitfields to 1 and ignore it or if equal to 2 process
25046 it. Pass it to recursive calls. Add wrapper
25047 with old arguments and diagnose ABI differences for C structures
25048 with zero width bitfields. Formatting fixes.
25050 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25052 PR rtl-optimization/98782
25053 * ira-int.h (ira_soft_conflict): Declare.
25054 * ira-color.c (max_soft_conflict_loop_depth): New constant.
25055 (ira_soft_conflict): New function.
25056 (spill_soft_conflicts): Likewise.
25057 (assign_hard_reg): Use them to handle the case described by
25058 the comment above ira_soft_conflict.
25059 (improve_allocation): Likewise.
25060 * ira.c (check_allocation): Allow allocnos with "soft" conflicts
25061 to share the same register.
25063 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25065 PR rtl-optimization/98782
25066 * ira-int.h (ira_caller_save_cost): New function.
25067 (ira_caller_save_loop_spill_p): Likewise.
25068 * ira-build.c (ira_propagate_hard_reg_costs): Test whether it is
25069 cheaper to spill a call-clobbered register throughout a loop rather
25070 than spill it around each individual call. If so, treat all
25071 call-clobbered registers as conflicts and...
25072 (propagate_allocno_info): ...do not propagate call information
25073 from the child to the parent.
25074 * ira-color.c (move_spill_restore): Update accordingly.
25075 * ira-costs.c (ira_tune_allocno_costs): Use ira_caller_save_cost.
25077 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25079 PR rtl-optimization/98782
25080 * ira-int.h (ira_allocno::might_conflict_with_parent_p): New field.
25081 (ALLOCNO_MIGHT_CONFLICT_WITH_PARENT_P): New macro.
25082 (ira_single_region_allocno_p): New function.
25083 (ira_total_conflict_hard_regs): Likewise.
25084 * ira-build.c (ira_create_allocno): Initialize
25085 ALLOCNO_MIGHT_CONFLICT_WITH_PARENT_P.
25086 (ira_propagate_hard_reg_costs): New function.
25087 (propagate_allocno_info): Use it. Try to avoid propagating
25088 hard register conflicts to parent allocnos if we can handle
25089 the conflicts by spilling instead. Limit the propagated
25090 register costs to the cost of spilling throughout the child loop.
25091 * ira-color.c (color_pass): Use ira_single_region_allocno_p to
25092 test whether a child and parent allocno can share the same
25094 (move_spill_restore): Adjust for the new behavior of
25095 propagate_allocno_info.
25097 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25099 PR rtl-optimization/98782
25100 * ira-int.h (ira_subloop_allocnos_can_differ_p): New function,
25102 * ira-color.c (color_pass): ...here.
25104 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25106 PR rtl-optimization/98782
25107 * ira-color.c (color_pass): Add comments to describe the spill costs.
25108 (move_spill_restore): Likewise. Fix reversed calculation.
25110 2022-01-10 Richard Sandiford <richard.sandiford@arm.com>
25112 PR rtl-optimization/98782
25113 * ira-int.h (ira_loop_border_costs): New class.
25114 * ira-color.c (ira_loop_border_costs::ira_loop_border_costs):
25116 (calculate_allocno_spill_cost): Use ira_loop_border_costs.
25117 (color_pass): Likewise.
25118 (move_spill_restore): Likewise.
25120 2022-01-10 Eric Botcazou <ebotcazou@adacore.com>
25123 * coretypes.h (unwind_info_type): Swap UI_SEH and UI_TARGET.
25125 2022-01-10 Richard Biener <rguenther@suse.de>
25127 PR tree-optimization/100359
25128 * tree-ssa-loop-ivcanon.c (try_unroll_loop_completely):
25129 Allow non-growing peeling with !allow_peel and UL_ALL.
25131 2022-01-08 Roger Sayle <roger@nextmovesoftware.com>
25133 * config/i386/i386-expand.c (ix86_expand_vector_move): Add
25134 special case for TImode to V1TImode moves, going via V2DImode.
25136 2022-01-08 Jakub Jelinek <jakub@redhat.com>
25139 * match.pd ((ptr) (x p+ y) p+ z -> (ptr) (x p+ (y + z))): New GENERIC
25142 2022-01-08 David Malcolm <dmalcolm@redhat.com>
25144 * doc/analyzer.texi
25145 (Special Functions for Debugging the Analyzer): Document
25146 __analyzer_dump_escaped.
25148 2022-01-08 David Malcolm <dmalcolm@redhat.com>
25150 * doc/analyzer.texi (Other Debugging Techniques): Document
25151 region::is_named_decl_p.
25153 2022-01-07 Andrew Pinski <apinski@marvell.com>
25156 * config/arm/aarch-common.c (arm_md_asm_adjust):
25157 Use a temp if !REG_P.
25159 2022-01-07 Uroš Bizjak <ubizjak@gmail.com>
25161 * config/i386/mmx.md (*move<V_32:mode>_internal): Add isa attribute.
25162 (*movv2qi_internal): Remve sse2 requirement for alternatives 4,5.
25164 2022-01-07 liuhongt <hongtao.liu@intel.com>
25166 PR rtl-optimization/103750
25167 * fwprop.c (forward_propagate_into): Allow propagations from
25168 inner loop to outer loop.
25170 2022-01-07 Roger Sayle <roger@nextmovesoftware.com>
25172 * config/nvptx/nvptx.md (*cnot<mode>2): New define_insn.
25174 2022-01-07 Haochen Gui <guihaoc@gcc.gnu.org>
25176 * config/rs6000/rs6000.md (rs6000_mffscrni): Define.
25177 (rs6000_set_fpscr_rn): Change the type of operand[0] from DI to SI.
25178 Call gen_rs6000_mffscrni when operand[0] is a const_0_to_3_operand.
25180 2022-01-07 liuhongt <hongtao.liu@intel.com>
25182 * config/i386/sse.md
25183 (*<avx512>_eq<mode>3<mask_scalar_merge_name>_1): Extend to
25184 UNSPEC_PCMP_UNSIGNED.
25186 2022-01-07 liuhongt <hongtao.liu@intel.com>
25189 * config/i386/i386-expand.c (ix86_expand_vector_set): Not use
25190 gen_avx2_pblendph_1 when elt == 0.
25191 * config/i386/sse.md (avx2_pblendph): Rename to ..
25192 (avx2_pblend<ssemodesuffix>_1).. this, and extend to V16HI.
25193 (*avx2_pblendw): Rename to ..
25194 (*avx2_pblend<ssemodesuffix>): .. this, and extend to V16HF.
25195 (avx2_pblendw): Rename to ..
25196 (*avx2_pblend<ssemodesuffix>): .. this, and extend to V16HF.
25197 (blendsuf): Removed.
25198 (sse4_1_pblend<blendsuf>): Renamed to ..
25199 (sse4_1_pblend<ssemodesuffix>): .. this.
25201 2022-01-06 H.J. Lu <hjl.tools@gmail.com>
25204 * config/i386/i386.c (ix86_output_indirect_function_return):
25205 Generate INT3 after indirect jmp for -mharden-sls=indirect-jmp.
25207 2022-01-06 H.J. Lu <hjl.tools@gmail.com>
25210 * config/i386/i386-opts.h (harden_sls): Replace
25211 harden_sls_indirect_branch with harden_sls_indirect_jmp.
25212 * config/i386/i386.c (ix86_output_jmp_thunk_or_indirect):
25214 (ix86_output_indirect_jmp): Likewise.
25215 (ix86_output_call_insn): Likewise.
25216 * config/i386/i386.opt: Replace indirect-branch with
25217 indirect-jmp. Replace harden_sls_indirect_branch with
25218 harden_sls_indirect_jmp.
25219 * doc/invoke.texi (-harden-sls=): Replace indirect-branch with
25222 2022-01-06 Uroš Bizjak <ubizjak@gmail.com>
25224 * config/i386/i386.c (ix86_output_ssemov) <MODE_DI>:
25225 Add %q modifier for operands in general registers.
25226 <MODE_SI>: Add %q modifier for operands in general registers.
25227 * config/i386/i386.md (*movhi_internal): Change type attribute of
25228 xmm-gpr interunit alternatives 9,10 to ssemov and mode attribute
25229 to SImode for non-avx512fp16 targets.
25230 (*movhf_internal): Ditto for xmm-gpr interunit alternatives 6,8.
25231 * config/i386/mmx.md (*movv2qi_internal):
25232 Ditto for xmm-gpr interunit alternatives 8,9.
25234 2022-01-06 Kito Cheng <kito.cheng@sifive.com>
25236 * common/config/riscv/riscv-common.c (riscv_implied_info): Add
25238 (riscv_ext_version_table): Add version info for vector extensions.
25239 (riscv_ext_flag_table): Add option mask for vector extensions.
25240 * config/riscv/riscv-opts.h (MASK_VECTOR_EEW_32): New.
25241 (MASK_VECTOR_EEW_64): New.
25242 (MASK_VECTOR_EEW_FP_32): New.
25243 (MASK_VECTOR_EEW_FP_64): New.
25244 (MASK_ZVL32B): New.
25245 (MASK_ZVL64B): New.
25246 (MASK_ZVL128B): New.
25247 (MASK_ZVL256B): New.
25248 (MASK_ZVL512B): New.
25249 (MASK_ZVL1024B): New.
25250 (MASK_ZVL2048B): New.
25251 (MASK_ZVL4096B): New.
25252 (MASK_ZVL8192B): New.
25253 (MASK_ZVL16384B): New.
25254 (MASK_ZVL32768B): New.
25255 (MASK_ZVL65536B): New.
25256 (TARGET_ZVL32B): New.
25257 (TARGET_ZVL64B): New.
25258 (TARGET_ZVL128B): New.
25259 (TARGET_ZVL256B): New.
25260 (TARGET_ZVL512B): New.
25261 (TARGET_ZVL1024B): New.
25262 (TARGET_ZVL2048B): New.
25263 (TARGET_ZVL4096B): New.
25264 (TARGET_ZVL8192B): New.
25265 (TARGET_ZVL16384B): New.
25266 (TARGET_ZVL32768B): New.
25267 (TARGET_ZVL65536B): New.
25268 * config/riscv/riscv.opt (Mask(VECTOR)): New.
25269 (riscv_vector_eew_flags): New.
25270 (riscv_zvl_flags): New.
25272 2022-01-06 Kito Cheng <kito.cheng@sifive.com>
25274 * common/config/riscv/riscv-common.c
25275 (riscv_subset_list::parse_multiletter_ext): Allow ext. name has
25278 2022-01-06 Jakub Jelinek <jakub@redhat.com>
25280 PR tree-optimization/103899
25281 * expr.c (expand_expr_real_1): Add a workaround for bogus uninit
25282 warning by moving context variable to the only spot where it is used
25283 and moving gcc_assert into if body.
25285 2022-01-06 Jakub Jelinek <jakub@redhat.com>
25287 PR rtl-optimization/103908
25288 * ifcvt.c (bb_valid_for_noce_process_p): Punt on bbs ending with
25291 2022-01-05 Bill Schmidt <wschmidt@linux.ibm.com>
25294 * config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
25295 Skip over instances with undefined function types.
25297 2022-01-05 Andrew Pinski <apinski@marvell.com>
25300 * config/i386/i386.h (x86_mfence): Mark with GTY.
25302 2022-01-05 Uroš Bizjak <ubizjak@gmail.com>
25305 * config/i386/mmx.md (VI_16_32): New mode iterator.
25306 (VI1_16_32): Ditto.
25307 (mmxvecsize): Handle V2QI mode.
25308 (<smaxmin:code><mode>3): Rename from <smaxmin:code>v4qi3.
25309 Use VI1_16_32 mode iterator.
25310 (<umaxmin:code><mode>3): Rename from <umaxmin:code>v4qi3.
25311 Use VI1_16_32 mode iterator.
25312 (abs<mode>2): Use VI_16_32 mode iterator.
25313 (uavgv2qi3_ceil): New insn pattern.
25315 2022-01-05 Martin Sebor <msebor@redhat.com>
25317 * gimple-ssa-warn-access.cc (pass_waccess::maybe_warn_memmodel): Use
25318 %qs to avoid -Wformat-diag.
25320 2022-01-05 Uroš Bizjak <ubizjak@gmail.com>
25323 * config/i386/mmx.md (one_cmplv2qi2): Change
25324 alternatives 1,2 type from sselog to sselog1.
25326 2022-01-05 Uroš Bizjak <ubizjak@gmail.com>
25329 * config/i386/i386-expand.c (expand_vec_perm_pshufb): Fix number of
25330 narrow mode remapped elements for !one_operand_p case.
25332 2022-01-05 Richard Biener <rguenther@suse.de>
25334 PR tree-optimization/103816
25335 * tree-vect-data-refs.c (vect_analyze_group_access_1): Also
25336 check DR_GROUP_GAP compute for overflow and representability.
25338 2022-01-05 Jakub Jelinek <jakub@redhat.com>
25341 * gimple-fold.c (fold_stmt_1): Don't call maybe_fold_reference
25342 for DEBUG stmts with ADDR_EXPR gimple_debug_bind_get_value,
25343 it can do unwanted rhs folding like &a[0] into &2.0 etc.
25345 2022-01-05 Kewen Lin <linkw@linux.ibm.com>
25348 * config/rs6000/rs6000.c (TARGET_NEED_IPA_FN_TARGET_INFO): New macro.
25349 (TARGET_UPDATE_IPA_FN_TARGET_INFO): Likewise.
25350 (rs6000_need_ipa_fn_target_info): New function.
25351 (rs6000_update_ipa_fn_target_info): Likewise.
25352 (rs6000_can_inline_p): Adjust for ipa function summary target info.
25353 * config/rs6000/rs6000.h (RS6000_FN_TARGET_INFO_HTM): New macro.
25354 * ipa-fnsummary.c (ipa_dump_fn_summary): Adjust for ipa function
25355 summary target info.
25356 (analyze_function_body): Adjust for ipa function summary target info
25357 and call hook rs6000_need_ipa_fn_target_info and
25358 rs6000_update_ipa_fn_target_info.
25359 (ipa_merge_fn_summary_after_inlining): Adjust for ipa function summary
25361 (inline_read_section): Likewise.
25362 (ipa_fn_summary_write): Likewise.
25363 * ipa-fnsummary.h (ipa_fn_summary::target_info): New member.
25364 * doc/tm.texi: Regenerate.
25365 * doc/tm.texi.in (TARGET_UPDATE_IPA_FN_TARGET_INFO): Document new hook.
25366 (TARGET_NEED_IPA_FN_TARGET_INFO): Likewise.
25367 * target.def (update_ipa_fn_target_info): New hook.
25368 (need_ipa_fn_target_info): Likewise.
25369 * targhooks.c (default_need_ipa_fn_target_info): New function.
25370 (default_update_ipa_fn_target_info): Likewise.
25371 * targhooks.h (default_update_ipa_fn_target_info): New declare.
25372 (default_need_ipa_fn_target_info): Likewise.
25374 2022-01-04 Martin Sebor <msebor@redhat.com>
25376 PR middle-end/99612
25377 * builtins.c (get_memmodel): Move warning code to
25378 gimple-ssa-warn-access.cc.
25379 (expand_builtin_atomic_compare_exchange): Same.
25380 (expand_ifn_atomic_compare_exchange): Same.
25381 (expand_builtin_atomic_load): Same.
25382 (expand_builtin_atomic_store): Same.
25383 (expand_builtin_atomic_clear): Same.
25384 * doc/extend.texi (__atomic_exchange_n): Update valid memory
25386 * gimple-ssa-warn-access.cc (memmodel_to_uhwi): New function.
25387 (struct memmodel_pair): New struct.
25388 (memmodel_name): New function.
25389 (pass_waccess::maybe_warn_memmodel): New function.
25390 (pass_waccess::check_atomic_memmodel): New function.
25391 (pass_waccess::check_atomic_builtin): Handle memory model.
25392 * input.c (expansion_point_location_if_in_system_header): Return
25393 original location if expansion location is in a system header.
25395 2022-01-04 Uroš Bizjak <ubizjak@gmail.com>
25398 * config/i386/mmx.md (one_cmplv2qi3): New insn pattern.
25399 (one_cmplv2qi3 splitters): New post-reload splitters.
25400 (*andnotv2qi3): New insn pattern.
25401 (andnotv2qi3 splitters): New post-reload splitters.
25402 (<any_logic:code>v2qi3): New insn pattern.
25403 (<any_logic:insn>v2qi3 splitters): New post-reload splitters.
25405 2022-01-04 Richard Biener <rguenther@suse.de>
25407 PR tree-optimization/103800
25408 * tree-vect-loop.c (vectorizable_phi): Remove assert and
25411 2022-01-04 Richard Biener <rguenther@suse.de>
25413 PR tree-optimization/103690
25414 * tree-pass.h (tail_merge_optimize): Adjust.
25415 * tree-ssa-tail-merge.c (tail_merge_optimize): Pass in whether
25416 to re-split critical edges, move CFG cleanup ...
25417 * tree-ssa-pre.c (pass_pre::execute): ... here, before
25418 simple_dce_from_worklist and delay freeing inserted_exprs from
25420 (fini_pre): .. here.
25422 2022-01-04 Roger Sayle <roger@nextmovesoftware.com>
25424 * config/nvptx/nvptx.h (STORE_FLAG_VALUE): Change to 1.
25425 * config/nvptx/nvptx.md (movbi): Use P1 constraint for true.
25426 (setcc_from_bi): Remove SImode specific pattern.
25427 (setcc<mode>_from_bi): Provide more general HSDIM pattern.
25428 (extendbi<mode>2, zeroextendbi<mode>2): Provide instructions
25429 for sign- and zero-extending BImode predicates to integers.
25430 (setcc_int<mode>): Remove previous (-1-based) instructions.
25431 (cstorebi4): Remove BImode to SImode specific expander.
25432 (cstore<mode>4): Fix indentation. Expand using setccsi_from_bi.
25433 (cstore<mode>4): For both integer and floating point modes.
25435 2022-01-04 Olivier Hainque <hainque@adacore.com>
25437 * gcc.c (driver_handle_option): do_save --sysroot.
25439 2022-01-04 Richard Biener <rguenther@suse.de>
25441 PR tree-optimization/103864
25442 PR tree-optimization/103544
25443 * tree-vect-slp.c (vect_analyze_slp_instance): Exclude
25444 reductions wrapped in conversions from SLP handling.
25445 (vect_analyze_slp): Revert PR103544 change.
25447 2022-01-04 Jakub Jelinek <jakub@redhat.com>
25449 PR rtl-optimization/103860
25450 * shrink-wrap.c (try_shrink_wrapping): Don't call can_get_prologue
25451 uselessly for blocks for which it has been called already.
25453 2022-01-04 Cui,Lili <lili.cui@intel.com>
25455 * common/config/i386/cpuinfo.h (get_intel_cpu): Add new model values
25456 to Alderlake and Rocketlake.
25458 2022-01-04 Chung-Lin Tang <cltang@codesourcery.com>
25460 PR middle-end/103643
25461 * gimplify.c (gimplify_omp_affinity): Adjust gimplify_expr of entire
25462 OMP_CLAUSE_DECL to use 'is_gimple_lvalue, fb_lvalue'
25464 2022-01-04 liuhongt <hongtao.liu@intel.com>
25467 * config/i386/sse.md (*bit_and_float_vector_all_ones):
25468 Force_reg operand 1 to avoid ICE.
25470 2022-01-04 Jason Merrill <jason@redhat.com>
25472 * tree-pretty-print.c (do_niy): Add spc parameter.
25474 (print_call_name): Add spc local variable.
25476 2022-01-03 Uroš Bizjak <ubizjak@gmail.com>
25479 * config/i386/mmx.md (mov<V_32:mode>): Remove TARGET_SSE2 constraint.
25480 (mov<V_32:mode>_internal): Ditto.
25481 (*push<V_32:mode>_rex64): Ditto.
25482 (movmisalign<V_32:mode>): Ditto.
25483 (*push<V_32:mode>_rex64 splitter): Enable for
25484 TARGET_64BIT && TARGET_SSE.
25485 (*push<V_32:mode>2): Remove insn pattern.
25487 2022-01-03 Andrew Pinski <apinski@marvell.com>
25490 * doc/extend.texi: Extend the documentation about Complex
25491 types for casting and also rewrite the __real__/__imag__
25492 expression portion to use tables.
25493 Move __builtin_complex to the Complex type section.
25495 2022-01-03 Jakub Jelinek <jakub@redhat.com>
25498 * internal-fn.def (ATOMIC_ADD_FETCH_CMP_0, ATOMIC_SUB_FETCH_CMP_0,
25499 ATOMIC_AND_FETCH_CMP_0, ATOMIC_OR_FETCH_CMP_0, ATOMIC_XOR_FETCH_CMP_0):
25501 * internal-fn.h (ATOMIC_OP_FETCH_CMP_0_EQ, ATOMIC_OP_FETCH_CMP_0_NE,
25502 ATOMIC_OP_FETCH_CMP_0_LT, ATOMIC_OP_FETCH_CMP_0_LE,
25503 ATOMIC_OP_FETCH_CMP_0_GT, ATOMIC_OP_FETCH_CMP_0_GE): New enumerators.
25504 * internal-fn.c (expand_ATOMIC_ADD_FETCH_CMP_0,
25505 expand_ATOMIC_SUB_FETCH_CMP_0, expand_ATOMIC_AND_FETCH_CMP_0,
25506 expand_ATOMIC_OR_FETCH_CMP_0, expand_ATOMIC_XOR_FETCH_CMP_0): New
25508 * optabs.def (atomic_add_fetch_cmp_0_optab,
25509 atomic_sub_fetch_cmp_0_optab, atomic_and_fetch_cmp_0_optab,
25510 atomic_or_fetch_cmp_0_optab, atomic_xor_fetch_cmp_0_optab): New
25512 * builtins.h (expand_ifn_atomic_op_fetch_cmp_0): Declare.
25513 * builtins.c (expand_ifn_atomic_op_fetch_cmp_0): New function.
25514 * tree-ssa-ccp.c: Include internal-fn.h.
25515 (optimize_atomic_bit_test_and): Add . before internal fn call
25516 in function comment. Change return type from void to bool and
25517 return true only if successfully replaced.
25518 (optimize_atomic_op_fetch_cmp_0): New function.
25519 (pass_fold_builtins::execute): Use optimize_atomic_op_fetch_cmp_0
25520 for BUILT_IN_ATOMIC_{ADD,SUB,AND,OR,XOR}_FETCH_{1,2,4,8,16} and
25521 BUILT_IN_SYNC_{ADD,SUB,AND,OR,XOR}_AND_FETCH_{1,2,4,8,16},
25522 for *XOR* ones only if optimize_atomic_bit_test_and failed.
25523 * config/i386/sync.md (atomic_<plusminus_mnemonic>_fetch_cmp_0<mode>,
25524 atomic_<logic>_fetch_cmp_0<mode>): New define_expand patterns.
25525 (atomic_add_fetch_cmp_0<mode>_1, atomic_sub_fetch_cmp_0<mode>_1,
25526 atomic_<logic>_fetch_cmp_0<mode>_1): New define_insn patterns.
25527 * doc/md.texi (atomic_add_fetch_cmp_0<mode>,
25528 atomic_sub_fetch_cmp_0<mode>, atomic_and_fetch_cmp_0<mode>,
25529 atomic_or_fetch_cmp_0<mode>, atomic_xor_fetch_cmp_0<mode>): Document
25530 new named patterns.
25532 2022-01-03 Richard Biener <rguenther@suse.de>
25534 PR middle-end/103851
25535 * tree-cfg.c (move_sese_region_to_fn): Always release SSA names.
25537 2022-01-03 Jakub Jelinek <jakub@redhat.com>
25540 * symtab.c: Include fold-const.h.
25541 (symtab_node::equal_address_to): If folding_initializer is true,
25542 handle it like memory_accessed. Simplify.
25544 2022-01-03 Martin Liska <mliska@suse.cz>
25546 * doc/extend.texi: Use ; for function declarations.
25548 2022-01-03 Jakub Jelinek <jakub@redhat.com>
25551 * symtab.c (symtab_node::equal_address_to): Return 0 if one of
25552 VAR_DECLs has "non overlapping" attribute and rs1 != rs2.
25554 2022-01-03 Jakub Jelinek <jakub@redhat.com>
25556 * gcc.c (process_command): Update copyright notice dates.
25557 * gcov-dump.c (print_version): Ditto.
25558 * gcov.c (print_version): Ditto.
25559 * gcov-tool.c (print_version): Ditto.
25560 * gengtype.c (create_file): Ditto.
25561 * doc/cpp.texi: Bump @copying's copyright year.
25562 * doc/cppinternals.texi: Ditto.
25563 * doc/gcc.texi: Ditto.
25564 * doc/gccint.texi: Ditto.
25565 * doc/gcov.texi: Ditto.
25566 * doc/install.texi: Ditto.
25567 * doc/invoke.texi: Ditto.
25569 2022-01-02 Uroš Bizjak <ubizjak@gmail.com>
25572 * config/i386/i386.h (VALID_SSE2_REG_MODE): Add V2QImode.
25573 (VALID_INT_MODE_P): Ditto.
25574 * config/i386/i386.c (ix86_secondary_reload): Handle
25575 V2QImode reloads from SSE register to memory.
25576 (vector_mode_supported_p): Always return true for V2QImode.
25577 * config/i386/i386.md (*subqi_ext<mode>_2): New insn pattern.
25578 (*negqi_ext<mode>_2): Ditto.
25579 * config/i386/mmx.md (movv2qi): New expander.
25580 (movmisalignv2qi): Ditto.
25581 (*movv2qi_internal): New insn pattern.
25582 (*pushv2qi2): Ditto.
25583 (negv2qi2 and splitters): Ditto.
25584 (<plusminus:insn>v2qi3 and splitters): Ditto.
25586 2022-01-02 John David Anglin <danglin@gcc.gnu.org>
25588 * config/pa/pa.md (atomic_storeq): Use optab_libfunc to access
25589 sync_lock_test_and_set libfunc. Call convert_memory_address to
25590 convert memory address to Pmode.
25591 (atomic_storehi, atomic_storesi, atomic_storedi): Likewise.
25594 Copyright (C) 2022 Free Software Foundation, Inc.
25596 Copying and distribution of this file, with or without modification,
25597 are permitted in any medium without royalty provided the copyright
25598 notice and this notice are preserved.