]> git.ipfire.org Git - thirdparty/gcc.git/commitdiff
Daily bump.
authorGCC Administrator <gccadmin@gcc.gnu.org>
Thu, 21 Dec 2023 00:18:03 +0000 (00:18 +0000)
committerGCC Administrator <gccadmin@gcc.gnu.org>
Thu, 21 Dec 2023 00:18:03 +0000 (00:18 +0000)
contrib/ChangeLog
gcc/ChangeLog
gcc/DATESTAMP
gcc/c-family/ChangeLog
gcc/c/ChangeLog
gcc/cp/ChangeLog
gcc/fortran/ChangeLog
gcc/testsuite/ChangeLog
libgcc/ChangeLog
libgomp/ChangeLog

index cfd23862732f749bf4bb1e1231b0eec590733734..a521553ae4cb25ceb379b26d7e30f781538f5a64 100644 (file)
@@ -1,3 +1,7 @@
+2023-12-20  Jason Merrill  <jason@redhat.com>
+
+       * gcc-git-customization.sh: Add git gcc-style alias.
+
 2023-12-14  Pierre-Emmanuel Patry  <pierre-emmanuel.patry@embecosm.com>
            Arthur Cohen  <arthur.cohen@embecosm.com>
 
index 5c20d1a8dd32737f394825ed2983bcbd46a48670..cefddcbc640251aa3ae99bbf07e19edb7e60b7e0 100644 (file)
@@ -1,3 +1,264 @@
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       PR target/113087
+       * config/riscv/riscv-v.cc (expand_select_vl): Optimize SELECT_VL.
+
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       PR target/113087
+       * config/riscv/riscv-vsetvl.cc: Disallow fusion when VL modification pollutes non AVL use.
+
+2023-12-20  Rimvydas Jasinskas  <rimvydas.jas@gmail.com>
+
+       * doc/invoke.texi: Document the new file extensions
+
+2023-12-20  Richard Sandiford  <richard.sandiford@arm.com>
+
+       PR rtl-optimization/111702
+       * cse.cc (set::mode): Move earlier.
+       (set::src_in_memory, set::src_volatile): Convert to bitfields.
+       (set::is_fake_set): New member variable.
+       (add_to_set): Add an is_fake_set parameter.
+       (find_sets_in_insn): Update calls accordingly.
+       (cse_insn): Do not apply REG_EQUAL notes to fake sets.  Do not
+       try to optimize them either, or validate changes to them.
+
+2023-12-20  Kuan-Lin Chen  <rufus@andestech.com>
+
+       * config/riscv/predicates.md (move_operand): Reject symbolic operands
+       with a type SYMBOL_FORCE_TO_MEM.
+       (call_insn_operand): Support for CM_Large.
+       (pcrel_symbol_operand): New.
+       * config/riscv/riscv-c.cc (riscv_cpu_cpp_builtins): Add builtin_define
+       "__riscv_cmodel_large".
+       * config/riscv/riscv-opts.h (riscv_code_model): Add CM_LARGE.
+       * config/riscv/riscv-protos.h (riscv_symbol_type): Add
+       SYMBOL_FORCE_TO_MEM.
+       * config/riscv/riscv.cc (riscv_classify_symbol) Support CM_LARGE model.
+       (riscv_symbol_insns) Add SYMBOL_FORCE_TO_MEM.
+       (riscv_cannot_force_const_mem): Ditto.
+       (riscv_split_symbol): Ditto.
+       (riscv_force_address): Check pseudo reg available before force_reg.
+       (riscv_size_ok_for_small_data_p): Disable in CM_LARGE model.
+       (riscv_can_use_per_function_literal_pools_p): New.
+       (riscv_elf_select_rtx_section): Handle per-function literal pools.
+       (riscv_output_mi_thunk): Add riscv_in_thunk_func.
+       (riscv_option_override): Support CM_LARGE model.
+       (riscv_function_ok_for_sibcall): Disable sibcalls in CM_LARGE model.
+       (riscv_in_thunk_func): New static.
+       * config/riscv/riscv.md (unspec): Define UNSPEC_FORCE_FOR_MEM.
+       (*large_load_address): New.
+       * config/riscv/riscv.opt (code_model): New.
+
+2023-12-20  Wang Pengcheng  <wangpengcheng.pp@bytedance.com>
+
+       * config/riscv/riscv.cc (riscv_macro_fusion_pair_p): Fix condition.
+
+2023-12-20  Andre Vieira  <andre.simoesdiasvieira@arm.com>
+
+       PR target/112787
+       * tree-vect-generic.cc (type_for_widest_vector_mode): Change function to
+       use original vector type and check widest vector mode has at most the
+       same number of elements.
+       (get_compute_type): Pass original vector type rather than the element
+       type to type_for_widest_vector_mode and remove now obsolete check for
+       the number of elements.
+
+2023-12-20  Siddhesh Poyarekar  <siddhesh@gotplt.org>
+
+       * tree-object-size.cc (object_size_info): Remove UNKNOWNS.
+       Drop all references to it.
+       (object_sizes_set): Move unknowns propagation code to...
+       (gimplify_size_expressions): ... here.  Also free reexamine
+       bitmap.
+       (propagate_unknowns): New parameter UNKNOWNS.  Update callers.
+
+2023-12-20  Thomas Schwinge  <thomas@codesourcery.com>
+
+       * config/gcn/gcn.h (LIBSTDCXX): Define to "gcc".
+
+2023-12-20  Richard Biener  <rguenther@suse.de>
+
+       * tree-vect-slp.cc (vect_transform_slp_perm_load_1): Also handle
+       CTOR and VIEW_CONVERT up to the load when performing chain DCE.
+
+2023-12-20  Xi Ruoyao  <xry111@xry111.site>
+
+       * config/loongarch/loongarch.cc
+       (loongarch_expand_vector_init_same): Remove "temp2" and reuse
+       "temp" instead.
+       (loongarch_expand_vector_init): Use gcc_unreachable () instead
+       of gcc_assert (0), and fix the comment for it.
+
+2023-12-20  Xi Ruoyao  <xry111@xry111.site>
+
+       PR target/113033
+       * config/loongarch/loongarch.cc
+       (loongarch_expand_vector_init_same): Replace gen_reg_rtx +
+       emit_move_insn with force_reg.
+       (loongarch_expand_vector_init): Likewise.
+
+2023-12-20  Xi Ruoyao  <xry111@xry111.site>
+
+       PR target/113034
+       * config/loongarch/lasx.md (UNSPEC_LASX_XVFCMP_*): Remove.
+       (lasx_xvfcmp_caf_<flasxfmt>): Remove.
+       (lasx_xvfcmp_cune_<FLASX:flasxfmt>): Remove.
+       (FSC256_UNS): Remove.
+       (fsc256): Remove.
+       (lasx_xvfcmp_<vfcond:fcc>_<FLASX:flasxfmt>): Remove.
+       (lasx_xvfcmp_<fsc256>_<FLASX:flasxfmt>): Remove.
+       * config/loongarch/lsx.md (UNSPEC_LSX_XVFCMP_*): Remove.
+       (lsx_vfcmp_caf_<flsxfmt>): Remove.
+       (lsx_vfcmp_cune_<FLSX:flsxfmt>): Remove.
+       (vfcond): Remove.
+       (fcc): Remove.
+       (FSC_UNS): Remove.
+       (fsc): Remove.
+       (lsx_vfcmp_<vfcond:fcc>_<FLSX:flsxfmt>): Remove.
+       (lsx_vfcmp_<fsc>_<FLSX:flsxfmt>): Remove.
+       * config/loongarch/simd.md
+       (fcond_simd): New define_code_iterator.
+       (<simd_isa>_<x>vfcmp_<fcond:fcond_simd>_<simdfmt>):
+       New define_insn.
+       (fcond_simd_rev): New define_code_iterator.
+       (fcond_rev_asm): New define_code_attr.
+       (<simd_isa>_<x>vfcmp_<fcond:fcond_simd_rev>_<simdfmt>):
+       New define_insn.
+       (fcond_inv): New define_code_iterator.
+       (fcond_inv_rev): New define_code_iterator.
+       (fcond_inv_rev_asm): New define_code_attr.
+       (<simd_isa>_<x>vfcmp_<fcond_inv>_<simdfmt>): New define_insn.
+       (<simd_isa>_<x>vfcmp_<fcond_inv:fcond_inv_rev>_<simdfmt>):
+       New define_insn.
+       (UNSPEC_SIMD_FCMP_CAF, UNSPEC_SIMD_FCMP_SAF,
+       UNSPEC_SIMD_FCMP_SEQ, UNSPEC_SIMD_FCMP_SUN,
+       UNSPEC_SIMD_FCMP_SUEQ, UNSPEC_SIMD_FCMP_CNE,
+       UNSPEC_SIMD_FCMP_SOR, UNSPEC_SIMD_FCMP_SUNE): New unspecs.
+       (SIMD_FCMP): New define_int_iterator.
+       (fcond_unspec): New define_int_attr.
+       (<simd_isa>_<x>vfcmp_<fcond_unspec>_<simdfmt>): New define_insn.
+       * config/loongarch/loongarch.cc (loongarch_expand_lsx_cmp):
+       Remove unneeded special cases.
+
+2023-12-20  demin.han  <demin.han@starfivetech.com>
+
+       * config/riscv/riscv-vector-costs.cc (max_number_of_live_regs): Fix
+       max live vregs calc
+       (preferred_new_lmul_p): Ditto
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       PR target/112962
+       * config/i386/i386-builtins.cc (ix86_builtins): Increase by one
+       element.
+       (def_builtin): If not -fnon-call-exceptions, set TREE_NOTHROW on
+       the builtin FUNCTION_DECL.  Add leaf attribute to DECL_ATTRIBUTES.
+       (ix86_add_new_builtins): Likewise.
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       PR tree-optimization/112941
+       * gimple-lower-bitint.cc (bitint_large_huge::handle_cast): If
+       save_cast_conditional, instead of adding assignment of t4 to
+       m_data[save_data_cnt + 1] before m_gsi, add phi nodes such that
+       t4 propagates to m_bb loop.  For constant idx, use
+       m_data[save_data_cnt] rather than m_data[save_data_cnt + 1] if inside
+       of the m_bb loop.
+       (bitint_large_huge::lower_mergeable_stmt): Clear m_bb when no longer
+       expanding inside of that loop.
+       (bitint_large_huge::lower_comparison_stmt): Likewise.
+       (bitint_large_huge::lower_addsub_overflow): Likewise.
+       (bitint_large_huge::lower_mul_overflow): Likewise.
+       (bitint_large_huge::lower_bit_query): Likewise.
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       * doc/invoke.texi (-Walloc-size): Add to the list of
+       warning options, remove unnecessary line-break.
+       (-Wcalloc-transposed-args): Document new warning.
+
+2023-12-20  Alex Coplan  <alex.coplan@arm.com>
+
+       PR target/113062
+       * config/aarch64/aarch64-ldp-fusion.cc
+       (ldp_bb_info::track_access): Punt on accesses with invalid
+       register operands, move definition of mem_size closer to its
+       first use.
+
+2023-12-20  Pan Li  <pan2.li@intel.com>
+
+       * config/riscv/riscv-v.cc (rvv_builder::npatterns_vid_diff_repeated_p):
+       New function to predicate the diff to vid is repeated or not.
+       (expand_const_vector): Add restriction
+       for the vid-diff code gen and implement general one.
+
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       * config/riscv/riscv.cc (riscv_legitimize_move): Fix ICE.
+
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       PR middle-end/112917
+       * builtins.cc (expand_bultin_stack_address): Add
+       STACK_POINTER_OFFSET.
+       * doc/extend.texi (__builtin_stack_address): Adjust.
+
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       PR rtl-optimization/113002
+       * cfgrtl.cc (commit_one_edge_insertion): Tolerate jumps in the
+       inserted sequence during expand.
+
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       * builtins.cc (delta_type): New template class.
+       (set_apply_args_size, get_apply_args_size): Replace with...
+       (saved_apply_args_size): ... this.
+       (set_apply_result_size, get_apply_result_size): Replace with...
+       (saved_apply_result_size): ... this.
+       (apply_args_size, apply_result_size): Adjust.
+
+2023-12-20  Jeff Law  <jlaw@ventanamicro.com>
+
+       * config/mcore/mcore.h (CC1_SPEC): Do not set -funsigned-bitfields.
+
+2023-12-20  Haochen Jiang  <haochen.jiang@intel.com>
+
+       * config/i386/avx512bwintrin.h: Allow 64 bit mask intrin usage
+       for -mno-evex512.
+       * config/i386/i386-builtin.def: Remove OPTION_MASK_ISA2_EVEX512
+       for 64 bit mask builtins.
+       * config/i386/i386.cc (ix86_hard_regno_mode_ok): Allow 64 bit
+       mask register for -mno-evex512.
+       * config/i386/i386.md (SWI1248_AVX512BWDQ_64): Remove
+       TARGET_EVEX512.
+       (*zero_extendsidi2): Change isa attribute to avx512bw.
+       (kmov_isa): Ditto.
+       (*anddi_1): Ditto.
+       (*andn<mode>_1): Remove TARGET_EVEX512.
+       (*one_cmplsi2_1_zext): Change isa attribute to avx512bw.
+       (*ashl<mode>3_1): Ditto.
+       (*lshr<mode>3_1): Ditto.
+       * config/i386/sse.md (SWI1248_AVX512BWDQ): Remove TARGET_EVEX512.
+       (SWI1248_AVX512BW): Ditto.
+       (SWI1248_AVX512BWDQ2): Ditto.
+       (*knotsi_1_zext): Ditto.
+       (kunpckdi): Ditto.
+       (SWI24_MASK): Removed.
+       (vec_pack_trunc_<mode>): Change iterator from SWI24_MASK to SWI24.
+       (vec_unpacks_lo_di): Remove TARGET_EVEX512.
+       (SWI48x_MASK): Removed.
+       (vec_unpacks_hi_<mode>): Change iterator from SWI48x_MASK to SWI48x.
+
+2023-12-20  Siddhesh Poyarekar  <siddhesh@gotplt.org>
+
+       PR tree-optimization/113012
+       * tree-object-size.cc (compute_builtin_object_size): Expand
+       comment for dynamic object sizes.
+       (collect_object_sizes_for): Always set COMPUTED bitmap for
+       dynamic object sizes.
+
 2023-12-20  Alexandre Oliva  <oliva@adacore.com>
 
        * ipa-strub.cc (gsi_insert_finally_seq_after_call): Likewise.
index 150569894451721d686b2316abef4d65c444e07c..1e1f595fc12de3d6864523e226fc89919cbffb01 100644 (file)
@@ -1 +1 @@
-20231220
+20231221
index 0401d893e877f7009c0dec83e330bd754e095110..38f39932d4a239a0a06464e3303fd91376eb20b6 100644 (file)
@@ -1,3 +1,16 @@
+2023-12-20  Florian Weimer  <fweimer@redhat.com>
+
+       PR c/113050
+       * c-common.cc (get_atomic_generic_size): Use
+       OPT_Wdiscarded_qualifiers instead of
+       OPT_Wincompatible_pointer_types.
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       * c.opt (Wcalloc-transposed-args): New warning.
+       * c-common.h (warn_for_calloc, warn_for_alloc_size): Declare.
+       * c-warn.cc (warn_for_calloc, warn_for_alloc_size): New functions.
+
 2023-12-16  Andrew Carlotti  <andrew.carlotti@arm.com>
 
        * c-attribs.cc (attr_target_exclusions): Make
index 04b703f791e8d2999e46f8d2e75d496ac922a276..678a3fbbb4bfb6875bd14d84fa26841786e7b47a 100644 (file)
@@ -1,3 +1,14 @@
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       * c-parser.cc (c_parser_postfix_expression_after_primary): Grow
+       sizeof_arg and sizeof_arg_loc arrays to 6 elements.  Call
+       warn_for_calloc if warn_calloc_transposed_args for functions with
+       alloc_size type attribute with 2 arguments.
+       (c_parser_expr_list): Use 6 instead of 3.
+       * c-typeck.cc (build_c_cast): Call warn_for_alloc_size for casts
+       of calls to functions with alloc_size type attribute.
+       (convert_for_assignment): Likewise.
+
 2023-12-19  Sandra Loosemore  <sandra@codesourcery.com>
 
        * c-parser.cc (omp_construct_selectors): Delete.
index 617f3c6aceb68b82ce188feab547815c30c88343..5db76e8bd2878fd41a2bef7bcec3d7482de07b71 100644 (file)
@@ -1,3 +1,14 @@
+2023-12-20  Jason Merrill  <jason@redhat.com>
+
+       PR c++/113088
+       PR c++/33799
+       * except.cc (maybe_splice_retval_cleanup): Handle an empty block.
+
+2023-12-20  Jason Merrill  <jason@redhat.com>
+
+       PR c++/103185
+       * typeck.cc (cp_build_array_ref): Handle swapped operands.
+
 2023-12-19  Patrick Palka  <ppalka@redhat.com>
 
        PR c++/113063
index ab9c0ae954c32c2e398c8996c0aed140a483fdbc..5ed35f785b61287ee4c6ee337d91dd95453289fe 100644 (file)
@@ -1,3 +1,10 @@
+2023-12-20  Rimvydas Jasinskas  <rimvydas.jas@gmail.com>
+
+       PR fortran/81615
+       * lang-specs.h (F951_CPP_OPTIONS): Do not hardcode ".f90" extension
+       (F951_CPP_EXTENSION): Use .fi/.fii for fixed/free form sources
+       * options.cc (form_from_filename): Handle the new extensions
+
 2023-12-19  Sandra Loosemore  <sandra@codesourcery.com>
 
        * gfortran.h: Include omp-selectors.h.
index eb6e900adc622222aae712755d10686e0889f423..8dbf8cba129e12c2b47cc684a4a8624e92a73bf4 100644 (file)
@@ -1,3 +1,106 @@
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       PR target/113087
+       * gcc.target/riscv/rvv/autovec/pr113087-2.c: New test.
+
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       PR target/113087
+       * gcc.target/riscv/rvv/autovec/pr113087-1.c: New test.
+
+2023-12-20  Richard Sandiford  <richard.sandiford@arm.com>
+
+       PR rtl-optimization/111702
+       * gcc.dg/rtl/aarch64/pr111702.c: New test.
+
+2023-12-20  Kuan-Lin Chen  <rufus@andestech.com>
+
+       * gcc.target/riscv/large-model.c: New test.
+
+2023-12-20  Jason Merrill  <jason@redhat.com>
+
+       PR c++/113088
+       PR c++/33799
+       * g++.dg/eh/return2.C: New test.
+
+2023-12-20  Jason Merrill  <jason@redhat.com>
+
+       PR c++/103185
+       * g++.dg/cpp1z/array-prvalue2.C: New test.
+       * g++.dg/cpp1z/eval-order3.C: Test swapped operands.
+
+2023-12-20  Andre Vieira  <andre.simoesdiasvieira@arm.com>
+
+       * gcc.target/aarch64/pr112787.c: New test.
+
+2023-12-20  Xi Ruoyao  <xry111@xry111.site>
+
+       PR target/113033
+       * gcc.target/loongarch/pr113033.c: New test.
+
+2023-12-20  Xi Ruoyao  <xry111@xry111.site>
+
+       PR target/113034
+       * gcc.target/loongarch/vfcmp-f.c: New test.
+       * gcc.target/loongarch/vfcmp-d.c: New test.
+       * gcc.target/loongarch/xvfcmp-f.c: New test.
+       * gcc.target/loongarch/xvfcmp-d.c: New test.
+       * gcc.target/loongarch/vector/lasx/lasx-vcond-2.c: Scan for cune
+       instead of cne.
+       * gcc.target/loongarch/vector/lsx/lsx-vcond-2.c: Likewise.
+
+2023-12-20  demin.han  <demin.han@starfivetech.com>
+
+       * gcc.dg/vect/costmodel/riscv/rvv/dynamic-lmul2-7.c: Moved to...
+       * gcc.dg/vect/costmodel/riscv/rvv/dynamic-lmul4-10.c: ...here.
+       * gcc.dg/vect/costmodel/riscv/rvv/dynamic-lmul4-4.c: Moved to...
+       * gcc.dg/vect/costmodel/riscv/rvv/dynamic-lmul8-11.c: ...here.
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       PR tree-optimization/112941
+       * gcc.dg/bitint-55.c: New test.
+
+2023-12-20  Jakub Jelinek  <jakub@redhat.com>
+
+       * gcc.dg/Walloc-size-4.c: New test.
+       * gcc.dg/Walloc-size-5.c: New test.
+       * gcc.dg/Wcalloc-transposed-args-1.c: New test.
+
+2023-12-20  Alex Coplan  <alex.coplan@arm.com>
+
+       PR target/113062
+       * gcc.dg/pr113062.c: New test.
+
+2023-12-20  Pan Li  <pan2.li@intel.com>
+
+       * gcc.target/riscv/rvv/autovec/bug-7.c: New test.
+
+2023-12-20  Juzhe-Zhong  <juzhe.zhong@rivai.ai>
+
+       * gcc.target/riscv/rvv/autovec/bug-8.c: New test.
+
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       PR rtl-optimization/113002
+       * gcc.dg/vect/pr113002.c: New.
+
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       * gcc.dg/torture/inline-mem-cmp-1.c: Copy timeout factor from
+       mem-cmp-1.c.
+       * gcc.dg/torture/inline-mem-cpy-1.c: Likewise.
+
+2023-12-20  Haochen Jiang  <haochen.jiang@intel.com>
+
+       * gcc.target/i386/avx10_1-6.c: Remove check for errors.
+       * gcc.target/i386/noevex512-2.c: Diito.
+
+2023-12-20  Siddhesh Poyarekar  <siddhesh@gotplt.org>
+
+       PR tree-optimization/113012
+       * gcc.dg/ubsan/pr113012.c: New test case.
+
 2023-12-19  Patrick Palka  <ppalka@redhat.com>
 
        PR c++/113063
index 0fc99e69b584691fdc02c06e5ba7fc8e6217ae4a..20da88228a91dce8e3069c5440bbdaf049a026dc 100644 (file)
@@ -1,3 +1,9 @@
+2023-12-20  Alexandre Oliva  <oliva@adacore.com>
+
+       PR middle-end/112917
+       * config.host (sparc, sparc64): Enable...
+       * config/sparc/t-sparc: ... this new fragment.
+
 2023-12-20  Alexandre Oliva  <oliva@adacore.com>
 
        * strub.c (ATTRIBUTE_NOINLINE): New.
index f88ff9f1005f6ede15aaddaba51ad5dec3912715..13167d74ffec4670f66f4b8314d5158f1503d1b6 100644 (file)
@@ -1,3 +1,10 @@
+2023-12-20  Julian Brown  <julian@codesourcery.com>
+
+       * plugin/plugin-nvptx.c (GOMP_OFFLOAD_memcpy2d): Adjust parameters to
+       avoid out-of-bounds array checks in CUDA runtime.
+       (GOMP_OFFLOAD_memcpy3d): Likewise.
+       * testsuite/libgomp.c-c++-common/memcpyxd-bias-1.c: New test.
+
 2023-12-18  Jakub Jelinek  <jakub@redhat.com>
 
        * testsuite/libgomp.c/declare-variant-1.c: Restrict the test to x86,