gcc/ChangeLog

   1 2020-01-07  Martin Liska  <mliska@suse.cz>
   2
   3         * params.opt: Add Optimization for various parameters.
   4
   5 2020-01-07  Martin Liska  <mliska@suse.cz>
   6
   7         PR ipa/83411
   8         * doc/extend.texi: Explain cloning for target_clone
   9         attribute.
  10
  11 2020-01-07  Martin Liska  <mliska@suse.cz>
  12
  13         PR tree-optimization/92860
  14         * common.opt: Make in Optimization option
  15         as it is affected by -O0, which is an Optimization
  16         option.
  17         * tree-inline.c (tree_inlinable_function_p):
  18         Use opt_for_fn for warn_inline.
  19         (expand_call_inline): Likewise.
  20
  21 2020-01-07  Martin Liska  <mliska@suse.cz>
  22
  23         PR tree-optimization/92860
  24         * common.opt: Make flag_ree as optimization
  25         attribute.
  26
  27 2020-01-07  Martin Liska  <mliska@suse.cz>
  28
  29         PR optimization/92860
  30         * params.opt: Mark param_min_crossjump_insns with Optimization
  31         keyword.
  32
  33 2020-01-07  Luo Xiong Hu  <luoxhu@linux.ibm.com>
  34
  35         * ipa-inline-analysis.c (estimate_growth): Fix typo.
  36         * ipa-inline.c (caller_growth_limits): Use OR instead of AND.
  37
  38 2020-01-06  Michael Meissner  <meissner@linux.ibm.com>
  39
  40         * config/rs6000/rs6000.c (hard_reg_and_mode_to_addr_mask): New
  41         helper function to return the valid addressing formats for a given
  42         hard register and mode.
  43         (rs6000_adjust_vec_address): Call hard_reg_and_mode_to_addr_mask.
  44
  45         * config/rs6000/constraints.md (Q constraint): Update
  46         documentation.
  47         * doc/md.texi (RS/6000 constraints): Update 'Q' cosntraint
  48         documentation.
  49
  50         * config/rs6000/vsx.md (vsx_extract_<mode>_var, VSX_D iterator):
  51         Use 'Q' for doing vector extract from memory.
  52         (vsx_extract_v4sf_var): Use 'Q' for doing vector extract from
  53         memory.
  54         (vsx_extract_<mode>_var, VSX_EXTRACT_I iterator): Use 'Q' for
  55         doing vector extract from memory.
  56         (vsx_extract_<mode>_<VS_scalar>mode_var): Use 'Q' for doing vector
  57         extract from memory.
  58
  59         * config/rs6000/rs6000.c (rs6000_adjust_vec_address): Add support
  60         for the offset being 34-bits when -mcpu=future is used.
  61
  62 2020-01-06  John David Anglin  <danglin@gcc.gnu.org>
  63
  64         * config/pa/pa.md: Revert change to use ordered_comparison_operator
  65         instead of cmpib_comparison_operator in cmpib patterns.
  66         * config/pa/predicates.md (cmpib_comparison_operator): Revert removal
  67         of cmpib_comparison_operator.  Revise comment.
  68
  69 2020-01-06  Richard Sandiford  <richard.sandiford@arm.com>
  70
  71         * tree-vect-slp.c (vect_build_slp_tree_1): Require all shifts
  72         in an IFN_DIV_POW2 node to be equal.
  73
  74 2020-01-06  Richard Sandiford  <richard.sandiford@arm.com>
  75
  76         * tree-vect-stmts.c (vect_check_load_store_mask): Rename to...
  77         (vect_check_scalar_mask): ...this.
  78         (vectorizable_store, vectorizable_load): Update call accordingly.
  79         (vectorizable_call): Use vect_check_scalar_mask to check the mask
  80         argument in calls to conditional internal functions.
  81
  82 2020-01-06  Andrew Stubbs  <ams@codesourcery.com>
  83
  84         * config/gcn/gcn-valu.md (subv64di3): Use separate alternatives for
  85         '0' matching inputs.
  86         (subv64di3_exec): Likewise.
  87
  88 2020-01-06  Bryan Stenson  <bryan@siliconvortex.com>
  89
  90         * config/mips/mips.c (vr4130_align_insns): Fix typo.
  91         * doc/md.texi (movstr): Likewise.
  92
  93 2020-01-06  Andrew Stubbs  <ams@codesourcery.com>
  94
  95         * config/gcn/gcn-valu.md (vec_extract<mode><scalar_mode>): Add early
  96         clobber.
  97
  98 2020-01-06  Richard Sandiford  <richard.sandiford@arm.com>
  99
 100         * config/aarch64/t-aarch64 ($(srcdir)/config/aarch64/aarch64-tune.md):
 101         Depend on...
 102         (s-aarch64-tune-md): ...this new stamp file.  Pipe the new contents
 103         to a temporary file and use move-if-change to update the real
 104         file where necessary.
 105
 106 2020-01-06  Richard Sandiford  <richard.sandiford@arm.com>
 107
 108         * config/aarch64/aarch64-sve.md (@aarch64_sel_dup<mode>): Use Upl
 109         rather than Upa for CPY /M.
 110
 111 2020-01-06  Andrew Stubbs  <ams@codesourcery.com>
 112
 113         * config/gcn/gcn.c (gcn_inline_constant_p): Allow 64 as an inline
 114         immediate.
 115
 116 2020-01-06  Martin Liska  <mliska@suse.cz>
 117
 118     PR tree-optimization/92860
 119     * params.opt: Mark param_max_combine_insns with Optimization
 120     keyword.
 121
 122 2020-01-05  Jakub Jelinek  <jakub@redhat.com>
 123
 124         PR target/93141
 125         * config/i386/i386.md (SWIDWI): New mode iterator.
 126         (DWI, dwi): Add TImode variants.
 127         (addv<mode>4): Use SWIDWI iterator instead of SWI.  Use
 128         <general_hilo_operand> instead of <general_operand>.  Use
 129         CONST_SCALAR_INT_P instead of CONST_INT_P.
 130         (*addv<mode>4_1): Rename to ...
 131         (addv<mode>4_1): ... this.
 132         (QWI): New mode attribute.
 133         (*addv<dwi>4_doubleword, *addv<dwi>4_doubleword_1): New
 134         define_insn_and_split patterns.
 135         (*addv<mode>4_overflow_1, *addv<mode>4_overflow_2): New define_insn
 136         patterns.
 137         (uaddv<mode>4): Use SWIDWI iterator instead of SWI.  Use
 138         <general_hilo_operand> instead of <general_operand>.
 139         (*addcarry<mode>_1): New define_insn.
 140         (*add<dwi>3_doubleword_cc_overflow_1): New define_insn_and_split.
 141
 142 2020-01-03  Konstantin Kharlamov  <Hi-Angel@yandex.ru>
 143
 144         * gdbinit.in (pr, prl, pt, pct, pgg, pgq, pgs, pge, pmz, pdd, pbs, pbm):
 145         Use "call" instead of "set".
 146
 147 2020-01-03  Martin Jambor  <mjambor@suse.cz>
 148
 149         PR ipa/92917
 150         * ipa-cp.c (print_all_lattices): Skip functions without info.
 151
 152 2020-01-03  Jakub Jelinek  <jakub@redhat.com>
 153
 154         PR target/93089
 155         * config/i386/i386-options.c (ix86_simd_clone_adjust): If
 156         TARGET_PREFER_AVX128, use prefer-vector-width=256 for 'c' and 'd'
 157         simd clones.  If TARGET_PREFER_AVX256, use prefer-vector-width=512
 158         for 'e' simd clones.
 159
 160         PR target/93089
 161         * config/i386/i386.opt (x_prefer_vector_width_type): Remove TargetSave
 162         entry.
 163         (mprefer-vector-width=): Add Save.
 164         * config/i386/i386-options.c (ix86_target_string): Add PVW argument, print
 165         -mprefer-vector-width= if non-zero.  Fix up -mfpmath= comment.
 166         (ix86_debug_options, ix86_function_specific_print): Adjust
 167         ix86_target_string callers.
 168         (ix86_valid_target_attribute_inner_p): Handle prefer-vector-width=.
 169         (ix86_valid_target_attribute_tree): Likewise.
 170         * config/i386/i386-options.h (ix86_target_string): Add PVW argument.
 171         * config/i386/i386-expand.c (ix86_expand_builtin): Adjust
 172         ix86_target_string caller.
 173
 174         PR target/93110
 175         * config/i386/i386.md (abs<mode>2): Use expand_simple_binop instead of
 176         emitting ASHIFTRT, XOR and MINUS by hand.  Use gen_int_mode with QImode
 177         instead of gen_int_shift_amount + convert_modes.
 178
 179         PR rtl-optimization/93088
 180         * loop-iv.c (find_single_def_src): Punt after looking through
 181         128 reg copies for regs with single definitions.  Move definitions
 182         to first uses.
 183
 184 2020-01-02  Dennis Zhang  <dennis.zhang@arm.com>
 185
 186         * config/arm/arm-c.c (arm_cpu_builtins): Define
 187         __ARM_FEATURE_MATMUL_INT8, __ARM_FEATURE_BF16_VECTOR_ARITHMETIC,
 188         __ARM_FEATURE_BF16_SCALAR_ARITHMETIC, and
 189         __ARM_BF16_FORMAT_ALTERNATIVE when enabled.
 190         * config/arm/arm-cpus.in (armv8_6, i8mm, bf16): New features.
 191         * config/arm/arm-tables.opt: Regenerated.
 192         * config/arm/arm.c (arm_option_reconfigure_globals): Initialize
 193         arm_arch_i8mm and arm_arch_bf16 when enabled.
 194         * config/arm/arm.h (TARGET_I8MM): New macro.
 195         (TARGET_BF16_FP, TARGET_BF16_SIMD): Likewise.
 196         * config/arm/t-aprofile: Add matching rules for -march=armv8.6-a.
 197         * config/arm/t-arm-elf (all_v8_archs): Add armv8.6-a.
 198         * config/arm/t-multilib: Add matching rules for -march=armv8.6-a.
 199         (v8_6_a_simd_variants): New.
 200         (v8_*_a_simd_variants): Add i8mm and bf16.
 201         * doc/invoke.texi (armv8.6-a, i8mm, bf16): Document new options.
 202
 203 2020-01-02  Jakub Jelinek  <jakub@redhat.com>
 204
 205         PR ipa/93087
 206         * predict.c (compute_function_frequency): Don't call
 207         warn_function_cold on functions that already have cold attribute.
 208
 209 2020-01-01  John David Anglin  <danglin@gcc.gnu.org>
 210
 211         PR target/67834
 212         * config/pa/pa.c (pa_elf_select_rtx_section): New.  Put references to
 213         COMDAT group function labels in .data.rel.ro.local section.
 214         * config/pa/pa32-linux.h (TARGET_ASM_SELECT_RTX_SECTION): Define.
 215
 216         PR target/93111
 217         * config/pa/pa.md (scc): Use ordered_comparison_operator instead of
 218         comparison_operator in B and S integer comparisons.  Likewise, use
 219         ordered_comparison_operator instead of cmpib_comparison_operator in
 220         cmpib patterns.
 221         * config/pa/predicates.md (cmpib_comparison_operator): Remove.
 222
 223 2020-01-01  Jakub Jelinek  <jakub@redhat.com>
 224
 225         Update copyright years.
 226
 227         * gcc.c (process_command): Update copyright notice dates.
 228         * gcov-dump.c (print_version): Ditto.
 229         * gcov.c (print_version): Ditto.
 230         * gcov-tool.c (print_version): Ditto.
 231         * gengtype.c (create_file): Ditto.
 232         * doc/cpp.texi: Bump @copying's copyright year.
 233         * doc/cppinternals.texi: Ditto.
 234         * doc/gcc.texi: Ditto.
 235         * doc/gccint.texi: Ditto.
 236         * doc/gcov.texi: Ditto.
 237         * doc/install.texi: Ditto.
 238         * doc/invoke.texi: Ditto.
 239
 240 2020-01-01  Jan Hubicka  <hubicka@ucw.cz>
 241
 242         * ipa.c (walk_polymorphic_call_targets): Fix updating of overall
 243         summary.
 244
 245 2020-01-01  Jakub Jelinek  <jakub@redhat.com>
 246
 247         PR tree-optimization/93098
 248         * match.pd (popcount): For shift amounts, use integer_onep
 249         or wi::to_widest () == cst instead of tree_to_uhwi () == cst
 250         tests.  Make sure that precision is power of two larger than or equal
 251         to 16.  Ensure shift is never negative.  Use HOST_WIDE_INT_UC macro
 252         instead of ULL suffixed constants.  Formatting fixes.
 253 \f
 254 Copyright (C) 2020 Free Software Foundation, Inc.
 255
 256 Copying and distribution of this file, with or without modification,
 257 are permitted in any medium without royalty provided the copyright
 258 notice and this notice are preserved.