gcc.git
6 years agore PR target/81959 (PowerPC __float128 optimization fails with integer PRE_INC addresses)
Michael Meissner [Fri, 1 Dec 2017 23:52:20 +0000 (23:52 +0000)]
re PR target/81959 (PowerPC __float128 optimization fails with integer PRE_INC addresses)

[gcc]
2017-12-01  Michael Meissner  <meissner@linux.vnet.ibm.com>

PR target/81959
* config/rs6000/rs6000.c (rs6000_address_for_fpconvert): Check for
whether we can allocate pseudos before trying to fix an address.
* config/rs6000/rs6000.md (float_<mode>si2_hw): Make sure the
memory address is indexed or indirect.
(floatuns_<mode>si2_hw2): Likewise.

[gcct/testsuite]
2017-12-01  Michael Meissner  <meissner@linux.vnet.ibm.com>

PR target/81959
* gcc.target/powerpc/pr81959.c: New test.

From-SVN: r255341

6 years agocompiler: introduce size threshold for nil checks
Than McIntosh [Fri, 1 Dec 2017 23:12:13 +0000 (23:12 +0000)]
compiler: introduce size threshold for nil checks

    Add a new control variable to the Gogo class that stores the size
    threshold for nil checks. This value can be used to control the
    policy for deciding when a given deference operation needs a check and
    when it does not. A size threshold of -1 means that every potentially
    faulting dereference needs an explicit check (and branch to error
    call). A size threshold of K (where K > 0) means that if the size of
    the object being dereferenced is >= K, then we need a check.

    Reviewed-on: https://go-review.googlesource.com/80996

* go-c.h (go_create_gogo_args): Add nil_check_size_threshold
field.
* go-lang.c (go_langhook_init): Set nil_check_size_threshold.

From-SVN: r255340

6 years ago* doc/extend.texi: Mention C++14 library complex suffix.
Jason Merrill [Fri, 1 Dec 2017 21:17:32 +0000 (16:17 -0500)]
* doc/extend.texi: Mention C++14 library complex suffix.

From-SVN: r255339

6 years agoGive #include hints for <complex>.
Jason Merrill [Fri, 1 Dec 2017 20:19:13 +0000 (15:19 -0500)]
Give #include hints for <complex>.

* name-lookup.c (get_std_name_hint): Add <complex>.
* parser.c (cp_parser_diagnose_invalid_type_name): Call
suggest_alternative_in_explicit_scope.
(cp_parser_namespace_name): Likewise.

From-SVN: r255336

6 years agoPR c++/79228 - extensions hide C++14 complex literal operators
Jason Merrill [Fri, 1 Dec 2017 20:19:07 +0000 (15:19 -0500)]
PR c++/79228 - extensions hide C++14 complex literal operators

libcpp/
* expr.c (interpret_float_suffix): Ignore 'i' in C++14 and up.
(interpret_int_suffix): Likewise.
gcc/cp/
* parser.c (cp_parser_userdef_numeric_literal): Be helpful about
'i' in C++14 and up.

From-SVN: r255335

6 years ago* Makefile.in (TAGS): Add c-family/*.cc.
Jason Merrill [Fri, 1 Dec 2017 20:19:01 +0000 (15:19 -0500)]
* Makefile.in (TAGS): Add c-family/*.cc.

From-SVN: r255334

6 years ago[AArch64] Fix address printing on ILP32
Wilco Dijkstra [Fri, 1 Dec 2017 19:54:29 +0000 (19:54 +0000)]
[AArch64] Fix address printing on ILP32

Fix address printing for ILP32.  The md file uses 'a' in assembler
templates for symbolic addresses in adrp/add, which end up calling
aarch64_print_operand_address.  However in ILP32 these are not valid
memory addresses (being ptr_mode rather than Pmode), so the assert
triggers.  Since it is incorrect to use symbols in memory addresses
(besides literal pool accesses), change the 'a' to 'c' in the md file.

Skip one failing test in ILP32 which combines the 'p' modifier with the 'a'
assembler template to fake a memory reference.

    gcc/
* config/aarch64/aarch64.md (call_insn): Use %c rather than %a.
(call_value_insn): Likewise.
(sibcall_insn): Likewise.
(sibcall_value_insn): Likewise.
(movsi_aarch64): Likewise.
(movdi_aarch64): Likewise.
(add_losym_): Likewise.
(ldr_got_small_): Likewise.
(ldr_got_small_sidi): Likewise.
(ldr_got_small_28k_): Likewise.
(ldr_got_small_28k_sidi): Likewise.
* config/aarch64/aarch64.c (aarch64_print_address_internal):
Move output_addr_const to symbolic case. Add error check.
    testsuite/
* gcc.dg/asm-4.c: Skip on AArch64 with ILP32 as test is incorrect.

From-SVN: r255333

6 years agoFix an undefined behavior in fortran/decl.c
Qing Zhao [Fri, 1 Dec 2017 18:20:45 +0000 (18:20 +0000)]
Fix an undefined behavior in fortran/decl.c

From-SVN: r255332

6 years agore PR fortran/83224 (creating character array from elements shorter than declared...
Thomas Koenig [Fri, 1 Dec 2017 18:06:31 +0000 (18:06 +0000)]
re PR fortran/83224 (creating character array from elements shorter than declared does not pad with whitespace properly and aborts)

2017-12-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83224
* frontend-passes.c (create_var): Also handle
character arrays, handling deferred lenghts.

2017-12-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83224
* gfortran.dg/dependency_51.f90: New test.

From-SVN: r255331

6 years agoFix narrowing conversions in string_view types
Jonathan Wakely [Fri, 1 Dec 2017 16:10:31 +0000 (16:10 +0000)]
Fix narrowing conversions in string_view types

* include/experimental/string_view (basic_string_view::_S_compare):
Use value-init so narrowing conversions are not ill-formed.
* include/std/string_view (basic_string_view::_S_compare): Likewise.

From-SVN: r255321

6 years agoDisable -Wliteral-suffix for standard UDLs
Jonathan Wakely [Fri, 1 Dec 2017 16:10:25 +0000 (16:10 +0000)]
Disable -Wliteral-suffix for standard UDLs

* include/bits/basic_string.h (operator""s): Add pragmas to disable
-Wliteral-suffix warnings.
* include/experimental/string_view (operator""sv): Likewise.
* include/std/chrono (operator""h, operator""min, operator""s)
(operator""ms, operator""us, operator""ns): Likewise.
* include/std/complex (operator""if, operator""i, operator""il):
Likewise.
* include/std/string_view (operator""sv): Likewise.
* testsuite/20_util/duration/literals/range.cc: Adjust dg-error.

From-SVN: r255320

6 years agoS/390: Split MVC instruction for better forwarding
Andreas Krebbel [Fri, 1 Dec 2017 15:33:15 +0000 (15:33 +0000)]
S/390: Split MVC instruction for better forwarding

Certain lengths used in an MVC instruction might disable operand
forwarding.  Split MVCs into up to 2 forwardable ones if possible.

gcc/ChangeLog:

2017-12-01  Andreas Krebbel  <krebbel@linux.vnet.ibm.com>

* config/s390/predicates.md (plus16_Q_operand): New predicate.
* config/s390/s390.md: Disable MVC merging peephole if it would
disable operand forwarding.
(new peephole2): Split MVCs if it would turn them into up to 2
forwardable MVCs.

From-SVN: r255319

6 years agoAdd comment to fix -Wfallthrough warning
Jonathan Wakely [Fri, 1 Dec 2017 15:10:23 +0000 (15:10 +0000)]
Add comment to fix -Wfallthrough warning

* include/bits/locale_facets_nonio.tcc (money_get::_M_extract): Add
fallthrough comment.

From-SVN: r255318

6 years agoFix -Wempty-body warnings for debug assertions
Jonathan Wakely [Fri, 1 Dec 2017 15:10:20 +0000 (15:10 +0000)]
Fix -Wempty-body warnings for debug assertions

* include/bits/node_handle.h (_Node_handle_common::operator=)
(_Node_handle_common::_M_swap): Add braces around debug assertions.

From-SVN: r255317

6 years agoUse const char* to fix -Wwrite-strings warning
Jonathan Wakely [Fri, 1 Dec 2017 15:10:16 +0000 (15:10 +0000)]
Use const char* to fix -Wwrite-strings warning

* include/ext/ropeimpl.h (rope::_S_dump): Use const char*.

From-SVN: r255316

6 years agoAdd [[noreturn]] attributes to fix warning
Jonathan Wakely [Fri, 1 Dec 2017 15:10:13 +0000 (15:10 +0000)]
Add [[noreturn]] attributes to fix warning

* libsupc++/nested_exception.h (__throw_with_nested_impl): Add
noreturn attribute.

From-SVN: r255315

6 years agoRemove stray semi-colons at namespace scope
Jonathan Wakely [Fri, 1 Dec 2017 15:10:09 +0000 (15:10 +0000)]
Remove stray semi-colons at namespace scope

* include/bits/regex_executor.tcc (_Executor::_M_rep_once_more):
Remove semi-colon after function body.
* include/bits/uniform_int_dist.h (_Power_of_2): Likewise.

From-SVN: r255314

6 years agoUse value-init syntax in std::__detail::__airy
Jonathan Wakely [Fri, 1 Dec 2017 15:10:06 +0000 (15:10 +0000)]
Use value-init syntax in std::__detail::__airy

* include/tr1/modified_bessel_func.tcc (__airy): Use value-init not
list-init.

From-SVN: r255313

6 years agoAdd noexcept to std::integral_constant members
Jonathan Wakely [Fri, 1 Dec 2017 15:09:57 +0000 (15:09 +0000)]
Add noexcept to std::integral_constant members

* include/std/type_traits (integral_constant): Make member functions
noexcept (LWG 2346).
* include/std/utility (integer_sequence): Likewise.

From-SVN: r255312

6 years agore PR fortran/82605 ([PDT] ICE in insert_parameter_exprs, at fortran/decl.c:3154)
Paul Thomas [Fri, 1 Dec 2017 15:05:55 +0000 (15:05 +0000)]
re PR fortran/82605 ([PDT] ICE in insert_parameter_exprs, at fortran/decl.c:3154)

2017-12-01  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/82605
* resolve.c (get_pdt_constructor): Initialize 'cons' to NULL.
(resolve_pdt): Correct typo in prior comment. Emit an error if
any parameters are deferred and the object is neither pointer
nor allocatable.

PR fortran/82606
* decl.c (gfc_get_pdt_instance): Continue if the parameter sym
is not present or has no name. Select the parameter by name
of component, rather than component order. Remove all the other
manipulations of 'tail' when building the pdt instance.
(gfc_match_formal_arglist): Emit and error if a star is picked
up in a PDT decl parameter list.

PR fortran/82622
* trans-array.c (set_loop_bounds): If a GFC_SS_COMPONENT has an
info->end, use it rather than falling through to
gcc_unreachable.
(structure_alloc_comps): Check that param->name is non-null
before comparing with the component name.
* trans-decl.c (gfc_get_symbol_decl): Do not use the static
initializer for PDT symbols.
(gfc_init_default_dt): Do nothing for PDT symbols.
* trans-io.c (transfer_array_component): Parameterized array
components use the descriptor ubound since the shape is not
available.

PR fortran/82719
PR fortran/82720
* trans-expr.c (gfc_conv_component_ref): Do not use the charlen
backend_decl of pdt strings. Use the hidden component instead.
* trans-io.c (transfer_expr): Do not do IO on "hidden" string
lengths. Use the hidden string length for pdt string transfers
by adding it to the se structure. When finished nullify the
se string length.

PR fortran/82866
* decl.c (gfc_match_formal_arglist): If a name is not found or
star is found, while reading a type parameter list, emit an
immediate error.
(gfc_match_derived_decl): On reading a PDT parameter list, on
failure to match call gfc_error_recovery.

PR fortran/82978
* decl.c (build_struct): Character kind defaults to 1, so use
kind_expr whatever is the set value.
(gfc_get_pdt_instance): Ditto.
* trans-array.c (structure_alloc_comps): Copy the expression
for the PDT string length before parameter substitution. Use
this expression for evaluation and free it after use.

2017-12-01  Paul Thomas  <pault@gcc.gnu.org>

PR fortran/82605
* gfortran.dg/pdt_4.f03 : Incorporate the new error.

PR fortran/82606
* gfortran.dg/pdt_19.f03 : New test.
* gfortran.dg/pdt_21.f03 : New test.

PR fortran/82622
* gfortran.dg/pdt_20.f03 : New test.
* gfortran.dg/pdt_22.f03 : New test.

PR fortran/82719
PR fortran/82720
* gfortran.dg/pdt_23.f03 : New test.

PR fortran/82866
* gfortran.dg/pdt_24.f03 : New test.

PR fortran/82978
* gfortran.dg/pdt_10.f03 : Correct for error in coding the for
kind 4 component and change the kind check appropriately.
* gfortran.dg/pdt_25.f03 : New test.

From-SVN: r255311

6 years agoFix bug in an OpenACC async test case
Cesar Philippidis [Fri, 1 Dec 2017 14:26:07 +0000 (06:26 -0800)]
Fix bug in an OpenACC async test case

libgomp/
* testsuite/libgomp.oacc-c-c++-common/data-2-lib.c: Add missing
call to acc_wait (1).

From-SVN: r255308

6 years agore PR tree-optimization/83232 (fma3d spec2000 regression on zen with -Ofast (generic...
Richard Biener [Fri, 1 Dec 2017 13:33:40 +0000 (13:33 +0000)]
re PR tree-optimization/83232 (fma3d spec2000 regression on zen with -Ofast (generic tuning) after r255268 by missed SLP oppurtunity)

2017-12-01  Richard Biener  <rguenther@suse.de>

PR tree-optimization/83232
* tree-vect-data-refs.c (vect_analyze_data_ref_accesses): Fix
detection of same access. Instead of breaking the group here
do not consider the duplicate.  Add comment explaining real fix.

* gfortran.dg/vect/pr83232.f90: New testcase.

From-SVN: r255307

6 years agoFix "central flowgraph" typo in machine desc docs
Jonathan Wakely [Fri, 1 Dec 2017 13:31:23 +0000 (13:31 +0000)]
Fix "central flowgraph" typo in machine desc docs

* doc/md.texi (Insn Splitting): Fix "central flowgraph" typo.

From-SVN: r255305

6 years agoRTEMS/Ada: Fix some POSIX types
Sebastian Huber [Fri, 1 Dec 2017 11:00:52 +0000 (11:00 +0000)]
RTEMS/Ada: Fix some POSIX types

gcc/ada
* libgnarl/s-osinte__rtems.ads (pthread_cond_t): Use correct size and
alignment.
(pthread_mutex_t): Likewise.
(pthread_rwlock_t): Likewise.

From-SVN: r255302

6 years agoFix wrong code by arm_final_prescan with fp16 move instructions
Sudakshina Das [Fri, 1 Dec 2017 10:12:55 +0000 (10:12 +0000)]
Fix wrong code by arm_final_prescan with fp16 move instructions

ChangeLog entry are as follow:

*** gcc/ChangeLog ***

2017-12-01  Sudakshina Das  <sudi.das@arm.com>

* config/arm/vfp.md (*movhf_vfp_fp16): Add conds attribute.

*** gcc/testsuite/ChangeLog ***

2017-12-01  Sudakshina Das  <sudi.das@arm.com>

* gcc.target/arm/armv8_2-fp16-move-2.c: New test.

From-SVN: r255301

6 years agofunction.h (struct function): Remove cilk_frame_decl, is_cilk_function and calls_cilk...
Jakub Jelinek [Fri, 1 Dec 2017 08:20:41 +0000 (09:20 +0100)]
function.h (struct function): Remove cilk_frame_decl, is_cilk_function and calls_cilk_spawn fields.

* function.h (struct function): Remove cilk_frame_decl,
is_cilk_function and calls_cilk_spawn fields.
* tree-inline.h (struct copy_body_data): Remove remap_var_for_cilk
field.
* omp-simd-clone.c (simd_clone_clauses_extract): Don't clear
cilk_elemental field.
* cgraph.h (struct cgraph_simd_clone): Remove cilk_elemental field.
* target.def: Adjust comment.
* config/i386/i386.c (ix86_simd_clone_compute_vecsize_and_simdlen):
Don't test cilk_elemental.
c-family/
* c-attribs.c (c_common_attribute_table): Remove "cilk simd function"
attribute.
(handle_simd_attribute): Don't check for "cilk simd function"
attribute.  Reindent, formatting changes.
cp/
* parser.c (cp_parser_new): Don't clear cilk_simd_fn_info.
(parsing_nsdmi): Adjust comment.
(cp_parser_omp_for_loop_init): Likewise.
* parser.h (struct cp_omp_declare_simd_data): Adjust comment.
(struct cp_parser): Remove cilk_simd_fn_info field.
* cp-tree.h (cilk_valid_spawn): Remove.

From-SVN: r255300

6 years agore PR tree-optimization/83233 (gimple-ssa-store-merging.c:183:3: warning: unnamed...
Jakub Jelinek [Fri, 1 Dec 2017 08:18:52 +0000 (09:18 +0100)]
re PR tree-optimization/83233 (gimple-ssa-store-merging.c:183:3: warning: unnamed type with no linkage used to declare variable with linkage)

PR tree-optimization/83233
* gimple-ssa-store-merging.c (nop_stats, bswap_stats): Use
bswap_stat name for the struct.

From-SVN: r255299

6 years agore PR c/79153 (-Wimplicit-fallthrough missed warning)
Jakub Jelinek [Fri, 1 Dec 2017 08:17:06 +0000 (09:17 +0100)]
re PR c/79153 (-Wimplicit-fallthrough missed warning)

PR c/79153
* tree.h (SWITCH_BREAK_LABEL_P): Define.
* gimplify.c (collect_fallthrough_labels): Handle GIMPLE_BIND
starting with a GIMPLE_SWITCH and ending with GIMPLE_LABEL with
SWITCH_BREAK_LABEL_P set on the label.
(gimplify_switch_expr): Set SWITCH_BREAK_LABEL_P on the label
added for default case if it was missing and not all cases covered.
Wrap GIMPLE_SWITCH and the switch_body_seq into a GIMPLE_BIND if
switch_body_seq ends with a GIMPLE_LABEL with SWITCH_BREAK_LABEL_P
set on the label.
* tree-chrec.c (evolution_function_is_univariate_p): Add return true;
to avoid -Wimplicit-fallthrough warning.
* config/i386/i386.c (ix86_expand_special_args_builtin): Add
FALLTHRU comment to avoid -Wimplicit-fallthrough warning.
c/
* c-parser.c: Include tree-iterator.h.
(c_parser_switch_statement): Emit LABEL_EXPR for the break label
into SWITCH_BODY instead of after it and set SWITCH_BREAK_LABEL_P
on it.
cp/
* cp-gimplify.c (genericize_switch_stmt): Emit LABEL_EXPR for the
break label into SWITCH_BODY instead of after it and set
SWITCH_BREAK_LABEL_P on it.
* parser.c (cp_parser_objc_expression): Add FALLTHRU comment to avoid
-Wimplicit-fallthrough warning.
fortran/
* match.c (gfc_match): Add FALLTHRU comment to avoid
-Wimplicit-fallthrough warning.
testsuite/
* c-c++-common/Wimplicit-fallthrough-7.c: Adjust expected warning
line.
* c-c++-common/Wimplicit-fallthrough-36.c: New test.

From-SVN: r255298

6 years agore PR tree-optimization/83221 (qsort comparator not anti-commutative: -2147483648...
Jakub Jelinek [Fri, 1 Dec 2017 08:14:21 +0000 (09:14 +0100)]
re PR tree-optimization/83221 (qsort comparator not anti-commutative: -2147483648, -2147483648)

PR tree-optimization/83221
* tree-ssa-reassoc.c (sort_by_operand_rank): Shift bb_rank
down by 16.
(init_reassoc): Formatting fix.

From-SVN: r255297

6 years agore PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)
Jakub Jelinek [Fri, 1 Dec 2017 08:11:56 +0000 (09:11 +0100)]
re PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)

PR sanitizer/81275
* tree-cfg.c (group_case_labels_stmt): Don't optimize away
C++ FE implicitly added __builtin_unreachable () until -Wreturn-type
is diagnosed.

* c-c++-common/tsan/pr81275.c: Remove dg-skip-if.

From-SVN: r255296

6 years agore PR sanitizer/83219 (c-c++-common/ubsan/unreachable-2.c fails starting with r255201)
Jakub Jelinek [Fri, 1 Dec 2017 08:08:55 +0000 (09:08 +0100)]
re PR sanitizer/83219 (c-c++-common/ubsan/unreachable-2.c fails starting with r255201)

PR sanitizer/83219
* tree-cfg.c: Include asan.h.
(gimple_seq_unreachable_p): Return false for -fsanitize=unreachable.

From-SVN: r255295

6 years agore PR fortran/83224 (creating character array from elements shorter than declared...
Thomas Koenig [Fri, 1 Dec 2017 07:35:50 +0000 (07:35 +0000)]
re PR fortran/83224 (creating character array from elements shorter than declared does not pad with whitespace properly and aborts)

2017-12-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83224
* frontend-passes.c (realloc_string_callback): Handle
case for which the RHS is an array expression.

2017-12-01  Thomas Koenig  <tkoenig@gcc.gnu.org>

PR fortran/83224
* gfortran.dg/dependency_50.f90: New test.

From-SVN: r255294

6 years agoFix movdi_internal to return MODE_TI with AVX512
Sergey Shalnov [Fri, 1 Dec 2017 06:40:27 +0000 (06:40 +0000)]
Fix movdi_internal to return MODE_TI with AVX512

gcc/
* config/i386/i386.md: Fix AVX512 register width in AVX512 instruction.

From-SVN: r255293

6 years agoReplace __popcntq with two _popcnt32.
Kirill Yukhin [Fri, 1 Dec 2017 06:25:26 +0000 (06:25 +0000)]
Replace __popcntq with two _popcnt32.

gcc/testsuite/
* gcc.target/i386/avx512f-vpcompressb-2.c: Fix popcnt for 64b case.

From-SVN: r255292

6 years agoFix wrong instruction vpcmpeqd generation
Sergey Shalnov [Fri, 1 Dec 2017 06:22:06 +0000 (06:22 +0000)]
Fix wrong instruction vpcmpeqd generation

gcc/
* config/i386/i386.c (standard_sse_constant_opcode): Fix wrong form for
vpcmpeqd instruction.

From-SVN: r255291

6 years agoFix registers type for MODE_TI
Sergey Shalnov [Fri, 1 Dec 2017 05:59:33 +0000 (05:59 +0000)]
Fix registers type for MODE_TI

gcc/
* config/i386/i386.c (standard_sse_constant_opcode): Fix
registers type for 128bit mode.

From-SVN: r255290

6 years ago_mulkc3.c (__mulkc3): Add forward declaration.
Michael Meissner [Fri, 1 Dec 2017 05:32:39 +0000 (05:32 +0000)]
_mulkc3.c (__mulkc3): Add forward declaration.

2017-11-30  Michael Meissner  <meissner@linux.vnet.ibm.com>

* config/rs6000/_mulkc3.c (__mulkc3): Add forward declaration.
* config/rs6000/_divkc3.c (__divkc3): Likewise.

From-SVN: r255289

6 years agoDaily bump.
GCC Administrator [Fri, 1 Dec 2017 00:16:19 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r255288

6 years agore PR c/83222 (Inconsistent "initializer element is not constant" error)
Jakub Jelinek [Thu, 30 Nov 2017 23:43:10 +0000 (00:43 +0100)]
re PR c/83222 (Inconsistent "initializer element is not constant" error)

PR c/83222
* c-tree.h (decl_constant_value_1): Declare.
* c-typeck.c (decl_constant_value_1): New function.
(decl_constant_value): Use it.
* c-fold.c (c_fully_fold_internal): If in_init, use
decl_constant_value_1 instead of decl_constant_value.

* gcc.c-torture/compile/pr83222.c: New test.

From-SVN: r255285

6 years agospellcheck-tree.c (test_find_closest_identifier): Use ; instead of ;;.
Jakub Jelinek [Thu, 30 Nov 2017 22:47:51 +0000 (23:47 +0100)]
spellcheck-tree.c (test_find_closest_identifier): Use ; instead of ;;.

* spellcheck-tree.c (test_find_closest_identifier): Use ; instead
of ;;.
* gengtype-state.c (read_state_pair): Likewise.
* gimple-fold.c (gimple_fold_builtin_string_compare): Likewise.
* sel-sched-dump.c (dump_insn_rtx_1): Likewise.
* ipa-cp.c (intersect_aggregates_with_edge): Likewise.
* ifcvt.c (noce_try_store_flag_constants): Likewise.
* tree-ssa-ccp.c (ccp_finalize): Likewise.
* omp-grid.c (grid_process_kernel_body_copy): Likewise.
* builtins.c (fold_builtin_3): Likewise.
* graphite-scop-detection.c
(scop_detection::stmt_has_simple_data_refs_p): Likewise.
* hsa-gen.c (hsa_function_representation::hsa_function_representation):
Likewise.
c/
* c-parser.c (c_parser_postfix_expression): Use ; instead of ;;.
jit/
* jit-recording.c
(recording::memento_of_new_rvalue_from_const <long>::write_reproducer):
Use ; instead of ;;.
lto/
* lto.c (create_subid_section_table): Use ; instead of ;;.
objc/
* objc-next-runtime-abi-01.c (generate_dispatch_table): Use ; instead
of ;;.

From-SVN: r255284

6 years agore PR sanitizer/81697 (Incorrect ASan global variables alignment on arm)
Maxim Ostapenko [Thu, 30 Nov 2017 21:38:16 +0000 (21:38 +0000)]
re PR sanitizer/81697 (Incorrect ASan global variables alignment on arm)

gcc/

2017-11-30  Maxim Ostapenko  <m.ostapenko@samsung.com>

PR sanitizer/81697
* asan.c (asan_protect_global): Add new ignore_decl_rtl_set_p
parameter. Return true if ignore_decl_rtl_set_p is true and other
conditions are satisfied.
* asan.h (asan_protect_global): Add new parameter.
* varasm.c (categorize_decl_for_section): Pass true as second parameter
to asan_protect_global calls.

gcc/testsuite/

2017-11-30  Maxim Ostapenko  <m.ostapenko@samsung.com>

PR sanitizer/81697
* c-c++-common/asan/pr81697.c: New test.

From-SVN: r255283

6 years agore PR libgcc/83112 (Silence warnings from PowerPC libgcc float128-ifunc.c compilation)
Michael Meissner [Thu, 30 Nov 2017 20:52:27 +0000 (20:52 +0000)]
re PR libgcc/83112 (Silence warnings from PowerPC libgcc float128-ifunc.c compilation)

2017-11-30  Michael Meissner  <meissner@linux.vnet.ibm.com>

PR libgcc/83112
* config/rs6000/float128-ifunc.c (__addkf3_resolve): Use the
correct type for all ifunc resolvers to silence -Wattribute-alias
warnings.  Eliminate the forward declaration of the resolver
functions which is no longer needed.
(__subkf3_resolve): Likewise.
(__mulkf3_resolve): Likewise.
(__divkf3_resolve): Likewise.
(__negkf2_resolve): Likewise.
(__eqkf2_resolve): Likewise.
(__nekf2_resolve): Likewise.
(__gekf2_resolve): Likewise.
(__gtkf2_resolve): Likewise.
(__lekf2_resolve): Likewise.
(__ltkf2_resolve): Likewise.
(__unordkf2_resolve): Likewise.
(__extendsfkf2_resolve): Likewise.
(__extenddfkf2_resolve): Likewise.
(__trunckfsf2_resolve): Likewise.
(__trunckfdf2_resolve): Likewise.
(__fixkfsi_resolve): Likewise.
(__fixkfdi_resolve): Likewise.
(__fixunskfsi_resolve): Likewise.
(__fixunskfdi_resolve): Likewise.
(__floatsikf_resolve): Likewise.
(__floatdikf_resolve): Likewise.
(__floatunsikf_resolve): Likewise.
(__floatundikf_resolve): Likewise.
(__extendkftf2_resolve): Likewise.
(__trunctfkf2_resolve): Likewise.

PR libgcc/83103
* config/rs6000/quad-float128.h (TF): Don't define if long double
is IEEE 128-bit floating point.
(TCtype): Define as either TCmode or KCmode, depending on whether
long double is IEEE 128-bit floating point.
(__mulkc3_sw): Add declarations for software/hardware versions of
complex multiply/divide.
(__divkc3_sw): Likewise.
(__mulkc3_hw): Likewise.
(__divkc3_hw): Likewise.
* config/rs6000/_mulkc3.c (_mulkc3): If we are building ifunc
handlers to switch between using software emulation and hardware
float128 instructions, build the complex multiply/divide functions
for both software and hardware support.
* config/rs6000/_divkc3.c (_divkc3): Likewise.
* config/rs6000/float128-ifunc.c (__mulkc3_resolve): Likewise.
(__divkc3_resolve): Likewise.
(__mulkc3): Likewise.
(__divkc3): Likewise.
* config/rs6000/t-float128-hw (fp128_hardfp_src): Likewise.
(fp128_hw_src): Likewise.
(fp128_hw_static_obj): Likewise.
(fp128_hw_shared_obj): Likewise.
(_mulkc3-hw.c): Create _mulkc3-hw.c and _divkc3-hw.c from
_mulkc3.c and _divkc3.c, changing the function name.
(_divkc3-hw.c): Likewise.
* config/rs6000/t-float128 (clean-float128): Delete _mulkc3-hw.c
and _divkc3-hw.c.

From-SVN: r255282

6 years agoCleanup riscv option docs.
Jim Wilson [Thu, 30 Nov 2017 18:03:46 +0000 (18:03 +0000)]
Cleanup riscv option docs.

gcc/
* doc/invoke.texi (RISC-V Options): Delete nonexistent -mmemcpy and
-mno-memcpy options.  For -mplt, -mfdiv, -mdiv, -msave-restore, and
-mstrict-align, add info on default value.  Delete redundant lines for
-mabi.  Add missing -mexplicit-relocs docs.

From-SVN: r255280

6 years agoPR c++/82219 - bogus -Wignored-qualifiers with template
Jason Merrill [Thu, 30 Nov 2017 17:43:08 +0000 (12:43 -0500)]
PR c++/82219 - bogus -Wignored-qualifiers with template

* pt.c (tsubst_copy_and_build) [STATIC_CAST_EXPR]: Suppress
-Wignored-qualifiers.

From-SVN: r255279

6 years agoPR libstdc++/65927 Fix -Wunused-but-set-parameter warning
Jonathan Wakely [Thu, 30 Nov 2017 17:29:08 +0000 (17:29 +0000)]
PR libstdc++/65927 Fix -Wunused-but-set-parameter warning

PR libstdc++/65927
* include/bits/locale_facets_nonio.h (__timepunct::_M_am_pm_format):
Remove statement with no effect.

From-SVN: r255278

6 years agoMove assertion from to_address to __to_address
Glen Joseph Fernandes [Thu, 30 Nov 2017 15:07:21 +0000 (15:07 +0000)]
Move assertion from to_address to __to_address

2017-11-30  Glen Joseph Fernandes  <glenjofe@gmail.com>

* include/bits/ptr_traits.h (__to_address, to_address): Move static
assertion.
* testsuite/20_util/to_address/1_neg.cc: New test.

From-SVN: r255277

6 years ago[ARC] Add trap instruction.
Claudiu Zissulescu [Thu, 30 Nov 2017 14:42:22 +0000 (15:42 +0100)]
[ARC] Add trap instruction.

2017-11-07  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.md (trap): New pattern.

From-SVN: r255276

6 years ago[ARC] Don't allow the last ZOL insn to be in a delay slot.
Claudiu Zissulescu [Thu, 30 Nov 2017 14:42:12 +0000 (15:42 +0100)]
[ARC] Don't allow the last ZOL insn to be in a delay slot.

The ARC ZOL implementation doesn't allow the last instruction to be a
control instruction or part of a delay slot.  Thus, we add a note to
the last ZOL instruction which will prevent it to finish into a delay
slot.

2017-10-20  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (hwloop_optimize): Prevent the last
        ZOL instruction to end into a delay slot.
* config/arc/arc.md (cond_delay_insn): Check if the instruction
can be placed into a delay slot against reg_note.
(in_delay_slot): Likewise.

testsuite/
2017-10-20  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/loop-3.c: New test.
* gcc.target/arc/loop-4.c: Likewise.

[FIX][ZOL] fix checking for jumps

From-SVN: r255275

6 years ago[ARC] Update uses for hw-loop labels.
Claudiu Zissulescu [Thu, 30 Nov 2017 14:42:01 +0000 (15:42 +0100)]
[ARC] Update uses for hw-loop labels.

Make sure we mark the hw-loop labels as beeing used.

gcc/
2017-09-19  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (hwloop_optimize): Update hw-loop's end/start
labels number of usages.

gcc/testsuite
2017-09-19  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/loop-2.cpp: New test.

From-SVN: r255274

6 years ago[ARC] Use TARGET_CANNOT_SUBSTITUTE_MEM_EQUIV.
Claudiu Zissulescu [Thu, 30 Nov 2017 14:41:49 +0000 (15:41 +0100)]
[ARC] Use TARGET_CANNOT_SUBSTITUTE_MEM_EQUIV.

Sometimes the memory equivalent is not valid due to a large offset.
For example replacing the ap register with its fp/sp-equivalent during
LRA step. To solve this we introduced TARGET_CANNOT_SUBSTITUTE_MEM_EQUIV.

gcc/
2017-08-08  Claudiu Zissulescu  <claziss@synopsys.com>

* config/arc/arc.c (arc_cannot_substitute_mem_equiv_p): New function.
(TARGET_CANNOT_SUBSTITUTE_MEM_EQUIV_P): Define.

gcc/testsuite
2017-08-08  Claudiu Zissulescu  <claziss@synopsys.com>

* gcc.target/arc/lra-1.c: New test.

From-SVN: r255273

6 years agofold-vec-abs-char-fwrapv.c: Add xxspltib insn to expected output.
Will Schmidt [Thu, 30 Nov 2017 14:38:33 +0000 (14:38 +0000)]
fold-vec-abs-char-fwrapv.c: Add xxspltib insn to expected output.

[testsuite]

2017-11-29  Will Schmidt  <will_schmidt@vnet.ibm.com>

* gcc.target/powerpc/fold-vec-abs-char-fwrapv.c: Add xxspltib insn
to expected output.
* gcc.target/powerpc/fold-vec-abs-char.c: Add xxspltib insn
to expected output.

From-SVN: r255272

6 years agoPR libstdc++/83226 avoid forming pointer-to-reference type
Jonathan Wakely [Thu, 30 Nov 2017 13:45:37 +0000 (13:45 +0000)]
PR libstdc++/83226 avoid forming pointer-to-reference type

PR libstdc++/83226
* include/bits/node_handle.h (_Node_handle::__pointer): Avoid forming
pointer-to-reference types.
* testsuite/23_containers/map/modifiers/insert/83226.cc: New test.

From-SVN: r255271

6 years agoRemove inv_list.
Julia Koval [Thu, 30 Nov 2017 12:17:16 +0000 (13:17 +0100)]
Remove inv_list.

gcc/c-family/
* c-common.h (inv_list): Remove.

From-SVN: r255270

6 years agore PR target/83210 (__builtin_mul_overflow() generates suboptimal code when exactly...
Jakub Jelinek [Thu, 30 Nov 2017 10:29:58 +0000 (11:29 +0100)]
re PR target/83210 (__builtin_mul_overflow() generates suboptimal code when exactly one argument is the constant 2)

PR target/83210
* internal-fn.c (expand_mul_overflow): Optimize unsigned
multiplication by power of 2 constant into two shifts + comparison.

* gcc.target/i386/pr83210.c: New test.

From-SVN: r255269

6 years agore PR target/81616 (Update -mtune=generic for the current Intel and AMD processors)
Jan Hubicka [Thu, 30 Nov 2017 09:36:36 +0000 (10:36 +0100)]
re PR target/81616 (Update -mtune=generic for the current Intel and AMD processors)

PR target/81616
* x86-tnue-costs.h (generic_cost): Revise for modern CPUs
* gcc.target/i386/l_fma_double_1.c: Update count of fma instructions.
* gcc.target/i386/l_fma_double_2.c: Update count of fma instructions.
* gcc.target/i386/l_fma_double_3.c: Update count of fma instructions.
* gcc.target/i386/l_fma_double_4.c: Update count of fma instructions.
* gcc.target/i386/l_fma_double_5.c: Update count of fma instructions.
* gcc.target/i386/l_fma_double_6.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_1.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_2.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_3.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_4.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_5.c: Update count of fma instructions.
* gcc.target/i386/l_fma_float_6.c: Update count of fma instructions.

From-SVN: r255268

6 years agore PR tree-optimization/83202 (Try joining operations on consecutive array elements...
Richard Biener [Thu, 30 Nov 2017 07:53:31 +0000 (07:53 +0000)]
re PR tree-optimization/83202 (Try joining operations on consecutive array elements during tree vectorization)

2017-11-30  Richard Biener  <rguenther@suse.de>

PR tree-optimization/83202
* tree-ssa-loop-ivcanon.c (try_unroll_loop_completely): Add
allow_peel argument and guard peeling.
(canonicalize_loop_induction_variables): Likewise.
(canonicalize_induction_variables): Pass false.
(tree_unroll_loops_completely_1): Pass unroll_outer to disallow
peeling from cunrolli.

* gcc.dg/vect/pr83202-1.c: New testcase.
* gcc.dg/tree-ssa/pr61743-1.c: Adjust.

From-SVN: r255267

6 years agocompiler: don't make map zero value constant
Ian Lance Taylor [Thu, 30 Nov 2017 05:03:29 +0000 (05:03 +0000)]
compiler: don't make map zero value constant

    The map zero value is a common symbol, and it doesn't really make
    sense to have a constant common symbol. Current GCC has started to
    reject this case, probably as part of the fix for PR 83100.

    Reviewed-on: https://go-review.googlesource.com/80877

From-SVN: r255266

6 years agobaseline_symbols.txt: Update.
John David Anglin [Thu, 30 Nov 2017 00:53:16 +0000 (00:53 +0000)]
baseline_symbols.txt: Update.

* config/abi/post/hppa-linux-gnu/baseline_symbols.txt: Update.

From-SVN: r255265

6 years agoDaily bump.
GCC Administrator [Thu, 30 Nov 2017 00:16:18 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r255264

6 years agocombine: Print to dump if some insn cannot be combined into i3
Segher Boessenkool [Wed, 29 Nov 2017 22:51:16 +0000 (23:51 +0100)]
combine: Print to dump if some insn cannot be combined into i3

Eventually we should print the reason that any combination fails.
This is a good start (these happen often).

* combine.c (try_combine): Print a message to dump file whenever
I0, I1, or I2 cannot be combined into I3.

From-SVN: r255261

6 years agocombine: Do not throw away unneeded arms of parallels (PR83156)
Segher Boessenkool [Wed, 29 Nov 2017 22:42:37 +0000 (23:42 +0100)]
combine: Do not throw away unneeded arms of parallels (PR83156)

The fix for PR82621 makes us not split an I2 if one of the results of
those SETs is unused, since combine does not handle that properly.  But
this results in degradation for i386 (or more in general, for any
target that does not have patterns for parallels with an unused result
as a CLOBBER instead of a SET for that result).

This patch instead makes us not split only if one of the results is set
again before I3.  That fixes PR83156 and also fixes PR82621.

Unfortunately it undoes the nice optimisations that the previous patch
did on powerpc.

PR rtl-optimization/83156
PR rtl-optimization/82621
* combine.c (try_combine): Don't split an I2 if one of the dests is
set again before I3.  Allow unused dests.

From-SVN: r255260

6 years agors6000: Add second variant of adde
Segher Boessenkool [Wed, 29 Nov 2017 22:30:24 +0000 (23:30 +0100)]
rs6000: Add second variant of adde

This adds a second variant of the adde insn pattern, this one with the
CA register as the second operand.  The existing pattern has it as the
third operand.  It would be ideal if RTL was always canonicalised like
that, but it isn't (and that is not trivial), and this is a simple and
harmless patch.

* config/rs6000/rs6000.md (*add<mode>3_carry_in_internal2): New.

From-SVN: r255259

6 years agore PR rtl-optimization/80818 (LRA clobbers live hard reg clobbered during remateriali...
Vladimir Makarov [Wed, 29 Nov 2017 22:19:25 +0000 (22:19 +0000)]
re PR rtl-optimization/80818 (LRA clobbers live hard reg clobbered during rematerialization)

2017-11-29  Vladimir Makarov  <vmakarov@redhat.com>

PR rtl-optimization/80818
* lra.c (collect_non_operand_hard_regs): New arg insn.  Pass it
recursively.  Use insn code for clobber.
(lra_set_insn_recog_data): Pass the new arg to
collect_non_operand_hard_regs.
(add_regs_to_insn_regno_info): Pass insn instead of uid.  Use insn
code for clobber.
(lra_update_insn_regno_info): Pass insn to
add_regs_to_insn_regno_info.

From-SVN: r255258

6 years agoRiscv patterns to optimize away some redundant zero/sign extends.
Jim Wilson [Wed, 29 Nov 2017 22:14:17 +0000 (22:14 +0000)]
Riscv patterns to optimize away some redundant zero/sign extends.

gcc/
* config/riscv/riscv.c (SINGLE_SHIFT_COST): New.
(riscv_rtx_costs): Case ZERO_EXTRACT, match new pattern, and return
SINGLE_SHIFT_COST.  Case LT and ZERO_EXTEND, likewise.  Case ASHIFT,
use SINGLE_SHIFT_COST.
* config/riscv/riscv.md (lshrsi3_zero_extend_1): New.
(lshrsi3_zero_extend_2, lshrsi3_zero_extend_3): New.

gcc/testsuite/
* gcc.target/riscv/riscv.exp: New.
* gcc.target/riscv/zero-extend-1.c: New.
* gcc.target/riscv/zero-extend-2.c: New.
* gcc.target/riscv/zero-extend-3.c: New.
* gcc.target/riscv/zero-extend-4.c: New.

Co-Authored-By: Andrew Waterman <andrew@sifive.com>
From-SVN: r255257

6 years agoC++: improve location of static_assert errors
David Malcolm [Wed, 29 Nov 2017 22:07:32 +0000 (22:07 +0000)]
C++: improve location of static_assert errors

gcc/cp/ChangeLog:
* parser.c (cp_parser_unary_expression): Generate a location for
"noexcept".
(cp_parser_trait_expr): Generate and return a location_t,
converting the return type from tree to cp_expr.
(cp_parser_static_assert): Pass location of the condition to
finish_static_assert, rather than that of the "static_assert"
token, where available.

gcc/testsuite/ChangeLog:
* g++.dg/cpp1y/static_assert3.C: New test case.

libstdc++-v3/ChangeLog:
* testsuite/20_util/duration/literals/range.cc: Update expected
line of a static_assert failure.

From-SVN: r255255

6 years agore PR c++/82293 (ICE in nonlambda_method_basetype at gcc/cp/lambda.c:886)
Paolo Carlini [Wed, 29 Nov 2017 21:54:18 +0000 (21:54 +0000)]
re PR c++/82293 (ICE in nonlambda_method_basetype at gcc/cp/lambda.c:886)

/cp
2017-11-29  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/82293
* lambda.c (nonlambda_method_basetype): Don't use LAMBDA_TYPE_P
on a null type.

/testsuite
2017-11-29  Paolo Carlini  <paolo.carlini@oracle.com>

PR c++/82293
* g++.dg/cpp0x/lambda/lambda-ice24.C: New.

From-SVN: r255254

6 years agoPR c++/82760 - memory corruption with aligned new.
Jason Merrill [Wed, 29 Nov 2017 21:01:23 +0000 (16:01 -0500)]
PR c++/82760 - memory corruption with aligned new.

* call.c (build_operator_new_call): Update *args if we add the
align_arg.

From-SVN: r255253

6 years agofold-vec-ld-char.c: Add lxv insn to expected output.
Will Schmidt [Wed, 29 Nov 2017 20:10:52 +0000 (20:10 +0000)]
fold-vec-ld-char.c: Add lxv insn to expected output.

[testsuite]

2017-11-29  Will Schmidt  <will_schmidt@vnet.ibm.com>

    * gcc.target/powerpc/fold-vec-ld-char.c: Add lxv insn to expected output.
    * gcc.target/powerpc/fold-vec-ld-double.c: Add lxv insn to expected output.
    * gcc.target/powerpc/fold-vec-ld-float.c: Add lxv insn to expected output.
    * gcc.target/powerpc/fold-vec-ld-int.c: Add lxv insn to expected output.
    * gcc.target/powerpc/fold-vec-ld-longlong.c: Add lxv insn to expected output.
    * gcc.target/powerpc/fold-vec-ld-short.c: Add lxv insn to expected output.

From-SVN: r255252

6 years agofold-vec-splat-8.c: Add vspltisb to expected output.
Will Schmidt [Wed, 29 Nov 2017 19:58:29 +0000 (19:58 +0000)]
fold-vec-splat-8.c: Add vspltisb to expected output.

[testsuite]
2017-11-29  Will Schmidt  <will_schmidt@vnet.ibm.com>

    * gcc.target/powerpc/fold-vec-splat-8.c: Add vspltisb to expected output.
    * gcc.target/powerpc/fold-vec-splats-int.c: Add mtvsrws to expected output.
    * gcc.target/powerpc/fold-vec-splats-longlong.c: Add mtvsrdd to expected output.

From-SVN: r255251

6 years agonable VBMI2 support [7/7]
Julia Koval [Wed, 29 Nov 2017 18:45:28 +0000 (19:45 +0100)]
nable VBMI2 support [7/7]

gcc/
* config/i386/avx512vbmi2intrin.h (_mm512_shldv_epi16,
_mm512_mask_shldv_epi16, _mm512_maskz_shldv_epi16, _mm512_shldv_epi32,
_mm512_mask_shldv_epi32, _mm512_maskz_shldv_epi32, _mm512_shldv_epi64,
_mm512_mask_shldv_epi64, _mm512_maskz_shldv_epi64): New intrinsics.
* config/i386/avx512vbmi2vlintrin.h (_mm256_shldv_epi16,
_mm256_mask_shldv_epi16, _mm256_maskz_shldv_epi16, _mm256_shldv_epi32,
_mm256_mask_shldv_epi32, _mm256_maskz_shldv_epi32, _mm256_shldv_epi64,
_mm256_mask_shldv_epi64, _mm256_maskz_shldv_epi64, _mm_shldv_epi16,
_mm_mask_shldv_epi16, _mm_maskz_shldv_epi16, _mm_shldv_epi32,
_mm_mask_shldv_epi32, _mm_maskz_shldv_epi32, _mm_shldv_epi64,
_mm_mask_shldv_epi64, _mm_maskz_shldv_epi64): Ditto.
* config/i386/i386-builtin.def (__builtin_ia32_vpshldv_v32hi,
__builtin_ia32_vpshldv_v32hi_mask, __builtin_ia32_vpshldv_v32hi_maskz,
__builtin_ia32_vpshldv_v16hi, __builtin_ia32_vpshldv_v16hi_mask,
__builtin_ia32_vpshldv_v16hi_maskz, __builtin_ia32_vpshldv_v8hi,
__builtin_ia32_vpshldv_v8hi_mask, __builtin_ia32_vpshldv_v8hi_maskz,
__builtin_ia32_vpshldv_v16si, __builtin_ia32_vpshldv_v16si_mask,
__builtin_ia32_vpshldv_v16si_maskz, __builtin_ia32_vpshldv_v8si,
__builtin_ia32_vpshldv_v8si_mask, __builtin_ia32_vpshldv_v8si_maskz,
__builtin_ia32_vpshldv_v4si, __builtin_ia32_vpshldv_v4si_mask,
__builtin_ia32_vpshldv_v4si_maskz, __builtin_ia32_vpshldv_v8di,
__builtin_ia32_vpshldv_v8di_mask, __builtin_ia32_vpshldv_v8di_maskz,
__builtin_ia32_vpshldv_v4di, __builtin_ia32_vpshldv_v4di_mask,
__builtin_ia32_vpshldv_v4di_maskz, __builtin_ia32_vpshldv_v2di,
__builtin_ia32_vpshldv_v2di_mask,
__builtin_ia32_vpshldv_v2di_maskz): New builtins.
* config/i386/sse.md (vpshldv_<mode>, vpshldv_<mode>_mask,
vpshldv_<mode>_maskz, vpshldv_<mode>_maskz_1): New patterns.

gcc/testsuite/
* gcc.target/i386/avx512f-vpshldv-1.c: New test.
* gcc.target/i386/avx512f-vpshldvd-2.c: Ditto.
* gcc.target/i386/avx512f-vpshldvq-2.c: Ditto.
* gcc.target/i386/avx512f-vpshldvw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshldv-1.c: Ditto.
* gcc.target/i386/avx512vl-vpshldvd-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshldvq-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshldvw-2.c: Ditto.

From-SVN: r255250

6 years agoEnable VBMI2 support [6/7]
Julia Koval [Wed, 29 Nov 2017 18:19:34 +0000 (19:19 +0100)]
Enable VBMI2 support [6/7]

gcc/
* config/i386/avx512vbmi2intrin.h (_mm512_shrdv_epi16,
_mm512_mask_shrdv_epi16, _mm512_maskz_shrdv_epi16, _mm512_shrdv_epi32,
_mm512_mask_shrdv_epi32, _mm512_maskz_shrdv_epi32, _mm512_shrdv_epi64,
_mm512_mask_shrdv_epi64, _mm512_maskz_shrdv_epi64): New intrinsics.
* config/i386/avx512vbmi2vlintrin.h (_mm256_shrdv_epi16,
_mm256_mask_shrdv_epi16, _mm256_maskz_shrdv_epi16, _mm256_shrdv_epi32,
_mm256_mask_shrdv_epi32, _mm256_maskz_shrdv_epi32, _mm256_shrdv_epi64,
_mm256_mask_shrdv_epi64, _mm256_maskz_shrdv_epi64, _mm_shrdv_epi16,
_mm_mask_shrdv_epi16, _mm_maskz_shrdv_epi16, _mm_shrdv_epi32,
_mm_mask_shrdv_epi32, _mm_maskz_shrdv_epi32, _mm_shrdv_epi64,
_mm_mask_shrdv_epi64, _mm_maskz_shrdv_epi64): Ditto.
* config/i386/i386-builtin-types.def (V32HI_FTYPE_V32HI_V32HI_V32HI,
V32HI_FTYPE_V32HI_V32HI_V32HI_INT, V16HI_FTYPE_V16HI_V16HI_V16HI_INT,
V8HI_FTYPE_V8HI_V8HI_V8HI_INT, V8SI_FTYPE_V8SI_V8SI_V8SI_INT,
V4SI_FTYPE_V4SI_V4SI_V4SI_INT, V8DI_FTYPE_V8DI_V8DI_V8DI,
V8DI_FTYPE_V8DI_V8DI_V8DI_INT, V4DI_FTYPE_V4DI_V4DI_V4DI_INT,
V16SI_FTYPE_V16SI_V16SI_V16SI, V16SI_FTYPE_V16SI_V16SI_V16SI_INT,
V2DI_FTYPE_V2DI_V2DI_V2DI_INT): New types.
* config/i386/i386.c (ix86_expand_args_builtin): Handle new types.
* config/i386/sse.md (vpshrdv_<mode>, vpshrdv_<mode>_mask,
vpshrdv_<mode>_maskz, vpshrdv_<mode>_maskz_1): New pattern.

gcc/testsuite/
* gcc.target/i386/avx512f-vpshrdv-1.c: New test.
* gcc.target/i386/avx512f-vpshrdvd-2.c: Ditto.
* gcc.target/i386/avx512f-vpshrdvq-2.c: Ditto.
* gcc.target/i386/avx512f-vpshrdvw-2.c: Ditto.
* gcc.target/i386/avx512f-vpshrdw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdv-1.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdvd-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdvq-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdvw-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdw-2.c: Ditto.

From-SVN: r255249

6 years agoAdd myself as GCC maintainer
Qing Zhao [Wed, 29 Nov 2017 17:55:42 +0000 (17:55 +0000)]
Add myself as GCC maintainer

From-SVN: r255248

6 years agore PR tree-optimization/83195 (pr82929.c scan for "Merging successful" fail)
Jakub Jelinek [Wed, 29 Nov 2017 17:38:54 +0000 (18:38 +0100)]
re PR tree-optimization/83195 (pr82929.c scan for "Merging successful" fail)

PR tree-optimization/83195
* gcc.dg/pr82929.c: Don't check for "Merging successful" on arm.
* gcc.dg/pr82929-2.c: New test.

From-SVN: r255247

6 years ago[SPARC] Recognize the load when accessing the GOT
Daniel Cederman [Wed, 29 Nov 2017 15:45:41 +0000 (15:45 +0000)]
[SPARC] Recognize the load when accessing the GOT

Needed for the UT699 errata workaround to function correctly when
compiling with -fPIC.

2017-11-29  Daniel Cederman  <cederman@gaisler.com>

gcc/
* config/sparc/sparc.c (sparc_do_work_around_errata): Treat the
        movsi_pic_gotdata_op instruction as a load for the UT699 errata
        workaround.

From-SVN: r255239

6 years ago[SPARC] Prevent -mfix-ut699 from generating b2bst errata sequences
Martin Aberg [Wed, 29 Nov 2017 15:41:45 +0000 (15:41 +0000)]
[SPARC] Prevent -mfix-ut699 from generating b2bst errata sequences

The sequence
  st
  fdivd / fsqrtd
  std
was generated in some cases with -mfix-ut699 when there was
a st before the div/sqrt. This sequence could trigger the b2bst errata.

Now the following safe sequence is generated instead:
  st
  nop
  fdivd / fsqrtd
  std

2017-11-29  Martin Aberg  <maberg@gaisler.com>

gcc/
* config/sparc/sparc.md (divdf3_fix): Add NOP and adjust length
        to prevent b2bst errata sequence.
        (sqrtdf2_fix): Likewise.

From-SVN: r255238

6 years ago[SPARC] Errata workaround for GRLIB-TN-0013
Daniel Cederman [Wed, 29 Nov 2017 15:37:19 +0000 (15:37 +0000)]
[SPARC] Errata workaround for GRLIB-TN-0013

This patch provides a workaround for the errata described in GRLIB-TN-0013.

If the workaround is enabled it will:

* Prevent div and sqrt instructions in the delay slot.

* Insert NOPs to prevent the sequence (div/sqrt) -> (two or three floating
  point operations or loads) -> (div/sqrt).

* Not insert NOPs if any of the floating point operations have a dependency
  on the destination register of the first (div/sqrt).

* Not insert NOPs if one of the floating point operations is a (div/sqrt).

* Insert NOPs to prevent (div/sqrt) followed by a branch.

It is applicable to GR712RC, UT700, and UT699.

2017-11-29  Daniel Cederman  <cederman@gaisler.com>

gcc/
* config/sparc/sparc.c (fpop_reg_depend_p): New function.
(div_sqrt_insn_p): New function.
(sparc_do_work_around_errata): Insert NOP instructions to
prevent sequences that could trigger the TN-0013 errata for
certain LEON3 processors.
(pass_work_around_errata::gate): Also test sparc_fix_lost_divsqrt.
(sparc_option_override): Set sparc_fix_lost_divsqrt appropriately.
* config/sparc/sparc.md (fix_lost_divsqrt): New attribute.
(in_branch_delay): Prevent div and sqrt in delay slot if
fix_lost_divsqrt.
* config/sparc/sparc.opt (sparc_fix_lost_divsqrt): New variable.

From-SVN: r255237

6 years ago[SPARC] Errata workaround for GRLIB-TN-0010
Daniel Cederman [Wed, 29 Nov 2017 15:23:40 +0000 (15:23 +0000)]
[SPARC] Errata workaround for GRLIB-TN-0010

This patch provides a workaround for the errata described in GRLIB-TN-0010.

If the workaround is enabled it will:

 * Insert a NOP between load instruction and atomic
   instruction (swap, ldstub, casa).

 * Insert a NOP at branch target if load in delay slot
   and atomic instruction at branch target.

It is applicable to UT700.

2017-11-29  Daniel Cederman  <cederman@gaisler.com>

gcc/
* config/sparc/sparc.c (atomic_insn_p): New function.
(sparc_do_work_around_errata): Insert NOP instructions to
prevent sequences that could trigger the TN-0010 errata for
UT700.
* config/sparc/sync.md (atomic_compare_and_swap_leon3_1): Make
instruction referable in atomic_insns_p.

From-SVN: r255236

6 years ago[SPARC] Errata workaround for GRLIB-TN-0011
Daniel Cederman [Wed, 29 Nov 2017 15:20:48 +0000 (15:20 +0000)]
[SPARC] Errata workaround for GRLIB-TN-0011

This patch provides a workaround for the errata described in GRLIB-TN-0011.

If the workaround is enabled it will:

 * Insert .align 16 before atomic instructions (swap, ldstub, casa).

It is applicable to GR712RC.

2017-11-29  Daniel Cederman  <cederman@gaisler.com>

gcc/
* config/sparc/sync.md (swapsi): 16-byte align if sparc_fix_gr712rc.
(atomic_compare_and_swap_leon3_1): Likewise.
(ldstub): Likewise.

From-SVN: r255235

6 years ago[SPARC] Errata workaround for GRLIB-TN-0012
Daniel Cederman [Wed, 29 Nov 2017 15:15:48 +0000 (15:15 +0000)]
[SPARC] Errata workaround for GRLIB-TN-0012

This patch provides a workaround for the errata described in GRLIB-TN-0012.

If the workaround is enabled it will:

 * Prevent any floating-point operation from being placed in the
   delay slot of an annulled integer branch.

 * Place a NOP at the branch target of an integer branch if it is
   a floating-point operation or a floating-point branch.

It is applicable to GR712RC.

2017-11-29  Daniel Cederman  <cederman@gaisler.com>

gcc/
* config/sparc/sparc.c (fpop_insn_p): New function.
(sparc_do_work_around_errata): Insert NOP instructions to
prevent sequences that could trigger the TN-0012 errata for
GR712RC.
(pass_work_around_errata::gate): Also test sparc_fix_gr712rc.
* config/sparc/sparc.md (fix_gr712rc): New attribute.
(in_branch_annul_delay): Prevent floating-point instructions
in delay slot of annulled integer branch.

From-SVN: r255234

6 years agore PR tree-optimization/83202 (Try joining operations on consecutive array elements...
Richard Biener [Wed, 29 Nov 2017 14:38:06 +0000 (14:38 +0000)]
re PR tree-optimization/83202 (Try joining operations on consecutive array elements during tree vectorization)

2017-11-29  Richard Biener  <rguenther@suse.de>

PR tree-optimization/83202
* tree-vect-slp.c (scalar_stmts_set_t): New typedef.
(bst_fail): Use it.
(vect_analyze_slp_cost_1): Add visited set, do not account SLP
nodes vectorized to the same stmts multiple times.
(vect_analyze_slp_cost): Allocate a visited set and pass it down.
(vect_analyze_slp_instance): Adjust.
(scalar_stmts_to_slp_tree_map_t): New typedef.
(vect_schedule_slp_instance): Add a map recording the SLP node
representing the vectorized stmts for a set of scalar stmts.
Avoid code-generating redundancies.
(vect_schedule_slp): Allocate map and pass it down.

* gcc.dg/vect/costmodel/x86_64/costmodel-pr83202.c: New testcase.

From-SVN: r255233

6 years agoFix PR number 83817->83187
Nathan Sidwell [Wed, 29 Nov 2017 13:39:23 +0000 (13:39 +0000)]
Fix PR number 83817->83187

From-SVN: r255232

6 years ago[PATCH] complex type canonicalization
Nathan Sidwell [Wed, 29 Nov 2017 13:22:44 +0000 (13:22 +0000)]
[PATCH] complex type canonicalization

https://gcc.gnu.org/ml/gcc-patches/2017-11/msg02453.html
PR c++/83817
* tree.c (build_complex_type): Fix canonicalization.  Only fill in
type if it is new.

PR c++/83187
* g++.dg/opt/pr83187.C: New.

From-SVN: r255231

6 years ago[AArch64] Fix ICE due to store_pair_lanes
Wilco Dijkstra [Wed, 29 Nov 2017 12:22:06 +0000 (12:22 +0000)]
[AArch64] Fix ICE due to store_pair_lanes

The recently added store_pair_lanes causes ICEs in output_operand.
This is due to aarch64_classify_address treating it like a 128-bit STR
rather than a STP.  The valid immediate offsets don't fully overlap,
causing it to return false.  Eg. offset 264 is a valid 8-byte STP offset
but not a valid 16-byte STR offset since it isn't a multiple of 16.

The original instruction isn't passed in the printing code, so the context
is unclear.  The solution is to add a new operand formatting specifier
which is used for LDP/STP instructions like this.  This, like the Uml
constraint that applies to store_pair_lanes, uses PARALLEL when calling
aarch64_classify_address so that it knows it is an STP.
Also add the 'z' specifier for future use by load/store pair instructions.

    gcc/
* config/aarch64/aarch64.c (aarch64_print_operand): Add new
cases for printing LDP/STP memory addresses.
(aarch64_print_address_internal): Renamed from
aarch64_print_operand_address, added parameter, add Pmode check.
(aarch64_print_ldpstp_address): New function for LDP/STP addresses.
(aarch64_print_operand_address): Indirect to
aarch64_print_address_internal.
* config/aarch64/aarch64-simd.md (store_pair_lanes): Use new
'y' operand output specifier.

From-SVN: r255230

6 years agore PR middle-end/83185 (ICE with -fsanitize=address in build_simple_mem_ref_loc,...
Jakub Jelinek [Wed, 29 Nov 2017 10:33:06 +0000 (11:33 +0100)]
re PR middle-end/83185 (ICE with -fsanitize=address in build_simple_mem_ref_loc, at tree.c:4696)

PR middle-end/83185
* tree.c (build_simple_mem_ref_loc): Handle
get_addr_base_and_unit_offset returning a MEM_REF.

* gcc.dg/asan/pr83185.c: New test.

From-SVN: r255229

6 years agore PR middle-end/80929 (Division with constant no more optimized to mult highpart)
Jakub Jelinek [Wed, 29 Nov 2017 10:31:36 +0000 (11:31 +0100)]
re PR middle-end/80929 (Division with constant no more optimized to mult highpart)

PR middle-end/80929
* rtlanal.c (seq_cost): For non-single_set insns try to use insn_cost.

From-SVN: r255228

6 years agofix merge conflict in libgfortran/ChangeLog
Janne Blomqvist [Wed, 29 Nov 2017 09:01:49 +0000 (11:01 +0200)]
fix merge conflict in libgfortran/ChangeLog

From-SVN: r255227

6 years agore PR target/80819 (Useless store to the stack in _mm_set_epi64x with SSE4 -mno...
Jakub Jelinek [Wed, 29 Nov 2017 08:40:13 +0000 (09:40 +0100)]
re PR target/80819 (Useless store to the stack  in _mm_set_epi64x with SSE4 -mno-avx)

PR target/80819
* config/i386/sse.md (vec_concatv2di): Remove * from (=Yr,0,*rm)
alternative.

* gcc.target/i386/pr80819-1.c: New test.
* gcc.target/i386/pr80819-2.c: New test.

From-SVN: r255226

6 years agore PR libfortran/83168 (FAIL: gfortran.dg/fmt_f0_2.f90 with a sanitized libgfortran)
Jerry DeLisle [Wed, 29 Nov 2017 03:07:43 +0000 (03:07 +0000)]
re PR libfortran/83168 (FAIL: gfortran.dg/fmt_f0_2.f90 with a sanitized libgfortran)

2017-11-28  Jerry DeLisle  <jvdelisle@gcc.gnu.org>

PR libgfortran/83168
* io/write.c (select_string): Bump size by one to avoid
overrun.

From-SVN: r255225

6 years agoDaily bump.
GCC Administrator [Wed, 29 Nov 2017 00:16:15 +0000 (00:16 +0000)]
Daily bump.

From-SVN: r255224

6 years agoFix some incorrectly formatted ChangeLog entries.
Jim Wilson [Tue, 28 Nov 2017 21:47:36 +0000 (13:47 -0800)]
Fix some incorrectly formatted ChangeLog entries.

From-SVN: r255220

6 years agoHandle very long lines when printing fix-it hints
David Malcolm [Tue, 28 Nov 2017 21:45:56 +0000 (21:45 +0000)]
Handle very long lines when printing fix-it hints

When fixing PR c/82050 I noticed a bug in how we print fix-it hints
for very long lines: we weren't taking into account the x-offset for
the line when printing the fix-it hint.

This could lead to output where instead of printing:

foo.c:14:3944: error: etc
  = foo.field
        ^~~~~
        replacement

where the lines have been offset to start printing at about column 3900,
the "replacement" line was erroneously *not* offset, and was thus
prefixed by thousands of spaces, leading to large whitespace gaps in
the output, and the replacement failing to line up with the source to be
replaced.

Fixed thusly.

gcc/ChangeLog:
* diagnostic-show-locus.c (layout::print_trailing_fixits): Handle
m_x_offset.
(layout::move_to_column): Likewise.

gcc/testsuite/ChangeLog:
* gcc.dg/plugin/diagnostic-test-show-locus-bw.c
(test_very_wide_line): Update expected output to include a
fix-it hint.
* gcc.dg/plugin/diagnostic-test-show-locus-color.c
(test_very_wide_line): Likewise.
* gcc.dg/plugin/diagnostic_plugin_test_show_locus.c
(test_show_locus): Add a fix-it hint to "test_very_wide_line".

From-SVN: r255219

6 years agore PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)
Jakub Jelinek [Tue, 28 Nov 2017 21:24:32 +0000 (22:24 +0100)]
re PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)

PR sanitizer/81275
* cp-tree.h (SWITCH_STMT_ALL_CASES_P): Define.
(SWITCH_STMT_NO_BREAK_P): Define.
(note_break_stmt, note_iteration_stmt_body_start,
note_iteration_stmt_body_end): Declare.
* decl.c (struct cp_switch): Add has_default_p, break_stmt_seen_p
and in_loop_body_p fields.
(push_switch): Clear them.
(pop_switch): Set SWITCH_STMT_CANNOT_FALLTHRU_P if has_default_p
and !break_stmt_seen_p.  Assert in_loop_body_p is false.
(note_break_stmt, note_iteration_stmt_body_start,
note_iteration_stmt_body_end): New functions.
(finish_case_label): Set has_default_p when both low and high
are NULL_TREE.
* parser.c (cp_parser_iteration_statement): Use
note_iteration_stmt_body_start and note_iteration_stmt_body_end
around parsing iteration body.
* pt.c (tsubst_expr): Likewise.
* cp-objcp-common.c (cxx_block_may_fallthru): Return false for
SWITCH_STMT which contains no BREAK_STMTs, contains a default:
CASE_LABEL_EXPR and where SWITCH_STMT_BODY isn't empty and
can't fallthru.
* semantics.c (finish_break_stmt): Call note_break_stmt.
* cp-gimplify.c (genericize_switch_stmt): Copy SWITCH_STMT_ALL_CASES_P
bit to SWITCH_ALL_CASES_P.  Assert that if SWITCH_STMT_NO_BREAK_P then
the break label is not TREE_USED.

* g++.dg/warn/pr81275-1.C: New test.
* g++.dg/warn/pr81275-2.C: New test.
* g++.dg/warn/pr81275-3.C: New test.
* c-c++-common/tsan/pr81275.c: Skip for C++ and -O2.

From-SVN: r255218

6 years agore PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)
Jakub Jelinek [Tue, 28 Nov 2017 21:22:52 +0000 (22:22 +0100)]
re PR sanitizer/81275 (-fsanitize=thread produce incorrect -Wreturn-type warning)

PR sanitizer/81275
* tree.c (block_may_fallthru): Return false if SWITCH_ALL_CASES_P
is set on SWITCH_EXPR and !block_may_fallthru (SWITCH_BODY ()).
c/
* c-typeck.c (c_finish_case): Set SWITCH_ALL_CASES_P if
c_switch_covers_all_cases_p returns true.
c-family/
* c-common.c (c_switch_covers_all_cases_p_1,
c_switch_covers_all_cases_p): New functions.
* c-common.h (c_switch_covers_all_cases_p): Declare.
testsuite/
* c-c++-common/tsan/pr81275.c: New test.

From-SVN: r255217

6 years agoPR 53796 Improve INQUIRE(RECL=...) handling
Janne Blomqvist [Tue, 28 Nov 2017 19:28:50 +0000 (21:28 +0200)]
PR 53796 Improve INQUIRE(RECL=...) handling

The current F2018 draft (N2137) specifies behavior of the RECL=
specifier in the INQUIRE statement, where it previously was left as
undefined. Namely:

- If the unit is not connected, RECL= should be given the value -1.
- If the unit is connected with stream access, RECL= should be given
  the value -2.

Further, as PR 53796 describes, the handling of RECL= is poor in other
ways as well. When the recl is set to the maximum possible
(GFC_INTEGER_8_HUGE / LLONG_MAX), which it does by default except for
preconnected units, and when INQUIRE(RECL=) is used with a 4 byte
integer, the value is truncated and the 4 byte value is thus
-1. Fixing this to generate an error is a lot of work, as currently
the truncation is done by the frontend, the library sees only an 8
byte value with no indication that the frontend is going to copy it to
a 4 byte one. Instead, this patch does a bit twiddling trick such that
the truncated 4 byte value is GFC_INTEGER_4_HUGE while still being
0.99999999 * GFC_INTEGER_8_HUGE which is large enough for all
practical purposes.

Finally, the patch removes GFORTRAN_DEFAULT_RECL which was used only
for preconnected units, and instead uses the same approach as describe
above.

Regtested on x86_64-pc-linux-gnu, Ok for trunk.

gcc/fortran/ChangeLog:

2017-11-28  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/53796
* gfortran.texi: Remove mentions of GFORTRAN_DEFAULT_RECL.

libgfortran/ChangeLog:

2017-11-28  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/53796
* io/inquire.c (inquire_via_unit): Set recl to -1 for unconnected
units.
* io/io.h (default_recl): New variable.
* io/open.c (new_unit): Set recl to default_recl for sequential,
-2 for stream access.
* io/transfer.c (read_block_form): Test against default_recl
instead of DEFAULT_RECL.
(write_block): Likewise.
* io/unit.c (init_units): Calculate max_offset, default_recl.
* libgfortran.h (DEFAULT_RECL): Remove.
* runtime/environ.c: Remove GFORTRAN_DEFAULT_RECL.

gcc/testsuite/ChangeLog:

2017-11-28  Janne Blomqvist  <jb@gcc.gnu.org>

PR fortran/53796
* gfortran.dg/inquire_recl_f2018.f90: New test.

From-SVN: r255215

6 years agoReject fix-it hints for various awkward boundary cases (PR c/82050)
David Malcolm [Tue, 28 Nov 2017 19:24:35 +0000 (19:24 +0000)]
Reject fix-it hints for various awkward boundary cases (PR c/82050)

PR c/82050 reports a failed assertion deep within diagnostic_show_locus's
code for printing fix-it hints.

The root cause is a fix-it hint suggesting a textual replacement,
where the affected column numbers straddle the LINE_MAP_MAX_COLUMN_NUMBER
boundary, so that the start of the range has a column number, but the
end of the range doesn't.

The fix is to verify that the column numbers are sane when adding fix-it
hints to a rich_location, rejecting fix-it hints where they are not.

libcpp/ChangeLog:
PR c/82050
* include/line-map.h (LINE_MAP_MAX_COLUMN_NUMBER): Move here.
* line-map.c (LINE_MAP_MAX_COLUMN_NUMBER): ...from here.
(rich_location::maybe_add_fixit): Reject fix-it hints in which
the start column exceeds the next column.

From-SVN: r255214

6 years ago[PR 82808] Use proper result types for arithmetic jump functions
Prathamesh Kulkarni [Tue, 28 Nov 2017 18:52:49 +0000 (18:52 +0000)]
[PR 82808] Use proper result types for arithmetic jump functions

2017-11-28  Prathamesh Kulkarni  <prathamesh.kulkarni@linaro.org>
    Martin Jambor  <mjambor@suse.cz>

PR ipa/82808
* tree.h (expr_type_first_operand_type_p): Declare
* tree.c (expr_type_first_operand_type_p): New function.
* ipa-prop.h (ipa_get_type): Allow i to be out of bounds.
(ipa_value_from_jfunc): Adjust declaration.
* ipa-cp.c (ipa_get_jf_pass_through_result): New parameter RES_TYPE.
Use it as result type for arithmetics, unless it is NULL in which case
be more conservative.
(ipa_value_from_jfunc): New parameter PARM_TYPE, pass it to
ipa_get_jf_pass_through_result.
(propagate_vals_across_pass_through): Likewise.
(propagate_scalar_across_jump_function): New parameter PARM_TYPE, pass
is to propagate_vals_across_pass_through.
(propagate_constants_across_call): Pass PARM_TYPE to
propagate_scalar_across_jump_function.
(find_more_scalar_values_for_callers_subset): Pass parameter type to
ipa_value_from_jfunc.
(cgraph_edge_brings_all_scalars_for_node): Likewise.
* ipa-fnsummary.c (evaluate_properties_for_edge): Renamed parms_info
to caller_parms_info, pass parameter type to ipa_value_from_jfunc.
* ipa-prop.c (try_make_edge_direct_simple_call): New parameter
target_type, pass it to ipa_value_from_jfunc.
(update_indirect_edges_after_inlining): Pass parameter type to
try_make_edge_direct_simple_call.

testsuite/
* gcc.dg/ipa/pr82808.c: New test.

Co-Authored-By: Martin Jambor <mjambor@suse.cz>
From-SVN: r255212

6 years ago* gimple-ssa-evrp-analyze.c
Jeff Law [Tue, 28 Nov 2017 18:21:10 +0000 (11:21 -0700)]
* gimple-ssa-evrp-analyze.c
(evrp_range_analyzer::record_ranges_from_phis): Only use SCEV to
refine ranges if scev_initialized_p returns true.
* vr-values.c (vr_values::extract_range_from_phi_node): Likewise.

From-SVN: r255211

6 years ago* g++.dg/cpp0x/fntmpdefarg2a.C: Move expected warning.
Jason Merrill [Tue, 28 Nov 2017 18:09:15 +0000 (13:09 -0500)]
* g++.dg/cpp0x/fntmpdefarg2a.C: Move expected warning.

From-SVN: r255210

6 years agoEnable VBMI2 support [5/7]
Julia Koval [Tue, 28 Nov 2017 17:59:39 +0000 (18:59 +0100)]
Enable VBMI2 support [5/7]

gcc/
* config/i386/avx512vbmi2intrin.h (_mm512_shrdi_epi16,
_mm512_mask_shrdi_epi16, _mm512_maskz_shrdi_epi16, _mm512_shrdi_epi32,
_mm512_mask_shrdi_epi32, _mm512_maskz_shrdi_epi32, _mm512_shrdi_epi64,
_mm512_mask_shrdi_epi64, _mm512_maskz_shrdi_epi64): New intrinsics.
* config/i386/avx512vbmi2vlintrin.h (_mm256_shrdi_epi16,
_mm256_mask_shrdi_epi16, _mm256_maskz_shrdi_epi16,
_mm256_mask_shrdi_epi32, _mm256_maskz_shrdi_epi32, _mm256_shrdi_epi32,
_mm256_mask_shrdi_epi64, _mm256_maskz_shrdi_epi64, _mm256_shrdi_epi64,
_mm_mask_shrdi_epi16, _mm_maskz_shrdi_epi16, _mm_shrdi_epi16,
_mm_mask_shrdi_epi32, _mm_maskz_shrdi_epi32, _mm_shrdi_epi32,
_mm_mask_shrdi_epi64, _mm_maskz_shrdi_epi64, _mm_shrdi_epi64): Ditto.
* config/i386/i386-builtin.def (__builtin_ia32_vpshrd_v32hi,
__builtin_ia32_vpshrd_v32hi_mask, __builtin_ia32_vpshrd_v16hi,
__builtin_ia32_vpshrd_v16hi_mask, __builtin_ia32_vpshrd_v8hi,
__builtin_ia32_vpshrd_v8hi_mask, __builtin_ia32_vpshrd_v16si,
__builtin_ia32_vpshrd_v16si_mask, __builtin_ia32_vpshrd_v8si,
__builtin_ia32_vpshrd_v8si_mask, __builtin_ia32_vpshrd_v4si,
__builtin_ia32_vpshrd_v4si_mask, __builtin_ia32_vpshrd_v8di,
__builtin_ia32_vpshrd_v8di_mask, __builtin_ia32_vpshrd_v4di,
__builtin_ia32_vpshrd_v4di_mask, __builtin_ia32_vpshrd_v2di,
__builtin_ia32_vpshrd_v2di_mask): New builtins.
* config/i386/sse.md (vpshrd_<mode><mask_name>): New pattern.

gcc/testsuite/
* gcc.target/i386/avx-1.c: Handle new intrinsics.
* gcc.target/i386/sse-13.c: Ditto.
* gcc.target/i386/sse-23.c: Ditto.
* gcc.target/i386/avx512f-vpshrdd-2.c: New.
* gcc.target/i386/avx512f-vpshrdq-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrd-1.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdd-2.c: Ditto.
* gcc.target/i386/avx512vl-vpshrdq-2.c: Ditto.

From-SVN: r255209

6 years agoEnable VBMI2 support [4/7]
Julia Koval [Tue, 28 Nov 2017 17:51:12 +0000 (18:51 +0100)]
Enable VBMI2 support [4/7]

gcc/
config/i386/avx512vbmi2intrin.h (_mm512_shldi_epi16,
_mm512_mask_shldi_epi16, _mm512_maskz_shldi_epi16, _mm512_shldi_epi32,
_mm512_mask_shldi_epi32, _mm512_maskz_shldi_epi32, _mm512_shldi_epi64,
_mm512_mask_shldi_epi64, _mm512_maskz_shldi_epi64): New intrinsics.
config/i386/avx512vbmi2vlintrin.h (_mm256_shldi_epi16,
_mm256_mask_shldi_epi16, _mm256_maskz_shldi_epi16,
_mm256_mask_shldi_epi32, _mm256_maskz_shldi_epi32, _mm256_shldi_epi32,
_mm256_mask_shldi_epi64, _mm256_maskz_shldi_epi64, _mm256_shldi_epi64,
_mm_mask_shldi_epi16, _mm_maskz_shldi_epi16, _mm_shldi_epi16,
_mm_mask_shldi_epi32, _mm_maskz_shldi_epi32, _mm_shldi_epi32,
_mm_mask_shldi_epi64, _mm_maskz_shldi_epi64, _mm_shldi_epi64): Ditto.
config/i386/i386-builtin-types.def (V32HI_FTYPE_V32HI_V32HI_INT,
V32HI_FTYPE_V32HI_V32HI_INT_V32HI_INT, V16SI_FTYPE_V16SI_V16SI_INT,
V16SI_FTYPE_V16SI_V16SI_INT_V16SI_INT,
V8DI_FTYPE_V8DI_V8DI_INT_V8DI_INT, V8SI_FTYPE_V8SI_V8SI_INT_V8SI_INT,
V16HI_FTYPE_V16HI_V16HI_INT_V16HI_INT,
V4DI_FTYPE_V4DI_V4DI_INT_V4DI_INT,
V8HI_FTYPE_V8HI_V8HI_INT_V8HI_INT,
V4SI_FTYPE_V4SI_V4SI_INT_V4SI_INT,
V2DI_FTYPE_V2DI_V2DI_INT_V2DI_INT): New types.
config/i386/i386-builtin.def (__builtin_ia32_vpshld_v32hi,
__builtin_ia32_vpshld_v32hi_mask, __builtin_ia32_vpshld_v16hi,
__builtin_ia32_vpshld_v16hi_mask, __builtin_ia32_vpshld_v8hi,
__builtin_ia32_vpshld_v8hi_mask, __builtin_ia32_vpshld_v16si,
__builtin_ia32_vpshld_v16si_mask, __builtin_ia32_vpshld_v8si,
__builtin_ia32_vpshld_v8si_mask, __builtin_ia32_vpshld_v4si,
__builtin_ia32_vpshld_v4si_mask, __builtin_ia32_vpshld_v8di,
__builtin_ia32_vpshld_v8di_mask, __builtin_ia32_vpshld_v4di,
__builtin_ia32_vpshld_v4di_mask, __builtin_ia32_vpshld_v2di,
__builtin_ia32_vpshld_v2di_mask): New builtins.
config/i386/i386.c (ix86_expand_args_builtin): Handle new types.
config/i386/sse.md (vpshld_<mode><mask_name>): New pattern.

gcc/testsuite/
gcc.target/i386/avx-1.c: Handle new intrinics.
gcc.target/i386/sse-13.c: Ditto.
gcc.target/i386/sse-23.c: Ditto.
gcc.target/i386/avx512f-vpshld-1.c: New test.
gcc.target/i386/avx512f-vpshldd-2.c: Ditto.
gcc.target/i386/avx512f-vpshldq-2.c: Ditto.
gcc.target/i386/avx512vl-vpshld-1.c: Ditto.
gcc.target/i386/avx512vl-vpshldd-2.c: Ditto.
gcc.target/i386/avx512vl-vpshldq-2.c: Ditto.

From-SVN: r255208