Richard Biener [Mon, 18 Mar 2019 13:59:11 +0000 (13:59 +0000)]
re PR tree-optimization/88945 (ICE in fold_convert_loc in FRE when using -fdump-tree-fre-details)
2019-03-18 Richard Biener <rguenther@suse.de>
PR middle-end/88945
* tree-ssanames.c (release_ssa_name_fn): For released SSA names
use a TREE_TYPE of error_mark_node to avoid ICEs when dumping
basic-blocks that are removed. Remove restoring SSA_NAME_VAR.
* tree-outof-ssa.c (eliminate_useless_phis): Remove redundant checking.
Andrew Stubbs [Mon, 18 Mar 2019 13:13:27 +0000 (13:13 +0000)]
Implement circular print buffer.
2019-03-18 Andrew Stubbs <ams@codesourcery.com>
gcc/
* config/gcn/gcn-run.c (struct output): Make next_output unsigned.
Extend queue to 1024 entries.
Add "consumed" field.
(gomp_print_output): Remove print_index parameter.
Add final parameter.
Change limit to unsigned.
Use consumed field to implement circular buffer.
Detect interrupted print in final pass.
Flush output at the end.
(run): Update gomp_print_output usage.
(main): Initialize kernargs->output_data.consumed.
Fix a case in which the vector cost model was ignored
This patch fixes a case in which we vectorised something with a
fully-predicated loop even after the cost model had rejected it.
E.g. the loop in the testcase has the costs:
and we can see that the loop executes at most three times, but we
decided to vectorise it anyway.
(The costs here are equal for three iterations, but the same thing
happens even when the vector code is strictly more expensive.)
The problem is the handling of "/VF" in:
/* Calculate number of iterations required to make the vector version
profitable, relative to the loop bodies only. The following condition
must hold true:
SIC * niters + SOC > VIC * ((niters-PL_ITERS-EP_ITERS)/VF) + VOC
where
SIC = scalar iteration cost, VIC = vector iteration cost,
VOC = vector outside cost, VF = vectorization factor,
PL_ITERS = prologue iterations, EP_ITERS= epilogue iterations
SOC = scalar outside cost for run time cost model check. */
We treat the "/VF" as truncating, but for fully-predicated loops, it's
closer to a ceil division, since fractional iterations are handled by a
full iteration with some predicate bits set to false.
The easiest fix seemed to be to calculate the minimum number of vector
iterations first, then use that to calculate the minimum number of scalar
iterations.
Calculating the minimum number of vector iterations might make sense for
unpredicated loops too, since calculating the scalar niters directly
doesn't take into account the fact that the VIC multiple has to be an
integer. But the handling of PL_ITERS and EP_ITERS for unpredicated
loops is a bit hand-wavy anyway, so maybe vagueness here cancels out
vagueness there?
Either way, changing this for unpredicated loops would be much too
invasive for stage 4, so the patch keeps it specific to fully-predicated
loops (i.e. SVE) for now. There's no functional change for other targets.
2019-03-18 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* tree-vect-loop.c (vect_estimate_min_profitable_iters): Fix the
calculation of the minimum number of scalar iterations for
fully-predicated loops.
gcc/testsuite/
* gcc.target/aarch64/sve/cost_model_1.c: New test.
Andrew Burgess [Mon, 18 Mar 2019 10:42:53 +0000 (10:42 +0000)]
gcc/riscv: Correctly ignore empty C++ structs when flattening for ABI
This fixes PR target/89627.
The RISC-V ABI document[1] says:
For the purposes of this section, "struct" refers to a C struct
with its hierarchy flattened, including any array fields. That is,
struct { struct { float f[1]; } g[2]; } and struct { float f; float
g; } are treated the same. Fields containing empty structs or
unions are ignored while flattening, even in C++, unless they have
nontrivial copy constructors or destructors.
However, this flattening only applies when one of the fields of the
flattened structure can be placed into a floating point register,
otherwise no flattening occurs.
Currently GCC fails to correctly consider that empty C++ structures
have a non-zero size when constructing the arguments from a flattened
structure, and as a result, trying to pass a C++ structure like this:
struct sf { struct {} e; float f; };
Doesn't work correctly, GCC fails to take the offset of 'f' within
'sf' into account and will actually pass the space backing 'e' as the
contents of 'f'.
This patch fixes this so that 'f' will be passed correctly. A couple
of new tests are added to cover this functionality.
PR target/89627
* config/riscv/riscv.c (riscv_pass_fpr_single): Add offset
parameter, and make use of it.
(riscv_get_arg_info): Pass offset to riscv_pass_fpr_single.
gcc/testsuite/ChangeLog:
PR target/89627
* g++.target/riscv/call-with-empty-struct-float.C: New file.
* g++.target/riscv/call-with-empty-struct-int.C: New file.
* g++.target/riscv/call-with-empty-struct.H: New file.
* g++.target/riscv/riscv.exp: New file.
Paolo Carlini [Mon, 18 Mar 2019 09:38:27 +0000 (09:38 +0000)]
re PR c++/85014 (internal compiler error: in lookup_base, at cp/search.c:185)
/cp
2019-03-18 Paolo Carlini <paolo.carlini@oracle.com>
PR c++/85014
* semantics.c (finish_non_static_data_member): Check return value
of context_for_name_lookup and immediately return error_mark_node
if isn't a type.
/testsuite
2019-03-18 Paolo Carlini <paolo.carlini@oracle.com>
* config/arc/arc.opt (mcode-density-frame): Get the inital value
from TARGET_CODE_DENSITY_FRAME_DEFAULT.
* config/arc/elf.h (TARGET_CODE_DENSITY_FRAME_DEFAULT): Define.
* config/arc/linux.h (TARGET_CODE_DENSITY_FRAME_DEFAULT): Define.
* config/arc/arc.md (pop_multi_fp_blink): Adjust constraints to
match what the ops is doing.
(push_multi_fp_blink): Likewise.
* config/arc/arc.c (arc_override_options): Enable enter/leave when
compiling for size and elf target.
(arc_save_callee_enter): Adjust note to match what enter/leave
operation does.
The ARC port is changing the allocation order in the
arc_conditional_register_usage function, but this is not the proper
way. Thus, we employ ADJUST_REG_ALLOC_ORDER hook for this task.
Jason Merrill [Sun, 17 Mar 2019 20:07:26 +0000 (16:07 -0400)]
PR c++/89571 - ICE with ill-formed noexcept on constructor.
Earlier changes to defer instantiating a defaulted noexcept-specifier that
depends on yet-unparsed default member initializers broke this testcase,
where instantiation fails for another reason. In this case there's no
reason to defer and try again later, so let's not.
* pt.c (maybe_instantiate_noexcept): Only return false if defaulted.
(regenerate_decl_from_template): Use it for noexcept-specs.
Jakub Jelinek [Sat, 16 Mar 2019 21:18:49 +0000 (22:18 +0100)]
re PR fortran/89724 (Fortran diagnostics give wrong line number because of math-vector-fortran.h header file)
PR fortran/89724
* scanner.c (load_line): Remove linenum and current_line static
variables, add warned_tabs automatic variable. Use current_file->line
instead of current_line and warned_tabs boolean to avoid diagnosing
tabs multiple times on the same line.
* gfortran.dg/continuation_15.f90: New test.
* gfortran.dg/continuation_16.f90: New test.
Thomas Koenig [Sat, 16 Mar 2019 11:50:03 +0000 (11:50 +0000)]
re PR fortran/84394 (compiler error when using modules with derived types in block data subprograms)
2019-03-16 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/84394
* symbol.c (gfc_add_subroutine): If we are encountering a
subrtoutine within a BLOCK DATA and the name starts with an
underscore, do not check.
2019-03-16 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/84394
* gfortran.dg/blockdata_11.f90: New test.
Kelvin Nilsen [Fri, 15 Mar 2019 19:52:43 +0000 (19:52 +0000)]
re PR target/87532 (bad results from vec_extract(unsigned char, foo) dependent upon function inline)
gcc/ChangeLog:
2019-03-15 Kelvin Nilsen <kelvin@gcc.gnu.org>
PR target/87532
* config/rs6000/rs6000-c.c (altivec_resolve_overloaded_builtin):
When handling vec_extract, use modular arithmetic to allow
constant selectors greater than vector length.
* config/rs6000/rs6000.c (rs6000_expand_vector_extract): Allow
V1TImode vectors to have constant selector values greater than 0.
Use modular arithmetic to compute vector index.
(rs6000_split_vec_extract_var): Use modular arithmetic to compute
index for in-memory vectors. Correct code generation for
in-register vectors.
(altivec_expand_vec_ext_builtin): Use modular arithmetic to
compute index.
gcc/testsuite/ChangeLog:
2019-03-15 Kelvin Nilsen <kelvin@gcc.gnu.org>
PR target/87532
* gcc.target/powerpc/fold-vec-extract-char.p8.c: Modify expected
instruction selection.
* gcc.target/powerpc/fold-vec-extract-int.p8.c: Likewise.
* gcc.target/powerpc/fold-vec-extract-short.p8.c: Likewise.
* gcc.target/powerpc/pr87532-mc.c: New test.
* gcc.target/powerpc/pr87532.c: New test.
* gcc.target/powerpc/vec-extract-v16qiu-v2.h: New test.
* gcc.target/powerpc/vec-extract-v16qiu-v2a.c: New test.
* gcc.target/powerpc/vec-extract-v16qiu-v2b.c: New test.
* gcc.target/powerpc/vsx-builtin-10a.c: New test.
* gcc.target/powerpc/vsx-builtin-10b.c: New test.
* gcc.target/powerpc/vsx-builtin-11a.c: New test.
* gcc.target/powerpc/vsx-builtin-11b.c: New test.
* gcc.target/powerpc/vsx-builtin-12a.c: New test.
* gcc.target/powerpc/vsx-builtin-12b.c: New test.
* gcc.target/powerpc/vsx-builtin-13a.c: New test.
* gcc.target/powerpc/vsx-builtin-13b.c: New test.
* gcc.target/powerpc/vsx-builtin-14a.c: New test.
* gcc.target/powerpc/vsx-builtin-14b.c: New test.
* gcc.target/powerpc/vsx-builtin-15a.c: New test.
* gcc.target/powerpc/vsx-builtin-15b.c: New test.
* gcc.target/powerpc/vsx-builtin-16a.c: New test.
* gcc.target/powerpc/vsx-builtin-16b.c: New test.
* gcc.target/powerpc/vsx-builtin-17a.c: New test.
* gcc.target/powerpc/vsx-builtin-17b.c: New test.
* gcc.target/powerpc/vsx-builtin-18a.c: New test.
* gcc.target/powerpc/vsx-builtin-18b.c: New test.
* gcc.target/powerpc/vsx-builtin-19a.c: New test.
* gcc.target/powerpc/vsx-builtin-19b.c: New test.
* gcc.target/powerpc/vsx-builtin-20a.c: New test.
* gcc.target/powerpc/vsx-builtin-20b.c: New test.
* gcc.target/powerpc/vsx-builtin-9a.c: New test.
* gcc.target/powerpc/vsx-builtin-9b.c: New test.
Alexandre Oliva [Fri, 15 Mar 2019 13:56:55 +0000 (13:56 +0000)]
[PR88534] accept VAR_DECL in class literal template parms
P0732R2 / C++ 2a introduce class literals as template parameters. The
front-end uses VAR_DECLs constructed from such literals to bind the
template PARM_DECLs, but dwarf2out.c used to reject such VAR_DECLs.
Taking DECL_INITIAL from such VAR_DECLs enables the generation of
DW_AT_const_value for them, at least when the class literal can
actually be represented as such.
As of recently the -march,-mcpu,-mtune strings in the error messages are
now quoted.
This patch adjusts the testcases in gcc.target/aarch64/ that had started
failing due to that change.
Ian Lance Taylor [Fri, 15 Mar 2019 04:34:43 +0000 (04:34 +0000)]
compiler: eliminate bound checks in append expression
The compiler generates two array index expressions when lowering
an append expression. Currently they generate bound checks.
Bound checks are not necessary in this case, as we know the slice
has, or will grow to, enough length and capacity. Eliminate them.
Jason Merrill [Thu, 14 Mar 2019 22:47:01 +0000 (18:47 -0400)]
hash-table.h (remove_elt_with_hash): Return if slot is NULL rather than if is_empty (*slot).
* hash-table.h (remove_elt_with_hash): Return if slot is NULL rather
than if is_empty (*slot).
* hash-set-tests.c (test_set_of_strings): Add tests for addition of
existing elt and for elt removal.
* hash-map-tests.c (test_map_of_strings_to_int): Add test for removal
of already removed elt.
* hashtab.c (htab_remove_elt_with_hash): Return if slot is NULL rather
than if *slot is HTAB_EMPTY_ENTRY.
Co-Authored-By: Jakub Jelinek <jakub@redhat.com>
From-SVN: r269695
we must copy the REG_EH_REGION note to the first insn and split the block
after the newly added insn. The REG_EH_REGION on the second insn will be
removed later since it no longer traps.
Richard Biener [Thu, 14 Mar 2019 14:05:26 +0000 (14:05 +0000)]
re PR target/89711 (ICE in insert_vi_for_tree, at tree-ssa-structalias.c:2832)
2019-03-14 Richard Biener <rguenther@suse.de>
PR target/89711
* config/i386/i386.c (make_resolver_func): Properly set
DECL_CONTEXT on the RESULT_DECL.
* config/rs6000/rs6000.c (make_resolver_func): Likewise.
Richard Biener [Thu, 14 Mar 2019 13:25:27 +0000 (13:25 +0000)]
c-tree.h (enum c_declspec_il): New.
2019-03-14 Richard Biener <rguenther@suse.de>
c/
* c-tree.h (enum c_declspec_il): New.
(struct c_declspecs): Merge gimple_p and rtl_p into declspec_il
enum bitfield.
* c-parser.c (c_parser_declaration_or_fndef): Adjust accordingly.
Pass start pass and declspec_il to c_parser_parse_gimple_body.
(c_parser_declspecs): Adjust.
* gimple-parser.c: Include cfg.h, cfghooks.h, cfganal.h, tree-cfg.h,
gimple-iterator.h, cfgloop.h, tree-phinodes.h, tree-into-ssa.h
and bitmap.h.
(struct gimple_parser): New.
(gimple_parser::push_edge): New method.
(c_parser_gimple_parse_bb_spec): New helper.
(c_parser_parse_gimple_body): Get start pass and IL specification.
Initialize SSA and CFG.
(c_parser_gimple_compound_statement): Handle CFG and SSA build.
Build a gimple_parser parsing state and pass it along.
(c_parser_gimple_statement): Change intermittend __PHI internal
function argument for the edge.
(c_parser_gimple_or_rtl_pass_list): Handle ssa, cfg flags.
(c_parser_gimple_goto_stmt): Record edges to build.
(c_parser_gimple_if_stmt): Likewise.
* gimple-parser.h (c_parser_parse_gimple_body): Adjust.
(c_parser_gimple_or_rtl_pass_list): Likewise.
* gimple-pretty-print.c: Include cfgloop.h.
(dump_gimple_phi): Adjust.
(dump_gimple_bb_header): Dump loop header for GIMPLE.
(pp_cfg_jump): Adjust.
(dump_implicit_edges): Dump fallthru to next block for GIMPLE as well.
* tree-cfg.c (build_gimple_cfg): Remove lower_phi_internal_fn call.
(lower_phi_internal_fn): Remove.
(verify_gimple_call): Remove IFN_PHI special-casing.
(dump_function_to_file): Dump IL state.
* tree-into-ssa.c (rewrite_add_phi_arguments): Revert changes
done to deal with PHI nodes being present in non-SSA state.
Jakub Jelinek [Thu, 14 Mar 2019 13:05:34 +0000 (14:05 +0100)]
re PR ipa/89684 (ICE in gsi_for_stmt, at gimple-iterator.c:613)
PR ipa/89684
* multiple_target.c (create_dispatcher_calls): Change
references_to_redirect from vector of ipa_ref * to vector of ipa_ref.
In the node->iterate_referring loop, push *ref rather than ref, call
ref->remove_reference () and always pass 0 to iterate_referring.
Jakub Jelinek [Thu, 14 Mar 2019 08:56:47 +0000 (09:56 +0100)]
re PR tree-optimization/89703 (ICE in compare_values_warnv, at tree-vrp.c:997)
PR tree-optimization/89703
* tree-ssa-strlen.c (valid_builtin_call): Punt if stmt call types
aren't compatible also with builtin_decl_explicit. Check pure
or non-pure status of BUILT_IN_STR{{,N}CMP,N{LEN,{CAT,CPY}{,_CHK}}}
and BUILT_IN_STPNCPY{,_CHK}.
* gcc.c-torture/compile/pr89703-1.c: New test.
* gcc.c-torture/compile/pr89703-2.c: New test.
H.J. Lu [Thu, 14 Mar 2019 08:49:54 +0000 (08:49 +0000)]
x32: Add addr32 prefix to VSIB address
32-bit indices in VSIB address are sign-extended to 64 bits. In x32,
when 32-bit indices are used as addresses, like in
vgatherdps %ymm7, 0(,%ymm9,1), %ymm6
32-bit indices, 0xf7fa3010, is sign-extended to 0xfffffffff7fa3010 which
is invalid address. Add addr32 prefix to UNSPEC_VSIBADDR instructions
for x32 if there is no base register nor symbol.
This fixes 175.vpr and 254.gap in SPEC CPU 2000 on x32 with
-Ofast -funroll-loops -march=haswell
gcc/
PR target/89523
* config/i386/i386.c (ix86_print_operand): Handle 'M' to add
addr32 prefix to VSIB address for X32.
* config/i386/sse.md (*avx512pf_gatherpf<mode>sf_mask): Prepend
"%M2" to opcode.
(*avx512pf_gatherpf<mode>df_mask): Likewise.
(*avx512pf_scatterpf<mode>sf_mask): Likewise.
(*avx512pf_scatterpf<mode>df_mask): Likewise.
(*avx2_gathersi<mode>): Prepend "%M3" to opcode.
(*avx2_gathersi<mode>_2): Prepend "%M2" to opcode.
(*avx2_gatherdi<mode>): Prepend "%M3" to opcode.
(*avx2_gatherdi<mode>_2): Prepend "%M2" to opcode.
(*avx2_gatherdi<mode>_3): Prepend "%M3" to opcode.
(*avx2_gatherdi<mode>_4): Prepend "%M2" to opcode.`
(*avx512f_gathersi<mode>): Prepend "%M4" to opcode.
(*avx512f_gathersi<mode>_2): Prepend "%M3" to opcode.
(*avx512f_gatherdi<mode>): Prepend "%M4" to opcode.
(*avx512f_gatherdi<mode>_2): Prepend "%M3" to opcode.
(*avx512f_scattersi<mode>): Prepend "%M0" to opcode.
(*avx512f_scatterdi<mode>): Likewise.
Jakub Jelinek [Thu, 14 Mar 2019 08:13:09 +0000 (09:13 +0100)]
re PR c++/89652 (ICE during constexpr evaluation)
PR c++/89652
* constexpr.c (struct constexpr_ctx): Change save_exprs type from
hash_set<tree> to vec<tree>.
(cxx_eval_call_expression): Adjust for save_exprs being a vec instead
of hash_set.
(cxx_eval_loop_expr): Likewise. Truncate the vector after each
removal of SAVE_EXPRs from values.
(cxx_eval_constant_expression) <case SAVE_EXPR>: Call safe_push
method on save_exprs instead of add.
Jason Merrill [Wed, 13 Mar 2019 23:34:51 +0000 (19:34 -0400)]
PR c++/86521 - C++17 copy elision in initialization by constructor.
This is an overlooked case in C++17 mandatory copy elision: We want overload
resolution to reflect that initializing an object from a prvalue does not
involve a copy or move constructor even when [over.match.ctor] says that
only constructors are candidates. Here I implement that by looking through
the copy/move constructor in joust.
* call.c (joust_maybe_elide_copy): New.
(joust): Call it.
Harald Anlauf [Wed, 13 Mar 2019 21:33:27 +0000 (21:33 +0000)]
re PR fortran/87045 (pointer to array of character)
2019-03-13 Harald Anlauf <anlauf@gmx.de>
PR fortran/87045
* trans-expr.c (gfc_trans_pointer_assignment): Move check for same
string length so that we do not get false errors for deferred
length.
PR fortran/87045
* gfortran.dg/pr87045.f90: New test.
Jakub Jelinek [Wed, 13 Mar 2019 20:22:15 +0000 (21:22 +0100)]
re PR debug/89498 (ICE in AT_loc_list, at dwarf2out.c:4871)
PR debug/89498
* dwarf2out.c (size_of_die): For dw_val_class_view_list always use
DWARF_OFFSET_SIZE.
(value_format): For dw_val_class_view_list never use DW_FORM_loclistx.
Marek Polacek [Wed, 13 Mar 2019 20:04:33 +0000 (20:04 +0000)]
PR c++/89660 - bogus error with -Wredundant-move.
* typeck.c (maybe_warn_pessimizing_move): Only accept (T &) &arg
as the std::move's argument. Don't call convert_for_initialization
when warn_redundant_move isn't on.
* g++.dg/cpp0x/Wredundant-move8.C: New test.
* g++.dg/cpp0x/Wredundant-move9.C: New test.
Janus Weil [Wed, 13 Mar 2019 19:52:23 +0000 (20:52 +0100)]
re PR fortran/89601 ([PDT] ICE: Segmentation fault (in resolve_component))
fix PR 89601
2019-03-13 Janus Weil <janus@gcc.gnu.org>
PR fortran/89601
* decl.c (gfc_match_formal_arglist): Reject empty type parameter lists.
(gfc_match_derived_decl): Mark as PDT only if type parameter list was
matched successfully.
2019-03-13 Janus Weil <janus@gcc.gnu.org>
PR fortran/89601
* gfortran.dg/pdt_16.f03: Modified to avoid follow-up errors.
* gfortran.dg/pdt_30.f90: New test case.
Eric Botcazou [Wed, 13 Mar 2019 09:15:02 +0000 (09:15 +0000)]
re PR sanitizer/80953 (Support libsanitizer on Solaris)
PR sanitizer/80953
Merge from LLVM revision 355979
* asan/asan_globals.c (GetGlobalsForAddress): Use internal_memcpy to
copy Global objects for SPARC with GCC.
Eric Botcazou [Wed, 13 Mar 2019 09:11:46 +0000 (09:11 +0000)]
re PR sanitizer/80953 (Support libsanitizer on Solaris)
PR sanitizer/80953
Merge from LLVM revision 355978
* sanitizer_common/sanitizer_allocator_primary32.h
(class SizeClassAllocator32): Assert that kSpaceSize is power of 2 if
SANITIZER_SIGN_EXTENDED_ADDRESSES is set.
(PointerIsMine): Deal with SANITIZER_SIGN_EXTENDED_ADDRESSES.
(ComputeRegionId): Likewise.
* sanitizer_common/sanitizer_linux.cc (GetMaxVirtualAddress): Return
appropriate value for SPARC 64-bit.
* sanitizer_common/sanitizer_platform.h (SANITIZER_MMAP_RANGE_SIZE):
Define for SPARC.
(SANITIZER_SIGN_EXTENDED_ADDRESSES): Define to 1 for SPARC 64-bit.
Eric Botcazou [Wed, 13 Mar 2019 09:05:43 +0000 (09:05 +0000)]
re PR sanitizer/80953 (Support libsanitizer on Solaris)
PR sanitizer/80953
Merge from LLVM revision 355965
* sanitizer_common/sanitizer_linux.cc (GetWriteFlag): Implement for
SPARC/Linux.
(GetPcSpBp): Likewise.
* sanitizer_common/sanitizer_stacktrace.cc (GetNextInstructionPc):
Adjust for SPARC.
* sanitizer_common/sanitizer_stacktrace.h (SANITIZER_CAN_FAST_UNWIND):
Define to 1 for SPARC.
* sanitizer_common/sanitizer_stacktrace_sparc.cc: Rewrite.
* sanitizer_common/sanitizer_unwind_linux_libcdep.cc (SlowUnwindStack):
Adjust the PC address for SPARC with GCC.
Thomas Koenig [Wed, 13 Mar 2019 07:21:33 +0000 (07:21 +0000)]
[multiple changes]
2019-03-13 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/66695
PR fortran/77746
PR fortran/79485
* gfortran.h (gfc_symbol): Add bind_c component.
(gfc_get_gsymbol): Add argument bind_c.
* decl.c (add_global_entry): Add bind_c argument to
gfc_get_symbol.
* parse.c (parse_block_data): Likewise.
(parse_module): Likewise.
(add_global_procedure): Likewise.
(add_global_program): Likewise.
* resolve.c (resolve_common_blocks): Likewise.
(resolve_global_procedure): Likewise.
(gfc_verify_binding_labels): Likewise.
* symbol.c (gfc_get_gsymbol): Add argument bind_c. Set bind_c
in gsym.
* trans-decl.c (gfc_get_module_backend_decl): Add bind_c argument
to gfc_get_symbol.
(gfc_get_extern_function_decl): If the sym has a binding label
and it cannot be found in the global symbol tabel, it is the wrong
one and vice versa.
2019-03-13 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/66695
PR fortran/77746
PR fortran/79485
* gfortran.dg/binding_label_tests_30.f90: New test.
* gfortran.dg/binding_label_tests_31.f90: New test.
* gfortran.dg/binding_label_tests_32.f90: New test.
* gfortran.dg/binding_label_tests_33.f90: New test.
Ian Lance Taylor [Wed, 13 Mar 2019 05:12:05 +0000 (05:12 +0000)]
compiler: compare parse methods when indexing interface types for export
This change fixes a bug in which two interface types were being
incorrectly commoned (considered identical) in the initial stages of
writing out types to export data. The indexer does a walk to collect
candidates for export, inserting types into a table to eliminate
duplicates; as part of this process a local interface type T1 was
being commoned with a different interface type T2. This caused a cycle
in the exported type graph due to the way embedded interfaces are
handled.
The fix was to add a new flag to the Type::is_identical utility
routine to request that interface type comparison be done by examining
the original parse methods, as opposed to the expanded method set,
then use the new flag when creating the hash map for the exporter.
Ian Lance Taylor [Wed, 13 Mar 2019 04:43:33 +0000 (04:43 +0000)]
compiler: add new debugging helper function debug_go_type()
Add a new debugging utility routine debug_go_type(), intended to
display the contents of a Type object in a way useful to debugging
a run of the compiler. Prior to this the only useful alternative
for debugging types was invoking the mangled_name() method, which
has problems (for example, won't work on interface types prior
to finalizing of methods).
This is a "deep" dump, meaning that all types reachable from the
type passed to debug_go_type() will be printed out. Example:
Robin Dapp [Tue, 12 Mar 2019 17:12:56 +0000 (17:12 +0000)]
S/390: Perform more aggressive inlining
This patch sets the inlining parameters for z13 and later to rather
aggressive values in response to PR85103 that caused performance
regressions in SPEC2006's sjeng and gobmk benchmarks.
Robin Dapp [Tue, 12 Mar 2019 17:08:27 +0000 (17:08 +0000)]
S/390: Rework instruction scheduling.
This patch set adds new pipeline descriptions for z13 and z14. Based
on that, the scoring and some properties are handled differently in
the scheduler hooks.
The patch before (r269620) also belongs to this set but I accidentally
applied it separately.
Iain Buclaw [Tue, 12 Mar 2019 16:31:48 +0000 (16:31 +0000)]
Use libiberty's lrealpath to avoid portability problems.
The dmd front-end function FileName::canonicalName could be called
during the semantic pass of import("file") expressions, so still
requires that realpath() be redefined.
Initial patch by Rainer Orth.
gcc/d/ChangeLog:
2019-03-12 Iain Buclaw <ibuclaw@gdcproject.org>
PR d/87866
* d-system.h (realpath): Redefine as lrealpath.
Paul Thomas [Tue, 12 Mar 2019 13:40:51 +0000 (13:40 +0000)]
re PR fortran/89363 (RANK incorrect for unallocated allocatable)
2019-03-12 Paul Thomas <pault@gcc.gnu.org>
PR fortran/89363
PR fortran/89364
* trans-expr.c (set_dtype_for_unallocated): New function.
(gfc_conv_gfc_desc_to_cfi_desc): Call it for allocatable and
pointer arguments.
(gfc_conv_procedure_call): Likewise. Also, set the ubound of
the final dimension to -1 for assumed rank formal args that are
associated with assumed size arrays.
* trans-intrinsic.c (gfc_conv_intrinsic_bound): Return -1 for
the final dimension of assumed rank entities that are argument
associated with assumed size arrays.
(gfc_conv_intrinsic_shape): Likewise return -1 for the final
dimension of the shape intrinsic.
2019-03-12 Paul Thomas <pault@gcc.gnu.org>
PR fortran/89363
* gfortran.dg/assumed_rank_16.f90: New test.
PR fortran/89364
* gfortran.dg/assumed_rank_17.f90: New test.
Fixes C++ mangling for substituted basic types that are target-specific.
Introduces a new method that currently does nothing, but could in future
make use of flag_abi_version as extern(C++) integration improves in
latter versions of the D front-end.
projects / thirdparty / gcc.git / log
