Wilco Dijkstra [Tue, 5 Mar 2019 15:04:01 +0000 (15:04 +0000)]
[ARM] Fix PR89222
The GCC optimizer can generate symbols with non-zero offset from simple
if-statements. Bit zero is used for the Arm/Thumb state bit, so relocations
with offsets fail if it changes bit zero and the relocation forces bit zero
to true. The fix is to disable offsets on function pointer symbols.
gcc/
PR target/89222
* config/arm/arm.md (movsi): Use targetm.cannot_force_const_mem
to decide when to split off a non-zero offset from a symbol.
* config/arm/arm.c (arm_cannot_force_const_mem): Disallow offsets
in function symbols.
testsuite/
PR target/89222
* gcc.target/arm/pr89222.c: Add new test.
Jakub Jelinek [Tue, 5 Mar 2019 13:38:59 +0000 (14:38 +0100)]
re PR tree-optimization/89487 (ICE in expand_expr_addr_expr_1, at expr.c:7993)
PR tree-optimization/89487
* gcc.dg/tree-ssa/pr89487.c: Include ../pr87600.h.
(caml_interprete): Ifdef the whole body out if REG1 or REG2 macros
aren't defined. Use REG1 instead of "%r15" and REG2 instead of
"%r14".
Jakub Jelinek [Tue, 5 Mar 2019 08:43:16 +0000 (09:43 +0100)]
re PR tree-optimization/89566 (ICE on compilable C++ code: in gimple_call_arg, at gimple.h:3166)
PR tree-optimization/89566
* gimple-ssa-sprintf.c (sprintf_dom_walker::handle_gimple_call):
Set info.fncode to BUILT_IN_NONE if gimple_call_builtin_p failed.
Punt if get_user_idx_format succeeds, but idx_format argument is
not provided or doesn't have pointer type, or if idx_args is above
number of provided arguments.
2019-03-04 Edward Smith-Rowland <3dw4rd@verizon.net>
PR libstdc++/88996 Implement P0439R0
Make std::memory_order a scoped enumeration.
* include/bits/atomic_base.h: For C++20 make memory_order a scoped enum,
add variables for the old enumerators. Adjust calls.
* testsuite/29_atomics/headers/atomic/types_std_c++2a.cc: New test.
* testsuite/29_atomics/headers/atomic/types_std_c++2a_neg.cc: New test.
Jonathan Wakely [Mon, 4 Mar 2019 12:21:06 +0000 (12:21 +0000)]
Implement polymorphic_allocator<byte> for C++20 (P0339R6)
* include/std/memory_resource (polymorphic_allocator): Add default
template argument for C++20.
(polymorphic_allocator::allocate_bytes)
(polymorphic_allocator::deallocate_bytes)
(polymorphic_allocator::allocate_object)
(polymorphic_allocator::deallocate_object)
(polymorphic_allocator::new_object)
(polymorphic_allocator::delete_object): New member functions for
C++20.
* testsuite/20_util/polymorphic_allocator/allocate_object.cc: New
test.
Bin Cheng [Mon, 4 Mar 2019 01:38:25 +0000 (01:38 +0000)]
re PR tree-optimization/89487 (ICE in expand_expr_addr_expr_1, at expr.c:7993)
PR tree-optimization/89487
* tree-loop-distribution.c (has_nonaddressable_dataref_p): New.
(create_rdg_vertices): Compute has_nonaddressable_dataref_p.
(distribute_loop): Don't do runtime alias check if there is non-
addressable data reference.
* tree-ssa-loop-ivopts.c (may_be_nonaddressable_p): Check if VAR_DECL
is a register variable.
* gcc/testsuite/gcc.dg/tree-ssa/pr89487.c: New test.
Jakub Jelinek [Sat, 2 Mar 2019 08:06:57 +0000 (09:06 +0100)]
re PR c++/71446 (Incorrect overload resolution when using designated initializers)
PR c++/71446
* cp-tree.h (CONSTRUCTOR_IS_DESIGNATED_INIT): Define.
* parser.c (cp_parser_braced_list): Adjust cp_parser_initializer_list
caller, set CONSTRUCTOR_IS_DESIGNATED_INIT.
(cp_parser_initializer_list): Add designated parameter, set *designated
to a bool whether any designators were parsed.
* decl.c (reshape_init): Copy over CONSTRUCTOR_IS_DESIGNATED_INIT if
needed.
* pt.c (tsubst_copy_and_build): Likewise.
* call.c (implicit_conversion): If CONSTRUCTOR_IS_DESIGNATED_INIT,
don't call build_list_conv, nor build_complex_conv, nor attempt to
convert a single element initializer to scalar.
* g++.dg/cpp2a/desig10.C: New test.
* g++.dg/cpp2a/desig11.C: New test.
* g++.dg/ext/desig4.C: Expect 4 new errors.
Jakub Jelinek [Sat, 2 Mar 2019 08:05:10 +0000 (09:05 +0100)]
re PR target/89506 (ICE: in decompose, at rtl.h:2266 with -Og -g)
PR target/89506
* config/arm/arm.md (cmpsi2_addneg): Use
trunc_int_for_mode (-INTVAL (...), SImode) instead of -INTVAL (...).
If operands[2] is 0 or INT_MIN, force use of subs.
(*compare_scc splitter): Use gen_int_mode.
(*negscc): Likewise.
* config/arm/thumb2.md (*thumb2_negscc): Likewise.
This is the result of these two oneliners:
perl -ni -e 'print unless /dg-skip-if "do not override -mcpu"/' \
$(find gcc/testsuite/gcc.target/powerpc/ -type f)
perl -pi -e 's/(dg-options.*)-mcpu=/\1-mdejagnu-cpu=/' \
$(find gcc/testsuite/gcc.target/powerpc/ -type f)
gcc/testsuite/
* gcc.target/powerpc/ throughout: Delete dg-skip-if "do not override
-mcpu". Use -mdejagnu-cpu= in dg-options instead of -mcpu=.
This adds an option -mdejagnu-cpu=. This option simply overrides what
is given in -mcpu=. The reason for this is that with older versions
of DejaGnu the value given in the RUNTESTFLAGS will override the value
a testcase wants to have.
* config/rs6000/rs6000.c (rs6000_option_override_internal): If
rs6000_dejagnu_cpu_index is set, use that to override rs6000_cpu_index.
* config/rs6000/rs6000.opt (mdejagnu-cpu=): New option.
Jakub Jelinek [Fri, 1 Mar 2019 14:20:03 +0000 (15:20 +0100)]
Implement P1002R1, Try-catch blocks in constexpr functions PR c++/89513
Implement P1002R1, Try-catch blocks in constexpr functions
PR c++/89513
* parser.c (cp_parser_ctor_initializer_opt_and_function_body):
Diagnose constexpr ctor or function with function-try-block with
pedwarn for c++17 and earlier. Formatting fix.
(cp_parser_try_block): Use pedwarn instead of error and only for
c++17 and earlier when try block appears in constexpr function.
* constexpr.c (build_constexpr_constructor_member_initializers):
Handle TRY_BLOCK here instead of erroring on it.
* g++.dg/cpp2a/constexpr-try1.C: New test.
* g++.dg/cpp2a/constexpr-try2.C: New test.
* g++.dg/cpp2a/constexpr-try3.C: New test.
* g++.dg/cpp2a/constexpr-try4.C: New test.
* g++.dg/cpp2a/constexpr-try5.C: New test.
* g++.dg/cpp0x/constexpr-ctor10.C: Don't expect error for C++2a.
Jonathan Wakely [Fri, 1 Mar 2019 13:50:36 +0000 (13:50 +0000)]
C++2a Utility functions to implement uses-allocator construction (P0591R4)
* include/std/memory (uses_allocator_construction_args): New set of
overloaded functions.
(make_obj_using_allocator, uninitialized_construct_using_allocator):
New functions.
* include/std/memory_resource (polymorphic_allocator::construct)
[__cplusplus > 201703l]: Replace all overloads with a single function
using uses_allocator_construction_args.
* testsuite/20_util/polymorphic_allocator/construct_c++2a.cc: New
test.
* testsuite/20_util/uses_allocator/make_obj.cc: New test.
Tamar Christina [Fri, 1 Mar 2019 13:34:14 +0000 (13:34 +0000)]
AArch64: Make every option in options.def one line
Due to config.gcc all the options need to be on one line because of the grep
lines which would select only the first line of the option.
This causes it not to select the right bits on options that are spread over
multiple lines when the --with-arch configure option is used. The issue happens
silently and you just get a compiler with an incorrect set of default flags.
This solution just collapses everything back to one line as they were in GCC7.
Unfortunately this does make some lines quite long.
I do have an alternate patch which used the pre-processors to first flatten the
file in config.gcc. I will send that one out for GCC 10.
Fix mask type choice in vectorizable_call (PR 89535)
This is another case in which we were failing to pass the expected
mask vector type to vect_get_vec_def_for_operand.
2019-02-28 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR tree-optimization/89535
* tree-vect-stmts.c (vectorizable_call): Record the vector types
for each operand. Calculate the fallback choice for mask operands
and pass it to vect_get_vec_def_for_operand.
gcc/testsuite/
PR tree-optimization/89535
* gfortran.dg/vect/pr89535.f90: New test.
Richard Biener [Fri, 1 Mar 2019 12:52:19 +0000 (12:52 +0000)]
re PR tree-optimization/89541 (ICE in VN_INFO(tree_node*))
2019-03-01 Richard Biener <rguenther@suse.de>
PR middle-end/89541
* tree-ssa-operands.c (add_stmt_operand): CONST_DECL may
get virtual operands.
(get_expr_operands): Handle CONST_DECL like other decls.
Backports support for extern(C++, "namespace"), which makes the
core.stdcpp package compilable.
Added predefined condition for CppRuntime_Gcc unconditionally, as it is
unlikely that D code will be linking to anything other than libstdc++
when extern(C++) is used.
Richard Biener [Fri, 1 Mar 2019 09:21:30 +0000 (09:21 +0000)]
re PR middle-end/89497 (ICE caused by Segmentation Fault when compiling cups 2.2.10 with LTO flags enabled)
2019-03-01 Richard Biener <rguenther@suse.de>
PR middle-end/89497
* tree-cfgcleanup.h (cleanup_tree_cfg): Add SSA update flags
argument, defaulted to zero.
* passes.c (execute_function_todo): Pass down SSA update flags
to cleanup_tree_cfg.
* tree-cfgcleanup.c: Include tree-into-ssa.h and tree-cfgcleanup.h.
(cleanup_tree_cfg_noloop): After cleanup_control_flow_pre update SSA
form if requested.
(cleanup_tree_cfg): Get and pass down SSA update flags.
* gcc.dg/tree-ssa/reassoc-43.c: Avoid false match in regex.
* g++.dg/tree-prof/devirt.C: Scan tracer dump for foldings
that happen now earlier.
When using the go tool with gccgo, this changes the default
compilation to use -O2. The -gccgoflags option can be used to
override this default. I think this change better corresponds to what
people expect when using the go tool.
Jason Merrill [Fri, 1 Mar 2019 00:08:58 +0000 (19:08 -0500)]
PR c++/88183 - ICE with .* fold-expression.
build_m_component_ref can't handle type-dependent operands, so let's use the
default case; tsubst_copy_and_build also uses build_x_binary_op for
substituting a DOTSTAR_EXPR.
* pt.c (fold_expression) [DOTSTAR_EXPR]: Remove special handling.
Eric Botcazou [Thu, 28 Feb 2019 23:03:01 +0000 (23:03 +0000)]
re PR tree-optimization/89536 (wrong code at -Os and above on x86_64-linux-gnu)
PR tree-optimization/89536
* tree-ssa-dom.c (edge_info::derive_equivalences) <BIT_NOT_EXPR>: Test
only whether bit #0 of the value is 0 instead of the entire value.
Marek Polacek [Thu, 28 Feb 2019 22:29:42 +0000 (22:29 +0000)]
PR c++/87068 - missing diagnostic with fallthrough statement.
* gimplify.c (expand_FALLTHROUGH_r): If IFN_FALLTHROUGH was found
at the end of a seq, save its location to walk_stmt_info.
(expand_FALLTHROUGH): Warn if IFN_FALLTHROUGH is at the end of
a switch.
* c-c++-common/Wimplicit-fallthrough-37.c: New test.
Jason Merrill [Thu, 28 Feb 2019 17:29:48 +0000 (12:29 -0500)]
PR c++/88049 - ICE with undefined destructor and anon namespace.
A type in an anonymous namespace can never be merged with one from
another translation unit, so a member of such a type is always its own
prevailing decl.
* lto-symtab.c (lto_symtab_prevailing_virtual_decl): Return early
for a type in an anonymous namespace.
Tamar Christina [Thu, 28 Feb 2019 10:43:41 +0000 (10:43 +0000)]
AArch64: Have empty HWCAPs string ignored during native feature detection
This patch makes the feature detection code for AArch64 GCC not add features
automatically when the feature had no hwcaps string to match against.
This means that -mcpu=native no longer adds feature flags such as +profile.
The behavior wasn't noticed before because at the time +profile was added a bug
was preventing any feature bits from being added by native detections.
The loop has also been changed as Jakub specified in order to avoid a memory
leak that was present in the existing code and to be slightly more efficient.
gcc/ChangeLog:
PR target/88530
* config/aarch64/aarch64-option-extensions.def: Document it.
* config/aarch64/driver-aarch64.c (host_detect_local_cpu): Skip feature
if empty hwcaps.
gcc/testsuite/ChangeLog:
PR target/88530
* gcc.target/aarch64/options_set_10.c: New test.
Jakub Jelinek [Thu, 28 Feb 2019 07:16:38 +0000 (08:16 +0100)]
re PR c/89520 (ICE tree check: accessed operand 4 of call_expr with 3 operands in convert_to_integer_1, at convert.c:668)
PR c/89520
* convert.c (convert_to_real_1, convert_to_integer_1): Punt for
builtins if they don't have a single scalar floating point argument.
Formatting fixes.
* gcc.dg/pr89520-1.c: New test.
* gcc.dg/pr89520-2.c: New test.
Ian Lance Taylor [Wed, 27 Feb 2019 22:35:10 +0000 (22:35 +0000)]
re PR go/89172 (FAIL: runtime/pprof)
PR go/89172
internal/cpu, runtime, runtime/pprof: handle function descriptors
When using PPC64 ELF ABI v1 a function address is not a PC, but is the
address of a function descriptor. The first field in the function
descriptor is the actual PC (see
http://refspecs.linuxfoundation.org/ELF/ppc64/PPC-elf64abi.html#FUNC-DES).
The libbacktrace library knows about this, and libgo uses actual PC
values consistently except for the helper function funcPC that appears
in both runtime and runtime/pprof.
This patch fixes funcPC by recording, in the internal/cpu package,
whether function descriptors are being used. We have to check for
function descriptors using a C compiler check, because GCC can be
configured using --with-abi to select the ELF ABI to use.
Jason Merrill [Wed, 27 Feb 2019 21:54:25 +0000 (16:54 -0500)]
PR c++/86969 - ICE with constexpr if and recursive generic lambdas.
Here, the problem was that extract_local_specs wasn't seeing that we use
'self' inside the lambda in the else of the inner constexpr if, because we
don't walk into lambda bodies and we didn't capture it in the lambda because
'self' is still dependent. Marek recently changed process_outer_var_ref to
do more implicit capture in templates; this example shows that we should
always capture non-packs, so that we can continue to not walk into lambda
bodies. We do walk into lambda bodies for pack expansions, so we can delay
deciding whether we're capturing a single element or the entire pack.
Immediately capturing a VLA means we need to create a dependent VLA capture
type, and not in the context of the lambda op(), since trying to look up the
instantiation of the op() while we're substituting into the capture list
would crash. So I force TYPE_CONTEXT and the binding level out to the
enclosing function before pushtag, avoid adding a TAG_DEFN, and instead
force the type to be complete in tsubst_lambda_expr.
* semantics.c (process_outer_var_ref): Do capture dependent vars.
* class.c (finish_struct): Only add TAG_DEFN if T is in
current_function_decl.
* lambda.c (vla_capture_type): Force the capture type out into the
lambda's enclosing function.
(add_capture): Pass in the lambda.
* pt.c (tsubst_lambda_expr): complete_type a VLA capture type.
PR rtl-optimization/89490
* varasm.c (get_block_for_section): Bail out for mergeable sections.
(default_use_anchors_for_symbol_p, output_object_block): Assert the
block section is not mergeable.
Jakub Jelinek [Wed, 27 Feb 2019 14:50:35 +0000 (15:50 +0100)]
re PR target/70341 (cost model for addresses is incorrect, slsr is using reg + reg + CST for arm)
PR target/70341
* config/arm/arm.md (arm_casesi_internal): New define_expand. Rename
old define_insn to ...
(*arm_casesi_internal): ... this. Add mode to LABEL_REFs.
* config/arm/thumb2.md (thumb2_casesi_internal): New define_expand.
Rename old define_insn to ...
(*thumb2_casesi_internal): ... this. Add mode to LABEL_REFs.
(thumb2_casesi_internal_pic): New define_expand. Rename old
define_insn to ...
(*thumb2_casesi_internal_pic): ... this. Add mode to LABEL_REFs.
Richard Biener [Wed, 27 Feb 2019 12:51:43 +0000 (12:51 +0000)]
re PR debug/88878 (.debug_pubnames/types empty with -flto)
2019-02-27 Richard Biener <rguenther@suse.de>
PR debug/88878
* dwarf2out.c (output_comdat_type_unit): Add early_lto_debug
parameter, prefix section name with .gnu.debuglto_ if true.
(dwarf2out_finish): Pass false to output_comdat_type_unit.
(dwarf2out_early_finish): Pass true to output_comdat_type_unit.
Richard Biener [Wed, 27 Feb 2019 11:59:07 +0000 (11:59 +0000)]
re PR lto/89514 (-g -fdebug-types-section -flto gives 'Dwarf Error: bad length' in gdb)
2019-02-27 Richard Biener <rguenther@suse.de>
PR debug/89514
* dwarf2out.c (size_of_die): Key on AT_ref (a)->comdat_type_p
rather than on use_debug_types, doing what output_die does.
(value_format): Likewise.
Jonathan Wakely [Wed, 27 Feb 2019 11:25:44 +0000 (11:25 +0000)]
PR libstdc++/89466 avoid slow xsltproc command in configure
Certain broken versions of xsltproc ignore the --nonet option and will
attempt to fetch the docbook stylesheet from the WWW when it isn't in
the local XML catalog.
This patch checks for the local stylesheet directory first, and doesn't
use xsltproc if no local stylesheets are found. Checking for the local
directory is done using xmlcatalog if available, only checking the
hardcoded list of directories if xmlcatalog fails. The right directory
for Suse is added to the hardcoded list.
This should avoid doing an xsltproc check that would need to download
the stylesheet, so no network connection is made even if a broken
xsltproc is present.
PR libstdc++/89466
* acinclude.m4 (GLIBCXX_CONFIGURE_DOCBOOK): Reorder check for local
stylesheet directories before check for xsltproc. Try to use
xmlcatalog to find local stylesheet directory before trying hardcoded
paths. Add path used by suse to hardcoded paths. Adjust xsltproc
check to look for the same stylesheet as doc/Makefile.am uses. Don't
use xsltproc if xmlcatalog fails to find a local stylesheet.
* configure.ac: Check for xmlcatalog.
* Makefile.in: Regenerate.
* configure: Likewise.
* doc/Makefile.in: Likewise.
* include/Makefile.in: Likewise.
* libsupc++/Makefile.in: Likewise.
* po/Makefile.in: Likewise.
* python/Makefile.in: Likewise.
* src/Makefile.in: Likewise.
* src/c++11/Makefile.in: Likewise.
* src/c++17/Makefile.in: Likewise.
* src/c++98/Makefile.in: Likewise.
* src/filesystem/Makefile.in: Likewise.
* testsuite/Makefile.in: Likewise.
Jakub Jelinek [Wed, 27 Feb 2019 08:41:01 +0000 (09:41 +0100)]
re PR tree-optimization/89280 (ICE: Segmentation fault (in is_gimple_reg_type))
PR tree-optimization/89280
* tree-cfgcleanup.c (maybe_dead_abnormal_edge_p,
builtin_setjmp_setup_bb): New functions.
(cleanup_control_flow_pre): Ignore maybe_dead_abnormal_edge_p edges.
When visiting __builtin_setjmp_setup block, queue in special
setjmp_vec vector edges from .ABNORMAL_DISPATCHER to corresponding
__builtin_setjmp_receiver. Remove .ABNORMAL_DISPATCHER basic blocks
from visited after the loop if they don't have any visited successor
blocks.
* gcc.c-torture/compile/pr89280.c: New test.
* gcc.dg/torture/pr57147-2.c: Don't expect a setjmp after noreturn
function. Skip the test for -O0.