The immediate cause of the problem is that the hash function usually
returns different hashes for these types, yet the equality function
thinks they are equal. This then raises the question of how the types
are supposed to be treated.
I think the answer is that the GNU vector type should be treated as
distinct from float32x4_t, not least because the two types mangle
differently. However, each type should implicitly convert to the other.
This would mean that, as far as the PR is concerned, the hashing
function is right to (sometimes) treat the types differently and
the equality function is wrong to treat them as the same.
The most obvious way to enforce the type difference is to use a
target-specific type attribute. That on its own is enough to fix
the PR. The difficulty is deciding whether the knock-on effects
are acceptable.
on the basis that the types are no longer reference-compatible.
I think that's again the correct behaviour, and consistent with
current Clang.
A trickier question is whether:
vecf x;
float32x4_t y;
… c ? x : y …
should be valid, and if so, what its type should be [PR92789].
As explained in the comment in the testcase, GCC and Clang both
accepted this, but GCC chose the “then” type while Clang chose
the “else” type. This can lead to different mangling for (probably
artificial) corner cases, as seen for “sel1” and “sel2” in the
testcase.
Adding the attribute makes GCC reject the conditional expression
as ambiguous. I think that too is the correct behaviour, for the
reasons described in the testcase. However, it does seem to have
the potential to break existing code.
It looks like aarch64_comp_type_attributes is missing cases for
the SVE attributes, but I'll handle that in a separate patch.
2020-06-30 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR target/92789
PR target/95726
* config/aarch64/aarch64.c (aarch64_attribute_table): Add
"Advanced SIMD type".
(aarch64_comp_type_attributes): Check that the "Advanced SIMD type"
attributes are equal.
* config/aarch64/aarch64-builtins.c: Include stringpool.h and
attribs.h.
(aarch64_mangle_builtin_vector_type): Use the mangling recorded
in the "Advanced SIMD type" attribute.
(aarch64_init_simd_builtin_types): Add an "Advanced SIMD type"
attribute to each Advanced SIMD type, using the mangled type
as the attribute's single argument.
gcc/testsuite/
PR target/92789
PR target/95726
* g++.target/aarch64/pr95726.C: New test.
Nathan Sidwell [Tue, 30 Jun 2020 17:43:08 +0000 (10:43 -0700)]
c++: Tweak function cloning names
On the modules branch I need to expose an intermediate step of the
function cloning, but before that it'd be nice to rationalize the
names somewhat, now that we also use that API for copying the equality
operator. Jason's recent patch caused me some pain by altering the
same code. I can only blame myself for not pushing some bits sooner.
Anyway, this patch makes the newly renamed copy_fndecl_with_name
static, and adds a wrapper copy_operator_fn, that takes an operator
code. The cdtor cloning functions are renamed to explicitly note they
expect cdtors. A followup patch will move some of the logic from
copy_fndecl_with_name to build_cdtor_clones.
This expects to scan the PTX offloading compilation assembler code (not host
code!), expecting that nvptx offloading code assembly is produced after the
host code, and thus overwrites the latter file. (Yes, that's certainly
ugly/fragile...)
Christophe Lyon [Tue, 30 Jun 2020 13:51:20 +0000 (13:51 +0000)]
arm: Warn if IRQ handler is not compiled with -mgeneral-regs-only [PR target/94743]
The interrupt attribute does not guarantee that the FP registers are
saved, which can result in problems difficult to debug.
Saving the FP registers and status registers can be a large penalty,
so it's probably not desirable to do that all the time.
If the handler calls other functions, we'd likely need to save all of
them, for lack of knowledge of which registers they actually clobber.
This is even more obscure for the end-user when the compiler inserts
calls to helper functions such as memcpy (some multilibs do use FP
registers to speed it up).
In the PR, we discussed adding routines in libgcc to save the FP
context and saving only locally-clobbered FP registers, but this seems
to be too much work for the purpose, given that in general such
handlers try to avoid this kind of penalty.
I suspect we would also want new attributes to instruct the compiler
that saving the FP context is not needed.
In the mean time, emit a warning to suggest re-compiling with
-mgeneral-regs-only. Note that this can lead to errors if the code
uses floating-point and -mfloat-abi=hard, eg:
argument of type 'double' not permitted with -mgeneral-regs-only
This can be troublesome for the user, but at least this would make
him aware of the latent issue.
The patch adds several testcases:
- pr94734-1-hard.c checks that a warning is emitted when using
-mfloat-abi=hard. Function IRQ_HDLR_Test can make implicit calls to
runtime floating-point routines (or direct use of FP instructions),
IRQ_HDLR_Test2 doesn't. We emit a warning in both cases, though.
- pr94734-1-softfp.c: same as above wih -mfloat-abi=softfp.
- pr94734-1-soft.c checks that no warning is emitted when using
-mfloat-abi=soft when the same code as above.
- pr94734-2.c checks that no warning is emitted when using
-mgeneral-regs-only.
- pr94734-3.c checks that no warning is emitted when using
-mgeneral-regs-only even using float-point data.
Yang Yang [Tue, 30 Jun 2020 10:04:31 +0000 (18:04 +0800)]
tree-optimization/95855 - Add checks to avoid spoiling if-conversion if
Add some checks in pass_splits_paths, so that pass_split_paths can
recognize the missed if-conversion opportunity and do not duplicate the
corresponding block.
2020-06-30 Yang Yang <yangyang305@huawei.com>
gcc/ChangeLog:
PR tree-optimization/95855
* gimple-ssa-split-paths.c (is_feasible_trace): Add extra
checks to recognize a missed if-conversion opportunity when
judging whether to duplicate a block.
gcc/testsuite/ChangeLog
PR tree-optimization/95855
* gcc.dg/tree-ssa/split-path-12.c: New testcase.
David Edelsohn [Tue, 30 Jun 2020 00:41:55 +0000 (20:41 -0400)]
testsuite: fix constexpr-is_literal.C deprecated c++17 and above.
std::is_literal_type is deprecated in C++17 and above.
This was introduced in c++11, but the testsuite list of standards now only
includes c++98 and c++14. This patch limits the test to c++14_only to
prevent deprecation warnings.
gcc/testsuite/ChangeLog
2020-06-30 David Edelsohn <dje.gcc@gmail.com>
* g++.dg/cpp0x/constexpr-is_literal.C: Limit test to -std=c++14.
Thomas Koenig [Tue, 30 Jun 2020 11:01:36 +0000 (13:01 +0200)]
Use CHARACTER(kind) string for calculating the type hash.
This regression came about because of a change in the way
types are displayed in error messages. The character
representation is also used to calculate the hashes for
our types, so this patch restores the old behavior if
we are indeed calculating a hash.
The test case also checks for the specific hash value because
changing that would be an ABI change, which we should not
be doing unintentionally.
gcc/fortran/ChangeLog:
2020-06-30 Thomas Koenig <tkoenig@gcc.gnu.org>
PR fortran/95355
* gfortran.h (gfc_typename): Add optional argument for_hash.
* misc.c (gfc_typename): When for_hash is true, just retur
CHARACTER(kind).
* class.c (gfc_intrinsic_hash_value): Call gfc_typename with
for_hash = true.
Jakub Jelinek [Tue, 30 Jun 2020 09:42:54 +0000 (11:42 +0200)]
c-family: Avoid ICEs on calls to internal functions [PR95963]
The following testcase ICEs since recent Martin's -Wnonnull changes,
we see a CALL_EXPR and ICE because CALL_EXPR_FN is NULL, which is
valid for internal function calls. Internal function calls don't have a
function type, and will never have format_arg attribute on them nor will
serve as the i18n routines -Wformat cares about.
2020-06-30 Jakub Jelinek <jakub@redhat.com>
PR c++/95963
* c-common.c (check_function_arguments_recurse): Don't crash on
calls to internal functions.
Iain Sandoe [Tue, 30 Jun 2020 06:26:51 +0000 (07:26 +0100)]
coroutines: Collect the function body rewrite code.
The standard describes a rewrite of the body of the user-authored
function (which wraps it in a try-catch block and provides the
initial and final suspend expressions). The exact arrangement of
this was still in flux right up until the DIS and as a consequence
was a bit of a moving target.
The net result was a fragmented implementation of the parts of
the rewrite which is now impeding progress in fixing other issues.
This patch collates the rewrite action into a single function and
carries this out earlier.
gcc/cp/ChangeLog:
* coroutines.cc (expand_one_await_expression): Remove
code dealing with initial suspend.
(build_actor_fn): Remove code special-casing initial
and final suspend. Handle the final suspend and marking
of the coroutine as done.
(coro_rewrite_function_body): New.
(bind_expr_find_in_subtree): Remove.
(coro_body_contains_bind_expr_p): Remove.
(morph_fn_to_coro): Split the rewrite of the original
function into coro_rewrite_function_body and call it.
Ian Lance Taylor [Mon, 29 Jun 2020 18:30:43 +0000 (11:30 -0700)]
compiler: remove some erroneous code that was never run
The code accidentally called Type::type_descriptor rather than the
do_type_descriptor method. Calling Type::type_descriptor with a second
argument of NULL would always crash. Since that never happened,
it revealed that this code was never actually executed.
Thomas Koenig [Mon, 29 Jun 2020 21:11:06 +0000 (23:11 +0200)]
Do not generate recursion check for compiler-generated procedures.
This one-line fix removes a check for recursion for procedures
which are compiler-generated, such as finalizers or deallocation.
These need to be recursive, even if the user code should not be.
gcc/fortran/ChangeLog:
PR fortran/95743
* trans-decl.c (gfc_generate_function_code): Do not generate
recursion check for compiler-generated procedures.
* doc/extend.texi: Change references to "future architecture" to
"ISA 3.1", "-mcpu=future" to "-mcpu=power10", and remove vaguer
references to "future" (because the future is now).
Roger Sayle [Mon, 29 Jun 2020 16:33:23 +0000 (17:33 +0100)]
middle-end: Optimize (A&C)^(B&C) to (A^B)&C in simplify_rtx (take 3).
2020-06-29 Roger Sayle <roger@nextmovesoftware.com>
gcc/ChangeLog:
* simplify-rtx.c (simplify_distributive_operation): New function
to un-distribute a binary operation of two binary operations.
(X & C) ^ (Y & C) to (X ^ Y) & C, when C is simple (i.e. a constant).
(simplify_binary_operation_1) <IOR, XOR, AND>: Call it from here
when appropriate.
(test_scalar_int_ops): New function for unit self-testing
scalar integer transformations in simplify-rtx.c.
(test_scalar_ops): Call test_scalar_int_ops for each integer mode.
(simplify_rtx_c_tests): Call test_scalar_ops.
Marek Polacek [Fri, 26 Jun 2020 16:40:59 +0000 (12:40 -0400)]
c++: Check uniqueness of concepts/variable templates [PR94553]
This patch wraps up PR94553. Variable template names have no C
compatibility implications so they should be unique in their
declarative region. It occurred to me that this applies to concepts
as well. This is not specified in [basic.scope.declarative]/4.2
but that seems like a bug in the standard.
I couldn't use variable_template_p because that uses PRIMARY_TEMPLATE_P
which uses DECL_PRIMARY_TEMPLATE and that might not have been set up yet
(push_template_decl hasn't yet been called). PRIMARY_TEMPLATE_P is
important to distinguish between a variable template and a variable in a
function template. But I think we don't have to worry about that in
duplicate_decls: a template declaration cannot appear at block scope,
and additional checks in duplicate_decls suggest that it won't ever
see a TEMPLATE_DECL for a variable in a function template. So
checking that the DECL_TEMPLATE_RESULT is a VAR_DECL seems to be fine.
I could have added a default argument to variable_template_p too to
avoid checking PRIMARY_TEMPLATE_P but it didn't seem worth the effort.
gcc/cp/ChangeLog:
PR c++/94553
* decl.c (duplicate_decls): Make sure a concept or a variable
template is unique in its declarative region.
gcc/testsuite/ChangeLog:
PR c++/94553
* g++.dg/cpp1y/pr68578.C: Adjust dg-error.
* g++.dg/cpp1y/var-templ66.C: New test.
* g++.dg/cpp2a/concepts-redecl1.C: New test.
Marek Polacek [Tue, 23 Jun 2020 22:07:34 +0000 (18:07 -0400)]
c++: Fix CTAD for aggregates in template [PR95568]
95568 complains that CTAD for aggregates doesn't work within
requires-clause and it turned out that it doesn't work when we try
the deduction in a template. The reason is that maybe_aggr_guide
creates a guide that can look like this
template<class T> X(decltype (X<T>::x))-> X<T>
where the parameter is a decltype, which is a non-deduced context. So
the subsequent build_new_function_call fails because unify_one_argument
can't deduce anything from it ([temp.deduct.type]: "If a template
parameter is used only in non-deduced contexts and is not explicitly
specified, template argument deduction fails.")
Those decltypes come from finish_decltype_type. We can just use
TREE_TYPE instead. I pondered using unlowered_expr_type, but that
didn't make any difference for the FIELD_DECLs I saw in
class-deduction-aggr6.C.
gcc/cp/ChangeLog:
PR c++/95568
* pt.c (collect_ctor_idx_types): Use TREE_TYPE.
gcc/testsuite/ChangeLog:
PR c++/95568
* g++.dg/cpp2a/class-deduction-aggr5.C: New test.
* g++.dg/cpp2a/class-deduction-aggr6.C: New test.
Andrew Stubbs [Fri, 26 Jun 2020 16:10:55 +0000 (17:10 +0100)]
amdgcn: Support basic DWARF
This is enough DWARF support for "-O0 -g" to work OK, within a single frame,
using the new rocgdb debugger from AMD. Debugging with optimization enabled
also works when the values are located in SGPRs or memory.
Scalars in VGPRs are problematic, EXEC_HI and VCC_HI are unmappable, and
CFI remains unimplemented.
gcc/ChangeLog:
* config/gcn/gcn-hsa.h (DBX_REGISTER_NUMBER): New macro.
* config/gcn/gcn-protos.h (gcn_dwarf_register_number): New prototype.
* config/gcn/gcn.c (gcn_expand_prologue): Add RTX_FRAME_RELATED_P
and REG_FRAME_RELATED_EXPR to stack and frame pointer adjustments.
(gcn_dwarf_register_number): New function.
(gcn_dwarf_register_span): New function.
(TARGET_DWARF_REGISTER_SPAN): New hook macro.
Kaipeng Zhou [Wed, 24 Jun 2020 14:48:45 +0000 (22:48 +0800)]
store-merging: ICE in find_bswap_or_nop_1 PR95854.
The patch add a judgement in find_bswap_or_nop_1 to make sure both
operand1 and operand2 cannot be converted to unsigned HOST_WIDE_INT.
If not, return NULL.
gcc/ChangeLog:
2020-06-24 Kaipeng Zhou <zhoukaipeng3@huawei.com>
PR tree-optimization/95854
* gimple-ssa-store-merging.c (find_bswap_or_nop_1): Return NULL
if operand 1 or 2 of a BIT_FIELD_REF cannot be converted to
unsigned HOST_WIDE_INT.
gcc/testsuite/ChangeLog:
2020-06-24 Kaipeng Zhou <zhoukaipeng3@huawei.com>
PR tree-optimization/95854
* gcc.dg/pr95854.c: New test.
Rainer Orth [Mon, 29 Jun 2020 09:51:07 +0000 (11:51 +0200)]
sparc: Remove register storage class in sparc.c
The switch to C++17 broke SPARC bootstrap:
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:8887:34: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
8887 | epilogue_renumber (register rtx *where, int test)
| ^~~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c: In function 'int epilogue_renumber(rtx_def**, int)':
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:8889:24: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
8889 | register const char *fmt;
| ^~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:8890:16: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
8890 | register int i;
| ^
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:8891:26: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
8891 | register enum rtx_code code;
| ^~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:8948:17: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
8948 | register int j;
| ^
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c: In function 'void sparc_print_operand_address(std::FILE*, machine_mode, rtx)':
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9671:16: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9671 | register rtx base, index = 0;
| ^~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9671:22: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9671 | register rtx base, index = 0;
| ^~~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9673:16: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9673 | register rtx addr = x;
| ^~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c: At global scope:
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9807:32: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9807 | sparc_type_code (register tree type)
| ^~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c: In function 'long unsigned int sparc_type_code(tree)':
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9809:26: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9809 | register unsigned long qualifiers = 0;
| ^~~~~~~~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:9810:21: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
9810 | register unsigned shift;
| ^~~~~
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c: In function 'int set_extends(rtx_insn*)':
/vol/gcc/src/hg/master/local/gcc/config/sparc/sparc.c:10306:16: error: ISO C++17 does not allow 'register' storage class specifier [-Werror=register]
10306 | register rtx pat = PATTERN (insn);
| ^~~
Fixed by removing the register keyword. Bootstrapped on
sparc-sun-solaris2.11.
The test commands for scanning optimization dump files
perform globbing on the argument that specifies the suffix
of the dump files to be scanned. This behavior is currently
undocumented. Furthermore, the current implementation of
"scan-dump" and similar procedures yields an error whenever
the file name globbing matches more than one file (due to an
attempt to call "open" on multiple files) while a failure to
match any file results in an unresolved test.
This commit documents the globbing behavior. The dump
scanning procedures are changed to make the test unresolved
if globbing matches more than one file.
gcc/ChangeLog:
2020-06-29 Frederik Harwath <frederik@codesourcery.com>
* doc/sourcebuild.texi: Describe globbing of the
dump file scanning commands "suffix" argument.
gcc/testsuite/ChangeLog:
2020-06-29 Frederik Harwath <frederik@codesourcery.com>
* lib/scandump.exp (glob-dump-file): New proc.
(scan-dump): Use glob-dump-file for file name expansion.
(scan-dump-times): Likewise.
(scan-dump-dem): Likewise.
(scan-dump-dem-not): Likewise.
Reviewed-by: Thomas Schwinge <thomas@codesourcery.com>
Martin Sebor [Sun, 28 Jun 2020 20:22:14 +0000 (14:22 -0600)]
Underline argument in -Wnonnull and in C++ extend warning to the this pointer [PR c++/86568].
Resolves:
PR c++/86568 - -Wnonnull warnings should highlight the relevant argument not the closing parenthesis
gcc/c-family/ChangeLog:
PR c++/86568
* c-common.c (struct nonnull_arg_ctx): Add members.
(check_function_nonnull): Use nonnull_arg_ctx as argument. Handle
C++ member functions specially. Consider the this pointer implicitly
nonnull.
(check_nonnull_arg): Use location of argument when available.
(check_function_arguments): Use nonnull_arg_ctx as argument.
gcc/ChangeLog:
PR c++/86568
* calls.c (maybe_warn_rdwr_sizes): Use location of argument if
available.
* tree-ssa-ccp.c (pass_post_ipa_warn::execute): Same. Adjust
indentation.
* tree.c (get_nonnull_args): Consider the this pointer implicitly
nonnull.
* var-tracking.c (deps_vec): New type.
(var_loc_dep_vec): New function.
(VAR_LOC_DEP_VEC): Use it.
Kewen Lin [Sun, 28 Jun 2020 02:17:04 +0000 (21:17 -0500)]
IFN: Fix mask_{load,store} optab support macros
When I am working on IFNs for vector with length, I noticed that the
current optab support query for mask_load/mask_store looks unexpected.
The mask_load/mask_store requires two modes for convert_optab query,
but the macros direct_mask_{load,store}_optab_supported_p uses
direct_optab_supported_p which asserts type pair should have the same mode.
I'm not sure whether we have some special reason here or just a typo,
since everything goes well now, mask_{load,store} optab check is mainly
handled by can_vec_mask_load_store_p.
But if we have some codes as below (eg: one checking for all IFNs finally)
* internal-fn.c (direct_mask_load_optab_supported_p): Use
convert_optab_supported_p instead of direct_optab_supported_p.
(direct_mask_store_optab_supported_p): Likewise.
Aldy Hernandez [Thu, 18 Jun 2020 08:37:14 +0000 (10:37 +0200)]
Move simplification of statements using ranges into its own class.
This moves all the simplification code from vr_values into a separate
class (simplify_using_ranges). In doing so, we get rid of a bunch of
dependencies on the internals of vr_values. The goal is to (a) remove
unnecessary interdependendcies (b) be able to use this engine with any
range infrastructure, as all it needs is a method to get the range for
an SSA name (get_value_range).
I also removed as many dependencies on value_range_equiv as possible,
preferring value_range. A few value_range_equiv uses remain, but for
cases where equivalences are actually used (folding conditionals, etc).
gcc/ChangeLog:
* gimple-ssa-evrp-analyze.h (vrp_visit_cond_stmt): Use
simplify_using_ranges class.
* gimple-ssa-evrp.c (class evrp_folder): New simplify_using_ranges
field. Adjust all methods to use new field.
* tree-ssa-dom.c (simplify_stmt_for_jump_threading): Use
simplify_using_ranges class.
* tree-vrp.c (class vrp_folder): New simplify_using_ranges
field. Adjust all methods to use new field.
(simplify_stmt_for_jump_threading): Use simplify_using_ranges class.
(vrp_prop::vrp_finalize): New vrp_folder argument.
(execute_vrp): Pass folder to vrp_finalize. Use
simplify_using_ranges class.
Remove cleanup_edges_and_switches call.
* vr-values.c (vr_values::op_with_boolean_value_range_p): Change
value_range_equiv uses to value_range.
(simplify_using_ranges::op_with_boolean_value_range_p): Use
simplify_using_ranges class.
(check_for_binary_op_overflow): Make static.
(vr_values::extract_range_basic): Pass this to
check_for_binary_op_overflow.
(compare_range_with_value): Change value_range_equiv uses to
value_range.
(vr_values::vr_values): Initialize simplifier field.
Remove uses of to_remove_edges and to_update_switch_stmts.
(vr_values::~vr_values): Remove uses of to_remove_edges and
to_update_switch_stmts.
(vr_values::get_vr_for_comparison): Move to simplify_using_ranges
class.
(vr_values::compare_name_with_value): Same.
(vr_values::compare_names): Same.
(vr_values::vrp_evaluate_conditional_warnv_with_ops): Same.
(vr_values::vrp_evaluate_conditional): Same.
(vr_values::vrp_visit_cond_stmt): Same.
(find_case_label_ranges): Change value_range_equiv uses to
value_range.
(vr_values::extract_range_from_stmt): Use simplify_using_ranges class.
(vr_values::simplify_truth_ops_using_ranges): Move to
simplify_using_ranges class.
(vr_values::simplify_div_or_mod_using_ranges): Same.
(vr_values::simplify_min_or_max_using_ranges): Same.
(vr_values::simplify_abs_using_ranges): Same.
(vr_values::simplify_bit_ops_using_ranges): Same.
(test_for_singularity): Change value_range_equiv uses to
value_range.
(range_fits_type_p): Same.
(vr_values::simplify_cond_using_ranges_1): Same.
(vr_values::simplify_cond_using_ranges_2): Make extern.
(vr_values::fold_cond): Move to simplify_using_ranges class.
(vr_values::simplify_switch_using_ranges): Same.
(vr_values::cleanup_edges_and_switches): Same.
(vr_values::simplify_float_conversion_using_ranges): Same.
(vr_values::simplify_internal_call_using_ranges): Same.
(vr_values::two_valued_val_range_p): Same.
(vr_values::simplify_stmt_using_ranges): Move to...
(simplify_using_ranges::simplify): ...here.
* vr-values.h (class vr_values): Move all the simplification of
statements using ranges methods and code from here...
(class simplify_using_ranges): ...to here.
(simplify_cond_using_ranges_2): New extern prototype.
Jakub Jelinek [Sat, 27 Jun 2020 10:43:36 +0000 (12:43 +0200)]
openmp: Non-rectangular loop support for non-composite worksharing loops and distribute
This implements the fallback mentioned in
https://gcc.gnu.org/pipermail/gcc/2020-June/232874.html
Special cases for triangular loops etc. to follow later, also composite
constructs not supported yet (need to check the passing of temporaries around)
and lastprivate might not give the same answers as serial loop if the last
innermost body iteration isn't the last one for some of the outer loops
(that will need to be solved separately together with rectangular loops that have no
innermost body iterations, but some of the outer loops actually iterate).
Also, simd needs work.
Jakub Jelinek [Sat, 27 Jun 2020 10:38:23 +0000 (12:38 +0200)]
c-family: Use TYPE_OVERFLOW_UNDEFINED instead of !TYPE_UNSIGNED in pointer_sum [PR95903]
For lp64 targets and int off ... ptr[off + 1]
is lowered in pointer_sum to *(ptr + ((sizetype) off + (sizetype) 1)).
That is fine when signed integer wrapping is undefined (and is not done
already if off has unsigned type), but changes behavior for -fwrapv, where
overflow is well defined. Runtime test could be:
int
main ()
{
char *p = __builtin_malloc (0x100000000UL);
if (!p) return 0;
char *q = p + 0x80000000UL;
int o = __INT_MAX__;
q[o + 1] = 1;
if (q[-__INT_MAX__ - 1] != 1) __builtin_abort ();
return 0;
}
with -fwrapv or so, not included in the testsuite because it requires 4GB
allocation (with some other test it would be enough to have something
slightly above 2GB, but still...).
2020-06-27 Jakub Jelinek <jakub@redhat.com>
PR middle-end/95903
gcc/c-family/
* c-common.c (pointer_int_sum): Use TYPE_OVERFLOW_UNDEFINED instead of
!TYPE_UNSIGNED check to see if we can apply distributive law and handle
smaller precision intop operands separately.
gcc/testsuite/
* c-c++-common/pr95903.c: New test.
Iain Sandoe [Sat, 27 Jun 2020 07:18:34 +0000 (08:18 +0100)]
coroutines: Handle awaiters that are sub-objects [PR95736]
Move deciding on initializers for awaitables to the build of the
co_await, this allows us to analyse cases that do not need
a temporary at that point.
As the PR shows, the late analysis meant that we were not
checking properly for the case that an awaiter is a sub-object
of an existing variable outside the current function scope (and
therefore does not need to be duplicated in the frame).
gcc/cp/ChangeLog:
PR c++/95736
* coroutines.cc (get_awaitable_var): New helper.
(build_co_await): Check more carefully before
copying an awaitable.
(expand_one_await_expression): No initializer
is required when the awaitable is not a temp.
(register_awaits): Remove handling that is now
completed when the await expression is built.
gcc/testsuite/ChangeLog:
PR c++/95736
* g++.dg/coroutines/pr95736.C: New test.
Iain Sandoe [Sat, 27 Jun 2020 07:54:39 +0000 (08:54 +0100)]
coroutines: Improve diagnostics for one allocator case.
If the user provides operator new and that is noexcept, this
implies that it can fail with a null return. At that point, we expect
to be able to call get_return_object_on_allocation_failure().
This diagnoses the case where such an operator new has been
provided, but the g-r-o-o-a-f is either missing or unusable.
Michael Meissner [Sat, 27 Jun 2020 05:40:48 +0000 (00:40 -0500)]
Add PowerPC tests for power10.
2020-06-27 Michael Meissner <meissner@linux.ibm.com>
* gcc.target/powerpc/prefix-add.c: New test.
* gcc.target/powerpc/prefix-si-constant.c: New test.
* gcc.target/powerpc/prefix-di-constant.c: New test.
* gcc.target/powerpc/prefix-ds-dq.c: New test.
* gcc.target/powerpc/prefix-no-update.c: New test.
* gcc.target/powerpc/prefix-large-dd.c: New test.
* gcc.target/powerpc/prefix-large-df.c: New test.
* gcc.target/powerpc/prefix-large-di.c: New test.
* gcc.target/powerpc/prefix-large-hi.c: New test.
* gcc.target/powerpc/prefix-large-kf.c: New test.
* gcc.target/powerpc/prefix-large-qi.c: New test.
* gcc.target/powerpc/prefix-large-sd.c: New test.
* gcc.target/powerpc/prefix-large-sf.c: New test.
* gcc.target/powerpc/prefix-large-si.c: New test.
* gcc.target/powerpc/prefix-large-udi.c: New test.
* gcc.target/powerpc/prefix-large-uhi.c: New test.
* gcc.target/powerpc/prefix-large-uqi.c: New test.
* gcc.target/powerpc/prefix-large-usi.c: New test.
* gcc.target/powerpc/prefix-large-v2df.c: New test.
* gcc.target/powerpc/prefix-large.h: Include file for new tests.
* gcc.target/powerpc/prefix-pcrel-dd.c: New test.
* gcc.target/powerpc/prefix-pcrel-df.c: New test.
* gcc.target/powerpc/prefix-pcrel-di.c: New test.
* gcc.target/powerpc/prefix-pcrel-hi.c: New test.
* gcc.target/powerpc/prefix-pcrel-kf.c: New test.
* gcc.target/powerpc/prefix-pcrel-qi.c: New test.
* gcc.target/powerpc/prefix-pcrel-sd.c: New test.
* gcc.target/powerpc/prefix-pcrel-sf.c: New test.
* gcc.target/powerpc/prefix-pcrel-si.c: New test.
* gcc.target/powerpc/prefix-pcrel-udi.c: New test.
* gcc.target/powerpc/prefix-pcrel-uhi.c: New test.
* gcc.target/powerpc/prefix-pcrel-uqi.c: New test.
* gcc.target/powerpc/prefix-pcrel-usi.c: New test.
* gcc.target/powerpc/prefix-pcrel-v2df.c: New test.
* gcc.target/powerpc/prefix-pcrel.h: Include file for new tests.
* gcc.target/powerpc/prefix-stack-protect.c: New test.
/* Several x86 os'es need a frame pointer for other reasons,
usually pertaining to setjmp. */
if (SUBTARGET_FRAME_POINTER_REQUIRED)
return true;
...
if (crtl->profile && !flag_fentry)
return true;
A frame pointer is needed only for -pg, not for -mfentry -pg. Remove
SUBTARGET_FRAME_POINTER_REQUIRED from gnu-user.h to make i386 GCC behave
the same as x86-64 GCC. This fixes
Yichao Yu [Fri, 26 Jun 2020 21:46:15 +0000 (15:46 -0600)]
Fix target clone indirection elimination
The current logic seems to be comparing the whole attribute tree between
the callee and caller (or at least the tree starting from the target attribute).
This is unnecessary and causes strange dependency of the indirection
elimination on unrelated properties like `noinline`(PR95780) and
`visibility`(PR95778).
This changes the comparison to be only on the `target` attribute which should
be the intent of the code.
gcc
* multiple_target.c (redirect_to_specific_clone): Fix tests
to check individual attribute rather than an attribute list.
gcc/testsuite
* gcc.target/i386/pr95778-1.c: New test.
* gcc.target/i386/pr95778-2.c: New test.
Marek Polacek [Fri, 15 May 2020 04:05:33 +0000 (00:05 -0400)]
c++: Change the default dialect to C++17.
Since GCC 9, C++17 support is no longer experimental. It was too late
to change the default C++ dialect to C++17 in GCC 10, but I think now
it's time to pull the trigger (C++14 was made the default in GCC 6.1).
We're still missing two C++17 library features, but that shouldn't stop
us. See
<https://gcc.gnu.org/onlinedocs/libstdc++/manual/status.html#status.iso.2017>
and
<https://gcc.gnu.org/projects/cxx-status.html#cxx17>
for the C++17 status.
I won't list all C++17 features here, but just a few heads-up:
- trigraphs were removed (hardly anyone cares, unless your keyboard is
missing the # key),
- operator++(bool) was removed (so some tests now run in C++14 and down
only),
- the keyword register was removed (some legacy code might trip on
this),
- noexcept specification is now part of the type system and C++17 does
not allow dynamic exception specifications anymore (the empty throw
specification is still available, but it is deprecated),
- the evaluation order rules are different in C++17,
- static constexpr data members are now implicitly inline (which makes
them definitions),
- C++17 requires guaranteed copy elision, meaning that a copy/move
constructor call might be elided completely. That means that if
something relied on a constructor being instantiated via e.g. copying
a function parameter, it might now fail.
I'll post an update for cxx-status.html and add a new caveat to changes.html
once this is in.
gcc/ChangeLog:
* doc/invoke.texi (C Dialect Options): Adjust -std default for C++.
* doc/standards.texi (C Language): Correct the default dialect.
(C++ Language): Update the default for C++ to gnu++17.
gcc/c-family/ChangeLog:
* c-opts.c (c_common_init_options): Default to gnu++17.
gcc/testsuite/ChangeLog:
* c-c++-common/torture/vector-subscript-3.c: In C++17, define away
the keyword register.
* g++.dg/cpp1z/attributes-enum-1a.C: Only run pre-C++17.
* g++.dg/cpp1z/fold7a.C: Likewise.
* g++.dg/cpp1z/nontype3a.C: Likewise.
* g++.dg/cpp1z/utf8-2a.C: Likewise.
* g++.dg/parse/error11.C: Update expected diagnostics for C++17.
* g++.dg/torture/pr34850.C: Add -Wno-attribute-warning.
* g++.dg/torture/pr49394.C: In C++17, use noexcept(false).
* g++.dg/torture/pr82154.C: Use -std=c++14.
* lib/target-supports.exp: Set to C++17.
* obj-c++.dg/try-catch-9.mm: Use -Wno-register.
libgomp/ChangeLog:
* testsuite/libgomp.c++/atomic-3.C: Use -std=gnu++14.
gcc/testsuite/
* gfortran.dg/gomp/combined-if.f90: Adjust expected number
of matches depending on whether nvptx offloading is supported.
* lib/target-supports.exp
(check_effective_target_offload_nvptx): New.
Eric Botcazou [Fri, 26 Jun 2020 13:23:05 +0000 (15:23 +0200)]
Take into account range info to optimize range tests into bit tests
The patch is aimed at addressing the following two issues:
1. In order to protect the shift operation from undefinedness, the
new bit test is guarded with a new test, but this new test uses the
range of the bit test values, not that of the shift operation so,
if the input is in the range of the shift operation but not of the
bit test values, then the subsequent VRP pass cannot eliminate the
new test. Moreover changing the new test to use the range of the
shift operation, instead of that of the bit test values, in the
general case would pessimize the cases which are in between.
2. If the new test can be eliminated, then it becomes profitable
to do the optimization into a bit test for one fewer comparison in
the source code.
Therefore the patch changes optimize_range_tests_to_bit_test to use
the range info of the input in order to eliminate the new test.
gcc/ChangeLog:
* tree-ssa-reassoc.c (dump_range_entry): New function.
(debug_range_entry): New debug function.
(update_range_test): Invoke dump_range_entry for dumping.
(optimize_range_tests_to_bit_test): Merge the entry test in the
bit test when possible and lower the profitability threshold.
gcc/ada/ChangeLog:
* exp_ch4.adb (Expand_Set_Membership): Expand the membership test
using left associativity instead of right associativity.
gcc/testsuite/ChangeLog:
* gnat.dg/opt86_pkg.ads: New helper.
* gnat.dg/opt86a.adb: New test.
* gnat.dg/opt86b.adb: Likewise.
* gnat.dg/opt86c.adb: Likewise.
The PR points out that the standard does not restrict promise
expressions to methods, but the current implementation does.
The patch factors out the building of a general promise expression,
and then uses it in a fairly mechanical replacement of each case
that we need such an expressions.
This extends the handling for p.xxxxxx() expressions to cover the
cases where the promise member is some form callable.
Tests are added for each of the promise expressions.
It's somewhat tortuous to find good uses for this for the
get-return-object and get-return-object-on-allocation-failure
cases, but they are included anyway.
gcc/cp/ChangeLog:
PR c++/95519
* coroutines.cc (struct coroutine_info):Add a field
to hold computed p.return_void expressions.
(coro_build_promise_expression): New.
(get_coroutine_return_void_expr): New.
(finish_co_yield_expr): Build the promise expression
using coro_build_promise_expression.
(finish_co_return_stmt): Likewise.
(build_init_or_final_await): Likewise.
(morph_fn_to_coro): Likewise, for several cases.
gcc/testsuite/ChangeLog:
PR c++/95519
* g++.dg/coroutines/torture/pr95519-00-return_void.C: New test.
* g++.dg/coroutines/torture/pr95519-01-initial-suspend.C: New test.
* g++.dg/coroutines/torture/pr95519-02-final_suspend.C: New test.
* g++.dg/coroutines/torture/pr95519-03-return-value.C: New test.
* g++.dg/coroutines/torture/pr95519-04-yield-value.C: New test.
* g++.dg/coroutines/torture/pr95519-05-gro.C: New test.
* g++.dg/coroutines/torture/pr95519-06-grooaf.C: New test.
* g++.dg/coroutines/torture/pr95519-07-unhandled-exception.C: New test.
Iain Sandoe [Fri, 26 Jun 2020 09:52:07 +0000 (10:52 +0100)]
coroutines: Handle bad g-r-o-o-a-f cases.
If we see a get_return_object_on_allocation_failure in the
promise, we expect to be able to use it. If this isn't
possible (because of some error in the declaration) then we
need to handle the erroneous return to allow following code
to complete.
gcc/cp/ChangeLog:
* coroutines.cc (morph_fn_to_coro): Handle error
returns in building g-r-o-o-a-f expressions.
gcc/testsuite/ChangeLog:
* g++.dg/coroutines/coro1-allocators.h (BAD_GROOAF_STATIC):
New.
* g++.dg/coroutines/coro-bad-grooaf-00-static.C: New test.
Richard Biener [Fri, 26 Jun 2020 08:08:58 +0000 (10:08 +0200)]
tree-optimization/95897 - fix fold-left SLP reduction insert place
This fixes computation of the insertion place for fold-left SLP
reductions where the PHIs do not have vectorized stmts. The
SLP representation isn't perfect here thus the following.
Nick Clifton [Fri, 26 Jun 2020 09:04:41 +0000 (10:04 +0100)]
This patch removes the use of the "register" keyword from the bsearch() and bsearch_r() functions supplied by libiberty. The register keyword is deprecated in C++17.
2020-06-25 Nick Clifton <nickc@redhat.com>
include/
* libiberty.h (bsearch_r): Remove use of the register keyword from
the prototype.
libiberty/
* bsearch.c (bsearch): Remove use of register keyword.
* bsearch_r.c (bsearch_r): Likewise.
Iain Buclaw [Thu, 25 Jun 2020 07:33:26 +0000 (09:33 +0200)]
d: Do not implicitly set DECL_DECLARED_INLINE_P on member functions.
This has been questionable behaviour since it was added, and though it
has no effect on wider discussions around what should be the correct
semantics of pragma(inline) within D modules, doing this tree-level
optimization has mostly zero benefit as cross-module inlining doesn't
happen anyway.
gcc/d/ChangeLog:
* decl.cc (get_symbol_decl): Do not implicitly set
DECL_DECLARED_INLINE_P on member functions.
Iain Buclaw [Tue, 23 Jun 2020 16:08:54 +0000 (18:08 +0200)]
d: Remove another dependency on the front-end OutBuffer type.
As the DMD front-end never frees allocated memory, the glue layer
between the DMD front-end and GCC should generally avoid using DMD types
and interfaces if the purpose is internal only.
gcc/d/ChangeLog:
* d-lang.cc (d_parse_file): Replace OutBuffer with obstack.
Backports the OutBuffer interface from upstream dmd master, removing
another difference this and the self-hosted D branch that is purely
refactoring, and doesn't introduce any mechanical changes.
Iain Buclaw [Tue, 23 Jun 2020 12:45:50 +0000 (14:45 +0200)]
d: Don't set DECL_INITIAL if initializer is 'void'.
Declarations initialized with `= void` were being default initialized.
That is not really the intent, and misses the small optimization that
should have been gained from using void initializations.
gcc/d/ChangeLog:
* decl.cc (DeclVisitor::visit (VarDeclaration *)): Don't set
DECL_INITIAL if initializer is 'void'.
Iain Buclaw [Tue, 23 Jun 2020 08:40:37 +0000 (10:40 +0200)]
d: Turn on deprecation warnings by default.
This is the default in the upstream reference compiler, and can reduce
some confusion when comparing warning/error messages of gdc and dmd side
by side.
Merges libphobos with upstream druntime d05ebaad and phobos 021ae0df7.
Richard Biener [Thu, 25 Jun 2020 10:47:20 +0000 (12:47 +0200)]
tree-optimization/95839 - allow CTOR vectorization without loads
This removes a premature check for enough datarefs in a basic-block
before we consider vectorizing it which leaves basic-blocks with
just vectorizable vector constructors unvectorized. The check
is effectively done by the following check for store groups
which then also include constructors.
2020-06-25 Richard Biener <rguenther@suse.de>
PR tree-optimization/95839
* tree-vect-slp.c (vect_slp_analyze_bb_1): Remove premature
check on the number of datarefs.
fortran: Apply if clause to all sub-constructs in combined OpenMP constructs
The unmodified 'if' clause should be applied to all the sub-constructs that
accept an 'if' clause in a combined OpenMP construct, and not just to the
'parallel' sub-construct.
Richard Biener [Thu, 25 Jun 2020 11:41:47 +0000 (13:41 +0200)]
Always use SLP vector defs to determine insertion place
With the last vectorizable_shift patch we can now always use the
SLP vector defs to determine the vectorized stmt insertion place,
paving the way for a "verifier" for pending restructuring and
BB vectorization of reductions and other live stmts.
2020-06-25 Richard Biener <rguenther@suse.de>
* tree-vect-slp.c (vect_schedule_slp_instance): Always use
vector defs to determine insertion place.
Richard Biener [Thu, 25 Jun 2020 09:21:20 +0000 (11:21 +0200)]
tree-optimization/95866 - avoid using scalar ops for vectorized shift
This avoids using the original scalar SSA operand when vectorizing
a shift with a vectorized shift operand where we know all vector
components have the same value and thus we can use a vector by
scalar shift. Using the scalar SSA operand causes a possibly
long chain of scalar computation to be retained so it's better
to simply extract lane zero from the available vectorized shift
operand.
2020-06-25 Richard Biener <rguenther@suse.de>
PR tree-optimization/95866
* tree-vect-stmts.c (vectorizable_shift): Reject incompatible
vectorized shift operands. For scalar shifts use lane zero
of a vectorized shift operand.
Martin Liska [Thu, 25 Jun 2020 09:20:52 +0000 (11:20 +0200)]
gcov-tool: fix merge operation for summary
libgcc/ChangeLog:
* libgcov-driver.c (merge_summary): Remove function as its name
is misleading and doing something different.
(dump_one_gcov): Add ATTRIBUTE_UNUSED for 2 args. Take read summary
in gcov-tool.
* libgcov-util.c (curr_object_summary): Remove.
(read_gcda_file): Remove unused curr_object_summary.
(gcov_merge): Merge summaries.
* libgcov.h: Add summary argument for gcov_info struct.
Martin Liska [Wed, 24 Jun 2020 06:08:00 +0000 (08:08 +0200)]
VEC_COND_EXPR: clean up first argument
gcc/ChangeLog:
PR tree-optimization/95745
PR middle-end/95830
* gimple-isel.cc (gimple_expand_vec_cond_exprs): Delete dead
SSA_NAMEs used as the first argument of a VEC_COND_EXPR. Always
return 0.
* tree-vect-generic.c (expand_vector_condition): Remove dead
SSA_NAMEs used as the first argument of a VEC_COND_EXPR.
Will Schmidt [Wed, 24 Jun 2020 20:28:24 +0000 (15:28 -0500)]
[PATCH, PR target/94954] Fix wrong codegen for vec_pack_to_short_fp32() builtin
Hi,
Fix codegen for builtin vec_pack_to_short_fp32. This includes adding
a define_insn for xvcvsphp, and adding a new define_expand for
convert_4f32_8f16.
2020-06-24 Will Schmidt <will_schmidt@vnet.ibm.com>
PR target/94954
gcc
* config/rs6000/altivec.h (vec_pack_to_short_fp32): Update.
* config/rs6000/altivec.md (UNSPEC_CONVERT_4F32_8F16): New unspec.
(convert_4f32_8f16): New define_expand
* config/rs6000/rs6000-builtin.def (convert_4f32_8f16): New builtin define
and overload.
* config/rs6000/rs6000-call.c (P9V_BUILTIN_VEC_CONVERT_4F32_8F16): New
overloaded builtin entry.
* config/rs6000/vsx.md (UNSPEC_VSX_XVCVSPHP): New unspec.
(vsx_xvcvsphp): New define_insn.