| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
gcc/ChangeLog:
* config/i386/i386-expand.cc (expand_vec_perm_blend): Generate
vpblendd instead of vpblendw for V4SI under avx2.
gcc/testsuite/ChangeLog:
* gcc.target/i386/pr88828-0.c: Adjust testcase.
|
|
|
|
Stepping through a gdb session inspecting costs that cause
gcc.dg/tree-ssa/slsr-13.c to fail, exposed that before this
patch, cris_rtx_costs told that a shift of 1 of a register
costs 5, while adding two registers costs 4.
Making the cost of a quick-immediate constant equal to using
a register (default 0) reflects actual performance and
size-cost better. It also happens to make
gcc.dg/tree-ssa/slsr-13.c pass with what looks like better
code being generated, and improves coremark performance by
0.4%.
But, blindly doing this for *all* valid operands that fit
the "quick-immediate" addressing mode, trips interaction
with other factors*, with the end result mixed at best. So,
do this only for MINUS and logical operations for the time
being, and only for modes that fit in one register.
*) Examples of "other factors":
- A bad default implementation of insn_cost or actually,
pattern_cost, that looks only at the set_src_cost and
furthermore sees such a cost of 0 as invalid. (Compare to
the more sane set_rtx_cost.) This naturally tripped up
combine and ifcvt, causing all sorts of changes, good and
bad.
- Having the same cost, to compare a register with 0 as with
-31..31, means a compare insn of an eliminable form no
longer looks preferable.
* config/cris/cris.cc (cris_rtx_costs) [CONST_INT]: Return 0
for many quick operands, for register-sized modes.
|
|
The -fmod= and -fdef= options do not work. After the linking
re-implementation and subsequent restructuring the -fmod= amd -fdef= are
now broken. This patch adds -fmod= and -fdef= processing into gm2spec.cc.
It also reduces the complexity of extension handling within M2Search
by storing the preceeding "." in the extension.
gcc/m2/ChangeLog:
PR modula2/109336
PR modula2/109315
* gm2-compiler/M2FileName.mod (CalculateFileName): Simplified by
ensuring the extension contains the ".".
(CalculateStemName): Re-formatted.
(ExtractExtension): Re-formatted.
(ExtractModule): Re-formatted.
* gm2-compiler/M2Options.def (setdefextension): Add block comment.
(setmodextension): Add block comment. Re-formatted.
* gm2-compiler/M2Options.mod (setdefextension): Add block comment.
(setmodextension): Add block comment. Re-formatted.
* gm2-compiler/M2Search.mod (FindSourceDefFile): Use
DefaultDefExt.
(DefaultDefExt): New constant.
(DefaultModExt): New constant.
(FindSourceModFile): Use DefaultModExt.
* gm2-gcc/m2decl.cc (m2decl_DeclareKnownVariable): Correct
spelling.
* gm2spec.cc (M2SOURCE): New constant.
(LANGSPEC): New value.
(MATHLIB): New value.
(WITHLIBC): New value.
(SKIPOPT): New value.
(lang_specific_driver): Replace seen_module_extension bool with
module_extension char *. Detect -fmod= and remember extension.
Use the extension to detect modula-2 source and mark it as such.
gcc/testsuite/ChangeLog:
PR modula2/109336
* gm2/link/nondefaultext/pass/hello.md: New test.
* gm2/link/nondefaultext/pass/liba.dm: New test.
* gm2/link/nondefaultext/pass/liba.md: New test.
* gm2/link/nondefaultext/pass/link-nondefaultext-pass.exp: New test.
Signed-off-by: Gaius Mulley <gaiusmod2@gmail.com>
|
|
Z*inx is conflict with float extensions, add incompatible check when
z*inx and f extension both enabled.
Since all float extension imply f extension and all z*inx extension
imply zfinx extension, so we just need to check f with zfinx extension
as the base case.
Co-Authored by: Kito Cheng <kito.cheng@gmail.com>
gcc/ChangeLog:
* common/config/riscv/riscv-common.cc (riscv_subset_list::parse):
New check.
gcc/testsuite/ChangeLog:
* gcc.target/riscv/arch-19.c: New test.
|
|
PR bootstrap/109310
gcc/ChangeLog:
* configure.ac: Emit a warning for deprecated option
--enable-link-mutex.
* configure: Regenerate.
|
|
When forwprop discovers unreachable code or makes decisions based
on unreachable edges make sure to cleanup the CFG since otherwise
SSA form can become invalid.
PR tree-optimization/109331
* tree-ssa-forwprop.cc (pass_forwprop::execute): When we
discover a taken edge make sure to cleanup the CFG.
* gcc.dg/torture/pr109331.c: New testcase.
|
|
The g++.dg/pr94920.C testcase looks for a specific number of
ABS_EXPRs but the vector example is prone to vector lowering so
the following instead of scanning the optimized dump scans the
forwprop1 dump which is before vector lowering and the point the
transforms should have happened.
* g++.dg/pr94920.C: Scan forwprop1 instead of optimized.
|
|
There's interfering between the to_removed queue and other mechanisms
removing stmts, in this case remove_prop_source_from_use. The following
makes the to_remove queue draining more permissive.
PR tree-optimization/109327
* tree-ssa-forwprop.cc (pass_forwprop::execute): Deal with
already removed stmts when draining to_remove.
* gcc.dg/pr109327.c: New testcase.
|
|
The testcases in this bug reveal cases where an early generated
type is collected because it was unused but gets attempted to
be recreated later when a late DIE for a function (an OpenMP
reduction) is created. That's unexpected and possibly the fault
of OpenMP but the following allows the re-creation of the context
type to succeed.
PR ipa/106124
* dwarf2out.cc (lookup_type_die): Reset TREE_ASM_WRITTEN
so we can re-create the DIE for the type if required.
* g++.dg/gomp/pr106124.C: New testcase.
|
|
[PR109301]
The following testcase ICEs since the sincos and vect pass order has
been swapped. It is not valid to replace vector sqrt (sqrt (x)) with
pow (x, 0.25) because build_real on vector type is invalid (could be
handled by using build_uniform_cst and adjusting type passed to
build_real) but more importantly because nothing checks if we can
actually do vector pow.
While we have pow_optab, apparently no target defines it, so it doesn't
seem to be worth bothering with for now and the patch just punts on
non-scalar sqrts.
I think the other simplifications next to it are fine, as they mostly
use CBRT which doesn't even have internal function (so is a builtin
only and therefore always scalar), or have already pow in the IL (which
doesn't have optab and shouldn't be thus vector either).
It is true that with <bits/math-vector.h> we do vectorize some calls to
pow or cbrt (but don't handle others strangely), but those aren't using
internal functions but simd clones and so match.pd doesn't know anything
about those (at least for now).
The following patch fixes it by mostly restoring the state before
r13-1763 where canonicalize_math_p () was true only until the end of the
pass which transformed pow or pow-like calls before vectorization (formerly
sincos pass, now it is powcabs pass).
powcabs is a pass in the spot sincos was happening before, so the
only change was defer the sin+cos simplification into cexpi to a later
new pass (except for the name moving with it) and none of the
canonicalize_math_p () guarded simplification in match.pd seem to rely
on those sin+cos -> cexpi simplifications and canonicalize_math_p is
the only user of this property.
2023-03-29 Jakub Jelinek <jakub@redhat.com>
Richard Biener <rguenther@suse.de>
PR tree-optimization/109301
* tree-ssa-math-opts.cc (pass_data_cse_sincos): Change
properties_provided from PROP_gimple_opt_math to 0.
(pass_data_expand_powcabs): Change properties_provided from 0 to
PROP_gimple_opt_math.
* gcc.dg/pr109301.c: New test.
|
|
With multi-argument PHIs and now doing VN on the if-converted blocks
the optimization of CSEing condition and negated condition doesn't
work well anymore. The following restores this a little bit for
the case of a single inverted condition into a COND_EXPR where
we can instead swap the COND_EXPR arms. The same optimization
is already done for the case of two-argument PHIs.
This avoids one comparison/mask for the testcase at hand.
PR tree-optimization/109154
* tree-if-conv.cc (gen_phi_arg_condition): Handle single
inverted condition specially by inverting at the caller.
(gen_phi_arg_condition): Swap COND_EXPR arms if requested.
|
|
The two hunks fix missing handling demonstrated by the two testcases: first,
if we omit one alias template parm but include another, we need to rewrite
the deduced template args to reflect the new position of the included parm.
Second, if we can't deduce any template args for a parameter pack, it is
deduced to an empty pack.
PR c++/109321
PR c++/109320
gcc/cp/ChangeLog:
* pt.cc (alias_ctad_tweaks): Rewrite deduced args.
(type_targs_deducible_from): Handle null pack deduction.
gcc/testsuite/ChangeLog:
* g++.dg/cpp2a/class-deduction-alias16.C: New test.
* g++.dg/cpp2a/class-deduction-alias17.C: New test.
|
|
In https://go.dev/cl/343873 we stopped padding zero-sized trailing
fields in functions that return multiple results where the last result
is zero-sized. This CL makes the corresponding change on the caller side.
The test case is https://go.dev/cl/479898.
Fixes golang/go#55242
Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/479955
|
|
|
|
* de.po: Update.
|
|
Follow-up to commit bcbeebc498126c50d73809ec8a4bd0bff27ee97b
"Fortran: Add support for WEAK attribute for variables".
gcc/testsuite/
* gfortran.dg/weak-2.f90: Enable for nvptx target.
|
|
PR c/107002 reports an assertion failure from deep inside the
diagnostic_shows_locus when attempting to print fix-it hints relating
to -Wxor-used-as-pow. The case involves macro expansions with
-ftrack-macro-expansion=0.
It doesn't seem to make much sense to emit this warning for macro
expansions, so this patch updates the warning not to (which seems
to also be clang's behavior). The patch also adds some bulletproofing
to diagnostic-show-locus.cc to be more robust against such invalid
fix-it hints.
Doing so fixes the ICE.
gcc/c-family/ChangeLog:
PR c/107002
* c-common.h (check_for_xor_used_as_pow): Add "rhs_loc" param.
* c-warn.cc (check_for_xor_used_as_pow): Add "rhs_loc" param.
Reject cases where involving macro expansions.
gcc/c/ChangeLog:
PR c/107002
* c-typeck.cc (parser_build_binary_op): Update for new param of
check_for_xor_used_as_pow.
gcc/cp/ChangeLog:
PR c/107002
* parser.cc (cp_parser_binary_expression): Update for new param of
check_for_xor_used_as_pow.
gcc/ChangeLog:
PR c/107002
* diagnostic-show-locus.cc (column_range::column_range): Factor
out assertion conditional into...
(column_range::valid_p): ...this new function.
(line_corrections::add_hint): Don't attempt to consolidate hints
if it would lead to invalid column_range instances.
gcc/testsuite/ChangeLog:
PR c/107002
* c-c++-common/Wxor-used-as-pow-1.c: Add macro test.
* c-c++-common/Wxor-used-as-pow-pr107002-0.c: New test.
* c-c++-common/Wxor-used-as-pow-pr107002-1.c: New test.
* c-c++-common/Wxor-used-as-pow-pr107002-2.c: New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
|
|
As mentioned in the PR, constructing a message from two parts by
concatenating them prevents translations, unless one of the parts
is a keyword which should be never translated.
The following patch fixes that.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
PR c++/109309
* contracts.cc: Include intl.h.
(check_postcondition_result): Don't form diagnostics from two halves
of an English message to allow translations.
|
|
The patch for PR91415 fixed -Wsequence-point to treat shifts and ARRAY_REF
as sequenced in C++17, and COMPONENT_REF as well. But this is unnecessary
for COMPONENT_REF, since the RHS is just a FIELD_DECL with no actual
evaluation, and in this testcase handling COMPONENT_REF as sequenced blows
up fast in a deep inheritance tree. Instead, look through it.
PR c++/107163
gcc/c-family/ChangeLog:
* c-common.cc (verify_tree): Don't use sequenced handling
for COMPONENT_REF.
gcc/testsuite/ChangeLog:
* g++.dg/warn/Wsequence-point-5.C: New test.
|
|
RVV intrinsic has defined a macro to identity the version of RVV
intrinsic spec, we missed that before, thanksful we are catch this
before release.
gcc/ChangeLog:
PR target/109312
* config/riscv/riscv-c.cc (riscv_ext_version_value): New.
(riscv_cpu_cpp_builtins): Define __riscv_v_intrinsic and
minor refactor.
gcc/testsuite/ChangeLog:
PR target/109312
* gcc.target/riscv/predef-__riscv_v_intrinsic.c: New test.
|
|
Do not attempt to use a plain subtraction for generating a three-way
comparison result in autopref_rank_for_schedule qsort comparator, as
offsets are not restricted and subtraction may overflow. Open-code
a safe three-way comparison instead.
gcc/ChangeLog:
PR rtl-optimization/109187
* haifa-sched.cc (autopref_rank_for_schedule): Avoid use of overflowing
subtraction in three-way comparison.
gcc/testsuite/ChangeLog:
PR rtl-optimization/109187
* gcc.dg/pr109187.c: New test.
|
|
First, class value_relation should not sanitize records. just create
what is asked.
Second., if there is not a relation record, compute_operand was
creating one for op1 == op2 if op1 and op2 were the same symbol. This
is not the correct way to communicate the information, as that record
will continue to be passed along the GORI unwind chain.
Instead, simply pass that information locally to the opX_range routine
for only the current statement.
PR tree-optimization/109265
PR tree-optimization/109274
gcc/
* gimple-range-gori.cc (gori_compute::compute_operand_range): Do
not create a relation record is op1 and op2 are the same symbol.
(gori_compute::compute_operand1_range): Pass op1 == op2 to the
handler for this stmt, but create a new record only if this statement
generates a relation based on the ranges.
(gori_compute::compute_operand2_range): Ditto.
* value-relation.h (value_relation::set_relation): Always create the
record that is requested.
gcc/testsuite/
* gcc.dg/pr109274.c: New.
* gfortran.dg/pr109265.f90: New.
|
|
When forwprop simplifies the CFG the 2nd order opportunities by
exposed degenerate PHIs are not realized. The following improves
this by properly tracking executable edges and thus handling this
for non-cyclic CFGs at least.
This avoids the bogus diagnostic reported for the testcase in this PR.
PR tree-optimization/107087
* tree-ssa-forwprop.cc (pass_forwprop::execute): Track
executable regions to avoid useless work and to better
propagate degenerate PHIs.
* g++.dg/pr107087.C: New testcase.
|
|
This is a comment-only change that I should have
done with the previous commit (304c7d44a) but
forgot to do so.
gcc/ChangeLog:
* config/i386/x-mingw32-utf8: update comments.
Signed-off-by: Jonathan Yong <10walls@gmail.com>
|
|
Before GCC 12, we would vectorize:
int32_t arr[] = { x, x, x, x };
at -O3. Vectorizing the store on its own is often a loss, particularly
for integers, so g:4963079769c99c4073adfd799885410ad484cbbe suppressed it.
This was necessary to fix regressions from enabling vectorisation at -O2,
However, the vectorisation is important if the code subsequently loads
from the array using vld1:
return vld1q_s32 (arr);
This approach of initialising an array and loading from it is the
recommend endian-agnostic way of constructing an ACLE vector.
As discussed in the PR notes, the general fix would be to fold the
store and load-back to a constructor (preferably before vectorisation).
But that's clearly not stage 4 material.
This patch instead delays folding vld1 until after inlining and
records which decls a vld1 loads from. It then treats vector
stores to those decls as free, on the optimistic assumption that
they will be removed later. The patch also brute-forces
vectorization of plain constructor+store sequences, since some
of the CPU costs make that (dubiously) expensive even when the
store is discounted.
Delaying folding showed that we were failing to update the vops.
The patch fixes that too.
Thanks to Tamar for discussion & help with testing.
gcc/
PR target/109072
* config/aarch64/aarch64-protos.h (aarch64_vector_load_decl): Declare.
* config/aarch64/aarch64.h (machine_function::vector_load_decls): New
variable.
* config/aarch64/aarch64-builtins.cc (aarch64_record_vector_load_arg):
New function.
(aarch64_general_gimple_fold_builtin): Delay folding of vld1 until
after inlining. Record which decls are loaded from. Fix handling
of vops for loads and stores.
* config/aarch64/aarch64.cc (aarch64_vector_load_decl): New function.
(aarch64_accesses_vector_load_decl_p): Likewise.
(aarch64_vector_costs::m_stores_to_vector_load_decl): New member
variable.
(aarch64_vector_costs::add_stmt_cost): If the function has a vld1
that loads from a decl, treat vector stores to those decls as
zero cost.
(aarch64_vector_costs::finish_cost): ...and in that case,
if the vector code does nothing more than a store, give the
prologue a zero cost as well.
gcc/testsuite/
PR target/109072
* gcc.target/aarch64/pr109072_1.c: New test.
* gcc.target/aarch64/pr109072_2.c: Likewise.
|
|
The following avoids duplicating matching code for (match ...)
in match.pd when possible. That's more easily possible for
(match ...) than simplify because we do not need to handle
common matches (those would be diagnosed only during compiling)
nor is the result able to inspect the active operator.
Specifically this reduces the size of the generated matches for
the atomic ops as noted in PR108129.
gimple-match.cc shrinks from 245k lines to 209k lines with this patch.
PR bootstrap/84402
PR tree-optimization/108129
* genmatch.cc (lower_for): For (match ...) delay
substituting into the match operator if possible.
(dt_operand::gen_gimple_expr): For user_id look at the
first substitute for determining how to access operands.
(dt_operand::gen_generic_expr): Likewise.
(dt_node::gen_kids): Properly sort user_ids according
to their substitutes.
(dt_node::gen_kids_1): Code-generate user_id matching.
|
|
Follow-up to the recently committed fix of fixing the "subcomand" typos
in diagnostics.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
Jonathan Wakely <jwakely@redhat.com>
* gcov-tool.cc (do_merge, do_merge_stream, do_rewrite, do_overlap):
Use subcommand rather than sub-command in function comments.
|
|
Trivial typo fix.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
PR fortran/109314
* openmp.cc (gfc_omp_absent_contains_clause): Fix typo in diagnostics
- composit -> composite.
|
|
As discussed in the PR, flushing denormals to zero on every frange::set
might be harmful for e.g. x < 0.0 comparisons, because we then on both
sides use ranges that include zero [-Inf, -0.0] on the true side, and
[-0.0, +Inf] NAN on the false side, rather than [-Inf, nextafter (-0.0, -Inf)]
on the true side.
The following patch does it only in range_operator_float::fold_range
which is right now used for +-*/ (both normal and reverse ops of those).
Though, I don't see any difference on the testcase in the PR, but not sure
what I should be looking at and the reduced testcase there has undefined
behavior.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/109154
* value-range.h (frange::flush_denormals_to_zero): Make it public
rather than private.
* value-range.cc (frange::set): Don't call flush_denormals_to_zero
here.
* range-op-float.cc (range_operator_float::fold_range): Call
flush_denormals_to_zero.
|
|
[PR106190]
The following testcase ICEs, because without optimization eh lowering
decides not to duplicate finally block of try/finally and so we end up
with variable guarded cleanup. The sanopt pass creates a cfg that ought
to be cleaned up (some IFN_UBSAN_* functions are lowered in this case with
constant conditions in gcond and when not allowing recovery some bbs which
end with noreturn calls actually have successor edges), but the cfg cleanup
is actually (it is -O0) done only during the optimized pass. We notice
there that the d[1][a] = 0; statement which has an EH edge is unreachable
(because ubsan would always abort on the out of bounds d[1] access), remove
the EH landing pad and block, but because that block just sets a variable
and jumps to another one which tests that variable and that one is reachable
from normal control flow, the __builtin_eh_pointer (1) later in there is
kept in the IL and we ICE during expansion of that statement because the
EH region has been removed.
The following patch fixes it by doing the cfg cleanup already during
sanopt pass if we create something that might need it, while the EH
landing pad is then removed already during sanopt pass, there is ehcleanup
later and we don't ICE anymore.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
PR middle-end/106190
* sanopt.cc (pass_sanopt::execute): Return TODO_cleanup_cfg if any
of the IFN_{UB,HWA,A}SAN_* internal fns are lowered.
* gcc.dg/asan/pr106190.c: New test.
|
|
On Wed, Mar 22, 2023 at 07:32:44AM +0100, Aldy Hernandez wrote:
> * value-range.cc (frange::set): Add nan_state argument.
> * value-range.h (class nan_state): New.
> (frange::get_nan_state): New.
The following patch makes use of those changes in float_widen_lhs_range.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
* range-op-float.cc (float_widen_lhs_range): Use pass get_nan_state
as 4th argument to set to avoid clear_nan and union_ calls.
|
|
gcc/testsuite/
* gcc.target/sparc/20230328-1.c: New test.
* gcc.target/sparc/20230328-2.c: Likewise.
* gcc.target/sparc/20230328-3.c: Likewise.
* gcc.target/sparc/20230328-4.c: Likewise.
|
|
-mpreferred-stack-boundary=2 DImode temporaries [PR109276]
The following testcase ICEs since r11-2259 because assign_386_stack_local
-> assign_stack_local -> ix86_local_alignment now uses 64-bit alignment
for DImode temporaries rather than 32-bit as before.
Most of the spots in the backend which ask for DImode temporaries are during
expansion and those apparently are handled fine with -m32
-mpreferred-stack-boundary=2, we dynamically realign the stack in that case
(most of the spots actually don't need that alignment but at least one
does), then 2 spots are in STV which I assume also work correctly.
But during splitting we can create a DImode slot which doesn't need to be
64-bit alignment (it is nicer for performance though), when we apparently
aren't able to detect it for dynamic stack realignment purposes.
The following patch just makes the slot 32-bit aligned in that rare case.
2023-03-28 Jakub Jelinek <jakub@redhat.com>
PR target/109276
* config/i386/i386.cc (assign_386_stack_local): For DImode
with SLOT_FLOATxFDI_387 and -m32 -mpreferred-stack-boundary=2 pass
align 32 rather than 0 to assign_stack_local.
* gcc.target/i386/pr109276.c: New test.
|
|
This is a regression present on the mainline and 12 branch at -O2, but the
issue is related to vectorization so was present at -O3 in earlier versions.
The vcondu expander that was added for VIS 3 more than a decade ago does not
fully work, because it does not filter out the unsigned condition codes (the
instruction is an UNSPEC that accepts only signed condition codes).
While I was at it, I also added the missing vcond and vcondu expanders for
the new comparison instructions that were added in VIS 4.
gcc/
PR target/109140
* config/sparc/sparc.cc (sparc_expand_vcond): Call signed_condition
on operand #3 to get the final condition code. Use std::swap.
* config/sparc/sparc.md (vcondv8qiv8qi): New VIS 4 expander.
(fucmp<gcond:code>8<P:mode>_vis): Move around.
(fpcmpu<gcond:code><GCM:gcm_name><P:mode>_vis): Likewise.
(vcondu<GCM:mode><GCM:mode>): New VIS 4 expander.
gcc/testsuite/
* gcc.target/sparc/20230328-1.c: New test.
* gcc.target/sparc/20230328-2.c: Likewise.
* gcc.target/sparc/20230328-3.c: Likewise.
* gcc.target/sparc/20230328-4.c: Likewise.
|
|
Nothing really serious this time.
gcc/
* doc/gm2.texi: Add missing Next, Previous and Top fields to most
top-level sections.
|
|
The patch that introduced the weak_undefined effective-target keyword
and corresponding dg-add-options support
commit 378ec7b87a5265dbe2d489c245fac98ef37fa638
Author: Alexandre Oliva <oliva@adacore.com>
Date: Thu Mar 23 00:45:05 2023 -0300
[testsuite] test for weak_undefined support and add options
badly broke the affected tests on macOS like so:
ERROR: gcc.dg/addr_equal-1.c: unknown dg option: 89 for " dg-add-options 5 weak_undefined "
ERROR: gcc.dg/addr_equal-1.c: unknown dg option: 89 for " dg-add-options 5 weak_undefined "
add_options_for_weak_undefined tries to call an non-existant proc "89".
Even after fixing this by escaping the brackets, two tests still failed to
link since they lacked the corresponding calls do dg-add-options
weak_undefined.
Tested on x86_64-apple-darwin20.6.0 and i386-pc-solaris2.11.
2023-03-27 Rainer Orth <ro@CeBiTec.Uni-Bielefeld.DE>
gcc/testsuite:
* lib/target-supports.exp (add_options_for_weak_undefined): Escape
brackets.
* gcc.dg/visibility-22.c: Add weak_undefined options.
libgomp:
* testsuite/libgomp.oacc-c-c++-common/routine-nohost-2.c: Add
weak_undefined options.
|
|
Prevent any name mangling in HOST_EXTRA_OBJS_SYMBOL
such that the linker always finds it by that name.
Also add the .manifest file as an explicit
dependency in the make rule such that the
object gets re-built if it changes.
gcc/ChangeLog:
* config.host: Pull in i386/x-mingw32-utf8 Makefile
fragment and reference utf8rc-mingw32.o explicitly
for mingw hosts.
* config/i386/sym-mingw32.cc: prevent name mangling of
stub symbol.
* config/i386/x-mingw32-utf8: Make utf8rc-mingw32.o
depend on manifest file explicitly.
Signed-off-by: Jonathan Yong <10walls@gmail.com>
|
|
This reverts commit 776a5bb5894315ab144dc74222fc580fde8fdd87.
PR rtl-optimization/109311
* cfgcleanup.cc (bb_is_just_return): Revert previous change.
|
|
Prior to the removal of STABS support the gdwarf, gstabs, ... options
formed a cycle with their Negative(..) option attribute. But that
didn't actually have any effect since most of the options also
are Joined or JoinedOrMissing for which there's no pruning of options
and so once ran into the set_debug_level diagnostics reporting
conflicting debug formats.
The following removes the remains of that cycle, which is a
Negative option from gdwarf to gdwarf-. With RejectNegative
added the expected effect of -gdwarf-4 -gdwarf would be to
enable DWARF5 support (but this doesn't happen for some reason).
I think the more sensible behavior is that seen and implemented
in opts.cc, the more specific -gdwarf-4 determines the DWARF level
and a later or earlier -gdwarf becomes a no-op. So the
Negative(..) annotation on gdwarf is just confusing.
* common.opt (gdwarf): Remove Negative(gdwarf-).
|
|
The following adds RejectNegative to the gdwarf, gdwarf-, ggdb and gvms
options since the current behavior is to treat the negative variant
the same as the positive variant. In particular -ggdb -gno-gdb
do not cancel, and plain -gno-dwarf will enable (dwarf!) debug output.
Rejecting the negative forms avoids interpreting sensible behavior
to combinations of options like -gdwarf-5 -gno-dwarf-3 and sticks to
the behavior that later -g options simply override earlier ones and
the only negative form is -g0.
* common.opt (gdwarf): Add RejectNegative.
(gdwarf-): Likewise.
(ggdb): Likewise.
(gvms): Likewise.
|
|
This patch has no effect on builds using reload of libgcc, newlib libc, my
own at-a-glance-testsuite and coremark. That somewhat surprisingly
also goes for LRA builds, even with all CRIS reload_in_progress
augmented to include lra_in_progress. I just noticed it when checking
because another port had a similar fix, where it mattered for LRA.
* config/cris/constraints.md ("T"): Correct to
define_memory_constraint.
|
|
The test-case gcc.target/cris/rld-legit1.c is a reduced
test-case that required defining LEGITIMIZE_RELOAD_ADDRESS
to stop the address from being decomposed into several insns
by reload. Valid but suboptimal code was generated.
(Before implementing that hook for CRIS, the same test-case
also exposed a bug in reload, and a fix was committed to
avoid an ICE; see e.g. git r0-71992-gff0d9879ab0f30 and
related commits. But, post-cc0, reload no longer handles
this test-case without LEGITIMIZE_RELOAD_ADDRESS helping and
there'd again an be ICE for CRIS (again: only if
LEGITIMIZE_RELOAD_ADDRESS is disabled). There's a patch to
reload to fix that, at
https://gcc.gnu.org/pipermail/gcc-patches/2023-February/612039.html)
But, LRA also does not handle that test-case gracefully, and
like reload without LEGITIMIZE_RELOAD_ADDRESS for CRIS,
decomposes the address into a suboptimal (but valid)
sequence, about as messy as that from reload, and
gcc.target/cris/rld-legit1.c would regress for LRA. There's
nothing equivalent to LEGITIMIZE_RELOAD_ADDRESS for LRA.
(Stepping through LRA, I can't find an obvious place where
to put such a hook. Granted, I haven't seen this kind of
messy decomposition in other code, so I'm not insisting a
LEGITIMIZE_RELOAD_ADDRESS-like hook is a good idea.)
These new peephole2's are required to not regress
gcc.target/cris/rld-legit1.c with LRA enabled for CRIS.
They don't appear to otherwise make a difference for neither
libgcc, newlib libc, my own at-a-glance tests nor coremark,
for neither LRA nor reload.
* config/cris/cris.md (BW2): New mode-iterator.
(lra_szext_decomposed, lra_szext_decomposed_indirect_with_offset): New
peephole2s.
|
|
This patch affects a post-reload define_split for CRIS that transforms
a condition-code-clobbering addition into a non-clobbering addition.
(A "two-operand" addition between registers is the only insn that has
both a condition-code-clobbering and a non-clobbering variant for
CRIS.) Many more "add.d":s are replaced by non-condition-code-
clobbering "addi":s after this patch, but most of the transformations
don't matter.
CRIS with LRA generated code that exposed a flaw with the original
patch: it bailed too easily, on *any* insn using the result of the
addition. To wit, more effort than simply applying reg_mentioned_p is
needed to inspect the user, in the code to avoid munging an insn
sequence that cmpelim is supposed to handle.
With this patch coremark score for CRIS (*with reload*) improves by
less than 0.01% (a single "nop" is eliminated in
core_state_transition, in an execution path that affects ~1/20 of all
of the 10240 calls). However, the original cause for this patch is to
not regress gcc.target/cris/pr93372-44.c for LRA, where otherwise a
needless "cmpq" is emitted. For CRIS with LRA, the performance effect
on coremark isn't even measurable, except by reducing the size of the
executable due to affecting non-called library code.
* config/cris/cris.md ("*add<mode>3_addi"): Improve to bail only
for possible eliminable compares.
|
|
* config/cris/constraints.md ("R"): Remove unused constraint.
|
|
|
|
gcc/ChangeLog:
PR gcov-profile/109297
* gcov-tool.cc (merge_usage): Fix "subcomand" typo.
(merge_stream_usage): Likewise.
(overlap_usage): Likewise.
|
|
I realized I never added a testcase for the fix of this bug. Now done
after verifying it still fails when reverting the fix.
PR tree-optimization/54498
* g++.dg/torture/pr54498.C: New testcase.
|
|
The following adds the testcase for the bug which was recently
fixed.
PR tree-optimization/108357
* gcc.dg/tree-ssa/pr108357.c: New testcase.
|
|
This patch adds missing mode specifiers for XTheadMemPair INSNs.
gcc/ChangeLog:
PR target/109296
* config/riscv/thead.md: Add missing mode specifiers.
Signed-off-by: Christoph Müllner <christoph.muellner@vrull.eu>
|
