Whilst trying to convert the add vendor branch script to work with
personal branches I encountered a minor issue where git would report
ambiguous refs when checking out the new branch.
It turns out that this is because git considers <me>/<topic> to be
ambiguous if both
refs/heads/<me>/<topic>
and
refs/remotes/<me>/<topic>
exist in the list of known branches.
Having thought about this a bit, I think the best solution is to adopt
something more like the vendors space and call the remote users/<me>
(this also works better if you want to set up remotes to track other
users branches as well).
So this patch has two parts.
1) It updates gcc-git-customization.sh to set up the new 'remote' and
converts any existing remote and branches tracking that to the new
format
2) It adds a new script to set up a personal branch on the gcc git repository.
* gcc-git-customization.sh: Use users/<pfx> for the personal remote
rather than just <pfx>. Convert any existing personal branches to the
new remote.
* git-add-user-branch.sh: New file.
I noticed, but ignored this code when addressing p80005, but having
fixed up defined(X) on the modules branch, I could see where it came
from, and it's obviously wrong as we've just pulled out a string
contant from the token.
* expr.c (parse_has_include): Remove bogus controlling macro code.
The testcase fails to link on targets without -pthread which is
implied by -fopenmp. Use dg-require-effective-target fopenmp to avoid
this problem.
2020-01-24 Christophe Lyon <christophe.lyon@linaro.org>
PR debug/92763
* g++.dg/debug/pr92763.C: Require fopenmp.
* Weaken expectation concerning acc_property_free_memory.
Do not expect the value returned by CUDA since that value might have
changed in the meantime.
* Use correct type for the results of calls to acc_get_property in tests.
libgomp/
* testsuite/libgomp.oacc-c-c++-common/acc_get_property-aux.c
(expect_device_properties): Remove "expected_free_mem" argument,
change "expected_total_mem" argument type to size_t;
change types of acc_get_property results to size_t,
adapt format strings.
* testsuite/libgomp.oacc-c-c++-common/acc_get_property.c:
Use %zu instead of %zd to print size_t values.
* testsuite/libgomp.oacc-c-c++-common/acc_get_property-2.c: Adapt and
rename to ...
* testsuite/libgomp.oacc-c-c++-common/acc_get_property-nvptx.c: ... this.
* testsuite/libgomp.oacc-c-c++-common/acc_get_property-3.c: Adapt and
rename to ...
* testsuite/libgomp.oacc-c-c++-common/acc_get_property-host.c: ... this.
Reviewed-by: Thomas Schwinge <thomas@codesourcery.com>
* target.def (flags_regnum): Also mention effect on delay slot filling.
* doc/tm.texi: Regenerate.
Previously only mentioned an effect on compare elimination.
PR analyzer/93367 reports a testsuite failure in abort.c on
hppa64-hp-hpux11.11 when detecting if the analyzer "knows" that the
condition holds after the assert.
The root cause is that the assertion failure function in that
configuration's <assert.h> is not marked with
__attribute__ ((__noreturn__)).
This patch reworks the test to avoid <assert.h> in favor of a custom
implementation of assert, so that the test demonstrates the idea without
relying on properties of <assert.h>.
gcc/testsuite/ChangeLog:
PR analyzer/93367
* gcc.dg/analyzer/abort.c: Remove include of <assert.h>.
Replace use of assert with a custom assertion implementation.
In libgo CacheLinePadSize is defined by the generated file cpugen.go.
Keep cpu_riscv64.go around, even though it is now empty, so that
we will pick up changes to it in future merges.
Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/216077
I've noticed this test failed on x86_64-linux with -m32 or -mx32 testing,
the triplet doesn't really say which actual multilib it is, and the test
really works with lp64.
2020-01-23 Jakub Jelinek <jakub@redhat.com>
PR inline-asm/93027
* gcc.target/i386/pr93027.c: Require lp64 target rather x86_64-*-*.
writeargv writes out empty arguments in a way that expandargv skips
them instead of preserving them. Fixed by writing out a pair of
quotes for them.
for libiberty/ChangeLog
* argv.c (writeargv): Output empty args as "".
The C++ headers #undef the functions we are testing for, just in case
they're implemented as macros, so do that in the cross math decl tests
as well.
for libstdc++-v3/ChangeLog
* crossconfig.m4 (GLIBCXX_CHECK_MATH_DECL): Reject macros.
* configure: Rebuild.
Padding in mbstate_t objects may get the memcmp to fail.
Attempt to avoid the failure with zero initialization.
for libstdc++-v3/ChangeLog
* testsuite/27_io/fpos/mbstate_t/1.cc: Zero-init mbstate_t.
The following testcase is miscompiled, because the postreload pass changes:
-(insn 14 13 23 2 (parallel [
- (set (reg:DI 1 dx [94])
- (plus:DI (reg:DI 1 dx [95])
- (reg:DI 5 di [92])))
- (clobber (reg:CC 17 flags))
- ]) "pr93402.c":8:30 186 {*adddi_1}
- (expr_list:REG_EQUAL (plus:DI (reg:DI 5 di [92])
- (const_int 111111111111 [0x19debd01c7]))
- (nil)))
-(insn 23 14 25 2 (set (reg:SI 0 ax)
+(insn 23 13 25 2 (set (reg:SI 0 ax)
(const_int 0 [0])) "pr93402.c":10:1 67 {*movsi_internal}
(nil))
(insn 25 23 26 2 (use (reg:SI 0 ax)) "pr93402.c":10:1 -1
(nil))
-(insn 26 25 35 2 (use (reg:DI 1 dx)) "pr93402.c":10:1 -1
+(insn 26 25 35 2 (use (plus:DI (reg:DI 1 dx [95])
+ (reg:DI 5 di [92]))) "pr93402.c":10:1 -1
(nil))
A USE insn is not a normal insn and verify_changes called from
apply_change_group is happy about any changes into it.
The following patch avoids this optimization if we were to change
the USE operand (this routine only changes a reg into (plus reg reg2)).
2020-01-23 Jakub Jelinek <jakub@redhat.com>
PR rtl-optimization/93402
* postreload.c (reload_combine_recognize_pattern): Don't try to adjust
USE insns.
* gcc.c-torture/execute/pr93402.c: New test.
A rather simple ICE where we failed to properly check for concept-ids
uses in nested-name-specifiers.
Tested x86_64-linux.
/cp
PR c++/92804
* parser.c (cp_parser_nested_name_specifier_opt): Properly
diagnose concept-ids.
/testsuite
PR c++/92804
* g++.dg/concepts/pr92804-1.C: New.
* g++.dg/concepts/pr92804-2.C: New.
If we can't change the argument to &"...", use the original arg instead of
the partially munged one.
PR c++/93331 - ICE with __builtin_strchr.
* constexpr.c (cxx_eval_builtin_function_call): Use the original
argument if we didn't manage to extract a STRING_CST.
In a template we don't instantiate a deferred noexcept-spec, and we don't
need it because we aren't going to do anything with the value of
throwing_cleanup in a template anyway.
PR c++/93345 - ICE with defaulted dtor and template.
PR c++/33799
* decl.c (cxx_maybe_build_cleanup): Don't try to set
throwing_cleanup in a template.
PR analyzer/93375 reports an ICE under certain circumstances
involving a call where the number of arguments at the callsite
is less than the parameter count of the callee,
Specifically, the ICE occurs when pruning a checker_path for a
diagnostic, when attempting to maintain which expression is of
interest through such a call.
The root cause is an assumption that there were enough arguments at
the callsite, within callgraph_superedge's methods for mapping
expressions between callee and caller.
This patch adds checks for this to the relevant methods, fixing the ICE.
gcc/analyzer/ChangeLog:
PR analyzer/93375
* supergraph.cc (callgraph_superedge::get_arg_for_parm): Fail
gracefully is the number of parameters at the callee exceeds the
number of arguments at the call stmt.
(callgraph_superedge::get_parm_for_arg): Likewise.
gcc/testsuite/ChangeLog:
PR analyzer/93375
* gcc.dg/analyzer/pr93375.c: New test.
My last patch didn't fix all the failures on unsignd char targets. We were
missing one warning because by suppressing -Wsign-conversion for the second
operand of + we missed an overflow that we want to warn about, and we
properly don't warn about unsigned / or %.
PR testsuite/93391 - PR 40752 test fails with unsigned plain char.
* c-warn.c (conversion_warning): Change -Wsign-conversion handling.
* lib/target-supports.exp (check_effective_target_unsigned_char):
New.
libgcc/ChangeLog:
2020-01-23 Dragan Mladjenovic <dmladjenovic@wavecomp.com>
* config/mips/gnustack.h: Check for TARGET_LIBC_GNUSTACK also.
gcc/ChangeLog:
2020-01-23 Dragan Mladjenovic <dmladjenovic@wavecomp.com>
* config.in: Regenerated.
* config/mips/linux.h (NEED_INDICATE_EXEC_STACK): Define to 1
for TARGET_LIBC_GNUSTACK.
* configure: Regenerated.
* configure.ac: Define TARGET_LIBC_GNUSTACK if glibc version is
found to be 2.31 or greater.
The following testcase ICEs, because during try_combine of i3:
(insn 18 17 19 2 (parallel [
(set (reg:CCO 17 flags)
(eq:CCO (plus:OI (sign_extend:OI (reg:TI 96))
(const_int 1 [0x1]))
(sign_extend:OI (plus:TI (reg:TI 96)
(const_int 1 [0x1])))))
(set (reg:TI 98)
(plus:TI (reg:TI 96)
(const_int 1 [0x1])))
]) "pr93376.c":8:10 223 {*addvti4_doubleword_1}
(expr_list:REG_UNUSED (reg:TI 98)
(expr_list:REG_DEAD (reg:TI 96)
(nil))))
and i2:
(insn 17 37 18 2 (set (reg:TI 96)
(const_wide_int 0x7fffffffffffffffffffffffffffffff)) "pr93376.c":8:10 65 {*movti_internal}
(nil))
the eq in there gets simplified into:
(eq:CCO (const_wide_int 0x080000000000000000000000000000000)
(const_wide_int 0x80000000000000000000000000000000))
and simplify-rtx.c tries to simplify it by simplifying MINUS
of the two operands.
Now, i386 defines MAX_BITSIZE_MODE_ANY_INT to 128, because OImode
and XImode are used mainly as a placeholder for the vector modes;
these new signed overflow patterns are an exception to that,
but what they really need is just TImode precision + 1 (maybe 2 worst case)
bits at any time.
wide-int.h defines WIDE_INT_MAX_ELTS in a way that it contains one more
HWI above number of HWIs to cover WIDE_INT_MAX_ELTS, so on i386 that is
3 HWIs, meaning that TImode precision + 1/2 bits is still representable in
there. Unfortunately, the way wi::sub_large is implemented, it needs
not just those 3 HWIs, but one HWI above the maximum of the lengths of
both operands, which means it buffer overflows, overwrites the following
precision in wide_int_storage and ICEs later on. The need for 4 HWIs is
only temporary, because canonize immediately after it canonicalizes it
back to 3 HWIs only.
The patch is something suggested by Richard S., avoid using OImode
for this and instead use a partial int mode that is smaller.
2020-01-23 Jakub Jelinek <jakub@redhat.com>
PR target/93376
* config/i386/i386-modes.def (POImode): New mode.
(MAX_BITSIZE_MODE_ANY_INT): Change from 128 to 160.
* config/i386/i386.md (DPWI): New mode attribute.
(addv<mode>4, subv<mode>4): Use <DPWI> instead of <DWI>.
(QWI): Rename to...
(QPWI): ... this. Use POI instead of OI for TImode.
(*addv<dwi>4_doubleword, *addv<dwi>4_doubleword_1,
*subv<dwi>4_doubleword, *subv<dwi>4_doubleword_1): Use <QPWI>
instead of <QWI>.
* gcc.dg/pr93376.c: New test.
The previous rule could leave an incomplete file if the build was
interrupted, which would then not be remade if make was run again.
This makes the rule more robust by writing to a temporary file and only
moving it into place as the final step. It also simplifies the rule so
that only the essential macro definitions are written to the file, not
the explanatory comments and commented out #undef lines.
Also, the macro for enabling LFS on Mac OS X 10.5 is now set
unconditionally, which is a bug fix from upstream autoconf.
PR libstdc++/91947
* include/Makefile.am (${host_builddir}/largefile-config.h): Simplify
rule.
* include/Makefile.in: Regenerate.
We can't yet represent the inverse of all conditions in rtl
(see g:865257c447cc50f5819e), triggering an ICE in the pass
that handles -mtrack-speculation. Since we don't expect these
insns to be optimised in any way, the easiest fix seemed to be
to add an insn that reverses the condition internally.
2020-01-23 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR target/93341
* config/aarch64/aarch64.md (UNSPEC_SPECULATION_TRACKER_REV): New
unspec.
(speculation_tracker_rev): New pattern.
* config/aarch64/aarch64-speculation.cc (aarch64_do_track_speculation):
Use speculation_tracker_rev to track the inverse condition.
gcc/testsuite/
PR target/93341
* gcc.target/aarch64/pr93341.c: New test.
gcc/testsuite/ChangeLog:
* gcc.dg/analyzer/data-model-3.c: Remove hardcoded "-O2" and move
to torture/conftest-1.c.
* gcc.dg/analyzer/torture/analyzer-torture.exp: New.
* gcc.dg/analyzer/torture/conftest-1.c: Move here from
analyzer/data-model-3.c.
* gcc.dg/analyzer/torture/poc.c: New test.
This fixes tracking of the alias-set of partial defs for use by
redundant store removal.
2020-01-23 Richard Biener <rguenther@suse.de>
PR tree-optimization/93381
* tree-ssa-sccvn.c (vn_walk_cb_data::push_partial_def): Take
alias-set of the def as argument and record the first one.
(vn_walk_cb_data::first_set): New member.
(vn_reference_lookup_3): Pass the alias-set of the current def
to push_partial_def. Fix alias-set used in the aggregate copy
case.
(vn_reference_lookup): Consistently set *last_vuse_ptr.
* real.c (clear_significand_below): Fix out-of-bound access.
* gcc.dg/torture/pr93354.c: New testcase.
The bzhi patterns are quite complicated because they need to accurately
describe the behavior of the instruction for all input values.
The following patterns are simple and make bzhi recognizable even for
cases where not all input values are valid, because the user used
a shift, in which case the low 8 bit of the last operand need to be in
between 0 and precision-1.
2020-01-23 Jakub Jelinek <jakub@redhat.com>
PR target/93346
* config/i386/i386.md (*bmi2_bzhi_<mode>3_2, *bmi2_bzhi_<mode>3_3):
New define_insn patterns.
* gcc.target/i386/pr93346.c: New test.
The definitions of the integer types for ILP32 newlib make the
resolution of some of the bool-related tests ambiguous.
2020-01-23 Richard Sandiford <richard.sandiford@arm.com>
gcc/testsuite/
* g++.target/aarch64/sve/acle/general-c++/whilele_1.C: Skip for ILP32.
check-function-bodies allows individual function tests to be
annotated with target/xfail selectors, but sometimes it's
useful to have the same selector for all functions.
2020-01-23 Richard Sandiford <richard.sandiford@arm.com>
gcc/
* doc/sourcebuild.texi (check-function-bodies): Add an
optional target/xfail selector.
gcc/testsuite/
* lib/scanasm.exp (check-function-bodies): Add an optional
target/xfail selector.
In this PR, auto-inc-dec was trying to turn:
(set (reg X) (plus (reg X) (const_int N)))
(clobber (mem (reg X)))
into:
(clobber (mem (pre_modify (reg X) ...)))
But bare clobber insns are just there to describe dataflow. They're
not supposed to generate any code.
2020-01-23 Richard Sandiford <richard.sandiford@arm.com>
gcc/
PR rtl-optimization/93124
* auto-inc-dec.c (merge_in_block): Don't add auto inc/decs to
bare USE and CLOBBER insns.
gcc/testsuite/
* gcc.dg/torture/pr93124.c: New test.
Type descriptors are normally weak and nm will report them as V,
so we will skip them when collecting the list of symbols.
But when not using GNU nm, they may be reported as D,
so also skip them in symstogo.
This fixes go/doc/check on Solaris.
Reviewed-on: https://go-review.googlesource.com/c/gofrontend/+/215900
The ICE in PR analyzer/93382 is a validation error.
The global variable "idx" acquires a "tainted" state from local array
n1[0]. When the frame is popped, the svalue for n1[0] is purged, but
the "taint" sm_state_map's entry for "idx" has a svalue_id referencing
the now-purged svalue. This is caught by program_state::validate as an
assertion failure.
This patch fixes the issue by resetting the origin id within
sm_state_map entries for the case where the origin id has been purged.
gcc/analyzer/ChangeLog:
PR analyzer/93382
* program-state.cc (sm_state_map::on_svalue_purge): If the
entry survives, but the origin is being purged, then reset the
origin to null.
gcc/testsuite/ChangeLog:
PR analyzer/93382
* gcc.dg/analyzer/pr93382.c: New test.
When adding namespaces to the analyzer in
r10-6151-g75038aa6aa5b562e6358108619d66ef2ccab9a53
I messed up the nesting of the #endif for #if CHECKING_P
and the closing of namespace ana.
This patch fixes it.
gcc/analyzer/ChangeLog:
* sm-signal.cc: Fix nesting of CHECKING_P and namespace ana.
The problem here is arc looks at current_output_insn unconditional
but sometimes current_output_insn is NULL. With patchable-function-entry,
it will be. This is similar to how the nios2, handles "%.".
Committed as obvious after a simple test with -fpatchable-function-entry=1.
ChangeLog:
* config/arc/arc.c (output_short_suffix): Check insn for nullness.
This reverts commit 9085381f1931cc3667412c8fff91878184835901 as it was
causing default dg-do to be set incorrectly on most targets.
Instead move testcases that are vect related testcase that
use "dg-require-effective-target vect_*" to the vect test area.
ChangeLog:
* gcc.dg/tree-ssa/pr88497-1.c: Move to ...
* gcc.dg/vect/pr88497-1.c: This.
* gcc.dg/tree-ssa/pr88497-2.c: Move to ...
* gcc.dg/vect/pr88497-2.c: This.
* gcc.dg/tree-ssa/pr88497-3.c: Move to ...
* gcc.dg/vect/pr88497-3.c: This.
* gcc.dg/tree-ssa/pr88497-4.c: Move to ...
* gcc.dg/vect/pr88497-4.c: This.
* gcc.dg/tree-ssa/pr88497-5.c: Move to ...
* gcc.dg/vect/pr88497-5.c: This.
* gcc.dg/tree-ssa/pr88497-6.c: Move to ...
* gcc.dg/vect/pr88497-6.c: This.
* gcc.dg/tree-ssa/pr88497-7.c: Move to ...
* gcc.dg/vect/pr88497-7.c: This.
Revert:
* tree-ssa.exp: Set DEFAULT_VECTCFLAGS and DEFAULT_VECTCFLAGS.
Call check_vect_support_and_set_flags also.
blender in SPEC is built with -funsigned-char, which is also the default on
PPC, and exposed -Wsign-conversion issues that weren't seen by the x86_64
testsuite. In blender we were complaining about operands to an expression
that we didn't't previously complain about as a whole. So only check
operands after we check the whole expression. Also, to fix the PR 40752
testcases on -funsigned-char targets, don't consider -Wsign-conversion for
the second operand of PLUS_EXPR, especially since fold changes
"x - 5" to "x + (-5)". And don't use SCHAR_MAX with plain char.
PR testsuite/93391 - PR 40752 test fails with unsigned plain char.
PR c++/40752
* c-warn.c (conversion_warning): Check operands only after checking
the whole expression. Don't check second operand of + for sign.
My earlier change removed the warning calls from this function, so the
location is no longer useful.
* c-common.c (unsafe_conversion_p): Remove location parm.
tree-ssa testcases sometimes check autovect effective target
but does not set it up. On MIPS, those testcases fail with
some TCL error messages. This fixes the issue by calling
check_vect_support_and_set_flags inside tree-ssa.exp.
There might be other .exp files which need to be done this
way too but I have not checked all of them.
Tested on x86_64-linux-gnu and a cross to mips64-octeon-linux-gnu.
Both full run of the testsuite and running tree-ssa.exp by itself.
testsuite/ChangeLog:
* tree-ssa.exp: Set DEFAULT_VECTCFLAGS and DEFAULT_VECTCFLAGS.
Call check_vect_support_and_set_flags also.
