This patch adds authentication for when the stack is unwound when an
exception is taken. All the changes here are done to the runtime code
in libgcc's unwinder code for Arm target. All the changes are guarded
under defined (__ARM_FEATURE_PAC_DEFAULT) and activated only if the
+pacbti feature is switched on for the architecture. This means that
switching on the target feature via -march or -mcpu is sufficient and
-mbranch-protection need not be enabled. This ensures that the
unwinder is authenticated only if the PACBTI instructions are
available in the non-NOP space as it uses AUTG. Just generating
PAC/AUT instructions using -mbranch-protection will not enable
authentication on the unwinder.
Pre-approved with the requested changes here
<https://gcc.gnu.org/pipermail/gcc-patches/2021-December/586555.html>.
gcc/ChangeLog:
* ginclude/unwind-arm-common.h (_Unwind_VRS_RegClass): Introduce
new pseudo register class _UVRSC_PAC.
libgcc/ChangeLog:
* config/arm/pr-support.c (__gnu_unwind_execute): Decode
exception opcode (0xb4) for saving RA_AUTH_CODE and authenticate
with AUTG if found.
* config/arm/unwind-arm.c (struct pseudo_regs): New.
(phase1_vrs): Introduce new field to store pseudo-reg state.
(phase2_vrs): Likewise.
(_Unwind_VRS_Get): Load pseudo register state from virtual reg set.
(_Unwind_VRS_Set): Store pseudo register state to virtual reg set.
(_Unwind_VRS_Pop): Load pseudo register value from stack into VRS.
Co-Authored-By: Tejas Belagod <tbelagod@arm.com>
Co-Authored-By: Srinath Parvathaneni <srinath.parvathaneni@arm.com>
This patch implements target feature macros when PACBTI is enabled
through the -march option or -mbranch-protection. The target feature
macros __ARM_FEATURE_PAC_DEFAULT and __ARM_FEATURE_BTI_DEFAULT are
specified in ARM ACLE
<https://developer.arm.com/documentation/101028/0012/5--Feature-test-macros?lang=en>
__ARM_FEATURE_PAUTH and __ARM_FEATURE_BTI are specified in the
pull-request <https://github.com/ARM-software/acle/pull/55>.
Approved here
<https://gcc.gnu.org/pipermail/gcc-patches/2021-December/586334.html>.
gcc/
* config/arm/arm-c.cc (arm_cpu_builtins): Define
__ARM_FEATURE_BTI_DEFAULT, __ARM_FEATURE_PAC_DEFAULT,
__ARM_FEATURE_PAUTH and __ARM_FEATURE_BTI.
gcc/testsuite/
* lib/target-supports.exp
(check_effective_target_mbranch_protection_ok): New function.
* gcc.target/arm/acle/pacbti-m-predef-2.c: New test.
* gcc.target/arm/acle/pacbti-m-predef-4.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-5.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-8.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-9.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-10.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-11.c: Likewise.
* gcc.target/arm/acle/pacbti-m-predef-12.c: Likewise.
Co-Authored-By: Tejas Belagod <tbelagod@arm.com>
Add -mbranch-protection option. This option enables the
code-generation of pointer signing and authentication instructions in
function prologues and epilogues.
gcc/ChangeLog:
* config/arm/arm.cc (arm_configure_build_target): Parse and validate
-mbranch-protection option and initialize appropriate data structures.
* config/arm/arm.opt (-mbranch-protection): New option.
* doc/invoke.texi (Arm Options): Document it.
Co-Authored-By: Tejas Belagod <tbelagod@arm.com>
Co-Authored-By: Richard Earnshaw <Richard.Earnshaw@arm.com>
This patch adds the -march feature +pacbti to Armv8.1-M Mainline.
This feature enables pointer signing and authentication instructions
on M-class architectures.
Pre-approved here
<https://gcc.gnu.org/pipermail/gcc-patches/2021-December/586144.html>.
gcc/Changelog:
* config/arm/arm.h (TARGET_HAVE_PACBTI): New macro.
* config/arm/arm-cpus.in (pacbti): New feature.
* doc/invoke.texi (Arm Options): Document it.
Co-Authored-By: Tejas Belagod <tbelagod@arm.com>
Hi all,
This change refactors all the mbranch-protection option parsing code and
types to make it common to both AArch32 and AArch64 backends.
This change also pulls in some supporting types from AArch64 to make
it common (aarch_parse_opt_result).
The significant changes in this patch are the movement of all branch
protection parsing routines from aarch64.c to aarch-common.c and
supporting data types and static data structures.
This patch also pre-declares variables and types required in the
aarch32 back-end for moved variables for function sign scope and key
to prepare for the impending series of patches that support parsing
the feature mbranch-protection in the aarch32 back-end.
gcc/ChangeLog:
* common/config/aarch64/aarch64-common.cc: Include aarch-common.h.
(all_architectures): Fix comment.
(aarch64_parse_extension): Rename return type, enum value names.
* config/aarch64/aarch64-c.cc (aarch64_update_cpp_builtins): Rename
factored out aarch_ra_sign_scope and aarch_ra_sign_key variables.
Also rename corresponding enum values.
* config/aarch64/aarch64-opts.h (aarch64_function_type): Factor
out aarch64_function_type and move it to common code as
aarch_function_type in aarch-common.h.
* config/aarch64/aarch64-protos.h: Include common types header,
move out types aarch64_parse_opt_result and aarch64_key_type to
aarch-common.h
* config/aarch64/aarch64.cc: Move mbranch-protection parsing types
and functions out into aarch-common.h and aarch-common.cc. Fix up
all the name changes resulting from the move.
* config/aarch64/aarch64.md: Fix up aarch64_ra_sign_key type name change
and enum value.
* config/aarch64/aarch64.opt: Include aarch-common.h to import
type move. Fix up name changes from factoring out common code and
data.
* config/arm/aarch-common-protos.h: Export factored out routines to both
backends.
* config/arm/aarch-common.cc: Include newly factored out types.
Move all mbranch-protection code and data structures from
aarch64.cc.
* config/arm/aarch-common.h: New header that declares types shared
between aarch32 and aarch64 backends.
* config/arm/arm-protos.h: Declare types and variables that are
made common to aarch64 and aarch32 backends - aarch_ra_sign_key,
aarch_ra_sign_scope and aarch_enable_bti.
* config/arm/arm.opt (config/arm/aarch-common.h): Include header.
(aarch_ra_sign_scope, aarch_enable_bti): Declare variable.
* config/arm/arm.cc: Add missing includes.
Co-Authored-By: Tejas Belagod <tbelagod@arm.com>
The previous change to avoid a duplicate multi directory
caused the m2/m2{cor,iso,log,min,pim} installs to happen
relative to the main library directory when not using
--enable-version-specific-runtime-libs which doesn't match
the drivers expectation where to find them. The following
fixes the original issue by simply dropping the duplicate
multi directory since the one in the $(inst_libdir) variable
now works.
Tested by building and installing with and without
--enable-version-specific-runtime-libs and compiling and
linking a modula-2 testcase successfully with the installed
compilers.
PR modula2/108144
libgm2/
* libm2cor/Makefile.am: Revert previous change, instead
drop the redundant $(MULTIDIR).
* libm2iso/Makefile.am: Likewise.
* libm2log/Makefile.am: Likewise.
* libm2min/Makefile.am: Likewise.
* libm2pim/Makefile.am: Likewise.
* libm2cor/Makefile.in: Regenerate.
* libm2iso/Makefile.in: Likewise.
* libm2log/Makefile.in: Likewise.
* libm2min/Makefile.in: Likewise.
* libm2pim/Makefile.in: Likewise.
Before, newlib 3.2 was required for amdgcn and 3.1 for nvptx.
Now recommended is 4.3.0 which was just released on 2023-01-20.
While currently the old versions would work fine, upcoming GCC
changes depend on a newer newlib. Thus, the minimal version is
bumped instead of just recommending the new version.
For GCN, the bump is in preparation for permitting non-threadlocal
stack variables and vectorized math functions - both scheduled for
GCC 13 and added to newlib in 4.3.0.
For nvptx, this includes an emulated clock (commit 6bb96d13a),
a calloc fix (5fca4e0f1) and changes to permit libgfortran to be
compiled with I/O support instead of only in minimal mode.
(Patch approved for GCC 13 but pending on a nvtpx patch,
which for which review is pending.)
gcc/ChangeLog:
* doc/install.texi (amdgcn, nvptx): Require newlib 4.3.0.
When we have a static declaration without definition we diagnose
that and turn it into an extern declaration. That can alter
the outcome of maybe_special_function_p here and there's really
no point in doing that, so don't.
PR tree-optimization/108449
* cgraphunit.cc (check_global_declaration): Do not turn
undefined statics into externs.
* gcc.dg/pr108449.c: New testcase.
The following adjusts libgm2 to properly use the multilib build
infrastructure, thereby fixing the install with
--enable-version-specific-runtime-libs
In particular config-ml.pl needs to be applied to generated Makefiles
as documented in the manual and we have to avoid clobbering the
variables via make arguments. The explicit install rules used different
ways to construct the multilib dir which isn't necessary and breaks
when MUTLIDIR is now finally set correctly. Instead use
$(toolexeclibdir).
This results in some dead variables in the Makefile.am (and there were
some before), I refrained from doing even more changes here.
Verified with an install with and without --enable-version-specific-runtime-libs
and checking the result.
PR modula2/108144
libgm2/
* configure.ac: Apply config-ml.pl to the generated Makefiles.
Set multilib_arg, use AM_PROG_LIBTOOL.
* configure: Regenerate.
* Makefile.am (AM_MAKEFLAGS): Do not override MULTI* flags.
* Makefile.in: Regenerate.
* libm2cor/Makefile.am: Install to $(toolexeclibdir)$(M2LIBDIR)
rather than $(inst_libdir)/$(MULTIDIR)$(M2LIBDIR).
* libm2iso/Makefile.am: Likewise.
* libm2log/Makefile.am: Likewise.
* libm2min/Makefile.am: Likewise.
* libm2pim/Makefile.am: Likewise.
* libm2cor/Makefile.in: Regenerate.
* libm2iso/Makefile.in: Likewise.
* libm2log/Makefile.in: Likewise.
* libm2min/Makefile.in: Likewise.
* libm2pim/Makefile.in: Likewise.
This PR points out a case where we are not extending the lifetime of a
temporary when the subobject is denoted by a pointer-to-member operation.
These rules were clarified in C++20 by CWG1299.
There are other cases that also need to be handled under CWG1299, but are
not fixed by this patch.
PR c++/53288
DR 1299
gcc/cp/ChangeLog:
* call.cc (extend_ref_init_temps_1): Handle ptrmem expression.
gcc/testsuite/ChangeLog:
* g++.dg/init/lifetime4.C: New test.
The original implementation for Modula-2 search paths allows things like
'-I/path/a:/path/b'. Such paths are not compatible with the rest of the
compiler (in particular, the preprocessor, which Modula-2 uses). In
preparation for removing that idiom, this patch rewrites the cases in the
testsuite and removes string processing where possible from lists of paths.
TODO: There are some apparent discrepancies and/or extraneous insertions
of partial include paths (especially in gm2_init_log, gm2_init_iso and
gm2_init_pim) - to be handled in some future patch.
Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
gcc/testsuite/ChangeLog:
* gm2/case/pass/case-pass.exp: Update for removal of concatenated
paths.
* gm2/complex/pass/complex-pass.exp: Likewise.
* gm2/coroutines/pim/run/pass/coroutines-pim-run-pass.exp: Likewise.
* gm2/iso/analysis/fail/iso-analysis-fail.exp: Likewise.
* gm2/iso/check/fail/iso-check-fail.exp: Likewise.
* gm2/iso/fail/iso-fail.exp: Likewise.
* gm2/iso/pass/iso-pass.exp: Likewise.
* gm2/isolib/run/pass/isolib-run-pass.exp: Likewise.
* gm2/pim/fail/pim-fail.exp: Likewise.
* gm2/pim/pass/pim-pass.exp: Likewise.
* gm2/pimlib/logitech/run/pass/pimlib-logitech-run-pass.exp: Likewise.
* gm2/pimlib/pass/pimlib-pass.exp: Likewise.
* gm2/pimlib/run/pass/pimlib-run-pass.exp: Likewise.
* gm2/projects/iso/run/pass/halma/projects-iso-run-pass-halma.exp: Likewise.
* gm2/projects/iso/run/pass/hello/projects-iso-run-pass-hello.exp: Likewise.
* gm2/projects/log/run/pass/hello/projects-log-run-pass-hello.exp: Likewise.
* gm2/projects/pim/run/pass/hello/projects-pim-run-pass-hello.exp: Likewise.
* gm2/recover/pass/recover-pass.exp: Likewise.
* gm2/switches/makeall/fail/switches-makeall-fail.exp: Likewise.
* gm2/switches/makeall/pass/switches-makeall-pass.exp: Likewise.
* gm2/switches/none/run/pass/gm2-none.exp: Likewise.
* gm2/switches/pim2/run/pass/switches-pim2-run-pass.exp: Likewise.
* gm2/ulmlib/pass/ulmlib-pass.exp: Likewise.
* gm2/ulmlib/std/pass/ulmlib-std-pass.exp: Likewise.
* gm2/ulmlib/sys/pass/ulmlib-sys-pass.exp: Likewise.
* lib/gm2.exp: Make the paths list a mandatory entry for each gm2_init_xxx
function. Remove the use of concatenated include and library paths.
Remove string processing where possible.
The recent gcc.dg/tree-ssa/clz-char.c test case failed for PRU target,
exposing a wrong code generation bug in the PRU backend. The "clz"
pattern did not produce correct output for QI and HI input operand
modes. SI mode is ok.
The "clz" pattern is expanded to an LMBD instruction to get the
left-most bit position having value "1". In turn, to get the correct
"clz" value, that bit position must be subtracted from the MSB bit
position of the input operand. The old behaviour of hard-coding 31
for MSB bit position is wrong.
The LMBD instruction returns 32 if input operand is zero, irrespective
of its register mode. This maps nicely for SI mode, where the "clz"
pattern outputs -1. It also leads to peculiar (but valid!) output
values from the "clz" pattern for QI and HI zero-valued inputs.
gcc/ChangeLog:
* config/pru/pru.h (CLZ_DEFINED_VALUE_AT_ZERO): Fix value for QI
and HI input modes.
* config/pru/pru.md (clz): Fix generated code for QI and HI
input modes.
gcc/testsuite/ChangeLog:
* gcc.target/pru/clz-hi-2.c: New test.
* gcc.target/pru/clz-hi.c: New test.
Signed-off-by: Dimitar Dimitrov <dimitar@dinux.eu>
For several Darwin linker versions, the link paths are first traversed
looking for shared libraries, and then again looking for archives.
We want the paths to be searched for both shared and archives in order
so that we will find the non-shared libffi in the testsuite (otherwise
the installed version in the system will be found instead).
Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
libffi/ChangeLog:
* testsuite/lib/libffi.exp: Search for both shared and archive
in each path in turn.
The macOS 13 SDK has unguarded Apple Blocks use in objc/runtime.h which
causes most of the objective-c tests to fail.
Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
fixincludes/ChangeLog:
* fixincl.x: Regenerate.
* inclhack.def (darwin_objc_runtime_1): New hack.
* tests/base/objc/runtime.h: New file.
The SDK for MacOS13 includes Apple-specific deprecations of some functions that
are not deprecated in Posix, C or C++ and widely used in GCC.
The fix makes the deprecation conditional on __APPLE_LOCAL_DEPRECATIONS so that
end users may still observe them but they are hidden from normal compilations.
Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
PR target/107568
fixincludes/ChangeLog:
* fixincl.x: Regenerate.
* inclhack.def: Add a fix for MacOS13 SDK function deprecations
in stdio.h.
* tests/base/stdio.h (__deprecated_msg): New test.
R-108136 modula-2 meets cppcheck part 2
This patch addresses the style warnings mentioned in the PR.
The array high style fixes to mc/keyc.mod and
gm2-libs/M2Dependent.mod provoke a rebuild of pge and mc.
gcc/m2/ChangeLog:
* gm2-libs/Args.mod (GetArg): Check index before
accessing array.
* gm2-libs/M2Dependent.mod (toCString): Check index
before accessing array.
* mc-boot/GArgs.c: Rebuilt.
* mc-boot/GM2Dependent.c: Rebuilt.
* mc-boot/Gkeyc.c: Rebuilt.
* mc/keyc.mod (seenUIntMin): Initialize once.
(seenUIntMax): Initialize once.
(checkLimits): Only test seenUIntMin and seenUIntMax
once.
* pge-boot/GArgs.c: Rebuilt.
* pge-boot/GM2Dependent.c: Rebuilt.
Signed-off-by: Gaius Mulley <gaiusmod2@gmail.com>
Change variables holding PC values from uint64_t to uintptr_t.
Patch by Björn Schäpers.
* dwarf.c (struct function_addrs): Change low and high fields to
uintptr_t.
(struct unit_addrs): Likewise.
(resolve_addr_index): Change address parameter to uintptr_t*.
(add_unit_addr): Change lowpc and highpc parameters to uintptr_t.
(add_function_range): Likewise.
(struct pcrange): Change lowpc and highpc fields to uintptr_t.
(add_low_high_range): Change add_range lowpc and highpc parameters
to uintptr_t.
(add_ranges_from_ranges): Likewise.
(add_ranges_from_rnglists): Likewise.
(add_low_high_range): Chnage lowpc and highpc variables to
uintpr_t.
(add_ranges_from_rnglists): Change some local variables to
uintptr_t.
(add_ranges_from_ranges): Change base parameter to uintptr_t.
(add_ranges_from_rnglists): Likewise.
(read_function_entry): Likewise.
(resolve_addr_index): Add explicit casts to uintptr_t.
(update_pcrange): Likewise.
(add_ranges_from_ranges): Likewise.
(add_ranges_from_rnglists): Likewise.
(read_function_entry): Likewise.
The 64-bit variant of PMULL{2} and AES instructions are available if FEAT_AES
is implemented according to the Arm ARM [1]. Similarly FEAT_SHA1 and
FEAT_SHA256 enable the use of SHA1 and SHA256 instruction variants.
This patch fixes arm_neon.h to correctly reflect the feature availability based
on '+aes' and '+sha2' as opposed to the ambiguous catch-all '+crypto'.
[1] Section D17.2.61, C7.2.215
2022-01-11 Tejas Belagod <tejas.belagod@arm.com>
gcc/ChangeLog:
* config/aarch64/arm_neon.h (vmull_p64, vmull_high_p64, vaeseq_u8,
vaesdq_u8, vaesmcq_u8, vaesimcq_u8): Gate under "nothing+aes".
(vsha1*_u32, vsha256*_u32): Gate under "nothing+sha2".
gcc/testsuite/ChangeLog:
* gcc.target/aarch64/acle/pmull64.c: New.
* gcc.target/aarch64/aes-fuse-1.c: Replace '+crypto' with corresponding
feature flag based on the intrinsic.
* gcc.target/aarch64/aes-fuse-2.c: Likewise.
* gcc.target/aarch64/aes_1.c: Likewise.
* gcc.target/aarch64/aes_2.c: Likewise.
* gcc.target/aarch64/aes_xor_combine.c: Likewise.
* gcc.target/aarch64/sha1_1.c: Likewise.
* gcc.target/aarch64/sha256_1.c: Likewise.
* gcc.target/aarch64/target_attr_crypto_ice_1.c: Likewise.
tree-ssa-loop-niter.cc (build_cltz_expr) gets unused variable mode
warning on some architectures where C[LT]Z_DEFINED_VALUE_AT_ZERO
macro(s) don't use the first argument (which includes the
defaults.h definitions of:
#define CLZ_DEFINED_VALUE_AT_ZERO(MODE, VALUE) 0
#define CTZ_DEFINED_VALUE_AT_ZERO(MODE, VALUE) 0
Other uses of this macro avoid this problem by avoiding temporaries
which are only used as argument to those macros, the following patch
does it the same way for consistency. Plus some formatting fixes
while at it.
2023-01-20 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/108457
* tree-ssa-loop-niter.cc (build_cltz_expr): Use
SCALAR_INT_TYPE_MODE (utype) directly as C[LT]Z_DEFINED_VALUE_AT_ZERO
argument instead of a temporary. Formatting fixes.
As the following testcases shows, when adding static operator[]
support I've missed that the 2 build_min_non_dep_op_overload functions
need to be adjusted. The first one we only use for the single index
case, but as cp_tree_code_length (ARRAY_REF) is 2, we were running
into an assertion there which compared nargs and expected_nargs.
For ARRAY_REF, the operator[] is either a non-static member or newly
static member, never out of class and for the static member case
if user uses single index the operator[] needs to have a single
argument as well, but the function is called with 2 - the object
it is invoked on and the index. We need to evaluate side-effects
of the object and use just a single argument in the call - the index.
The other build_min_non_dep_op_overload overload has been added
solely for ARRAY_REF - CALL_EXPR is the other operator that accepts
variable number of operands but that one goes through different
routines. There we asserted it is a METHOD_TYPE, so again
we shouldn't assert that but handle the case when it is not one
by making sure object's side-effects are evaluated if needed and
passing all the index arguments to the static operator[].
2023-01-19 Jakub Jelinek <jakub@redhat.com>
PR c++/108437
* cp-tree.h (keep_unused_object_arg): Declare.
* call.cc (keep_unused_object_arg): No longer static.
* tree.cc (build_min_non_dep_op_overload): Handle ARRAY_REF
with overload being static member function.
* g++.dg/cpp23/subscript12.C: New test.
* g++.dg/cpp23/subscript13.C: New test.
For anonymous union members we create artificial VAR_DECLs which
have DECL_VALUE_EXPR for the actual COMPONENT_REF. That works
just fine inside of functions (including global dynamic constructors),
because during gimplification such VAR_DECLs are gimplified as
their DECL_VALUE_EXPR. This is also done during regimplification.
But references to these artificial vars in DECL_INITIAL expressions
aren't ever replaced by the DECL_VALUE_EXPRs, so we end up either
with link failures like on the testcase below, or worse ICEs with
LTO.
The following patch fixes those during cp_fully_fold_init where we
already walk all the trees (!data->genericize means that
function rather than cp_fold_function).
2023-01-19 Jakub Jelinek <jakub@redhat.com>
PR c++/53932
* cp-gimplify.cc (cp_fold_r): During cp_fully_fold_init replace
DECL_ANON_UNION_VAR_P VAR_DECLs with their corresponding
DECL_VALUE_EXPR.
* g++.dg/init/pr53932.C: New test.
While looking at the PR, I've noticed one row in rr_union_table
is wrong. relation_union should be commutative, but due to that
bug is not. The following patch adds a self-test for that
property (fails without the first hunk) and fixes that line.
The actual floating point relation problem isn't fixed by this patch
though.
2023-01-19 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/108447
* value-relation.cc (rr_union_table): Fix VREL_UNDEFINED row order.
(relation_tests): Add self-tests for relation_{intersect,union}
commutativity.
* selftest.h (relation_tests): Declare.
* function-tests.cc (test_ranges): Call it.
Check invalid third argument to __builtin_ia32_prefetch when expaning
__builtin_ia32_prefetch to avoid ICE later.
gcc/
PR target/108436
* config/i386/i386-expand.cc (ix86_expand_builtin): Check
invalid third argument to __builtin_ia32_prefetch.
gcc/testsuite/
* gcc.target/i386/pr108436.c: New test.
expand_omp_for_init_counts was using for the case where collapse(2)
inner loop has init expression dependent on non-constant multiple of
the outer iterator and the condition upper bound expression doesn't
depend on the outer iterator fold_unary (NEGATE_EXPR, ...). This
will just return NULL if it can't be folded, we need fold_build1
instead.
2023-01-19 Jakub Jelinek <jakub@redhat.com>
PR middle-end/108459
* omp-expand.cc (expand_omp_for_init_counts): Use fold_build1 rather
than fold_unary for NEGATE_EXPR.
* testsuite/libgomp.c/pr108459.c: New test.
My integration testing [1] of -fanalyzer in GCC 13 is showing a lot of
diagnostics from the new -Wanalyzer-deref-before-check warning on
real-world C projects, and most of these seem to be false positives.
This patch updates the warning to make it much less likely to fire:
- only intraprocedural cases are now reported
- reject cases in which there are control flow paths to the check
that didn't come through the dereference, by looking at BB dominator
information. This fixes a false positive seen in git-2.39.0's
pack-revindex.c: load_revindex_from_disk (PR analyzer/108455), in
which a shared "cleanup:" section checks "data" for NULL, and
depending on how much of the function is executed "data" might or
might not have already been dereferenced.
The counts of -Wanalyzer-deref-before-check diagnostics in [1]
before/after this patch show this improvement:
Known false positives: 6 -> 0 (-6)
Known true positives: 1 -> 1
Unclassified positives: 123 -> 63 (-60)
[1] https://github.com/davidmalcolm/gcc-analyzer-integration-tests
gcc/analyzer/ChangeLog:
PR analyzer/108455
* analyzer.h (class checker_event): New forward decl.
(class state_change_event): Indent.
(class warning_event): New forward decl.
* checker-event.cc (state_change_event::state_change_event): Add
"enode" param.
(warning_event::get_desc): Update for new param of
evdesc::final_event ctor.
* checker-event.h (state_change_event::state_change_event): Add
"enode" param.
(state_change_event::get_exploded_node): New accessor.
(state_change_event::m_enode): New field.
(warning_event::warning_event): New "enode" param.
(warning_event::get_exploded_node): New accessor.
(warning_event::m_enode): New field.
* diagnostic-manager.cc
(state_change_event_creator::on_global_state_change): Pass
src_node to state_change_event ctor.
(state_change_event_creator::on_state_change): Likewise.
(null_assignment_sm_context::set_next_state): Pass NULL for
new param of state_change_event ctor.
* infinite-recursion.cc
(infinite_recursion_diagnostic::add_final_event): Update for new
param of warning_event ctor.
* pending-diagnostic.cc (pending_diagnostic::add_final_event):
Pass enode to warning_event ctor.
* pending-diagnostic.h (evdesc::final_event): Add reference to
warning_event.
* sm-malloc.cc: Include "analyzer/checker-event.h" and
"analyzer/exploded-graph.h".
(deref_before_check::deref_before_check): Initialize new fields.
(deref_before_check::emit): Reject warnings in which we were
unable to determine the enodes of the dereference and the check.
Reject warnings interprocedural warnings. Reject warnings in which
the dereference doesn't dominate the check.
(deref_before_check::describe_state_change): Set m_deref_enode.
(deref_before_check::describe_final_event): Set m_check_enode.
(deref_before_check::m_deref_enode): New field.
(deref_before_check::m_check_enode): New field.
gcc/testsuite/ChangeLog:
PR analyzer/108455
* gcc.dg/analyzer/deref-before-check-1.c: Add test coverage
involving dominance.
* gcc.dg/analyzer/deref-before-check-pr108455-1.c: New test.
* gcc.dg/analyzer/deref-before-check-pr108455-git-pack-revindex.c:
New test.
Signed-off-by: David Malcolm <dmalcolm@redhat.com>
In some case the libraries list was being set before gm2_init_xxx was
called. In some cases it was omitted - this could lead to a difference
between the link libs and the interfaces (the effect of this would be
dependent on the order in which the .exps were run, which makes it also
depend on the -j and the system).
To avoid a mismatch between the module include paths and the added libs
we now make sure that they are both added in the gm_init_xxxx functions
(if finer control over granularity is needed, then we should as a TODO
add a generic gm_init_xxx that takes a library list and ensures that the
imports and libs are matched in the same order).
Also we cannot use a default variable in tcl if the source for that
variable could be absent, but something else follows, there is no way
to put an empty placeholder in.
Signed-off-by: Iain Sandoe <iain@sandoe.co.uk>
gcc/testsuite/ChangeLog:
* gm2/complex/run/pass/complex-run-pass.exp: Remove gm2_link_lib.
* gm2/iso/run/pass/iso-run-pass.exp: Likewise.
* gm2/link/externalscaffold/pass/link-externalscaffold-pass.exp:
* gm2/pimlib/logitech/run/pass/pimlib-logitech-run-pass.exp: Likewise.
* gm2/pimlib/run/pass/pimlib-run-pass.exp: Likewise.
* gm2/projects/iso/run/pass/halma/projects-iso-run-pass-halma.exp:
Likewise.
* gm2/projects/iso/run/pass/hello/projects-iso-run-pass-hello.exp:
Likewise.
* gm2/projects/pim/run/pass/hello/projects-pim-run-pass-hello.exp:
Likewise.
* gm2/sets/run/pass/sets-run-pass.exp: Likewise.
* gm2/switches/none/run/pass/gm2-none.exp: Likewise.
* gm2/switches/pic/run/pass/switches-pic-run-pass.exp: Likewise.
* gm2/projects/pim/run/pass/random/projects-pim-run-pass-random.exp:
Likewise, and also ensure that the -g option is appended to avoid it
being taken as a path.
* lib/gm2.exp: Ensure for each gm2_init_xxxx function that the set of
libraries added matches the set of -I and -L options.
This patch adds the dependencies for automatically generated files used
by the Modula-2 front end.
contrib/ChangeLog:
* gcc_update (files_and_dependencies): Add dependencies for
gcc/m2/gm2config.h.in, gcc/m2/configure,
gcc/m2/gm2-libs/config-host, libgm2/Makefile.in,
libgm2/aclocal.m4, libgm2/libm2cor/Makefile.in,
libgm2/libm2pim/Makefile.in, libgm2/libm2iso/Makefile.in,
libgm2/libm2log/Makefile.in and libgm2/libm2min/Makefile.in.
Signed-off-by: Gaius Mulley <gaiusmod2@gmail.com>
The previous patch added an assert which should not be applied to PST
types (Pure Scalable Types) because alignment does not matter in this
case. This patch moves the assert after the PST case is handled to
avoid the ICE.
PR target/108411
gcc/
* config/aarch64/aarch64.cc (aarch64_layout_arg): Improve
comment. Move assert about alignment a bit later.
The following fixes --enable-version-specific-runtime-libs for
the modula2 target libraries. The issue is that the install
happens via for example
toolexeclib_LTLIBRARIES = libm2cor.la
and toolexeclibdir is set to $(toolexecdir)/$(gcc_version)$(MULTISUBDIR)
but the Makefile.am do not define $(gcc_version) but instead
$(version) which is used locally to define libsubdir. The fix
is to consistently define and use $(gcc_version), also properly
supporting --with-gcc-major-version-only
PR modula2/108144
libgm2/
* configure.ac: Add GCC_BASE_VER.
* configure: Re-generate.
* Makefile.am: Use @get_gcc_base_ver@ for gcc_version.
* libm2cor/Makefile.am: Likewise. Use gcc_version instead
of version.
* libm2iso/Makefile.am: Likewise.
* libm2log/Makefile.am: Likewise.
* libm2min/Makefile.am: Likewise.
* libm2pim/Makefile.am: Likewise.
* Makefile.in: Re-generate.
* libm2cor/Makefile.in: Likewise.
* libm2iso/Makefile.in: Likewise.
* libm2log/Makefile.in: Likewise.
* libm2min/Makefile.in: Likewise.
* libm2pim/Makefile.in: Likewise.
As mentioned in the simplify_rotate comment, for e.g.
((T) ((T2) X << (Y & (B - 1)))) | ((T) ((T2) X >> ((-Y) & (B - 1))))
we already emit
X r<< (Y & (B - 1))
as replacement. This PR is about the
((T) ((T2) X << Y)) OP ((T) ((T2) X >> (B - Y)))
((T) ((T2) X << (int) Y)) OP ((T) ((T2) X >> (int) (B - Y)))
forms if T2 is wider than T. Unlike e.g.
(X << Y) OP (X >> (B - Y))
which is valid just for Y in [1, B - 1], the above 2 forms are actually
valid and do the rotates for Y in [0, B] - for Y 0 the X value is preserved
by the left shift and right logical shift by B adds just zeros (but because
the shift is in wider precision B is still valid shift count), while for
Y equal to B X is preserved through the latter shift and the former adds
just zeros.
Now, it is unclear if we in the middle-end treat rotates with rotate count
equal or larger than precision as UB or not, unlike shifts there are less
reasons to do so, but e.g. expansion of X r<< Y if there is no rotate optab
for the mode is emitted as (X << Y) | (((unsigned) X) >> ((-Y) & (B - 1)))
and so with UB on Y == B.
The following patch does multiple things:
1) for the above 2, asks the ranger if Y could be equal to B and if so,
instead of using X r<< Y uses X r<< (Y & (B - 1))
2) for the
((T) ((T2) X << Y)) | ((T) ((T2) X >> ((-Y) & (B - 1))))
((T) ((T2) X << (int) Y)) | ((T) ((T2) X >> (int) ((-Y) & (B - 1))))
forms that were fixed 2 days ago it only punts if Y might be in the
[B,B2-1] range but isn't known to be in the
[0,B][2*B,2*B][3*B,3*B]... range. Because for Y which is a multiple
of B but smaller than B2 it acts as a rotate too, left shift provides
0 and (-Y) & (B - 1) is 0 and so preserves X. Though, for the cases
where Y is not known to be in [0,B-1] the patch also uses
X r<< (Y & (B - 1)) rather than X r<< Y
3) as discussed with Aldy, instead of using global ranger it uses a pass
specific copy but lazily created on first simplify_rotate that needs it;
this e.g. handles rotate inside of if body where the guarding condition
limits the shift count to some range which will not work with the
global ranger (unless there is some SSA_NAME to attach the range to).
Note, e.g. on x86 X r<< (Y & (B - 1)) and X r<< Y actually emit the
same assembly because rotates work the same even for larger rotate counts,
but that is handled only during combine.
2023-01-19 Jakub Jelinek <jakub@redhat.com>
PR tree-optimization/108440
* tree-ssa-forwprop.cc: Include gimple-range.h.
(simplify_rotate): For the forms with T2 wider than T and shift counts of
Y and B - Y add & (B - 1) masking for the rotate count if Y could be equal
to B. For the forms with T2 wider than T and shift counts of
Y and (-Y) & (B - 1), don't punt if range could be [B, B2], but only if
range doesn't guarantee Y < B or Y = N * B. If range doesn't guarantee
Y < B, also add & (B - 1) masking for the rotate count. Use lazily created
pass specific ranger instead of get_global_range_query.
(pass_forwprop::execute): Disable that ranger at the end of pass if it has
been created.
* c-c++-common/rotate-10.c: New test.
* c-c++-common/rotate-11.c: New test.
gcc/ChangeLog:
* config/aarch64/aarch64-simd.md (aarch64_simd_vec_set<mode>): Use
exact_log2 (INTVAL (operands[2])) >= 0 as condition for gating
the pattern.
(aarch64_simd_vec_copy_lane<mode>): Likewise.
(aarch64_simd_vec_copy_lane_<vswap_width_name><mode>): Likewise.
The testcase used to get scheduled differently depending on the
presence of debug insns with MEMs. It's not clear to me why those
MEMs affected scheduling, but the cselib pre-canonicalization of the
MEM address is not used at all when analyzing debug insns, so the
memory allocation and lookup are pure waste. Somehow, avoiding that
waste fixes the problem, or makes it go latent.
for gcc/ChangeLog
PR debug/106746
* sched-deps.cc (sched_analyze_2): Skip cselib address lookup
within debug insns.
for gcc/testsuite/ChangeLog
PR debug/106746
* gcc.target/i386/pr106746.c: New.
On Windows, 'system' is called - that fails with -1 if the command
interpreter could not be started; on POSIX systems, if the child
process could not be started by the shell, exit(127)/_exit(127) is
called/returned. On Windows, cmd.exe (and also the PowerShell) return
errorlevel 9009.
libgfortran/ChangeLog:
* intrinsics/execute_command_line.c (execute_command_line): On
Windows, regard system()'s return value of 9009 as EXEC_INVALIDCOMMAND.
In this ICE-on-invalid, we crash on
gcc_assert (INTEGRAL_TYPE_P (type));
in perform_integral_promotions, because a nullptr is an INTEGER_CST,
but not INTEGRAL_TYPE_P, and check_case_value is only checking the
former. In the test I'm testing other "shall be an integral constant
expression" contexts as well.
PR c/108424
gcc/c-family/ChangeLog:
* c-common.cc (check_case_value): Check INTEGRAL_TYPE_P.
gcc/testsuite/ChangeLog:
* gcc.dg/c2x-nullptr-6.c: New test.
P0482R6 deprecated these functions for C++20. There was a ballot comment
on the C++23 CD saying to un-deprecate it, but LEWG just rejected that,
so let's add attributes to deprecate them.
libstdc++-v3/ChangeLog:
* include/bits/fs_path.h (u8path): Add deprecated attribute.
* testsuite/27_io/filesystem/path/construct/90281.cc: Add
-Wno-deprecated-declarations for C++20 and later.
* testsuite/27_io/filesystem/path/factory/u8path-char8_t.cc:
Likewise.
* testsuite/27_io/filesystem/path/factory/u8path.cc: Likewise.
* testsuite/27_io/filesystem/path/native/string.cc: Likewise.
* testsuite/27_io/filesystem/path/factory/u8path-depr.cc: New test.
Since the r12-4515-g58f339fc5eaae7 change std::random_device::entropy()
returns non-zero for hardware sources such as RDRAND. However, the call
to the underlying _M_getentropy function is conditionally compiled
according to #if _GLIBCXX_USE_DEV_RANDOM which means it only happens for
targets that support /dev/random and /dev/urandom. This means entropy()
always returns zero for x86 Windows, even though the RDRAND and RDSEED
sources work there.
The _M_getentropy() function is always compiled into the library, it
just doesn't get called for targets without /dev/random. We can change
that just by removing the #if conditional. This is not an ABI change,
because new code will just start calling the existing _M_getentropy
function, old code that has inlined entropy() will not call it.
Similarly, the std::random_device destructor doesn't call the underlying
_M_fini function unless _GLIBCXX_USE_DEV_RANDOM is defined. That's less
of a problem because it's still true that the only resources that need
to be freed are when one of /dev/random or /dev/urandom has been opened
for reading, which is only possible when _GLIBCXX_USE_DEV_RANDOM is
defined. The _M_fini function does also destroy a random engine object
if a std::linear_congruential_engine object is used, but that destructor
is trivial and so no resources are leaked if it's not called. Remove the
preprocessor condition in the destructor too, so that we always call the
_M_fini function even if it doesn't have side effects. This makes the
destructor non-trivial for Windows and bare metal targets, but as the
class is non-copyable that shouldn't cause any ABI change in practice.
libstdc++-v3/ChangeLog:
* include/bits/random.h (random_device) [!_GLIBCXX_USE_DEV_RANDOM]:
Always call _M_fini and _M_getentropy.
