Replacing call to __kmp_msg(kmp_ms_fatal,...) with __kmp_fatal(...) caused an
issue when incomplete message is displayed in case an error message is followed
by another message (e.g. by a hint messa)ge. This is because __kmp_fatal()
passes incomplete list of arguments to __kmp_msg().

Patch by Olga Malysheva

Differential Revision: https://reviews.llvm.org/D39248

llvm-svn: 316623

"/DISCARD/" is a special section name in the linker script to discard
input sections. Previously, we handled it as if it were a real section,
so an input section can be assigned but dead. However, allowing sections
to be

 - assigned and alive (will be emitted),
 - not assigned and dead (removed), or
 - assigned but dead (???)

feels logically wrong and practically error-prone. This patch removes
the last combination of the states.

llvm-svn: 316622

The test is in clang-tools-extra/test/pp-trace

llvm-svn: 316621

Summary:
The 32-bit allocator is now on par with the 64-bit in terms of security (chunks
randomization is done, batches separation is done).

Unless objection, the comment can go away.

Reviewers: alekseyshl

Reviewed By: alekseyshl

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D39303

llvm-svn: 316620

Add the option to lookup an address in the debug information and print
out the file, function, block and line table details.

Differential revision: https://reviews.llvm.org/D38409

llvm-svn: 316619

Contrary to the deleted comment, in most cases CmpRuns.py produces a
fairly small amount of output, which is useful to see straight away to
see what has changed when executing the integration tests.

llvm-svn: 316618

Discussion at: https://reviews.llvm.org/D39220

llvm-svn: 316617

Summary:
On FreeBSD11.0 the FileCheck NOT string "1.0" will be matched by
`.amd_amdgpu_isa "amdgcn-unknown-freebsd11.0--gfx802"` at the end of the
file. Add a CHECK for that directive to avoid failing the test.

Reviewers: rampitec, kzhuravl

Reviewed By: rampitec, kzhuravl

Subscribers: emaste, kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye, llvm-commits, krytarowski

Differential Revision: https://reviews.llvm.org/D39306

llvm-svn: 316616

Max backedge taken count is always expected to be a constant; and this is
usually true by construction -- it is a SCEV expression with constant inputs.
However, if the max backedge expression ends up being computed to be a udiv with
a constant zero denominator[0], SCEV does not fold the result to a constant
since there is no constant it can fold it to (SCEV has no representation for
"infinity" or "undef").

However, in computeMaxBECountForLT we already know the denominator is positive,
and thus at least 1; and we can use this fact to avoid dividing by zero.

[0]: We can end up with a constant zero denominator if the signed range of the
stride is more precise than the unsigned range.

llvm-svn: 316615

llvm-svn: 316614

llvm-svn: 316613

Summary: This reverts commit r316582. It looks like this commit broke tests on one buildbot:
http://lab.llvm.org:8011/builders/llvm-clang-x86_64-expensive-checks-win/builds/5719

. . .
Failing Tests (1):
    LLVM :: Transforms/CalledValuePropagation/simple-arguments.ll

Reviewers:

Subscribers:

llvm-svn: 316612

Summary:
CUDA 9's minimum sm is sm_30.

Ideally we should also make sm_30 the default when compiling with CUDA
9, but that seems harder than it should be.

Subscribers: sanjoy

Differential Revision: https://reviews.llvm.org/D39109

llvm-svn: 316611

Add a CFI protection check that is implemented by building a graph and inspecting the output to deduce if the indirect CF instruction is CFI protected. Also added the output of this instruction to printIndirectInstructions().

Reviewers: vlad.tsyrklevich

Subscribers: llvm-commits, kcc, pcc, mgorny

Differential Revision: https://reviews.llvm.org/D38428

llvm-svn: 316610

This creates a new Architecture plugin and moves the stop info override
callback to this place. The motivation for this is to remove complex
dependencies from the ArchSpec class because it is used in a lot of
places that (should) know nothing about Process instances and StopInfo
objects.

I also add a test for the functionality covered by the override
callback.

Differential Revision: https://reviews.llvm.org/D31172

llvm-svn: 316609

This creates space for addidional arm-specific tests. I will be adding
one of those in a follow-up commit.

llvm-svn: 316608

Summary:
Also enable -no-pie on Gnu toolchain (previously available on Darwin only).

Non-PIE executables won't even start on recent Android, and DT_RPATH is ignored by the loader.

Reviewers: srhines, danalbert

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D38430

llvm-svn: 316606

The repro in https://bugs.llvm.org/show_bug.cgi?id=34362
caused the left nullptr to be cast to a int* implicitly, which
resulted diagnosing this falsely.

Differential Revision: https://reviews.llvm.org/D39301

llvm-svn: 316605

llvm-svn: 316604

Summary: Make sure shifts are legal/specified by the legalizerinfo before creating it

Reviewers: qcolombet, dsanders, rovka, t.p.northover

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D39264

llvm-svn: 316602

In getOffsetRange, Max can be set to 0 to force the extender replacement
to be at or below the original value. This would cause the new offset to
be non-negative, which is preferred for memory instructions (to reduce
the likelihood of it getting constant-extended due to predication). The
problem happens when the range is shifted by an offset (present in the
instruction being examined) and the offset is negative. The entire range
for the allowable deviation will then be strictly negative. This creates
a problem, since 0 is assumed to be a valid deviation.

llvm-svn: 316601

ArrayRef<T>() equals to ArrayRef<T>(nullptr, 0), so it looks like
we don't need to handle size 0 as a special case.

llvm-svn: 316600

Darwin uses char * for the variadic list type (va_list).  We use the PPC
SVR4 ABI for PPC, which uses a structure type for the va_list.  When
constructing the GEP, we would fail due to the incorrect handling for
the va_list.  Correct this to use the right type.

llvm-svn: 316599

Add fallback definition of internal_syscall_ptr and internal_syscall64
for Linux/x86_64 and Linux/aarch64.

llvm-svn: 316598

llvm-svn: 316597

Summary:
The 64-bit primary has had random shuffling of chunks for a while, this
implements it for the 32-bit primary. Scudo is currently the only user of
`kRandomShuffleChunks`.

This change consists of a few modifications:
- move the random shuffling functions out of the 64-bit primary to
  `sanitizer_common.h`. Alternatively I could move them to
  `sanitizer_allocator.h` as they are only used in the allocator, I don't feel
  strongly either way;
- small change in the 64-bit primary to make the `rand_state` initialization
  `UNLIKELY`;
- addition of a `rand_state` in the 32-bit primary's `SizeClassInfo` and
  shuffling of chunks when populating the free list.
- enabling the `random_shuffle.cpp` test on platforms using the 32-bit primary
  for Scudo.

Some comments on why the shuffling is done that way. Initially I just
implemented a `Shuffle` function in the `TransferBatch` which was simpler but I
came to realize this wasn't good enough: for chunks of 10000 bytes for example,
with a `CompactSizeClassMap`, a batch holds only 1 chunk, meaning shuffling the
batch has no effect, while a region is usually 1MB, eg: 104 chunks of that size.
So I decided to "stage" the newly gathered chunks in a temporary array that
would be shuffled prior to placing the chunks in batches.
The result is looping twice through n_chunks even if shuffling is not enabled,
but I didn't notice any significant significant performance impact.

Reviewers: alekseyshl

Reviewed By: alekseyshl

Subscribers: srhines, llvm-commits, kubamracek

Differential Revision: https://reviews.llvm.org/D39244

llvm-svn: 316596

ASan allocator stores the requested alignment for new and new[] calls
and on delete and delete[] verifies that alignments do match.

The representable alignments are: default alignment, 8, 16, 32, 64, 128,
256 and 512 bytes. Alignments > 512 are stored as 512, hence two
different alignments > 512 will pass the check (possibly masking the bug),
but limited memory requirements deemed to be a resonable tradeoff for
relaxed conditions.

The feature is controlled by new_delete_type_mismatch flag, the same one
protecting new/delete matching size check.

Differential revision: https://reviews.llvm.org/D38574

Issue: https://github.com/google/sanitizers/issues/799
llvm-svn: 316595

Resource file compilation requires a working resource compiler.
Unfortunately, llvm-rc isn't quite there yet [1], and cmake's rc
invocation only works on Windows [2]. Until both those issues are
addressed, disable resource file usage on non-Windows build hosts, to
unblock Windows cross-compilation. This is also consistent with the
existing comment, which says "If *on Windows* and building with MSVC".

[1] http://lists.llvm.org/pipermail/llvm-dev/2017-October/118433.html
[2] https://public.kitware.com/pipermail/cmake/2017-October/066441.html

Differential Revision: https://reviews.llvm.org/D39265

llvm-svn: 316594

As indicated by Table 1-1 in Intel Architecture Instruction Set Extensions and Future Features Programming Reference from October 2017.

llvm-svn: 316593

As indicated by Table 1-1 in Intel Architecture Instruction Set Extensions and Future Features Programming Reference from October 2017.

llvm-svn: 316592

Summary:
Changes:

 * Add initial msan stub support.
 * Handle NetBSD specific pthread_setname_np(3).
 * NetBSD supports __attribute__((tls_model("initial-exec"))),
   define it in SANITIZER_TLS_INITIAL_EXEC_ATTRIBUTE.
 * Add ReExec() specific bits for NetBSD.
 * Simplify code and add syscall64 and syscall_ptr for !NetBSD.
 * Correct bunch of syscall wrappers for NetBSD.
 * Disable test/tsan/map32bit on NetBSD as not applicable.
 * Port test/tsan/strerror_r to a POSIX-compliant OSes.
 * Disable __libc_stack_end on NetBSD.
 * Disable ReadNullSepFileToArray() on NetBSD.
 * Define struct_ElfW_Phdr_sz, detected missing symbol by msan.
 * Change type of __sanitizer_FILE from void to char. This helps
   to reuse this type as an array. Long term it will be properly
   implemented along with SANITIZER_HAS_STRUCT_FILE setting to 1.
 * Add initial NetBSD support in lib/tsan/go/buildgo.sh.
 * Correct referencing stdout and stderr in tsan_interceptors.cc
   on NetBSD.
 * Document NetBSD x86_64 specific virtual memory layout in
   tsan_platform.h.
 * Port tests/rtl/tsan_test_util_posix.cc to NetBSD.
 * Enable NetBSD tests in test/msan/lit.cfg.
 * Enable NetBSD tests in test/tsan/lit.cfg.

Sponsored by <The NetBSD Foundation>

Reviewers: joerg, vitalybuka, eugenis, kcc, dvyukov

Reviewed By: dvyukov

Subscribers: #sanitizers, llvm-commits, kubamracek

Tags: #sanitizers

Differential Revision: https://reviews.llvm.org/D39124

llvm-svn: 316591

llvm-svn: 316590

Summary:
They might not be mapped on some platforms such as Win64.  In
particular, this happens if the user address is null. There will not be
any shadow memory 5*16 bytes before the user address. This happens on
Win64 in the error_report_callback.cc test case. It's not clear why this
isn't a problem on Linux as well.

Fixes PR35058

Reviewers: vitalybuka

Subscribers: kubamracek, llvm-commits

Differential Revision: https://reviews.llvm.org/D39260

llvm-svn: 316589

Use llvm instrinsic by default
Provide amdgpu workaround

v2: drop old amd copyrights

Reviewer: Aaron Watry
Reviewed-by: Vedran Miletić <vedran@miletic.net>
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 316588

AMDGPU targets don't have insturction for it,
so it'll be expanded to C * log2 anyway.

v2: use native_log2 instead of the more precise sw implementation
v3: move to amdgpu
v4: drop old AMD copyright

Reviewer: Aaron Watry
Signed-off-by: Jan Vesely <jan.vesely@rutgers.edu>
llvm-svn: 316587

  - memory-legalizer-atomic-load.ll -> memory-legalizer-load.ll
  - memory-legalizer-atomic-store.ll -> memory-legalizer-store.ll

llvm-svn: 316586

expressions.

llvm-svn: 316585

llvm-svn: 316584

It is PR34946.

Spec (http://man7.org/linux/man-pages/man1/ld.1.html) tells about
--orphan-handling=MODE, option where MODE can be one of four:
"place", "discard", "warn", "error".
Currently we already report orphans when -verbose given,
what becomes excessive with option implemented.

Patch stops reporting orphans when -versbose is given,
and support "place", "warn" and "error" modes.
It is not yet clear that "discard" mode is useful so it is not supported.

Differential revision: https://reviews.llvm.org/D39000

llvm-svn: 316583

Summary: For some irreducible CFG the domtree nodes might be dead, do not update domtree for dead nodes.

Reviewers: kuhar, dberlin, hfinkel

Reviewed By: kuhar

Subscribers: llvm-commits, mcrosier

Differential Revision: https://reviews.llvm.org/D38960

llvm-svn: 316582