Summary:
Provide __hwasan_shadow_init that can be used to initialize shadow w/o touching libc.
It can be used to bootstrap an unusual case of fully-static executable with
hwasan-instrumented libc, which needs to run hwasan code before it is ready to serve
user calls like madvise().

Reviewers: vitalybuka, kcc

Subscribers: kubamracek, llvm-commits

Differential Revision: https://reviews.llvm.org/D50581

llvm-svn: 339606

NFC: Add a test to LV showing that reduction is not possible when reduction var is reset in the loop

Added a test case to reduction showing where it's illegal to identify
vectorize a loop.
Resetting the reduction var during loop iterations disallows us from
widening the dependency cycle to VF, thereby making it illegal to
vectorize the loop.

llvm-svn: 339605

This is a very partial fix for the reported problem. I suspect
we do not get this fold in most motivating cases because most of
the time, the libcall would have been replaced by an intrinsic,
and that optimization is handled elsewhere...but maybe it should
be handled here?

llvm-svn: 339604

The number of loops associated with the OpenMP loop constructs should
not be considered as the number loops to collapse.

llvm-svn: 339603

See D50465 for the actual opt itself.

Differential Revision: https://reviews.llvm.org/D50464

llvm-svn: 339602

[analyzer][UninitializedObjectChecker] Refactoring p5.: Handle pedantic mode in the checker class only

Differential Revision: https://reviews.llvm.org/D50508

llvm-svn: 339601

Fix SelectionDAG::computeKnownBits asserting when handling EXTRACT_SUBVECTOR
when zero extending the demanded elements mask if it is already as long as the
source vector.

Differential Revision: https://reviews.llvm.org/D49574

llvm-svn: 339600

[analyzer][UninitializedObjectChecker] Refactoring p4.: Wrap FieldRegions and reduce weight on FieldChainInfo

Before this patch, FieldChainInfo used a spaghetti: it took care of way too many cases,
even though it was always meant as a lightweight wrapper around
ImmutableList<const FieldRegion *>.
This problem is solved by introducing a lightweight polymorphic wrapper around const
FieldRegion *, FieldNode. It is an interface that abstracts away special cases like
pointers/references, objects that need to be casted to another type for a proper note
messages.

Changes to FieldChainInfo:

  * Now wraps ImmutableList<const FieldNode &>.
  * Any pointer/reference related fields and methods were removed
  * Got a new add method. This replaces it's former constructors as a way to create a
    new FieldChainInfo objects with a new element.

Changes to FindUninitializedField:

  * In order not to deal with dynamic memory management, when an uninitialized field is
    found, the note message for it is constructed and is stored instead of a
    FieldChainInfo object. (see doc around addFieldToUninits).

Some of the test files are changed too, from now on uninitialized pointees of references
always print "uninitialized pointee" instead of "uninitialized field" (which should've
really been like this from the beginning).

I also updated every comment according to these changes.

Differential Revision: https://reviews.llvm.org/D50506

llvm-svn: 339599

llvm-svn: 339598

Multiversioned member functions inside of a template type were 
not properly being emitted.  The solution to this is to simply 
ensure that their bodies are correctly evaluated/assigned during
template instantiation.

llvm-svn: 339597

This is a standalone part of the effort to reduce FieldChainInfos inteerface.

Differential Revision: https://reviews.llvm.org/D50505

llvm-svn: 339596

In this patch, the following classes and functions have been moved to a header file:

    FieldChainInfo
    FindUninitializedFields
    isPrimitiveType

This also meant that they moved from anonymous namespace to clang::ento.

Code related to pointer chasing now relies in its own file.

There's absolutely no functional change in this patch -- its literally just copy pasting.

Differential Revision: https://reviews.llvm.org/D50504

llvm-svn: 339595

Summary:
This patch introduces error handling to propagate the errors from llvm-mca library classes (or what will become library classes) up to the driver.  This patch also introduces an enum to make clearer the intention of the return value for Stage::execute.

This supports PR38101.

Reviewers: andreadb, courbet, RKSimon

Reviewed By: andreadb

Subscribers: llvm-commits, tschuett, gbedwell

Differential Revision: https://reviews.llvm.org/D50561

llvm-svn: 339594

Summary:
Fixes r339581 ("[SEMA] add more -Wfloat-conversion to
compound assigment analysis").

This test case was caught in postsubmit testing.

Reviewers: aaron.ballman, gkistanova

Reviewed By: aaron.ballman

Subscribers: cfe-commits, srhines

Differential Revision: https://reviews.llvm.org/D50647

llvm-svn: 339593

llvm-svn: 339592

This patch is the first part of a series of patches to refactor UninitializedObjectChecker. The goal of this effort is to

    Separate pointer chasing from the rest of the checker,
    Increase readability and reliability,
    Don't impact performance (too bad).

In this one, ImmutableList's factory is moved to FindUninitializedFields.

Differential Revision: https://reviews.llvm.org/D50503

llvm-svn: 339591

llvm-svn: 339590

llvm-svn: 339589

llvm-svn: 339588

llvm-svn: 339587

Also fixed a typo that wasn't discovered as `create` was never instantiated.

Differential Revision: https://reviews.llvm.org/D50646

llvm-svn: 339586

This change fixes an incorrect behavior of the omp_control_tool function when
called from Fortran applications.  A tool callback function for this event is
supposed to get NULL for the third argument according to the specification, but
the current implementation just passes a garbage value. A possible fix is to use
the OPTIONAL attribute for the third argument.

Patch by Hansang Bae

Differential Revision: https://reviews.llvm.org/D50565

llvm-svn: 339585

Fixes OSS-Fuzz: https://bugs.chromium.org/p/oss-fuzz/issues/detail?id=5223

llvm-svn: 339584

Summary:
Removing FastDemangle will greatly reduce maintenance efforts. This patch replaces the last point of use in LLDB. Semantics should be kept intact.

Once this is agreed upon, we can:
* Remove the FastDemangle sources
* Add more features e.g. substitutions in template parameters, considering all variations, etc.

Depends on LLVM patch https://reviews.llvm.org/D50586

Reviewers: erik.pilkington, friss, jingham, JDevlieghere

Subscribers: kristof.beyls, chrib, lldb-commits

Differential Revision: https://reviews.llvm.org/D50587

llvm-svn: 339583

The current static_assert only checks that ObjCObjectTypeBitfields
fits into an unsigned. However it turns out that FunctionTypeBitfields
do not currently fits into an unsigned. Therefore the anonymous
union containing the bit-fields always use 8 bytes instead of 4.

This patch removes the lone misguided static_assert and systematically
checks the size of each bit-field.

Reviewed By: erichkeane

Differential Revision: https://reviews.llvm.org/D50630

llvm-svn: 339582

Summary: Fixes Bug: https://bugs.llvm.org/show_bug.cgi?id=27061

Reviewers: aaron.ballman, acoomans

Reviewed By: aaron.ballman, acoomans

Subscribers: acoomans, cfe-commits, srhines, pirama

Differential Revision: https://reviews.llvm.org/D50467

llvm-svn: 339581

This function calls a callback whenever a <type> is parsed.

This is necessary to implement FindAlternateFunctionManglings in LLDB, which
uses a similar hack in FastDemangle. Once that function has been updated to use
this version, FastDemangle can finally be removed.

Differential revision: https://reviews.llvm.org/D50586

llvm-svn: 339580

llvm-svn: 339579

Also consider vector constants when simplifying `pow()`.

Differential revision: https://reviews.llvm.org/D50035

llvm-svn: 339578

This is a follow-up of r339552.

As pointed out by Craig in D50566, we don't need a formatted_raw_ostream to
indent strings. We can use instead raw_ostream::indent().

Internally, class PredicateExpander already keeps track of the current
indentation level. Also, the grammar for predicates is well parenthesized, and
therefore we don't need to use a formatted_raw_ostream to continuously track the
column number. Instead we can safely replace all the uses of
formatted_raw_ostream::PadToColumn() with uses of raw_ostream::indent().

By replacing formatted_raw_ostream with a simpler raw_ostream, we also avoid the
implicit check on the newline character on every print to stream.

No functional change intended.

llvm-svn: 339577

Patch by Kim Gräsman.

Differential Revision: https://reviews.llvm.org/D50623

llvm-svn: 339576

Summary:
When compiling with `WERROR=ON` & a recent clang, having the `st(?)` registers
in the clobber list produces a fatal error (except `st(7)` for some reason):
```
.../sanitizer_common/sanitizer_atomic_clang_x86.h:98:9: error: inline asm clobber list contains reserved registers: ST0, ST1, ST2, ST3, ST4, ST5, ST6 [-Werror,-Winline-asm]
        "movq %1, %%mm0;"  // Use mmx reg for 64-bit atomic moves
        ^
<inline asm>:1:1: note: instantiated into assembly here
        movq 8(%esp), %mm0;movq %mm0, (%esi);emms;
^
.../sanitizer_common/sanitizer_atomic_clang_x86.h:98:9: note: Reserved registers on the clobber list may not be preserved across the asm statement, and clobbering them may lead to undefined behaviour.
        "movq %1, %%mm0;"  // Use mmx reg for 64-bit atomic moves
        ^
<inline asm>:1:1: note: instantiated into assembly here
        movq 8(%esp), %mm0;movq %mm0, (%esi);emms;
^
```
As far as I can tell, they were in there due to the use of the `emms`
instruction, but removing the clobber doesn't appear to have a functional
impact. I am unsure if there is a better way to address this.

Reviewers: eugenis, vitalybuka

Reviewed By: vitalybuka

Subscribers: kubamracek, delcypher, jfb, llvm-commits, #sanitizers

Differential Revision: https://reviews.llvm.org/D50562

llvm-svn: 339575

This reverts commit r339568 because of the problems with the buildbots.

llvm-svn: 339574

llvm-svn: 339573

Reviewers: ilya-biryukov, hokein

Reviewed By: hokein

Subscribers: hokein, ioeric, jkorous, arphaman, cfe-commits

Differential Revision: https://reviews.llvm.org/D50635

llvm-svn: 339572

Summary:
This allows member functions to be marked as reinitializing the object. After a
moved-from object has been reinitialized, the check will no longer consider it
to be in an indeterminate state.

The patch that adds the attribute itself is at https://reviews.llvm.org/D49911

Reviewers: ilya-biryukov, aaron.ballman, alexfh, hokein, rsmith

Reviewed By: aaron.ballman

Subscribers: dblaikie, xazax.hun, cfe-commits

Tags: #clang-tools-extra

Differential Revision: https://reviews.llvm.org/D49910

llvm-svn: 339571

It breaks when using EXPENSIVE_CHECKS with the error message
"Bad machine code: Using an undefined physical register".

llvm-svn: 339570

Summary:
This is for use by clang-tidy's bugprone-use-after-move check -- see
corresponding clang-tidy patch at https://reviews.llvm.org/D49910.

Reviewers: aaron.ballman, rsmith

Reviewed By: aaron.ballman

Subscribers: cfe-commits

Differential Revision: https://reviews.llvm.org/D49911

llvm-svn: 339569

The number of loops associated with the OpenMP loop constructs should
not be considered as the number loops to collapse.

llvm-svn: 339568

Differential Revision: https://reviews.llvm.org/D50592

llvm-svn: 339567