r242520 was reverted in r244313 as the expected behaviour of the alias
attribute in C is that the alias has the same size as the aliasee. However
we can re-introduce adding the size on the alias when the aliasee does not,
from a source code or object perspective, exist as a discrete entity. This
happens when the aliasee is not a symbol, or when that symbol is private.

Differential Revision: http://reviews.llvm.org/D11943

llvm-svn: 244752

The version of libxml2 required by c-index-test must be at least 2.5.3. Considering that this version was released in 2003, you might think the check a bit ridiculous. Unfortunately, GnuWin32 ships with libxml2 2.4.12, which was released in 2001.

This allows us to have GnuWin32 on the PATH on Windows without causing compilation errors.

llvm-svn: 244751

http://reviews.llvm.org/D11899

llvm-svn: 244750

llvm-svn: 244749

On Mach-O emitting aliases for the variables that make up a MergedGlobals
variable can cause problems when linking with dead stripping enabled so don't
do that, except for external variables where we must emit an alias.

llvm-svn: 244748

llvm-svn: 244747

removes a corner case in tests that don't set the diagnostic consumer.
In tests, it is good, not to set the diagnostic consumer so that Clang's
parsing diagnostics are still displayed in the test output and only
ClangTidy's output is analyzed differently.

llvm-svn: 244745

[mips][microMIPS] Create microMIPS64r6 subtarget and implement DALIGN, DAUI, DAHI, DATI, DEXT, DEXTM and DEXTU instructions
Differential Revision: http://reviews.llvm.org/D10923

llvm-svn: 244744

Add PT_PHDR segment depending on its availability in linker script's
PHDRS command, fallback if no linker script is given.
Handle FILEHDR, PHDRS and FLAGS attributes of program header.

Differential Revision: http://reviews.llvm.org/D11589

llvm-svn: 244743

On x86/x86_64 read only watchpoints aren't supported. Fall back
to read/write watchpoints in that case.

Note: Logic should be added to ignore the watchpoint hit when
occurred because of a write.

llvm-svn: 244742

http://reviews.llvm.org/D11902

llvm-svn: 244741

Differential Revision: http://reviews.llvm.org/D11904

llvm-svn: 244740

The ASAN callbacks are public symbols so we can search for them
with reading only the symbol table (not the debug info). Whit this
change the attach time for big executables with debug symbols
decreased by a factor of ~4.

Differential revision: http://reviews.llvm.org/D11384

llvm-svn: 244739

On android .oat files (compiled java code) don't have symbol
information but on SDK 23+ it can be generated by the oatdump tool
(based on the dex information).

This CL adds logic to download this information and store it in the
module cache.

Differential revision: http://reviews.llvm.org/D11936

llvm-svn: 244738

The SDK version implies the features supported by a given android
device. This version number will be used in future changes to execute
the right command on the device.

Differential revision: http://reviews.llvm.org/D11935

llvm-svn: 244737

llvm-svn: 244736

llvm-svn: 244735

  While the compile time is not affected by this patch much it will
  allow us to look at all translated expressions after the SCoP is build
  in a convenient way. Additionally, bigger SCoPs or SCoPs with
  repeating complicated expressions might benefit from the cache later
  on.

Reviewers: grosser, Meinersbur

Subscribers: #polly

Differential Revision: http://reviews.llvm.org/D11975

llvm-svn: 244734

llvm-svn: 244733

llvm-svn: 244732

  Added missing documentation and linked to the correct revision.

Differential Revision: http://reviews.llvm.org/D11974

llvm-svn: 244731

  This change has three major advantages:
    - The ScopInfo becomes smaller.
    - It allows to use the SCEVAffinator from outside the ScopInfo.
    - A member object allows state which in turn allows e.g., caching.

Differential Revision: http://reviews.llvm.org/D9099

llvm-svn: 244730

This abstracts away the test for "when can we fold across a MachineInstruction"
into the the MI interface, and changes call-frame optimization use the same test
the peephole optimizer users.

Differential Revision: http://reviews.llvm.org/D11945

llvm-svn: 244729

llvm-svn: 244728

  In order to find the llvm-obj directory it has to be (or a soft link
  to it) at one of the following locations:

    ${POLLY_SRC_DIR}/build
    ${POLLY_SRC_DIR}.build
    ${POLLY_SRC_DIR}-build
    s/${POLLY_SRC_DIR}/src/build

  Alternatively, the environment variable $POLLY_BIN_DIR can point to it.

llvm-svn: 244727

llvm-svn: 244726

llvm-svn: 244725

This breaks the buildbot:
http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux/builds/19656

llvm-svn: 244724

As discussed in D11886, this patch moves the SSE/AVX vector blend folding to instcombiner from PerformINTRINSIC_WO_CHAINCombine (which allows us to remove this completely).

InstCombiner already had partial support for this, I just had to add support for zero (ConstantAggregateZero) masks and also the case where both selection inputs were the same (allowing us to ignore the mask).

I also moved all the relevant combine tests into InstCombine/blend_x86.ll

Differential Revision: http://reviews.llvm.org/D11934

llvm-svn: 244723

Reinstantiate better diagnostic, this time with a fatal error so we don't add a dependency onto gtest from the header.

llvm-svn: 244722

Before we only modeled PHI nodes if at least one incoming basic block was itself
part of the region, now we always model them except if all of their operands are
part of a single non-affine subregion which we model as a black-box.

This change only affects PHI nodes in the entry block, that have exactly one
incoming edge. Before this change, we did not model them and as a result code
generation would not know how to code generate them. With this change, code
generation can code generate them like any other PHI node.

This issue was exposed by r244606. Before this change simplifyRegion would have
moved these PHI nodes out of the SCoP, so we would never have tried to code
generate them. We could implement this behavior again, but changing the IR
after the scop has been modeled and transformed always adds a risk of us
invalidating earlier analysis results. It seems more save and overall also more
consistent to just model and handle this one-entry-edge PHI nodes like any
other PHI node in the scop.

Solution proposed by:  Michael Kruse  <llvm@meinersbur.de>

llvm-svn: 244721

This one was extracted from the test-suite's pifft and caused a
miscompilation because a scalar was not written to its alloca address.

llvm-svn: 244720

Add parsing of openmp directives inside structs/unions in C mode.

llvm-svn: 244719

llvm-svn: 244718

llvm-svn: 244717

major, minor, and patchlevel in the qHostInfo reply.  
Document that qHostInfo may report major/minor/patch
separately / in addition to the version: combination.

<rdar://problem/22125465> 

llvm-svn: 244716

All of the other docblocks for the CXCursor_* cursor kind enum values
include documentation that ends with a period. Add a period to the end
of the CXCursor_TypedefDecl documentation to follow this convention.

Patch by Brian Gesiak!

llvm-svn: 244715

llvm-svn: 244714

The same value is used multiple times through the function.  Hoist the condition
into a variable.  This should fix a silly static analysis warning where the
conditions flip around.  No functional change intended.

llvm-svn: 244713

[libFuzzer] add two flags, -tbm_depth and -tbm_width to control how the trace-based-mutations are applied

llvm-svn: 244712