syms before constructing the compile units so we're not emitting
section references to sections not there already.

llvm-svn: 174663

llvm-svn: 174662

Add #include <unistd.h> to OProfileWrapper.cpp.   This provides the declarations for 'read' and 'close' that are otherwise missing, and result in 'error: <foo> was not declared in this scope'.

This matches the issue as reported in bug 15055 "Can no longer compile LLVM with --with-oprofile"

llvm-svn: 174661

Certain vector operations don't vectorize well with the current
PowerPC implementation.  Element insert/extract performs poorly
without VSX support because Altivec requires going through memory.
SREM, UREM, and VSELECT all produce bad scalar code.

There's a lot of work to do for the cost model before
autovectorization will be tuned well, and this is not an attempt to
address the larger problem.

llvm-svn: 174660

llvm-svn: 174659

llvm-svn: 174658

original JALR instruction with one register operand to be a pseudo-instruction.

llvm-svn: 174657

Remove all the unused code.

Patch by: Christian König

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174656

Allows nexuiz to run with radeonsi.

Patch by: Michel Dänzer

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174655

20 more little piglits with radeonsi.

Patch by: Michel Dänzer

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174654

The _SGPR variants where wrong.

Patch by: Christian König

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174653

v2: rebased on current upstream

Patch by: Christian König

Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174652

This is for the case when no processor is passed to the backend.  This
prevents the

'' is not a recognized processor for this target (ignoring processor)

warning from being generated by clang.

llvm-svn: 174651

llvm-svn: 174650

llvm-svn: 174649

Retain all hidden methods in the global method pool, because they may become visible <rdar://problem/13172858>.

llvm-svn: 174648

ContainerUSR is not really a CXString, but it should own the underlying memory
buffer.  Thus, it is better to change the type to std::string.  This will not
introduce extra copying overhead, since the workaround that is being removed
was already making a copy.

llvm-svn: 174647

We don't want too many classes in a pass and the classes obscure the details. I
was going a little overboard with object modeling here. Replace classes by
generic code that handles both loads and stores.

No functionality change intended.

llvm-svn: 174646

llvm-svn: 174645

in an unevaluated context.

llvm-svn: 174644

The rewriter was previously reading the content buffer from the file itself.
Since we are now keeping the content in memory and writing to the file only
once, the rewriter's buffer (from the file) was not in sync with the
RefactoringTool's buffer. Adding an overrideFileContents call (similar to how
Clang-format handles for this) will resolve this issue.

Author: Jack Yang <jack.yang@intel.com>
Reviewers: gribozavr, klimek
llvm-svn: 174643

llvm-svn: 174642

PR15138 was opened because of a segfault in the Bitcode writer.
The actual issue ended up being a bug in APInt where calls to
APInt::getActiveWords returns a bogus value when the APInt value
is 0.  This patch fixes the problem by ensuring that getActiveWords
returns 1 for 0 valued APInts.

llvm-svn: 174641

llvm-svn: 174640

llvm-svn: 174639

SBValueList was backed by a ValueObjectList. This caused us to lose track of the additional metadata in the ValueImpl that backs SBValue.
This checkin fixes that by backing SBValueList with ValueListImpl (that essentially wraps a vector<SBValue>).

llvm-svn: 174638

llvm-svn: 174637

llvm-svn: 174636

llvm-svn: 174635

Patch by: Michel Dänzer

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174634

Handle vectors of 1 to 16 integers.

Change the intrinsic names to prevent the wrong one from being selected at
runtime due to the overloading.

Patch By: Michel Dänzer

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174633

v1i32, v2i32, v8i32 and v16i32.

Only add VGPR register classes for integer vector types, to avoid attempts
copying from VGPR to SGPR registers, which is not possible.

Patch By: Michel Dänzer

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 174632

Vector selects are cheap on NEON. They get lowered to a vbsl instruction.

radar://13158753

llvm-svn: 174631

llvm-svn: 174630

See https://code.google.com/p/address-sanitizer/issues/detail?id=10 for the context.

llvm-svn: 174629

Enable AddressSanitizer.AllocDeallocMismatch tests.

llvm-svn: 174628

llvm-svn: 174627

Also fix large_func_test.cc, which got broken when we switched to allocator2: the OOB access was too far from the original allocation, so ASan decided to describe the next one.

llvm-svn: 174626

llvm-svn: 174625

this is required to catch close of file descriptors created in getaddrinfo()

llvm-svn: 174624