<rdar://problem/12481949> Fixing SBValue.GetValueAsSigned() to do the right thing when dealing with a 32-bit negative value

llvm-svn: 166603

llvm-svn: 166602

Explicitly specify C++98 when building the C++ precompiled header for this test, so that the test will work on Windows.

llvm-svn: 166601

llvm-svn: 166600

llvm-svn: 166599

Cleanup some variable names to indicate auto pointers and also manager the llvm::Module memory more correctly.

llvm-svn: 166598

llvm-svn: 166597

Back out r166591, not sure why this made it through since I cancelled the command. Bleh, sorry about this!

llvm-svn: 166596

llvm-svn: 166595

source operand is a register.

llvm-svn: 166594

Implement a basic VectorTargetTransformInfo interface to be used by the loop and bb vectorizers for modeling the cost of instructions.

llvm-svn: 166593

see the offsetof operator.  Previously, we were matching something like MOVrm
in the front-end and later matching MOVrr in the back-end.  This change makes
things more consistent.  It also fixes cases where we can't match against a 
memory operand as the source (test cases coming).
Part of rdar://12470317

llvm-svn: 166592

llvm-svn: 166591

reference-counted.

llvm-svn: 166587

llvm-svn: 166586

llvm-svn: 166584

reference-counted, and hold a reference to it in HeaderSearch.

llvm-svn: 166583

llvm-svn: 166579

This checkin also adds in some tests that utilize these paths and updates some of the
clients.

llvm-svn: 166578

llvm-svn: 166577

llvm-svn: 166572

llvm-svn: 166571

llvm-svn: 166567

llvm-svn: 166566

llvm-svn: 166562

llvm-svn: 166561

This lets one build ASan runtime for ARM/Android by running
make -C tools/clang/runtime/ \
  LLVM_ANDROID_TOOLCHAIN_DIR=/path/to/ndk/toolchain
in an existing build tree.

llvm-svn: 166560

llvm-svn: 166559

llvm-svn: 166558

llvm-svn: 166557

llvm-svn: 166552

will be represented in the IR as a plain "i32" type.  This causes the
tests to spuriously fail on platforms where int is not a 32-bit type,
or where the ABI requires attributes like "signext" or "zeroext" to
be used.

This patch adds -triple or -target parameters to force those tests
to use the i386-unknown-unknown target.

llvm-svn: 166551

Replace the allocator only if the replace_cfallocator flag is set (in some cases it wasn't checked)

llvm-svn: 166550

llvm-svn: 166549

[ASan] Due to data races, ASan malloc stats are inaccurate, which may cause certain ASan interface functions returning negative values (casted to unsigned). Return a reasonable value if such a case is detected.

llvm-svn: 166548

llvm-svn: 166547

- If more than 1 elemennts are defined and target supports the vectorized
  conversion, use the vectorized one instead to reduce the strength on
  conversion operation.

llvm-svn: 166546

- As there's no 64-bit GPRs in 32-bit mode, a custom conversion from v2u32 to
  v2f32 is added to improve the efficiency of the code generated.

llvm-svn: 166545

According to r166543, it is not needed for now.

llvm-svn: 166544

Many of our tests specify triples that are not built into clang.
In this commit we allow clang to fail loading the triple if we are only
using clang to emit llvm ir.

llvm-svn: 166543