llvm-svn: 175538

HexagonInstrFormats.td.

llvm-svn: 175537

No intended functionality change.

llvm-svn: 175536

I don't want to rule out the possibility of linking to e.g. interesting
blog posts about uses of Clang, so avoid restricting the content to
"projects".

This breaks URL compatibility, but this document was committed less than
an hour ago so hopefully nobody has linked to it yet.

llvm-svn: 175535

llvm-svn: 175534

We can now format:
  #define A template <typename T>

Before this created a segfault :-/.

llvm-svn: 175533

    
Using a new metafunction for detecting the presence of the member
'getDecl' in a type T, added support to hasDeclaration for any such type
T. This allows hasDecl() to be replaced and enables several other
subclasses of clang::Type to use hasDeclaration.

Updated unittests and LibASTMatchersReference.html.

Reviewers: klimek
llvm-svn: 175532

PPCJITInfo.cpp in r175394.

llvm-svn: 175531

llvm-svn: 175530

A couple concrete examples are sure to be a win. If you know of any
other external projects using Clang, please let me know!

Patch by Laszlo Nagy! (with a title tweak by me)

llvm-svn: 175529

Due to the execution order of doFinalization functions, the GC information were
deleted before AsmPrinter::doFinalization was executed. Thus, the
GCMetadataPrinter::finishAssembly was never called.

The patch fixes that by moving the code of the GCInfoDeleter::doFinalization to
Printer::doFinalization.

llvm-svn: 175528

llvm-svn: 175527

llvm-svn: 175525

- rework the way SBDebugger.SetAsync() is used to avoid side effects (reset original value at TearDownHook)
- refactor expectedFailureClang (and add expectedFailureGcc decorator)
- mark TestChangeValueAPI.py as expectedFailureGcc due to PR-15039

llvm-svn: 175523

- do not pass (incompatible option) -stdlib=libstdc++ option if using GCC as test compiler

llvm-svn: 175522

- Should allow tests to run without 'clang' installed

llvm-svn: 175521

A vectorized sitfp on doubles will get scalarized to a sequence of an
extract_element of <2 x i32>, a bitcast to f32 and a sitofp.
Due to the the extract_element, and the bitcast we will uneccessarily generate
moves between scalar and vector registers.

The patch fixes this by using a COPY_TO_REGCLASS and a EXTRACT_SUBREG to extract
the element from the vector instead.

radar://13191881

llvm-svn: 175520

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175519

This stops the Machine Verifier from complaining about uses of undefined
physical registers.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175518

Kernel function arguments are lowered to loads from the PARAM_I address
space.  When creating these load instructions, we were initializing
their MachinePointerInfo with an Arguement object that was not attached
to any function.  This was causing the MachineScheduler to crash when
it tried to access the parent of the Arguement.

This has been fixed by initializing the MachinePointerInfo with a
UndefValue instead.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175517

In some cases, we were losing track of live implicit registers which
was creating dead defs and causing the scheduler to produce invalid
code.

NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175516

llvm-svn: 175515

llvm-svn: 175514

This way it fits on a random 1G device.

llvm-svn: 175513

[ASan] Allow ASan default runtime options be overriden at compile time by providing ASAN_DEFAULT_OPTIONS macro

llvm-svn: 175512

llvm-svn: 175511

llvm-svn: 175510

Fix initialization-order bug in llvm::Support::TimeValue. TimeValue::now() is explicitly called during module initialization of lib/Support/Process.cpp. It reads the field of global object PosixZeroTime, which is not guaranteed to be initialized at this point. Found by AddressSanitizer with -fsanitize=init-order option.

llvm-svn: 175509

This patch makes asan instrument memory accesses with unusual sizes (e.g. 5 bytes or 10 bytes), e.g. long double or
packed structures.
Instrumentation is done with two 1-byte checks
(first and last bytes) and if the error is found
__asan_report_load_n(addr, real_size) or
__asan_report_store_n(addr, real_size)
is called.

asan-rt part

Also fix lint.

llvm-svn: 175508

This patch makes asan instrument memory accesses with unusual sizes (e.g. 5 bytes or 10 bytes), e.g. long double or
packed structures.
Instrumentation is done with two 1-byte checks
(first and last bytes) and if the error is found
__asan_report_load_n(addr, real_size) or
__asan_report_store_n(addr, real_size)
is called.

Also, call these two new functions in memset/memcpy
instrumentation.

asan-rt part will follow.

llvm-svn: 175507

Fix a bug in mutex_try_to_lock.  This was previously trying to unlock a mutex that it didn't own, causing an assertion failure in mutex.cpp.  The issue was that the unique_lock went out of scope, releasing the lock on m, then m.unlock() was called on an already-unlocked mutex.  

This change removes the spurious m.unlock() call.  

If this test was previously passing for anyone with assertions enabled, then they should investigate bugs in their pthread implementation, as pthread_unlock() should not return 0 if the mutex is currently unlocked.

llvm-svn: 175506

[Sanitizer] If -fsanitize-blacklist= option is not explicitly specified, make Clang look for the default sanitizer-specific blacklist in the resource directory.

llvm-svn: 175505

execle(argv[0], ...) is wrong, because argv[0] may contain a PATH-resolvable
program name. Execute /proc/self/exe instead.

llvm-svn: 175504

llvm-svn: 175503

llvm-svn: 175502

llvm-svn: 175501

The key bug was

  if (Other.StartOfLineLevel < StartOfLineLevel) ..

instead of

  if (Other.StartOfLineLevel != StartOfLineLevel) ..

Also cleaned up the function to be more consistent in the comparisons.

llvm-svn: 175500

llvm-svn: 175499

llvm-svn: 175498

llvm-svn: 175497