llvm-svn: 157389

The idea isthat asan/tsan can survive if user intercepts the same functions. At the same time user has an ability to call back into asan/tsan runtime. See the following tests for examples:
asan/output_tests/interception_failure_test-linux.cc
asan/output_tests/interception_test-linux.cc
asan/output_tests/interception_malloc_test-linux.cc

llvm-svn: 157388

llvm-svn: 157387

LowerSwitch::Clusterify : main functinality was replaced with CRSBuilder::optimize, so big part of Clusterify's code was reduced.
test/Transform/LowerSwitch/feature.ll - this test was refactored: grep + count was replaced with FileCheck usage.

llvm-svn: 157384

llvm-svn: 157383

llvm-svn: 157381

llvm-svn: 157380

llvm-svn: 157379

cycle.

llvm-svn: 157378

llvm-svn: 157377

llvm-svn: 157375

llvm-svn: 157374

Make some opcode tables static and const. Allows code to avoid making copies to pass the tables around.

llvm-svn: 157373

http://llvm.org/bugs/show_bug.cgi?id=12924

This issue was that the source location was pointing to a
non-printable character and so CaretEnd was pointing one
_column_ past the caret but not one _character_ past the
caret. So the conversion between column and byte locations
wasn't working (because the conversion is only valid from
the first column or byte of a character).

llvm-svn: 157372

llvm-svn: 157371

volatile reference to a temporary is not viable. My interpretation is that
DR1152 was a bugfix, not a rule change for C++11, so this is not conditional on
the language mode. This matches g++'s behavior.

llvm-svn: 157370

llvm-svn: 157369

llvm-svn: 157368

Mark a static table as const. Shrink opcode size in static tables to uint16_t. Simplify loop iterating over one of those tables. No functional change intended.

llvm-svn: 157367

short-circuit of the Python SWIG building, rather than relying on the SDKROOT being set.

llvm-svn: 157363

Add a warning to diagnose statements in C++ like "*(volatile int*)x;".  Conceptually, this is part of -Wunused-value, but I added a separate flag -Wunused-volatile-lvalue so it doesn't get turned off by accident with -Wno-unused-value.  I also made a few minor improvements to existing unused value warnings in the process.  <rdar://problem/11516811>.

llvm-svn: 157362

Supports the use-case scenario of immediately continuing the process once attached.
Add a simple completion test case from "process attach --con" to "process attach --continue ".

llvm-svn: 157361

Portions of this patch by Xi Wang. Reviewed by Jordy Rose. Thank you both.

Codegen support will follow soon.

llvm-svn: 157360

statement. // rdar://11474836.

llvm-svn: 157359

llvm-svn: 157358

llvm-svn: 157357

rdar://11357807

llvm-svn: 157356

Live ranges with a constrained register class may benefit from splitting
around individual uses. It allows the remaining live range to use a
larger register class where it may allocate. This is like spilling to a
different register class.

This is only attempted on constrained register classes.

<rdar://problem/11438902>

llvm-svn: 157354

Add SBProcess::GetNumSupportedHardwareWatchpoints() API and export it through the Python scripting bridge.
Add/modify some test cases.

llvm-svn: 157353

clang/test/Tooling: Update comments in left 2 tests on XFAIL. They are incompatible to -fms-compatibility.

llvm-svn: 157352

llvm-svn: 157351

It fixes test/Tooling on Win32 hosts.

llvm-svn: 157350

llvm-svn: 157349

llvm-svn: 157348

idiom that is used commonly in setters:

   [backingValue autorelease];
   backingValue = [newValue retain]; // in general a +1 assign

rdar://9914061

llvm-svn: 157347

rdar://11518308

llvm-svn: 157346

Add the capability to display the number of supported hardware watchpoints to the "watchpoint list" command.

Add default Process::GetWatchpointSupportInfo() impl which returns an error of "not supported".
Add "qWatchpointSupportInfo" packet to the gdb communication layer to support this, and modify TestWatchpointCommands.py to test it.

llvm-svn: 157345

llvm-svn: 157342

CHECK. The latter error was hidden by the former, and the test harness
used by e.g. "make check" silently ignored that opt was printing an
error message about an unknown flag instead of running on the test file.

llvm-svn: 157341

Now that the coalescer keeps live intervals and machine code in sync at
all times, it needs to deal with identity copies differently.

When merging two virtual registers, all identity copies are removed
right away. This means that other identity copies must come from
somewhere else, and they are going to have a value number.

Deal with such copies by merging the value numbers before erasing the
copy instruction. Otherwise, we leave dangling value numbers in the live
interval.

This fixes PR12927.

llvm-svn: 157340