in an expression and doesn't ignore the stop.

Patch by Jim Ingham.

<rdar://problem/14583884>

llvm-svn: 187434

register i7 as a live-in if current function's return address is taken.

This revision fixes PR16269.

llvm-svn: 187433

Both the parameter itself and the argument to the parameter must be prefixed
with -Xlinker so that they are passed properly.

llvm-svn: 187432

llvm-svn: 187431

This is a follow up patch for r187390 to implement the parser for the
Windows-style command line. This should follow the rule as described
at http://msdn.microsoft.com/en-us/library/windows/desktop/17w5ykft(v=vs.85).aspx

Differential Revision: http://llvm-reviews.chandlerc.com/D1235

llvm-svn: 187430

tip-of-trunk clang has corrected some access checks for special members in a virtual inheritance hierarchy.  Change a few private inheritances to protected.  This change will not impact what the test was testing.  This fixes http://llvm.org/bugs/show_bug.cgi?id=16753.

llvm-svn: 187429

Committing r187204 with fixes for darwin. Note that one of the lit tests are
disabled on windows due to a bug in writing header replacements to file.

Header replacements are now written to disk in YAML format for an external tool
to merge. A unique file will be created in the same directory as the header
with all replacements that came from a source file that included the header
file. The YAML file will have:
 - Name of the file
 - Transform ID that generated the replacement
 - Offset
 - Length
 - Replacement text

Any tool reading these replacements should read them using the
TransformDocument struct.

llvm-svn: 187428

on the system, and report it when running the driver in verbose mode.
Without this it is essentially impossible to understand why a particular
GCC toolchain is used by Clang for libstdc++, libgcc, etc.

This also required threading a hook through the toolchain layers for
a specific toolchain implementation to print custom information under
'clang -v'. The naming here isn't spectacular. Suggestions welcome.

llvm-svn: 187427

This uses $(CXX) instead of $(CC) to generate the .d files for tests built
with Makefiles.

llvm-svn: 187426

Add format specifiers to various format ids so we can print thread ids in decimal on Linux and FreeBSD.

CC: emaste

Differential Revision: http://llvm-reviews.chandlerc.com/D1234

llvm-svn: 187425

- Thanks to Ilia Filippov for pointing out the inconsistency!

llvm-svn: 187424

to handle the case of an integer constant (DWARF 3 and later).

- Fixes tests that assert in RecordLayoutBuilder::updateExternalFieldOffset
because LLDB was providing an external AST source with missing member offsets.

llvm-svn: 187423

Instantiate RegisterContext... based on getOS() instead of with
compile-time #ifdef-ery.

The assert() here is unfortunate, but better than crashing with no
explanation.

This change is equivalent to r186865 for elf-core.

llvm-svn: 187422

llvm-svn: 187421

llvm-svn: 187420

llvm-svn: 187419

llvm-svn: 187418

llvm-svn: 187417

These tests fail on FreeBSD due to missing build support, the same reason
they fail on Linux.

llvm-svn: 187416

llvm-svn: 187415

llvm-svn: 187414

llvm-svn: 187413

llvm-svn: 187412

llvm-svn: 187411

This patch implements parsing of mips FCC register operands. The example instructions have been added to test files.

llvm-svn: 187410

llvm-svn: 187409

It's a sanity check, mostly, and we've seen threads with >256Mb stack.

llvm-svn: 187408

It needs interface that is missing from the NDK, and it is not used on Android anyway.

llvm-svn: 187407

llvm-svn: 187406

clang/test/Driver/qa_override.c: Resurrect a part of r187376. It still requires the feature 'clang-driver' for cygming.

llvm-svn: 187405

When simplifying a (or (and B A) (and C ~A)) to a (VBSL A B C) ensure that the
bitwidth of the second operands to both ands match before comparing the negation
of the values.

Split the check of the value of the second operands to the ands.  Move the cast
and variable declaration slightly higher to make it slightly easier to follow.

Bug-Id: 16700
Signed-off-by: Saleem Abdulrasool <compnerd@compnerd.org>
llvm-svn: 187404

llvm-svn: 187403

llvm-svn: 187402

Added the notion of Type and TargetSpecific attributes to the clang tablegen.  In turn, this fixes a mistake with Ptr32, Ptr64, UPtr and SPtr attribtues generating AST nodes that are never actually used.

llvm-svn: 187401

llvm-svn: 187400

This is the first of many upcoming patches for PowerPC fast
instruction selection support.  This patch implements the minimum
necessary for a functional (but extremely limited) FastISel pass.  It
allows the table-generated portions of the selector to be created and
used, but in most cases selection will fall back to the DAG selector.
None of the block terminator instructions are implemented yet, and
most interesting instructions require some special handling.
Therefore there aren't any new test cases with this patch.  There will
be quite a few tests coming with future patches.

This patch adds the make/CMake support for the new code (including
tablegen -gen-fast-isel) and creates the FastISel object for PPC64 ELF
only.  It instantiates the necessary virtual functions
(TargetSelectInstruction, TargetMaterializeConstant,
TargetMaterializeAlloca, tryToFoldLoadIntoMI, and FastLowerArguments),
but of these, only TargetMaterializeConstant contains any useful
implementation.  This is present since the table-generated code
requires the ability to materialize integer constants for some
instructions.

This patch has been tested by building and running the
projects/test-suite code with -O0.  All tests passed with the
exception of a couple of long-running tests that time out using -O0
code generation.

llvm-svn: 187399

Refactor some attributes to use checkFunctionOrMethodArgumentIndex instead of using custom logic.  No functional changes intended.

llvm-svn: 187398

build_vector is lowered to REG_SEQUENCE, which is something the register
allocator does a good job at optimizing.

llvm-svn: 187397

This patch prevents the following combine when the input vector is used more
than once.
insert_vector_elt (build_vector elt0, ..., eltN), NewEltIdx, idx
=>
build_vector elt0, ..., NewEltIdx, ..., eltN 

The reasons are:
- Building a vector may be expensive, so try to reuse the existing part of a
  vector instead of creating a new one (think big vectors).
- elt0 to eltN now have two users instead of one. This may prevent some other
  optimizations.

llvm-svn: 187396

Fixed a crasher when using memory threads where a thread is sticking around too long and was causing problems when it didn't have a thread plan. 

llvm-svn: 187395