R600 doesn't need to do any scheduling on the SelectionDAG now that it
has a very good MachineScheduler.  Also, using the VLIW SelectionDAG
scheduler was having a major impact on compile times. For example with
the phatk kernel here are the LLVM IR to machine code compile times:

With Sched::VLIW

Total Compile Time:                  1.4890 Seconds (User + System)
SelectionDAG Instruction Scheduling: 1.1670 Seconds (User + System)

With Sched::Source

Total Compile Time:                  0.3330 Seconds (User + System)
SelectionDAG Instruction Scheduling: 0.0070 Seconds (User + System)

The code ouput was identical with both schedulers.  This may not be true
for all programs, but it gives me confidence that there won't be much
reduction, if any, in code quality by using Sched::Source.

llvm-svn: 188215

llvm-svn: 188214

llvm-svn: 188213

incomplete types, courtesy of Luke Zarko.

llvm-svn: 188212

The cmake-clang-x86_64 was upset:
error: 'template<class ImplClass, class RetTy> class clang::ConstStmtVisitor' used without template parameters

llvm-svn: 188211

This is used to name the linked output file.

Differential Revision: http://llvm-reviews.chandlerc.com/D1344

llvm-svn: 188210

This is just a couple of minor fixes to account for the existence
of ElaboratedType.

llvm-svn: 188209

I'm not really satisfied with the ad-hoc nature of
Sema::diagnoseQualifiedDeclaration, but I'm not sure how to fix it.

Fixes <rdar://problem/14639501>.

llvm-svn: 188208

Reviewed by delesley, dblaikie.

Add the annotations and code needed to support a basic 'consumed' analysis.

Summary:
This new analysis is based on academic literature on linear types.  It tracks
the state of a value, either as unconsumed, consumed, or unknown.  Methods are
then annotated as CallableWhenUnconsumed, and when an annotated method is
called while the value is in the 'consumed' state a warning is issued.  A value
may be tested in the conditional statement of an if-statement; when this occurs
we know the state of the value in the different branches, and this information
is added to our analysis.  The code is still highly experimental, and the names
of annotations or the algorithm may be subject to change.

llvm-svn: 188206

llvm-svn: 188205

In order to appease people (in Apple) who accuse me for committing "huge change" (?) without proper review. 

Thank Eric for fixing a compile-warning. 

llvm-svn: 188204

llvm-svn: 188203

llvm-svn: 188202

llvm-svn: 188201

llvm-svn: 188200

has been finalized except for sizes and offsets. Update test accordingly.

llvm-svn: 188199

llvm-svn: 188198

llvm-svn: 188197

llvm-svn: 188196

llvm-svn: 188195

We support this libcall on all platforms except for OpenBSD (See
lib/Codegen/StackProtector.cpp).

llvm-svn: 188193

Nico Rieck:  this patch series fixes visibility issues on Windows as explained in <http://lists.cs.uiuc.edu/pipermail/cfe-dev/2013-August/031214.html>.

llvm-svn: 188192

llvm-svn: 188191

llvm-svn: 188190

llvm-svn: 188189

  1. Add some helper classes for partitions. They are designed in a
     way such that the top-level LTO driver will not see much difference 
     with or without partitioning.

  2. Introduce work-dir. Now all intermediate files generated during 
     LTO phases will be saved under work-dir. User can specify the workdir
     via -lto-workdir=/path/to/dir. By default the work-dir will be 
     erased before linker exit. To keep the workdir, do -lto-keep, or -lto-keep=1.

    TODO: Erase the workdir, if the linker exit prematurely.  
      We are currently not able to remove directory on signal. The support 
      routines simply ignore directory.

  3. Add one new API lto_codegen_get_files_need_remove().
     Linker and LTO plugin will communicate via this API about which files
    (including directories) need to removed before linker exit.

llvm-svn: 188188

Patch by Ron Ofir.

llvm-svn: 188187

Do not generate new vector values for the same entries because we know that the incoming values
from the same block must be identical.

llvm-svn: 188185

This patch adds -mmsa and -mno-msa to the options supported by 
clang to enable and disable support for MSA.

When MSA is enabled, a predefined macro '__mips_msa' is defined to 1.

Patch by Daniel Sanders

llvm-svn: 188184

This fixes the incorrect implementation of iterating on file/directory
paths.

Differential Review: http://llvm-reviews.chandlerc.com/D1277

llvm-svn: 188183

llvm-svn: 188182

Not hooked up to R600 yet due to current lack of support, at least on EG.

Signed-off-by: Aaron Watry <awatry@gmail.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 188181

It's supported by the R600 LLVM back-end now, at least for evergreen.

Signed-off-by: Aaron Watry <awatry@gmail.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 188180

Signed-off-by: Aaron Watry <awatry@gmail.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 188179

- Open files before calling stat on them.
- Go through FileManager for getting the buffer of named pipes. It has the
  necessary plumbing to deal with "volatile" files.
- Print the cause when stdin reading fails. The only case I can imagine where
  this happens is when stdin is wired to a device file, so no test case.

llvm-svn: 188178

llvm-svn: 188177

llvm-svn: 188176

Before:
  CHECK(controller->WriteProto(FLAGS_row_key, FLAGS_proto)) << "\""
                                                            << FLAGS_proto
                                                            << "\"";

After:
  SemaRef.Diag(Loc, diag::note_for_range_begin_end)
      << BEF << IsTemplate << Description << E->getType();

llvm-svn: 188175

Various tests had sprung up over the years which had --check-prefix=ABC on the
RUN line, but "CHECK-ABC:" later on. This happened to work before, but was
strictly incorrect. FileCheck is getting stricter soon though.

Patch by Ron Ofir.

llvm-svn: 188174

Various tests had sprung up over the years which had --check-prefix=ABC on the
RUN line, but "CHECK-ABC:" later on. This happened to work before, but was
strictly incorrect. FileCheck is getting stricter soon though.

Patch by Ron Ofir.

llvm-svn: 188173