llvm-svn: 60442

llvm-svn: 60441

llvm-svn: 60440

consistent with the way it's generally done in other places.

llvm-svn: 60439

- Incorporate Tilmann Scheller's ISD::TRUNCATE custom lowering patch
- Update SPU calling convention info, even if it's not used yet (but can be
  at some point or another)
- Ensure that any-extended f32 loads are custom lowered, especially when
  they're promoted for use in printf.

llvm-svn: 60438

- Added method "setPTHManager" that will be called by the driver to install
  a PTHManager for the Preprocessor.
- Fixed some comments.
- Added EnterSourceFileWithPTH to mirror EnterSourceFileWithLexer.

llvm-svn: 60437

Added PTHManager, a utility class that will be used by Preprocessor to lazily create PTHLexer objects for pre-tokenized files.

llvm-svn: 60436

- Output 32 bit integers using bit-shifting + write of individual bytes.
  This is motivated because we aren't guaranteed to load 32-bit ints of the mmaped PTH file at 4-byte offsets.
- Don't emit flags for IdentifierInfos.  These are lazily populated by the Preprocessor/Parser.
- Only write out tokens for files with absolute paths.  This is potentially temporary, but simplifies things for now.

llvm-svn: 60435

llvm-svn: 60434

llvm-svn: 60433

llvm-svn: 60432

llvm-svn: 60431

work in prgress.

llvm-svn: 60430

llvm-svn: 60429

llvm-svn: 60428

llvm-svn: 60427

llvm-svn: 60426

llvm-svn: 60425

llvm-svn: 60423

This may be the case on 64-bit systems. Whether that fact is a bug is a different question, but it's easy to cure the symptom.

llvm-svn: 60422

llvm-svn: 60421

llvm-svn: 60420

llvm-svn: 60418

llvm-svn: 60416

ScalarExprEmitter::VisitBinLOr.

llvm-svn: 60415

llvm-svn: 60414

llvm-svn: 60413

llvm-svn: 60409

straight-forward implementation.  This does not require any extra
alias analysis queries beyond what we already do for non-local loads.

Some programs really really like load PRE.  For example, SPASS triggers
this ~1000 times, ~300 times in 255.vortex, and ~1500 times on 403.gcc.

The biggest limitation to the implementation is that it does not split
critical edges.  This is a huge killer on many programs and should be
addressed after the initial patch is enabled by default.

The implementation of this should incidentally speed up rejection of 
non-local loads because it avoids creating the repl densemap in cases 
when it won't be used for fully redundant loads.

This is currently disabled by default.
Before I turn this on, I need to fix a couple of miscompilations in
the testsuite, look at compile time performance numbers, and look at
perf impact.  This is pretty close to ready though.

llvm-svn: 60408

llvm-svn: 60407

llvm-svn: 60406

llvm-svn: 60405

llvm-svn: 60404

llvm-svn: 60403

llvm-svn: 60402

llvm-svn: 60401

llvm-svn: 60400

  constant. If X is a constant, then this is folded elsewhere.

- Added a note to Target/README.txt to indicate that we'd like to implement
  this when we're able.

llvm-svn: 60399

llvm-svn: 60398

- No need to do a swap on a canonicalized pattern.

No functionality change.

llvm-svn: 60397