is necessary for eventual gcc commmand line compatibility.

llvm-svn: 63384

 - This brings us down to an 8% failure rate on the first 500 return
   types tests (from 12%).

llvm-svn: 63383

llvm-svn: 63382

an illegal type.

llvm-svn: 63380

llvm-svn: 63379

llvm-svn: 63378

Enable emitting of constant values in non-default address space as well. The APIs emitting constants now take an additional parameter signifying the address space in which to emit. The APIs like getData8BitsDirective() etc are made virtual enabling targets to be able to define appropirate directivers for various sizes and address spaces.

llvm-svn: 63377

llvm-svn: 63376

llvm-svn: 63375

llvm-svn: 63374

llvm-svn: 63373

llvm-svn: 63372

llvm-svn: 63371

If a MachineInstr doesn't have a memoperand but has an opcode that
is known to load or store, assume its memory reference may alias
*anything*, including stack slots which the compiler completely
controls.

To partially compensate for this, teach the ScheduleDAG building
code to do basic getUnderlyingValue analysis. This greatly
reduces the number of instructions that require restrictive
dependencies. This code will need to be revisited when we start
doing real alias analysis, but it should suffice for now.

llvm-svn: 63370

llvm-svn: 63369

llvm-svn: 63368

llvm-svn: 63367

llvm-svn: 63366

- Modify ReassociateOps so that the resulting SDValue is what the comment claims
  it is.

llvm-svn: 63365

llvm-svn: 63364

llvm-svn: 63363

the constructed APSInt, patch suggested by Ben Laurie!

llvm-svn: 63362

llvm-svn: 63361

llvm-svn: 63360

llvm-svn: 63359

llvm-svn: 63358

llvm-svn: 63357

llvm-svn: 63356

llvm-svn: 63355

LookupName et al. Instead, use an enum and a bool to describe its
contents.

Optimized the C/Objective-C path through LookupName, eliminating any
unnecessarily C++isms. Simplify IdentifierResolver::iterator, removing
some code and arguments that are no longer used.

Eliminated LookupDeclInScope/LookupDeclInContext, moving all callers
over to LookupName, LookupQualifiedName, or LookupParsedName, as
appropriate.

All together, I'm seeing a 0.2% speedup on Cocoa.h with PTH and
-disable-free. Plus, we're down to three name-lookup routines.

llvm-svn: 63354

llvm-svn: 63353

llvm-svn: 63352

(objc2 nonfragile-abi).

llvm-svn: 63351

llvm-svn: 63350

 - -Xclang always forwards to clang

 - -Xanalyzer replaces -WA,; it seems like the cleaner mechanism and
    is more readable.

llvm-svn: 63349

llvm-svn: 63348

- NonLoc::MakeVal() would use sizeof(unsigned) (literally) instead of consulting
  ASTContext for the size (in bits) of 'int'. While it worked, it was a
  conflation of concepts and using ASTContext.IntTy is 100% correct.
- RegionStore::getSizeInElements() no longer assumes that a VarRegion has the
  type "ConstantArray", and handles the case when uses use ordinary variables
  as if they were arrays.
- Fixed ElementRegion::getRValueType() to just return the rvalue type of its
  "array region" in the case the array didn't have ArrayType.
- All of this fixes <rdar://problem/6541136>

llvm-svn: 63347

This redoes the default mode that ccc runs in w.r.t. using clang. Now
ccc defaults to always using clang for any task clang can
handle. However, the following options exist to tweak this behavior:

 -ccc-no-clang: Don't use clang at all for compilation (still used for
  static analysis).
 
 -ccc-no-clang-cxx: Don't use clang for C++ and Objective-C++ inputs.

 -ccc-no-clang-cpp: Don't use clang as a preprocessor.

 -ccc-clang-archs <archs>: If present, only use clang for the given
  comma separated list of architectures. This only works on Darwin for
  now.

Note that all -ccc options must be first on the command line.

llvm-svn: 63346

cpp-output, not c-cpp-output.

llvm-svn: 63345

non-fragile abi.

llvm-svn: 63343