This fixes <rdar://problem/5988388> clang ObjC rewriter: mystery 'w' inserted in exception throw.

llvm-svn: 54017

llvm-svn: 54016

llvm-svn: 54015

regressions, such as PR2595. Also, there is a significant code-quality
issue in SPEC 464.h264ref and a few others.

llvm-svn: 54014

llvm-svn: 54013

llvm-svn: 54012

This fixes PR 2573: http://llvm.org/bugs/show_bug.cgi?id=2573

llvm-svn: 54009

llvm-svn: 54008

llvm-svn: 54007

 - Fix multiple issues with the way case ranges were emitted, see test
   cases details about the specific issues.

 - The root issue was not being careful about how basic blocks were
   emitted which led to them being chained together incorrectly,
   resulting in improper control flow.

 - Fixes <rdar://problem/6098585>

llvm-svn: 54006

llvm-svn: 54005

llvm-svn: 54004

llvm-svn: 54003

workaround a GCC 3.3 bug observed on OpenBSD.

llvm-svn: 54002

dependencies with constant load nodes. This allows them to be scheduled
freely.

llvm-svn: 54001

and knowledge of PseudoSourceValues. This unfortunately isn't sufficient to allow
constants to be rematerialized in PIC mode -- the extra indirection is a
complication.

llvm-svn: 54000

Use ASTContext::isObjCObjectPointerType() to check if an ivar is a reference to an Objective-C object.

llvm-svn: 53999

llvm-svn: 53998

command-line option, and disable it by default. It introduced performance
regressions because CodeGen is currently not able to remat such loads.

llvm-svn: 53997

llvm-svn: 53995

This fixes a critical rewriter bug (<rdar://problem/6096760> clang ObjC rewriter: 'self' not expected value in class method called with 'super').

Also added a couple FIXME's since I'm not happy with my fix to Sema. It would be nicer if the super handling for class/instance messages was the same (based on PreDefinedExpr).

llvm-svn: 53994

This fixes: http://llvm.org/bugs/show_bug.cgi?id=2593

llvm-svn: 53993

llvm-svn: 53992

llvm-svn: 53991

large numbers of CBE regressions caused by r53958.

llvm-svn: 53990

This fixes PR 2592: http://llvm.org/bugs/show_bug.cgi?id=2592

llvm-svn: 53987

llvm-svn: 53986

Store the predecessor MBB in the PHIUnion, rather than an index, since the indices will change after renumbering.

llvm-svn: 53985

llvm-svn: 53984

llvm-svn: 53983

Fix a catastrophic PPC64 ABI bug: i32 operands which are passed in memory (all of the parameter registers are used) are loaded from sp offsets that were off by 4.

llvm-svn: 53979

llvm-svn: 53977

Patch by Zhongxing Xu!

llvm-svn: 53976

 - Also cleaned up emission slightly 
 - Inspection of the code revealed several other bugs, however. Case
   ranges are not properly wired and can result in switch cases being
   dropped or even infinite loops. See: <rdar://problem/6098585>

Completes: <rdar://problem/6094119>
llvm-svn: 53975

 - For getting an APSInt from a known integer constant Expr.
 - TODO: Many users of Expr::isIntegerConstantExpr in codegen should
   probably be using this instead...

llvm-svn: 53974

llvm-svn: 53973

llvm-svn: 53972

llvm-svn: 53971

Rename instance variables, parameter argument names to eliminate a bunch of compilation warnings with -Wshadow.

llvm-svn: 53970

llvm-svn: 53969