llvm-svn: 43698

Removed all macro code for PIC (goodbye "la").
Support tested with shootout bench.

llvm-svn: 43697

-emitbitcode option and the addition of -preverify.

llvm-svn: 43696

llvm-svn: 43695

llvm-svn: 43694

llvm-svn: 43693

llvm-svn: 43692

that there is no padding.

llvm-svn: 43691

Remove HandleObjcMetaDataEmission(), I inadvertantly resurrected it.

llvm-svn: 43690

llvm-svn: 43689

should only effect x86 when using long double.  Now
12/16 bytes are output for long double globals (the
exact amount depends on the alignment).  This brings
globals in line with the rest of LLVM: the space
reserved for an object is now always the ABI size.
One tricky point is that only 10 bytes should be
output for long double if it is a field in a packed
struct, which is the reason for the additional
argument to EmitGlobalConstant.

llvm-svn: 43688

Two additions...

- Synthesize the funky cast for objc_msgSend(). For the basic case, it looks like...

((id (*)(id, SEL))(void *)objc_msgSend)(obj, sel);

The "void *" cast is needed to workaround a GCC "bandaid" (Chris says it has something to do with the inliner). Without the extra "void *" cast, we get spurious warnings/notes that look like...

xx.m:17: warning: function called through a non-compatible type
xx.m:17: note: if this code is reached, the program will abort

- Add prototypes for the ObjC functions we call, objc_msgSend/objc_getClass for now (don't depend on them being included).

llvm-svn: 43685

llvm-svn: 43684

llvm-svn: 43683

llvm-svn: 43682

Evan, please review this.

llvm-svn: 43680

llvm-svn: 43679

llvm-svn: 43678

removed. This document could still stand for significant improvement:

  * Editing the pass descriptions; most were lifted with minimal editing from
    comments. Although implementation details were elided, many of the were not
    written for the audience that would be interested in this document.
  * More "before and after" examples.
  * More implicit dependency details. (Perhaps listing transforms in
    -std-compile-opts order would help alleviate this.)
  * Adding documentation for how to invoke passes programmatically.
  * Rearranging the document into a more logical taxonomy. For instance, putting
    profiling passes together.

llvm-svn: 43677

llvm-svn: 43676

llvm-svn: 43675

Also cleaned up some comments in source files.

llvm-svn: 43674

llvm-svn: 43673

or getTypeSizeInBits as appropriate in ScalarReplAggregates.
The right change to make was not always obvious, so it would
be good to have an sroa guru review this.  While there I noticed
some bugs, and fixed them: (1) arrays of x86 long double have
holes due to alignment padding, but this wasn't being spotted
by HasStructPadding (renamed to HasPadding).  The same goes
for arrays of oddly sized ints.  Vectors also suffer from this,
in fact the problem for vectors is much worse because basic
vector assumptions seem to be broken by vectors of type with
alignment padding.   I didn't try to fix any of these vector
problems.  (2) The code for extracting smaller integers from
larger ones (in the "int union" case) was wrong on big-endian
machines for integers with size not a multiple of 8, like i1.
Probably this is impossible to hit via llvm-gcc, but I fixed
it anyway while there and added a testcase.  I also got rid of
some trailing whitespace and changed a function name which
had an obvious typo in it.

llvm-svn: 43672

If an interval is being undone clear its preference as well since the source interval may have been undone as well.

llvm-svn: 43670

regs on x86-64.

llvm-svn: 43669

metric is way off for these in general, and this works around
buggy code like that in PR1764.  we'll see if there is a big
performance impact of this.  If so, I'll revert it tomorrow.

llvm-svn: 43668

llvm-svn: 43667

Add a method prototype slot/getter to the ObjCMessageExpr AST.

llvm-svn: 43666

Implement rewrite rules for ObjC string constants.

llvm-svn: 43665

llvm-svn: 43663

can be eliminated by the allocator is the destination and source targets the
same register. The most common case is when the source and destination registers
are in different class. For example, on x86 mov32to32_ targets GR32_ which
contains a subset of the registers in GR32.

The allocator can do 2 things:
1. Set the preferred allocation for the destination of a copy to that of its source.
2. After allocation is done, change the allocation of a copy destination (if
   legal) so the copy can be eliminated.

This eliminates 443 extra moves from 403.gcc.

llvm-svn: 43662

ASTConsumer to process the AST before it is destroyed.
This allows elimination of HandleObjcMetaDataEmission.

llvm-svn: 43659

My previous patch did the same for @catch AST.

llvm-svn: 43654

llvm-svn: 43653

llvm-svn: 43652

llvm-svn: 43651

llvm-svn: 43649

key functions to implement.

llvm-svn: 43648

the target pointer to be passed by reference.  This can result in less
typing, as the object to be deserialized can be inferred from the
argument.

llvm-svn: 43647