llvm-svn: 130866

llvm-svn: 130865

llvm-svn: 130864

Teach MachineCSE how to do simple cross-block CSE involving physregs.  This allows, for example, eliminating duplicate cmpl's on x86. Part of rdar://problem/8259436 .

llvm-svn: 130862

Clang.

llvm-svn: 130861

llvm-svn: 130860

llvm-svn: 130859

These tests all follow the same pattern:

	mov	r2, r0
	movs	r0, #0
	$CMP	r2, r1
	it	eq
	moveq	r0, #1
	bx	lr

The first 'mov' can be eliminated by rematerializing 'movs r0, #0' below the
test instruction:

	$CMP	r0, r1
	mov.w	r0, #0
	it	eq
	moveq	r0, #1
	bx	lr

So far, only physreg coalescing can do that. The register allocators won't yet
split live ranges just to eliminate copies. They can learn, but this particular
problem is not likely to show up in real code. It only appears because r0 is
used for both the function argument and return value.

llvm-svn: 130858

llvm-svn: 130857

llvm-svn: 130856

llvm-svn: 130855

llvm-svn: 130854

llvm-svn: 130853

Adjacent bit fields are packed into the same 1-, 2-, or
4-byte allocation unit if the integral types are the same
size. // rdar://8823265.

llvm-svn: 130851

llvm-svn: 130850

llvm-svn: 130849

llvm-svn: 130848

Prevent instructions using $gp from being placed between a jalr and the instruction that restores the clobbered $gp.

llvm-svn: 130847

it is both inefficient and unexpected by dwarfdump. Change to
a DW_FORM_data4.

While in here, change the predicate name to reflect that the position
is not really absolute (it is an offset), just that the linker needs a
relocation.

llvm-svn: 130846

Physreg joining is still on by default, but I will turn it off shortly.

llvm-svn: 130844

Minor cleanup:  When asynchronous output comes along,
back up over and delete the prompt before writing out the
asynchronous output, rather than just going to a new line.

llvm-svn: 130843

Tighten up check for empty (i.e. no meaningful debug info) module. This fixes dwarf-die2.c test case from gcc test suite.

llvm-svn: 130842

but according to my super-optimizer there are only two missed simplifications
of -instsimplify kind when compiling bzip2, and this is one of them.  It amuses
me to have bzip2 be perfectly optimized as far as instsimplify goes!

llvm-svn: 130840

llvm-svn: 130839

Like in r126648, provide (empty) default implementation for pure virtual getMemoryBufferSizes(). Not all use cases have meaningful implementations.

llvm-svn: 130838

Add C++ include paths for libstdc++ with Ubuntu 11.04 and Debian's libstdc++6-4.5, as collected from http://packages.ubuntu.com/search?mode=exactfilename&suite=natty&section=all&arch=any&searchon=contents&keywords=c%2B%2Blocale.h and http://packages.debian.org/wheezy/armel/libstdc++6-4.5-dev/filelist
Fixes compilation with clang++ on Ubuntu 11.04.

llvm-svn: 130837

This is more efficient as it's all done at once at the end of the TU.
This could still get expensive, so a flag is provided to disable it. As
an added bonus, the diagnostics will now print out a cycle.

The PCH test is XFAILed because we currently can't deal with a note
emitted in the header and I, being tired, see no other way to verify the
serialization of delegating constructors. We should probably address
this problem /somehow/ but no good solution comes to mind.

llvm-svn: 130836

frontend, if applicable.

llvm-svn: 130835

llvm-svn: 130834

llvm-svn: 130833

Change the boolean "use_dynamic" over to a tri-state, no-dynamic, dynamic-w/o running target,
and dynamic with running target.

llvm-svn: 130832

llvm-svn: 130831

llvm-svn: 130830

This adds functionality to remove size/zero extension during indvars
without generating a canonical IV and rewriting all IV users. It's
disabled by default so should have no effect on codegen. Work in progress.

llvm-svn: 130829

llvm-svn: 130828

if they match that production, i.e. if they're template type parameters
or decltypes (or, as an obvious case not yet described in the ABI document,
if they're template template parameters applied to template arguments).

llvm-svn: 130824

Enhance clang_getCXTUResourceUsage() to return the amount of memory used by the Preprocessor's bump allocator as well as those from the PreprocessingRecord.

llvm-svn: 130823

llvm-svn: 130822

declarations.

llvm-svn: 130821

not, I'll just add them here and be done with it.

llvm-svn: 130819