llvm-svn: 153132

instead of skipping the current loop.

My prior fix was incomplete because of an overzealous compile-time optimization:
Better fix for: <rdar://problem/11049788> Segmentation fault: 11 in LoopStrengthReduce

llvm-svn: 153131

llvm-svn: 153130

llvm-svn: 153129

expression parser code.

llvm-svn: 153128

llvm-svn: 153127

llvm-svn: 153126

llvm-svn: 153125

llvm-svn: 153124

llvm-svn: 153123

llvm-svn: 153122

Fixes PR12284. The test case only triggered under asan/valgrind, but it's better than nothing.

llvm-svn: 153120

via functions for certain pointer initialization
fields. // rdar://11076938

llvm-svn: 153117

llvm-svn: 153116

TrackingVH notices when it gets RAUW'd. Fixes PR12305 and PR12315.

llvm-svn: 153115

precedence over the VINSERTF128 avx1 patterns.

llvm-svn: 153114

llvm-svn: 153113

ARMBaseRegisterInfo::canRealignStack was checking for variable-sized objects
but not for stack adjustments around calls.  Use hasReservedCallFrame() to
check for both.  The hasBasePointer function was already correctly checking
both conditions, so the effect of this was that a base pointer would be used
without checking whether the base pointer register could be reserved. I don't
have a small testcase for this.

<rdar://problem/11075906>

llvm-svn: 153110

ARMFrameLowering::hasReservedCallFrame is already checking for variable
sized objects, so there's no point in checking it twice.

llvm-svn: 153109

llvm-svn: 153108

llvm-svn: 153107

llvm-svn: 153106

llvm-svn: 153105

Platforms can now auto-select themselves if you specify a full target triple when doing a "target create" command.

Each platform now knows if it can handle an architecture and a platform can be found using an architecture. Each platform can look at the arch, vendor and OS and know if it should be used or not.

llvm-svn: 153104

whitespace from test case.  No functional change intended.

llvm-svn: 153103

llvm-svn: 153102

llvm-svn: 153101

llvm-svn: 153099

changes to how meta-data is declared. 
// rdar://11076938

llvm-svn: 153098

Needed when building -mdynamic-no-pic code.

rdar://10459256

llvm-svn: 153097

llvm-svn: 153096

llvm-svn: 153095

one place and use it throughout. Also, change ivar name to avoid
name collisions. // rdar://11079366

llvm-svn: 153093

This results in things such as

	vmovups	16(%rdi), %xmm0
	vinsertf128	$1, %xmm0, %ymm0, %ymm0

to be combined to

    vinsertf128	$1, 16(%rdi), %ymm0, %ymm0

rdar://11076953

llvm-svn: 153092

From the Intel Optimization Reference Manual, Section 11.6.2.  When data cannot
be aligned or alignment is not known, 16-byte memory accesses may provide better
performance.
rdar://11076953

llvm-svn: 153091

llvm-svn: 153090

The ARM instructions that have an unpredictable behavior when the pc register operand is given now fail with soft fail. Modified the regression tests to reflect this.

llvm-svn: 153089

llvm-svn: 153086

llvm-svn: 153085

llvm-svn: 153084