llvm-svn: 109581

llvm-svn: 109510

llvm-svn: 109509

llvm-svn: 109506

llvm-svn: 109502

llvm-svn: 109424

dependence on DominanceFrontier. Instead, add an explicit DominanceFrontier
pass in StandardPasses.h to ensure that it gets scheduled at the right
time.

Declare that loop unrolling preserves ScalarEvolution, and shuffle some
getAnalysisUsages.

This eliminates one LoopSimplify and one LCCSA run in the standard
compile opts sequence.

llvm-svn: 109413

llvm-svn: 109412

explicit inequality check.

llvm-svn: 109401

llvm-svn: 109400

build of llvm-gcc.

llvm-svn: 109355

it needs to be mapped to refer to the function in the new module, not the old
one. Fixes PR7700.

llvm-svn: 109353

llvm-svn: 109132

llvm-svn: 109122

A non function local MDNode can have an operand which is cloned by MapValue(). 

llvm-svn: 109117

llvm-svn: 109103

llvm-svn: 109101

llvm-svn: 109100

llvm-svn: 109099

llvm-svn: 109098

llvm-svn: 109092

for creating and populating NamedMDNodes.

llvm-svn: 109061

llvm-svn: 109045

llvm-svn: 108968

llvm-svn: 108967

llvm-svn: 108962

llvm-svn: 108961

avoiding MDNode overhead.

llvm-svn: 108909

use getIncomingValueForBlock instead of
LoopInfo::getCanonicalInductionVariableIncrement.

llvm-svn: 108865

llvm-svn: 108736

llvm-svn: 108687

llvm-svn: 108614

so there is no locking involved in type refinement.

llvm-svn: 108553

a redundant loopsimplify run from the default -O2 sequence.

llvm-svn: 108539

llvm-svn: 108530

llvm-svn: 108522

llvm-svn: 108517

Working on testcases for Owen.

llvm-svn: 108494

different widths. In a use with a narrower fixup, formulae
may be wider than the fixup, in which case the high bits
aren't necessarily meaningful, so it isn't safe to reuse
them for uses with wider fixups.

This fixes PR7618, though the testcase is too large for a
reasonable regression test, since it heavily dependes on
hitting LSR's heuristics in a certain way.

llvm-svn: 108455

llvm-svn: 108453