Patch by Matthijs Kooijman!

llvm-svn: 50861

also need to be checked for memory modifying instructions before we
can sink them.  THis fixes the second half of PR2297.

llvm-svn: 50860

delete the first store in:

store x -> p
load p
store y -> p

This is for PR2297.

llvm-svn: 50859

llvm-svn: 50836

llvm-svn: 50793

llvm-svn: 50775

We need to update PHIs containing the exiting block, not the exit block.  We really should come up with better names for these.

llvm-svn: 50770

llvm-svn: 50713

llvm-svn: 50699

llvm-svn: 50696

ComputeMaskedBits.

llvm-svn: 50692

themselves weren't in the anonymous namespace.

llvm-svn: 50673

DemoteRegToStack doesn't work with MRVs yet, because it relies on the
ability to load/store things.

This fixes PR2285.

llvm-svn: 50667

using dyn_cast<> on already freed memory.

llvm-svn: 50618

llvm-svn: 50600

llvm-svn: 50591

llvm-svn: 50586

llvm-svn: 50582

llvm-svn: 50564

from x86-64 calling convention work.

llvm-svn: 50545

fixes the 254.gap regression on x86 and the 403.gcc regression on x86-64.

llvm-svn: 50537

2) Return NULL instead of false in several places for tidiness.
3) fix a bug optimizing  sprintf(p, "%c", x)

llvm-svn: 50521

a FunctionPass.  This makes it simpler, fixes dozens of bugs, adds
a couple of minor features, and shrinks is considerably: from
2214 to 1437 lines.

llvm-svn: 50520

llvm-svn: 50473

to instcombine.

llvm-svn: 50472

Revert r50441.  The original code was correct.  Add some more comments so that I don't make the same mistake in the future.

llvm-svn: 50446

we were checking for it in the wrong order.  This caused a miscompilation because the
return slot optimization assumes that the call it is dealing with is NOT a memcpy.

llvm-svn: 50444

llvm-svn: 50442

of this was suggested by Chris.

llvm-svn: 50441

llvm-svn: 50437

llvm-svn: 50436

This fixes the second half of PR2262

llvm-svn: 50430

llvm-svn: 50428

llvm-svn: 50406

llvm-svn: 50390

llvm-svn: 50384

llvm-svn: 50383

the trip count.

llvm-svn: 50382

llvm-svn: 50370

ComputeMaskedBits knows about cttz, ctlz, and ctpop. Teach
SelectionDAG's ComputeMaskedBits what InstCombine's knows
about SRem. And teach them both some things about high bits
in Mul, UDiv, URem, and Sub. This allows instcombine and
dagcombine to eliminate sign-extension operations in
several new cases.

llvm-svn: 50358