llvm-svn: 24457

llvm-svn: 24456

port to properly use L for the bb prefix instead of .

llvm-svn: 24454

llvm-svn: 24453

llvm-svn: 24452

llvm-svn: 24451

llvm-svn: 24450

to specific sections.  Delete some dead functions copied from the X86 backend.

llvm-svn: 24449

llvm-svn: 24448

backend to implement global variables in sections.

llvm-svn: 24447

llvm-svn: 24446

asserts.

llvm-svn: 24445

change.

llvm-svn: 24443

llvm-svn: 24442

llvm-svn: 24441

darwin, use it when printing the constant pool indices so the labels are
appropriately private, emit cp entries to .const instead of .data on darwin
and only emit a single .section for the constant pool, not one for each
entry.

llvm-svn: 24440

llvm-svn: 24438

break control dependence.

llvm-svn: 24437

llvm-svn: 24435

llvm-svn: 24434

llvm-svn: 24430

llvm-svn: 24429

HAVE_GV case fixed up by me.

llvm-svn: 24428

from last night.

llvm-svn: 24427

llvm-svn: 24425

Patch by Evan Cheng

llvm-svn: 24423

llvm-svn: 24422

llvm-svn: 24421

llvm-svn: 24420

Cheng!

llvm-svn: 24419

llvm-svn: 24418

packed types with an element count of 1, although more generic support is
coming.  This allows LLVM to turn the following code:

void %foo(<1 x float> * %a) {
entry:
  %tmp1 = load <1 x float> * %a;
  %tmp2 = add <1 x float> %tmp1, %tmp1
  store <1 x float> %tmp2, <1 x float> *%a
  ret void
}

Into:

_foo:
        lfs f0, 0(r3)
        fadds f0, f0, f0
        stfs f0, 0(r3)
        blr

llvm-svn: 24416

half the problem.

llvm-svn: 24414

llvm-svn: 24413

llvm-svn: 24412

has a single def.  In this case, look for uses that are dominated by the def
and attempt to rewrite them to directly use the stored value.

This speeds up mem2reg on these values and reduces the number of phi nodes
inserted.  This should address PR665.

llvm-svn: 24411

llvm-svn: 24410

llvm-svn: 24409

llvm-svn: 24408

llvm-svn: 24406