Since LCSSA switched over to DenseMap, we have to be more careful to avoid iterator invalidation.  Fixes PR2385.

llvm-svn: 51777

once (ie, at two different places in the source, not two times on the
commandline).

llvm-svn: 51771

llvm-svn: 51770

insertvalue / extractvalue instructions.

llvm-svn: 51766

llvm-svn: 51761

llvm-svn: 51721

llvm-svn: 51718

llvm-svn: 51717

Make the renumbering correct in the face of deleted instructions that have been removed from the LiveIntervals maps.

llvm-svn: 51714

llvm-svn: 51709

cases due to an isel deficiency already noted in
lib/Target/X86/README.txt, but they can be matched in this fold-call.ll
testcase, for example.

This is interesting mainly because it exposes a tricky tblgen bug;
tblgen was incorrectly computing the starting index for variable_ops
in the case of a complex pattern.

llvm-svn: 51706

function to flush a specified std::ostream.

llvm-svn: 51705

llvm-svn: 51704

llvm-svn: 51701

definitions. This adds a new construct, "discard", for indicating
that a named node in the input matching pattern is to be discarded,
instead of corresponding to a node in the output pattern. This
allows tblgen to know where the arguments for the varaible_ops are
supposed to begin.

This fixes "rdar://5791600", whatever that is ;-).

llvm-svn: 51699

llvm-svn: 51698

memmove to a more plausible value, now that it's actually being used.

llvm-svn: 51696

llvm-svn: 51694

Renumbering needs to account for instruction slot offsets when performing lookups in the index maps.

llvm-svn: 51691

llvm-svn: 51686

llvm-svn: 51680

the one case that ADCE catches that normal DCE doesn't: non-induction variable
loop computations.

This implementation handles this problem without using postdominators.

llvm-svn: 51668

llvm-svn: 51667

llvm-svn: 51666

llvm-svn: 51664

llvm-svn: 51663

like.

llvm-svn: 51662

llvm-svn: 51661

instruction to execute. This can be used for transformations (like two-address
conversion) to remat an instruction instead of generating a "move"
instruction. The idea is to decrease the live ranges and register pressure and
all that jazz.

llvm-svn: 51660

that it's safe to remat an instruction.

llvm-svn: 51659

llvm-svn: 51658

Factor the numbering computation into a separate method, and add the slightest attempt at some renumbering logic, which is currently unused.

llvm-svn: 51652

llvm-svn: 51648

llvm-svn: 51647

llvm-svn: 51644

llvm-svn: 51636

llvm-svn: 51635

llvm-svn: 51634

on ia64.

llvm-svn: 51633

llvm-svn: 51632