instruction to insert before can be end(). getDebugLoc on
end() returns an invalid value, therefore use the debug
loc of the call instruction, and give it to InsertLabel.

llvm-svn: 81207

When remat'ing and destination virtual register has a sub-register index. Make sure the sub-register class matches the register class of the remat'ed instruction definition register class.

llvm-svn: 81204

null in the case of an empty struct, so don't try to call getNumValues
on it.

llvm-svn: 81180

from floating-point to integer first, and bitcast the result
back to floating-point. Previously, this test was passing by
falling back to SelectionDAG lowering. The resulting code isn't
as nice, but it's correct and CodeGen now stays on the fast path.

llvm-svn: 81171

llvm-svn: 81124

Patch by Erick Tryzelaar.

llvm-svn: 81116

about by icc (#593, partial).  Patch by Erick Tryzelaar.

llvm-svn: 81115

llvm-svn: 81112

icc (#177, partial).  Patch by Erick Tryzelaar.

llvm-svn: 81106

Do not use DenseMap operator[] because it inserts new entry if lookup fails. Use find() to check an entry in a DenseMap first.

llvm-svn: 81058

llvm-svn: 81055

a new class, MachineInstrIndex, which hides arithmetic details from
most clients. This is a step towards allowing the register allocator
to update/insert code during allocation.

llvm-svn: 81040

llvm-svn: 80997

llvm-svn: 80994

llvm-svn: 80993

llvm-svn: 80976

llvm-svn: 80971

FastISel to recognize this pattern and emit a floating-point
negation using xor.

llvm-svn: 80963

llvm-svn: 80962

Create our own block initializer for kill fixups as the scheduling one wasn't doing the right thing.

llvm-svn: 80958

for the complicated case where one register is tied to multiple destinations.
This avoids the extra scan of instruction operands that was introduced by
my recent change.  I also pulled some code out into a separate
TryInstructionTransform method, added more comments, and renamed some
variables.

Besides all those changes, this takes care of a FIXME in the code regarding
an assumption about there being a single tied use of a register when
converting to a 3-address form.  I'm not aware of cases where that assumption
is violated, but the code now only attempts to transform an instruction,
either by commuting its operands or by converting to a 3-address form,
for the simple case where there is a single pair of tied operands.

llvm-svn: 80945

swapping the operands if necessary.

llvm-svn: 80940

Fixed a test that ensures the LocalRewriter does not attempt to
avoid reloads by reusing clobbered registers.

This was causing issues in 256.bzip2 when compiled with PIC for
a while (starting at r78217), though the problem has since been masked. 

llvm-svn: 80872

llvm-svn: 80798

llvm-svn: 80773

to find the kill, since we already have the operand.

llvm-svn: 80736

llvm-svn: 80733

llvm-svn: 80731

Remove Offset from ExternalSybmol MachineOperands, this is unused (and at least partly unsupported, in X86 encoding at least).

llvm-svn: 80726

llvm-svn: 80717

llvm-svn: 80710

llvm-svn: 80704

llvm-svn: 80702

llvm-svn: 80689

EmitLabel()

llvm-svn: 80677

tied to different source registers, the TwoAddressInstructionPass needs to
be smarter.  Change it to check before replacing a source register whether
that source register is tied to a different destination register, and if so,
defer handling it until a subsequent iteration.

llvm-svn: 80654

makes an eggregious hack somewhat more palatable. Bringing the LSDA forward
and making it a GV available for reference would be even better, but is
beyond the scope of what I'm looking to solve at this point.

Objective C++ code could generate function names that broke the previous
scheme. This fixes that.

llvm-svn: 80649

llvm-svn: 80637

llvm-svn: 80633

llvm-svn: 80631