This is in preparation of supporting multiple global split candidates in
a single live range split operation.

llvm-svn: 135074

llvm-svn: 135072

llvm-svn: 135071

a complete misunderstanding of the code.

llvm-svn: 135070

llvm-svn: 135068

The immediate is of limited range and the operand type should reflect that.

llvm-svn: 135066

llvm-svn: 135064

llvm-svn: 135063

llvm-svn: 135062

llvm-svn: 135061

Keeping the instructions in alphabetical order, just like in the ARM ARM.
Adding FIXMEs for skipped instructions when adding tests out of order.

llvm-svn: 135060

llvm-svn: 135059

It's not safe to fold (fptrunc (sqrt (fpext x))) to (sqrtf x) if there is another use of sqrt. rdar://9763193

llvm-svn: 135058

llvm-svn: 135056

Catch potential cascading errors on a malformed so_reg operand and bail after
the first error.

Add some tests for the diagnostics we do want.

llvm-svn: 135055

llvm-svn: 135053

llvm-svn: 135052

Now works for parsing register shifted register and register shifted
immediate arithmetic instructions, including the 'rrx' rotate with extend.

llvm-svn: 135049

llvm-svn: 135047

Update the debug output interface for MCParsedAsmOperand to have a print()
method which takes an output stream argument, an << operator which invokes
the print method using the given stream, and a dump() method which prints
the operand to the dbgs() stream. This makes the interface more consistent
with the rest of LLVM, and more convenient to use at the debugger command
line.

llvm-svn: 135043

llvm-svn: 135042

llvm-svn: 135040

ExtractValueInst APIs to use ArrayRef: a new constructor taking a
(begin, end) range, and operators == and != for element-wise comparison.

llvm-svn: 135039

llvm-svn: 135037

to Benjamin Kramer for steering me in the right direction here.

llvm-svn: 135031

llvm-svn: 135024

llvm-svn: 135023

not the FDE

llvm-svn: 135020

llvm-svn: 135018

if (x != 0) x = 1
if (x == 1) x = 1

Previous codegen looks like this:
        mov     r1, r0
        cmp     r1, #1
        mov     r0, #0
        moveq   r0, #1

The naive lowering select between two different values. It should recognize the
test is equality test so it's more a conditional move rather than a select:
        cmp     r0, #1
        movne   r0, #0

rdar://9758317

llvm-svn: 135017

llvm-svn: 135015

llvm-svn: 135014

assert when the frame pointer is -1 (i.e., the function is "frameless").

Still to do: "frameless" unwind information.

llvm-svn: 135013

llvm-svn: 135006

llvm-svn: 135000

Leak found by valgrind.

llvm-svn: 134994

llvm-svn: 134987

Use memory barriers to force if-conversion off for these tests instead of
the internal llc command line option ifcvt-limit.

llvm-svn: 134986

1,$s/llvm::Type::getInt\(..\)Ty(builder.getContext())/builder.getInt\1Ty()/g
1,$s/builder.getInt\(..*\)Ty()->getPointerTo()/builder.getInt\1PtrTy()/g

vi sub commands (second one was not a reversion but requested by 
Tobias Grosser.

Mod was tested, but other examples have failed to build as they are currently 
being thrashed with the const qualifier removal change.

llvm-svn: 134985

StructType::get() and TargetData::getIntPtrType().

llvm-svn: 134982