llvm-svn: 134048

This function has to deal with a lot of special cases, and the old
version got it wrong sometimes. In particular, it would sometimes leave
multiple uses in the stack interval in a single block. That causes bad
code with multiple reloads in the same basic block.

The new version handles block entry and exit in a single pass. It first
eliminates all the easy cases, and then goes on to create a local
interval for the blocks with difficult interference. Previously, we
would only create the local interval for completely isolated blocks.

It can happen that the stack interval becomes completely empty because
we could allocate a register in all edge bundles, and the new local
intervals deal with the interference. The empty stack interval is
harmless, but we need to remove a SplitKit assertion that checks for
empty intervals.

llvm-svn: 134047

llvm-svn: 134046

wipe out stack blocks when they go out of scope.
// rdar://9227352

llvm-svn: 134045

llvm-svn: 134042

specifiers.  Fixes <rdar://problem/9607158>." because it causes false positives
on some code that uses CF toll free bridging.
 - I'll let Doug or Ted figure out the right fix here, possibly just to accept
   any pointer type.

llvm-svn: 134041

llvm-svn: 134040

This was part of Kaelyn's original patch that got dropped while I was
working on it, but after I ran my tests. =/ Sorry.

llvm-svn: 134039

to the same declaration when correcting typos. This is done by
essentially sorting the corrections as they're added.

Original patch by Kaelyn Uhrain, but modified for style and correctness
by accounting for more than just the textual spelling.

This still is a bit of a WIP hack to make this deterministic. Kaelyn
(and myself) are working on a more principled solution going forward.

llvm-svn: 134038

llvm-svn: 134037

llvm-svn: 134036

now (incorrectly) pass.  Once the appropriate fixes have been made this test
should be reverted.

llvm-svn: 134035

llvm-svn: 134033

up several places where we never expect to have NULL pointers to assert
early.

This fixes a valgrind error within CorrectTypo, but not the
non-determinism.

llvm-svn: 134032

llvm-svn: 134031

llvm-svn: 134030

llvm-svn: 134029

llvm-svn: 134028

llvm-svn: 134027

llvm-svn: 134026

llvm-svn: 134025

llvm-svn: 134024

remove @expectedFailure from TestObjCMethods.py's print_ivars_correctly()
function (it has been passing for a while).

llvm-svn: 134022

sink them into MC layer.
- Added MCInstrInfo, which captures the tablegen generated static data. Chang
TargetInstrInfo so it's based off MCInstrInfo.

llvm-svn: 134021

llvm-svn: 134020

two:

eOptionMarkPCSourceLine = (1u << 2), // Mark the source line that contains the current PC (mixed mode only)
eOptionMarkPCAddress    = (1u << 3)  // Mark the disassembly line the contains the PC

This allows mixed mode to show the line that contains the current PC, and it
allows us to mark the PC address in the disassembly if desired. Having these
be separate gives more control on the disassembly output. SBFrame::Disassemble()
doesn't enable any of these options.

llvm-svn: 134019

Drop the FpMov instructions, use plain COPY instead.

Drop the FpSET/GET instruction for accessing fixed stack positions.
Instead use normal COPY to/from ST registers around inline assembly, and
provide a single new FpPOP_RETVAL instruction that can access the return
value(s) from a call. This is still necessary since you cannot tell from
the CALL instruction alone if it returns anything on the FP stack. Teach
fast isel to use this.

This provides a much more robust way of handling fixed stack registers -
we can tolerate arbitrary FP stack instructions inserted around calls
and inline assembly. Live range splitting could sometimes break x87 code
by inserting spill code in unfortunate places.

As a bonus we handle floating point inline assembly correctly now.

llvm-svn: 134018

character in std::string was causing failures for a few ObjC and Obj-C++ tests
when -flto was enabled.  Revision 133999 resolved this issue. Thanks Jay!
rdar://9685235
PR10210

llvm-svn: 134017

Fixes rdar://9281377

llvm-svn: 134016

a vector for collection. Use iterators where needed instead.
// rdar://6817577

llvm-svn: 134015

llvm-svn: 134014

llvm-svn: 134013

llvm-svn: 134011

llvm-svn: 134010

don't rely on the existence of a copy constructor.

llvm-svn: 134009

llvm-svn: 134008

  vector<int>

to

  std::vector<int>

Patch by Kaelyn Uhrain, with minor tweaks + PCH support from me. Fixes
PR5776/<rdar://problem/8652971>.

Thanks Kaelyn!

llvm-svn: 134007

llvm-svn: 134006

llvm-svn: 134005

avoid getting embedded trailing null bytes in std::strings.

llvm-svn: 133999