We had stored both f64 values and v2f64, etc. values in the VSX registers. This
worked, but was suboptimal because we would always spill 16-byte values even
through we almost always had scalar 8-byte values. This resulted in an
increase in stack-size use, extra memory bandwidth, etc. To fix this, I've
added 64-bit subregisters of the Altivec registers, and combined those with the
existing scalar floating-point registers to form a class of VSX scalar
floating-point registers. The ABI code has also been enhanced to use this
register class and some other necessary improvements have been made.

llvm-svn: 205075

llvm-svn: 205074

llvm-svn: 205073

llvm-svn: 205072

[HeaderSearch] Make sure we clear the mapped name from the LookupFileCacheInfo when we reset the start point.

rdar://16462455

llvm-svn: 205071

Canonicalise the default triple that is used on Windows.  This should hopefully
fix the MSVC buildbots.

llvm-svn: 205070

Improve -Wunreachable-code to provide a means to indicate code is intentionally marked dead via if((0)).

Taking a hint from -Wparentheses, use an extra '()' as a sigil that
a dead condition is intentionally dead.  For example:

  if ((0)) { dead }

When this sigil is found, do not emit a dead code warning.  When the
analysis sees:

  if (0)

it suggests inserting '()' as a Fix-It.

llvm-svn: 205069

of TemplateArgumentLocs. 'uint64_t' has higher alignment requirements than a
pointer on some platforms.

llvm-svn: 205068

Emit 32-bit register names instead of 64-bit register names if the target does
not have 64-bit general purpose registers.

<rdar://problem/14653996>

llvm-svn: 205067

default until we provide a work-around for warning
suppression.

llvm-svn: 205066

[SemaObjC] For the semantics of the designated-initializer attribute, consider that the ObjC interface of the class
includes its class extensions.

This is follow-up for rdar://16305347

llvm-svn: 205065

unless the macro comes from a system header

llvm-svn: 205064

Committed this by accident before it was done last time.

Original message:

    Rather than rolling our own functions to read little endian data
    from a buffer, we can use the support in llvm's Endian.h.

    No functional change.

llvm-svn: 205062

Committed this by accident before it was done last time.

Original message:

    Rather than rolling our own functions to write little endian data
    to an ostream, we can use the support in llvm's EndianStream.h.

    No functional change.

llvm-svn: 205061

for expression evaluations that try one and then all threads.

<rdar://problem/15598528>

llvm-svn: 205060

Turns out debug_frame does use multiple fragments, so it doesn't
compress correctly with the current approach. Disable compressing it for
now while I figure out what's the best solution for it.

llvm-svn: 205059

WinCOFF cannot form PC relative relocations to support absolute
MCValues.  We should reenable this once WinCOFF supports emission of
IMAGE_REL_I386_REL32 relocations.

This fixes PR19272.

llvm-svn: 205058

llvm-svn: 205057

llvm-svn: 205056

Patch by: Jeroen Ketema

llvm-svn: 205055

Patch by: Jeroen Ketema

llvm-svn: 205054

llvm-svn: 205053

Also remove unused vector.

llvm-svn: 205052

llvm-svn: 205051

This is a bit of a stab in the dark, since I have zlib on my machine.
Just going to bounce it off the bots & see if it sticks.

Do we have some convention for negative REQUIRES: checks? Or do I just
need to add a feature like I've done here?

llvm-svn: 205050

The test was failing because clang-cl changes the default triple
to target MSVC-style Win32. This is kind of wonky, but hasn't been
a problem until we started warning:

  warning: unknown platform, assuming -mfloat-abi=soft

Some of the tests in cl-options.c were running with -Werror, causing them
to fail.

Fixing this by FileCheck-ifying those tests instead of using -Werror.

llvm-svn: 205049

This reverts commit r205045.

llvm-svn: 205048

This reverts commit r205044.

llvm-svn: 205047

Not only did I invert the indices when I wrote the code, but I also did the
same thing when I wrote the regression test. Oops.

llvm-svn: 205046

Rather than rolling our own functions to read little endian data from
a buffer, we can use the support in llvm's Endian.h.

No functional change.

llvm-svn: 205045

Rather than rolling our own functions to write little endian data to
an ostream, we can use the support in llvm's EndianStream.h.

No functional change.

llvm-svn: 205044

condition where we could end up killing debugserver (and thus the target) before it had a chance
to detach.
Also fix debugserver to send the OK AFTER it detaches to avoid the same race condition.

<rdar://problem/16202713>

llvm-svn: 205043

Unfortunately this one fails deep inside the mips backend, so xfail it.

llvm-svn: 205042

v2[fi]64 values need to be explicitly passed in VSX registers. This is because
the code in TRI that finds the minimal register class given a register and a
value type will assert if given an Altivec register and a non-Altivec type.

llvm-svn: 205041

llvm-svn: 205040

llvm-svn: 205039

Response file is a command line argument in the form of @file. The GNU-
compatible driver expands the file contents, replacing @file argument.

Differential Revision: http://llvm-reviews.chandlerc.com/D3210

llvm-svn: 205038

Responding to Justin's review of r205025.

llvm-svn: 205037

It was using "lc  -filetype=obj" just to pass the result to
"llvm-objdupm -disassemble" and then filecheck assembly.

The CHECK-NOT would never match anyway since it was missing $.

llvm-svn: 205036

llvm-svn: 205035