There hasn't been a separate cc1 option table since r155916.

llvm-svn: 187993

llvm-svn: 187992

The /Za flag should probably do more, but let's start with this.

Differential Revision: http://llvm-reviews.chandlerc.com/D1320

llvm-svn: 187991

Patch by Richard (legalize@xmission.com)

llvm-svn: 187990

llvm-svn: 187989

llvm-svn: 187988

llvm-svn: 187987

llvm-svn: 187986

Remove the now redundant FeatureFP16 from the Cortex-A15 feature list. It was made redundant when FeatureVFP4 was added which implies FP16.

llvm-svn: 187985

llvm-svn: 187984

Implementation uses __thread not available on certain platforms.

llvm-svn: 187983

llvm-svn: 187982

Contributed-by: Star Tan <tanmx_star@yeah.net>
llvm-svn: 187981

llvm-svn: 187980

Summary:
This patch adds tooling::deduplicate() which removes duplicates from and
looks for conflicts in a vector of Replacements.

Differential Revision: http://llvm-reviews.chandlerc.com/D1314

llvm-svn: 187979

llvm-svn: 187978

llvm-svn: 187977

llvm-svn: 187976

llvm-svn: 187975

llvm-svn: 187974

The name "tCDP" isn't used anywhere else in the source code, so renaming it for consistency doesn't cause any problems.
This is the only Thumb2 instruction defined with "t" prefix; all other Thumb2 instructions have "t2" prefix (e.g. "t2CDP2" which is defined immediately afterwards).

Patch by Artyom Skrobov.

llvm-svn: 187973

llvm-svn: 187972

r187967: Disable inlining between sanitized and non-sanitized functions.
llvm-svn: 187971

contains the vtable pointer for the class as the operand at index
12. Replace the i32 0 with a null MDNode.

llvm-svn: 187970

'-fno-unroll-loops'. The option to the backend is even called
'DisableUnrollLoops'. This is precisely the form that Clang *didn't*
support. We didn't recognize the flag, we didn't pass it to the CC1
layer, and even if we did we wouldn't use it. Clang only inspected the
positive form of the flag, and only did so to enable loop unrolling when
the optimization level wasn't high enough. This only occurs for an
optimization level that even has a chance of running the loop unroller
when optimizing for size.

This commit wires up the 'no' variant, and switches the code to actually
follow the standard flag pattern of using the last flag and allowing
a flag in either direction to override the default.

I think this is still wrong. I don't know why we disable the loop
unroller entirely *from Clang* when optimizing for size, as the loop
unrolling pass *already has special logic* for the case where the
function is attributed as optimized for size! We should really be
trusting that. Maybe in a follow-up patch, I don't really want to change
behavior here.

llvm-svn: 187969

llvm-svn: 187968

Inlining between functions with different values of sanitize_* attributes
leads to over- or under-sanitizing, which is always bad.

llvm-svn: 187967

be void and therefore not have a type entry. Only add the type if
it is non-void and provide a testcase.

llvm-svn: 187966

Remove AllUndef check from one of the loops in getVectorShuffle. It was already handled by the 'AllLHS && AllRHS' check after the previous loop.

llvm-svn: 187965

llvm-svn: 187964

the type exists.

Fix up cases where we weren't checking for optional types and add
an assert to addType to make sure we catch this in the future.

Fix up a testcase that was using the tag for DW_TAG_array_type
when it meant DW_TAG_enumeration_type.

llvm-svn: 187963

llvm-svn: 187962

llvm-svn: 187961

Making use of the recently-added ISD::FROUND, which allows for custom lowering
of round(), the PPC backend will now map frin to round(). Previously, we had
been using frin to lower nearbyint() (and rint() via some custom lowering to
handle the extra fenv flags requirements), but only in fast-math mode because
frin does not tie-to-even. Several users had complained about this behavior,
and this new mapping of frin to round is certainly more appropriate (and does
not require fast-math mode).

In effect, this reverts r178362 (and part of r178337, replacing the nearbyint
mapping with the round mapping).

llvm-svn: 187960

Zhihao Yuan: Replace operator& with addressof in reference_wrapper constructor.

It breaks a clang bootstrap.

llvm-svn: 187959

so that it can be toggled independently of other deprecated warnings.

llvm-svn: 187958

It broke a public build bot.

llvm-svn: 187957

This reverts commit r77814.

We were sticking global constants in the .data section instead of in the
.rdata section when emitting for COFF.

This fixes PR16831.

llvm-svn: 187956

Fixes PR16394.

llvm-svn: 187955

llvm-svn: 187954