llvm-svn: 232997

Simplify boolean expresions using true and false with clang-tidy.

Patch by Richard.

llvm-svn: 232996

llvm-svn: 232995

llvm-svn: 232994

llvm-svn: 232993

Don't use a separate table for compares anymore,
and use the same VOP2_REV class.

llvm-svn: 232992

Also don't count the class instructions as isCompare anymore.

llvm-svn: 232991

These are already set in the base class for the instruction.

llvm-svn: 232990

llvm-svn: 232989

This enables very common cases to switch to the
smaller encoding.

All of the standard LLVM canonicalizations of comparisons
are the opposite of what we want. Compares with constants
are moved to the RHS, but the first operand can be an inline
immediate, literal constant, or SGPR using the 32-bit VOPC
encoding.

There are additional bad canonicalizations that should
also be fixed, such as canonicalizing ge x, k to gt x, (k + 1)
if this makes k no longer an inline immediate value.

llvm-svn: 232988

Use VOPCX_F64 to not need the let Defs = [EXEC]

llvm-svn: 232987

It isn't used, and these will probably never be directly selected.

llvm-svn: 232986

llvm-svn: 232985

Simplify boolean expressions involving `true` and `false` with `clang-tidy`.

Actually upon inspection a bunch of these boolean variables could be
factored away entirely anyway - using find_if and then testing the
result before using it. This also helps reduce indentation in the code
anyway - and a bunch of other related simplification fell out nearby so
I just committed all of that.

Patch by Richard Thomson (legalize@xmission.com)

Differential Revision: http://reviews.llvm.org/D8517

llvm-svn: 232984

Summary:
Saw this while reading some code in DynamicLoader classes. Looks like this has
been a FIXME since 2011 at least.

Test Plan: Run unit tests.

Reviewers: clayborg

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D8558

llvm-svn: 232983

See detailed discussion at

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20140915/235418.html

and r217907, r217948:

 http://llvm.org/viewvc/llvm-project?view=revision&revision=217907
 http://llvm.org/viewvc/llvm-project?view=revision&revision=217948

llvm-svn: 232982

NFC.

llvm-svn: 232981

Also reorder includes a bit, NFC.

llvm-svn: 232980

Reviewers: clayborg

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D8554

llvm-svn: 232979

This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.

llvm-svn: 232978

This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.

llvm-svn: 232977

NFC.

llvm-svn: 232976

NFC.

llvm-svn: 232975

This was hardcoding some make rules instead of relying on
Makefile.rules, which was causing some of the logic to be
incorrect for Windows.

Patch by: Adrian McCarthy
Differential Revision: http://reviews.llvm.org/D8363

llvm-svn: 232974

llvm-svn: 232973

[sanitizer] Fix file access modes in SanitizerCommon.InternalMmapWithOffset and SanitizerCommon.FileOps

llvm-svn: 232972

On AArch64, the -fallow-half-args-and-returns option is the default.
With it, the half type is considered legal (rather than the i16 used
normally for __fp16), but no operation is, except conversions and
load/stores and such.

The previous behavior was tantamount to saying LangOpts.NativeHalfType
was implied by LangOpts.HalfArgsAndReturns, which isn't true.
Instead, teach the various parts of CodeGen that already know about
half (using the intrinsics or not) about this weird in-between case,
where the "half" type is legal, but operations on it aren't.

This is a smaller intermediate step to the end-goal of removing the
intrinsic, always using "half", and letting the backend legalize.

Builds on r232968.
rdar://20045970, rdar://17468714
Differential Revision: http://reviews.llvm.org/D8367

llvm-svn: 232971

llvm-svn: 232970

- replace hard tabs with 4-space indents
- delete EOL whitespace

llvm-svn: 232969

Fix the CodeGen so that for types bigger than float, instead of
converting to fp16 via the sequence "InTy -> float -> fp16", we
perform conversions in just one step.  This avoids the double
rounding which potentially changes results from a natural
IEEE-754 operation.

rdar://17594379, rdar://17468714
Differential Revision: http://reviews.llvm.org/D4602
Part of: http://reviews.llvm.org/D8367

llvm-svn: 232968

Patch by Geoff Berry<gberry@codeaurora.org>.

llvm-svn: 232967

Summary:
This has been discovered while experimenting with the gecko linker on android.
In general, assert()'ing on "user input" is a bad idea.

Test Plan: Run unit tests.

Reviewers: clayborg, tfiala

Subscribers: lldb-commits

Differential Revision: http://reviews.llvm.org/D8495

llvm-svn: 232966

This is a generic implementation which just calls sqrt.  Targets should
override this if they want a faster implementation.

v2:
  - Alphabetize SOURCES

llvm-svn: 232965

This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.

v2:
  - Remove unnecessary copyright.

llvm-svn: 232964

v2:
  - Move common code into a macro
  - Use the same constant for all vector types.

llvm-svn: 232963

This change is incorrect since it converts double rounding into single rounding,
which can produce different results. Instead this optimization will be done by
modifying Clang's codegen to not produce double rounding in the first place.

This reverts commit r232954.

llvm-svn: 232962

Patch by Richard (legalize@xmission.com)

Differential Revision: http://reviews.llvm.org/D8521

llvm-svn: 232961

This will help avoid naming conflicts with functions defined in
kernels linking with libclc.

llvm-svn: 232960

It is required because the name of the executable exceeded the maximum
allowed file name on android.

llvm-svn: 232959

Anton tried this 5 years ago but it was reverted due to extra VMOVs
being emitted. This can be easily fixed with a liberal application
of patterns - matching loads/stores and extractelts.

llvm-svn: 232958