oota-llvm.git
11 years agoSomehow this important part of the patch, where I actually check the Mask,
Joey Gouly [Thu, 12 Sep 2013 14:23:19 +0000 (14:23 +0000)]
Somehow this important part of the patch, where I actually check the Mask,
got lost during my iterations of review.

Thanks to Hal for spotting it!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190604 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[LTO] Fix the LTO tool, after my API breakage.
Joey Gouly [Thu, 12 Sep 2013 12:55:29 +0000 (12:55 +0000)]
[LTO] Fix the LTO tool, after my API breakage.

Thanks to Zonr Chang!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190602 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoLLVM interpreter: added a test for insert- extract- value
Elena Demikhovsky [Thu, 12 Sep 2013 10:52:03 +0000 (10:52 +0000)]
LLVM interpreter: added a test for insert- extract- value

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190600 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoLLVM Interpreter: implementation of "insertvalue" and "extractvalue";
Elena Demikhovsky [Thu, 12 Sep 2013 10:48:23 +0000 (10:48 +0000)]
LLVM Interpreter: implementation of "insertvalue" and "extractvalue";
undef constatnt for structure and test for these functions.

done by Yuri Veselov (mailto:Yuri.Veselov@intel.com)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190599 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd an instruction deprecation feature to TableGen.
Joey Gouly [Thu, 12 Sep 2013 10:28:05 +0000 (10:28 +0000)]
Add an instruction deprecation feature to TableGen.

The 'Deprecated' class allows you to specify a SubtargetFeature that the
instruction is deprecated on.

The 'ComplexDeprecationPredicate' class allows you to define a custom
predicate that is called to check for deprecation.
For example:
  ComplexDeprecationPredicate<"MCR">

would mean you would have to define the following function:
  bool getMCRDeprecationInfo(MCInst &MI, MCSubtargetInfo &STI,
                             std::string &Info)

Which returns 'false' for not deprecated, and 'true' for deprecated
and store the warning message in 'Info'.

The MCTargetAsmParser constructor was chaned to take an extra argument of
the MCInstrInfo class, so out-of-tree targets will need to be changed.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190598 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAVX-512: implemented extractelement with variable index.
Elena Demikhovsky [Thu, 12 Sep 2013 08:55:00 +0000 (08:55 +0000)]
AVX-512: implemented extractelement with variable index.
Added parsing of mask register and "zeroing" semantic, like {%k1} {z}.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190595 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFixup for r190409: add dep on LZMA only if CMake is cross-compiling
Alexey Samsonov [Thu, 12 Sep 2013 08:26:53 +0000 (08:26 +0000)]
Fixup for r190409: add dep on LZMA only if CMake is cross-compiling

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190591 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoPPC: Enable aggressive anti-dependency breaking
Hal Finkel [Thu, 12 Sep 2013 05:24:49 +0000 (05:24 +0000)]
PPC: Enable aggressive anti-dependency breaking

Aggressive anti-dependency breaking is enabled by default for all PPC cores.
This provides a general speedup on the P7 and other platforms (among other
factors, the instruction group formation for the non-embedded PPC cores is done
during post-RA scheduling). In order to do this safely, the incompatibility
between uses of the MFOCRF instruction and anti-dependency breaking are
resolved by marking MFOCRF with hasExtraSrcRegAllocReq. As noted in the removed
FIXME, the problem was that MFOCRF's output is sensitive to the identify of the
source register, and always paired with a shift to undo this effect. Because
anti-dependency breaking is unaware of this hidden dependency of the shift
amount on the source register of the MFOCRF instruction, changing that register
must be inhibited.

Two test cases were adjusted: The SjLj test was made more insensitive to
register choices and scheduling; the saveCR test disabled anti-dependency
breaking because part of what it is testing is proper register reuse.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190587 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix crash in AggressiveAntiDepBreaker with empty CriticalPathSet
Hal Finkel [Thu, 12 Sep 2013 04:22:31 +0000 (04:22 +0000)]
Fix crash in AggressiveAntiDepBreaker with empty CriticalPathSet

If no register classes are added to CriticalPathRCs, then the CriticalPathSet
bitmask will be empty. In that case, ExcludeRegs must remain NULL or else this
line will cause a segfault:

  } else if ((ExcludeRegs != NULL) && ExcludeRegs->test(AntiDepReg)) {

I have no in-tree test case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190584 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoR600/SI: expose TBUFFER_STORE_FORMAT_* for OpenGL transform feedback
Tom Stellard [Thu, 12 Sep 2013 02:55:14 +0000 (02:55 +0000)]
R600/SI: expose TBUFFER_STORE_FORMAT_* for OpenGL transform feedback

For _XYZ, the type of VDATA is v4i32, because v3i32 doesn't exist.

The ADDR64 bit is not exposed. A simpler intrinsic that doesn't take
a resource descriptor might be nicer.

The maximum number of input SGPRs is bumped to 17.

Signed-off-by: Marek Olšák <marek.olsak@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190575 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoR600: Don't use trans slot for instructions that read LDS source registers
Tom Stellard [Thu, 12 Sep 2013 02:55:06 +0000 (02:55 +0000)]
R600: Don't use trans slot for instructions that read LDS source registers

This fixes some regressions in the piglit local memory store tests
introduced by recent commits which made the scheduler aware of the trans
slot.

It's not possible to test this using lit, because there is no way to
determine from the assembly dumps whether or not an instruction is in
the trans slot.

Even if this were possible, the test would be highly sensitive to
changes in the scheduler and might generate confusing false negatives.

Reviewed-by: Vincent Lejeune<vljn at ovi.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190574 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTypo fixes.
Rui Ueyama [Thu, 12 Sep 2013 01:43:21 +0000 (01:43 +0000)]
Typo fixes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190569 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoMove variable under condition where it is used
Matt Arsenault [Thu, 12 Sep 2013 01:07:58 +0000 (01:07 +0000)]
Move variable under condition where it is used

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190567 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix comment to match what the assert actually enforces
Matt Arsenault [Thu, 12 Sep 2013 01:07:54 +0000 (01:07 +0000)]
Fix comment to match what the assert actually enforces

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190566 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove pointless assertion after r190376
Matt Arsenault [Thu, 12 Sep 2013 01:07:49 +0000 (01:07 +0000)]
Remove pointless assertion after r190376

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190565 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoGreatly simplify the PPC A2 scheduling itinerary
Hal Finkel [Wed, 11 Sep 2013 23:25:21 +0000 (23:25 +0000)]
Greatly simplify the PPC A2 scheduling itinerary

As Andy pointed out to me a long time ago, there are no structural hazards in
the later pipeline stages of the A2, and so modeling them is useless. Also,
modeling the top pre-dispatch stages is deceiving because, when multiple
hardware threads are active, those resources are shared among the threads. The
bypass definitions were mostly wrong, and so those have been removed. The
resulting itinerary is much simpler, and more accurate.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190562 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoEnable MI scheduling (and CodeGen AA) by default for embedded PPC cores
Hal Finkel [Wed, 11 Sep 2013 23:05:25 +0000 (23:05 +0000)]
Enable MI scheduling (and CodeGen AA) by default for embedded PPC cores

For embedded PPC cores (especially the A2 core), using the MI scheduler with AA
is far superior to the other scheduling options.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190558 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[CMake] Update GetSVN.cmake to use LLVM version control helper scripts.
Jordan Rose [Wed, 11 Sep 2013 22:46:41 +0000 (22:46 +0000)]
[CMake] Update GetSVN.cmake to use LLVM version control helper scripts.

This allows the logic to work with Git, and also uses the variable names
to match what Clang is actually looking for.

This changes the interface of GetSVN.cmake. Clang change to follow.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190556 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix anachronism/typo in syntax of declaration of llvm.va_start.
Nick Lewycky [Wed, 11 Sep 2013 22:04:52 +0000 (22:04 +0000)]
Fix anachronism/typo in syntax of declaration of llvm.va_start.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190552 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse the appropriate return type for the compact unwind encoding.
Bill Wendling [Wed, 11 Sep 2013 21:47:57 +0000 (21:47 +0000)]
Use the appropriate return type for the compact unwind encoding.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190551 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoImplement TTI getUnrollingPreferences for PowerPC
Hal Finkel [Wed, 11 Sep 2013 21:20:40 +0000 (21:20 +0000)]
Implement TTI getUnrollingPreferences for PowerPC

The PowerPC A2 core greatly benefits from aggressive concatenation unrolling;
use the new getUnrollingPreferences to enable this by default when targeting
the PPC A2 core.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190549 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoMove into an anonymous namespace and closer to where it's used.
Bill Wendling [Wed, 11 Sep 2013 20:38:09 +0000 (20:38 +0000)]
Move into an anonymous namespace and closer to where it's used.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190547 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug info: add more comments.
Manman Ren [Wed, 11 Sep 2013 19:40:28 +0000 (19:40 +0000)]
Debug info: add more comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190544 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd getUnrollingPreferences to TTI
Hal Finkel [Wed, 11 Sep 2013 19:25:43 +0000 (19:25 +0000)]
Add getUnrollingPreferences to TTI

Allow targets to customize the default behavior of the generic loop unrolling
transformation. This will be used by the PowerPC backend when targeting the A2
core (which is in-order with a deep pipeline), and using more aggressive
defaults is important.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190542 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTry to fix the atom buildbots by adding an explicit 'cpu' to the 'llc' command.
Bill Wendling [Wed, 11 Sep 2013 19:06:04 +0000 (19:06 +0000)]
Try to fix the atom buildbots by adding an explicit 'cpu' to the 'llc' command.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190541 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug Info: move class definition of DIRef.
Manman Ren [Wed, 11 Sep 2013 18:55:55 +0000 (18:55 +0000)]
Debug Info: move class definition of DIRef.

Definition of DIRef used to require the full definition of DIType because
of usage of DIType::isType in DIRef::resolve. We now use DIDescriptor::isType
instead to remove the requirement and move definition of DIRef before DIType.

With this, we can move the definition of DIType::getContext to the header
file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190540 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert "Give internal classes hidden visibility."
Benjamin Kramer [Wed, 11 Sep 2013 18:05:11 +0000 (18:05 +0000)]
Revert "Give internal classes hidden visibility."

It works with clang, but GCC has different rules so we can't make all of those
hidden. This reverts commit r190534.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190536 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[lit] Add support for attach arbitrary metrics to test results.
Daniel Dunbar [Wed, 11 Sep 2013 17:45:11 +0000 (17:45 +0000)]
[lit] Add support for attach arbitrary metrics to test results.

 - This is a work-in-progress and all details are subject to change, but I am
   trying to build up support for allowing lit to be used as a driver for
   performance tests (or other tests which might want to record information
   beyond simple PASS/FAIL).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190535 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoGive internal classes hidden visibility.
Benjamin Kramer [Wed, 11 Sep 2013 17:42:27 +0000 (17:42 +0000)]
Give internal classes hidden visibility.

Worth 100k on a linux/x86_64 Release+Asserts clang.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190534 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agollvm-tblgen: Mangle operand replacements into the strings in printAliasInstr.
Benjamin Kramer [Wed, 11 Sep 2013 15:42:16 +0000 (15:42 +0000)]
llvm-tblgen: Mangle operand replacements into the strings in printAliasInstr.

Cuts down the bloat in the AArch64 asm writer a bit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190527 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDon't expose symbols of lle_ functions.
Benjamin Kramer [Wed, 11 Sep 2013 12:42:39 +0000 (12:42 +0000)]
Don't expose symbols of lle_ functions.

+ formatting fixes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190523 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Added test cases that were supposed to be part of r190507, r190509, r1905...
Daniel Sanders [Wed, 11 Sep 2013 12:39:25 +0000 (12:39 +0000)]
[mips][msa] Added test cases that were supposed to be part of r190507, r190509, r190512, and r190518.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190522 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Added support for matching mulv, nlzc, sll, sra, srl, and subv from norma...
Daniel Sanders [Wed, 11 Sep 2013 11:58:30 +0000 (11:58 +0000)]
[mips][msa] Added support for matching mulv, nlzc, sll, sra, srl, and subv from normal IR (i.e. not intrinsics)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190518 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agollvm-tblgen: Stabilize subreg index tables.
Benjamin Kramer [Wed, 11 Sep 2013 11:22:26 +0000 (11:22 +0000)]
llvm-tblgen: Stabilize subreg index tables.

Otherwise SequenceToOffsetTable will sort by pointer and becomes non-deterministic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190514 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Added support for matching fadd, fdiv, flog2, fmul, frint, fsqrt, and...
Daniel Sanders [Wed, 11 Sep 2013 10:51:30 +0000 (10:51 +0000)]
[mips][msa] Added support for matching fadd, fdiv, flog2, fmul, frint, fsqrt, and fsub from normal IR (i.e. not intrinsics)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190512 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoPath: Add an in-place version of path::native.
Benjamin Kramer [Wed, 11 Sep 2013 10:45:21 +0000 (10:45 +0000)]
Path: Add an in-place version of path::native.

This reflects the common use case of nativizing a prepared path. The existing
version invokes undefined behavior if input = output, add an assert to catch
that case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190510 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Added support for matching div_[su] from normal IR (i.e. not intrinsics)
Daniel Sanders [Wed, 11 Sep 2013 10:38:58 +0000 (10:38 +0000)]
[mips][msa] Added support for matching div_[su] from normal IR (i.e. not intrinsics)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190509 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Added support for matching addv from normal IR (i.e. not intrinsics)
Daniel Sanders [Wed, 11 Sep 2013 10:28:16 +0000 (10:28 +0000)]
[mips][msa] Added support for matching addv from normal IR (i.e. not intrinsics)

The corresponding intrinsic is now lowered into equivalent IR (ISD::ADD) before instruction selection.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190507 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Separate the configuration of int/float vector types since they will...
Daniel Sanders [Wed, 11 Sep 2013 10:15:48 +0000 (10:15 +0000)]
[mips][msa] Separate the configuration of int/float vector types since they will diverge soon

No functional change

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190506 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Corrected the definition of the dotp_[su].[hwd] intrinsics
Daniel Sanders [Wed, 11 Sep 2013 09:59:17 +0000 (09:59 +0000)]
[mips][msa] Corrected the definition of the dotp_[su].[hwd] intrinsics

The elements of the operands should be half the width of the elements of
the result.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190505 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoconfig-ix.cmake: Tweak stray endif(...) to suppress a warning in CMake.
NAKAMURA Takumi [Wed, 11 Sep 2013 09:42:03 +0000 (09:42 +0000)]
config-ix.cmake: Tweak stray endif(...) to suppress a warning in CMake.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190502 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSimplify the checking of function attributes by using the simple methods.
Bill Wendling [Wed, 11 Sep 2013 08:35:09 +0000 (08:35 +0000)]
Simplify the checking of function attributes by using the simple methods.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190499 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse type form of getIntPtrType
Matt Arsenault [Wed, 11 Sep 2013 07:29:40 +0000 (07:29 +0000)]
Use type form of getIntPtrType

This doesn't change anything since malloc always returns
address space 0.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190498 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix typo
Rui Ueyama [Wed, 11 Sep 2013 05:22:01 +0000 (05:22 +0000)]
Fix typo

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190492 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTeach loop-idiom about address space pointer sizes
Matt Arsenault [Wed, 11 Sep 2013 05:09:42 +0000 (05:09 +0000)]
Teach loop-idiom about address space pointer sizes

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190491 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd braces
Matt Arsenault [Wed, 11 Sep 2013 05:09:35 +0000 (05:09 +0000)]
Add braces

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190490 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRe-submit r190469: YAMLIO: Fix string quoting logic.
Rui Ueyama [Wed, 11 Sep 2013 04:00:08 +0000 (04:00 +0000)]
Re-submit r190469: YAMLIO: Fix string quoting logic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190485 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert "YAMLIO: Fix string quoting logic." (r190469)
Hans Wennborg [Wed, 11 Sep 2013 01:59:32 +0000 (01:59 +0000)]
Revert "YAMLIO: Fix string quoting logic." (r190469)

It was turning the buildbots red.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190480 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added support for getting/setting operands of values and getting...
Michael Gottesman [Wed, 11 Sep 2013 01:38:12 +0000 (01:38 +0000)]
[python-bindings] Added support for getting/setting operands of values and getting the number of operands of a value.

Also in the process did some cleanups for BasicBlock.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190477 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added support for iterating over a basic blocks instructions, getti...
Michael Gottesman [Wed, 11 Sep 2013 01:17:38 +0000 (01:17 +0000)]
[python-bindings] Added support for iterating over a basic blocks instructions, getting their name/dumping them, f/b iteration.

Tests are included.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190475 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added support for iterating over a function's basic blocks, dumping...
Michael Gottesman [Wed, 11 Sep 2013 01:01:40 +0000 (01:01 +0000)]
[python-bindings] Added support for iterating over a function's basic blocks, dumping/getting names of those bb, f/w iteration.

Tests are included.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190473 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove trailing whitespace
Rui Ueyama [Wed, 11 Sep 2013 00:53:07 +0000 (00:53 +0000)]
Remove trailing whitespace

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190472 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added support for getting a module's functions, iterating f/b over...
Michael Gottesman [Wed, 11 Sep 2013 00:52:47 +0000 (00:52 +0000)]
[python-bindings] Added support for getting a module's functions, iterating f/b over said functions, dumping/print name of functions.

Tests are included as well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190471 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoYAMLIO: Fix string quoting logic.
Rui Ueyama [Wed, 11 Sep 2013 00:45:48 +0000 (00:45 +0000)]
YAMLIO: Fix string quoting logic.

YAMLIO printed a string as is without quotes unless it contains a newline
character. That did not suffice. We also need to quote a string if it starts
with a backquote, quote, double quote or atsign, or it's the empty string.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190469 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Export OpCode from core.py.
Michael Gottesman [Wed, 11 Sep 2013 00:41:07 +0000 (00:41 +0000)]
[python-bindings] Export OpCode from core.py.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190468 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added test for reading a module from bitcode.
Michael Gottesman [Wed, 11 Sep 2013 00:41:05 +0000 (00:41 +0000)]
[python-bindings] Added test for reading a module from bitcode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190467 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRename variables for consistency.
Eli Friedman [Wed, 11 Sep 2013 00:41:02 +0000 (00:41 +0000)]
Rename variables for consistency.

No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190466 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Fixed 3 test failures caused by typos.
Michael Gottesman [Wed, 11 Sep 2013 00:41:02 +0000 (00:41 +0000)]
[python-bindings] Fixed 3 test failures caused by typos.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190465 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoGet rid of unused isPodLike definitions.
Eli Friedman [Wed, 11 Sep 2013 00:36:54 +0000 (00:36 +0000)]
Get rid of unused isPodLike definitions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190461 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoSupport ANSI escape code on Windows
Nico Rieck [Wed, 11 Sep 2013 00:36:48 +0000 (00:36 +0000)]
Support ANSI escape code on Windows

In some cases (e.g. when a build system pipes stderr) the Windows console
API cannot be used to color output. For these, provide a way to switch to
ANSI escape codes. This is required for Clang's -fansi-escape-codes option.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190460 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFixed typo.
Michael Gottesman [Wed, 11 Sep 2013 00:27:23 +0000 (00:27 +0000)]
Fixed typo.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190459 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added code for loading a module from bitcode, getset its datalayout...
Michael Gottesman [Wed, 11 Sep 2013 00:23:14 +0000 (00:23 +0000)]
[python-bindings] Added code for loading a module from bitcode, getset its datalayout, getset its target, dump it, print it to a file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190458 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Added bindings for LLVMContextRef and a test for creating a new...
Michael Gottesman [Wed, 11 Sep 2013 00:23:10 +0000 (00:23 +0000)]
[python-bindings] Added bindings for LLVMContextRef and a test for creating a new context or getting the global context.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190457 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Implemented the PassRegistry class and the calls to initialize...
Michael Gottesman [Wed, 11 Sep 2013 00:23:08 +0000 (00:23 +0000)]
[python-bindings] Implemented the PassRegistry class and the calls to initialize/shutdown llvm. Also included an initialize_llvm declaration.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190456 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Removed unused import byref from llvm/disassembler.py.
Michael Gottesman [Wed, 11 Sep 2013 00:23:05 +0000 (00:23 +0000)]
[python-bindings] Removed unused import byref from llvm/disassembler.py.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190455 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDon't assert on invalid loop vectorization hint.
Eli Friedman [Tue, 10 Sep 2013 23:45:25 +0000 (23:45 +0000)]
Don't assert on invalid loop vectorization hint.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190450 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix another mistake in r190442.
Eli Friedman [Tue, 10 Sep 2013 23:22:56 +0000 (23:22 +0000)]
Fix another mistake in r190442.

Sorry about that; I'll try to be more careful about DEBUG mode.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190449 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix unused variables.
Eli Friedman [Tue, 10 Sep 2013 23:18:14 +0000 (23:18 +0000)]
Fix unused variables.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190448 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix mistake in r190442.
Eli Friedman [Tue, 10 Sep 2013 23:09:24 +0000 (23:09 +0000)]
Fix mistake in r190442.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190446 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove unused functions.
Eli Friedman [Tue, 10 Sep 2013 22:42:31 +0000 (22:42 +0000)]
Remove unused functions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190442 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoHoist section call out of loop.
Eric Christopher [Tue, 10 Sep 2013 21:49:37 +0000 (21:49 +0000)]
Hoist section call out of loop.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190440 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTry to unbreak mingw32 buildbot.
Rui Ueyama [Tue, 10 Sep 2013 21:32:42 +0000 (21:32 +0000)]
Try to unbreak mingw32 buildbot.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190438 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix missing CHECK-LABELs
Matt Arsenault [Tue, 10 Sep 2013 19:57:05 +0000 (19:57 +0000)]
Fix missing CHECK-LABELs

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190426 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoTeach ScalarEvolution about pointer address spaces
Matt Arsenault [Tue, 10 Sep 2013 19:55:24 +0000 (19:55 +0000)]
Teach ScalarEvolution about pointer address spaces

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190425 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoFix typo
Matt Arsenault [Tue, 10 Sep 2013 19:54:54 +0000 (19:54 +0000)]
Fix typo

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190424 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd getenv() wrapper that works on multibyte environment variable.
Rui Ueyama [Tue, 10 Sep 2013 19:45:51 +0000 (19:45 +0000)]
Add getenv() wrapper that works on multibyte environment variable.

On Windows, character encoding of multibyte environment variable varies
depending on settings. The only reliable way to handle it I think is to use
GetEnvironmentVariableW().

GetEnvironmentVariableW() works on wchar_t string, which is on Windows UTF16
string. That's not ideal because we use UTF-8 as the internal encoding in LLVM.
This patch defines a wrapper function which takes and returns UTF-8 string for
GetEnvironmentVariableW().

The wrapper function does not do any conversion and just forwards the argument
to getenv() on Unix.

Differential Revision: http://llvm-reviews.chandlerc.com/D1612

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190423 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoLoopVectorize: PHI nodes are always at the beginning of a block, no need to scan...
Benjamin Kramer [Tue, 10 Sep 2013 18:46:15 +0000 (18:46 +0000)]
LoopVectorize: PHI nodes are always at the beginning of a block, no need to scan the whole block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190422 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug Info: create scope children DIEs when the scope DIE is not null.
Manman Ren [Tue, 10 Sep 2013 18:40:41 +0000 (18:40 +0000)]
Debug Info: create scope children DIEs when the scope DIE is not null.

We try to create the scope children DIEs after we create the scope DIE. But
to avoid emitting empty lexical block DIE, we first check whether a scope
DIE is going to be null, then create the scope children if it is not null.
From the number of children, we decide whether to actually create the scope DIE.

This patch also removes an early exit which checks for a special condition.
It also removes deletion of un-used children DIEs that are generated
because we used to generate children DIEs before the scope DIE.

Deletion of un-used children DIEs may cause problem because we sometimes keep
created DIEs in a member variable of a CU.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190421 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agocmake: Install llvm-tblgen again
Hans Wennborg [Tue, 10 Sep 2013 18:35:14 +0000 (18:35 +0000)]
cmake: Install llvm-tblgen again

It was removed in r189130, but it turns out this makes life hard for
folks packaging LLVM and Clang and building the latter based on the
LLVM package.

Note that this only adds back the LLVM tblgen, and it's obviously
not included when LLVM_INSTALL_TOOLCHAIN_ONLY is set.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190419 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug Info: define a DIRef template.
Manman Ren [Tue, 10 Sep 2013 18:30:07 +0000 (18:30 +0000)]
Debug Info: define a DIRef template.

Specialize the constructors for DIRef<DIScope> and DIRef<DIType> to make sure
the Value is indeed a scope ref and a type ref.

Use DIScopeRef for DIScope::getContext and DIType::getContext and use DITypeRef
for getContainingType and getClassType.

DIScope::generateRef now returns a DIScopeRef instead of a "Value *" for
readability and type safety.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190418 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoARM: Use the PICADD opcode calculated.
Jim Grosbach [Tue, 10 Sep 2013 17:21:39 +0000 (17:21 +0000)]
ARM: Use the PICADD opcode calculated.

We were figuring out whether to use tPICADD or PICADD, then just using
tPICADD unconditionally anyway. Oops.

A testcase from someone familiar enough with ELF to produce one would
be appreciated. The existing PIC testcase correctly verifies the .s
generated, but that doesn't catch this bug, which only showed up in
direct-to-object mode.

http://llvm.org/bugs/show_bug.cgi?id=17180

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190417 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove unused private member in ARMAsmPrinter.cpp.
Logan Chien [Tue, 10 Sep 2013 15:10:02 +0000 (15:10 +0000)]
Remove unused private member in ARMAsmPrinter.cpp.

This commit removes the unused "AttributeItem" from
ObjectAttributeEmitter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190412 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdding LZMA as dep for XML2 on 2.8.0 or higher
Renato Golin [Tue, 10 Sep 2013 14:24:17 +0000 (14:24 +0000)]
Adding LZMA as dep for XML2 on 2.8.0 or higher

LibXML2 config doesn't specify lzma as a dependency, which breaks
cross-compilation builds using new linkers (ld 2.21 or higher).

There is a bug on libxml2 to fix that, but since it's going to take
a while for things to go round and back, so we should have a harmless
addition of the library until then.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190409 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[asan] refactor the use-after-return API so that the size class is computed at compil...
Kostya Serebryany [Tue, 10 Sep 2013 13:16:56 +0000 (13:16 +0000)]
[asan] refactor the use-after-return API so that the size class is computed at compile time instead of at run-time. llvm part

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190407 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[SystemZ] Update README.
Richard Sandiford [Tue, 10 Sep 2013 12:22:45 +0000 (12:22 +0000)]
[SystemZ] Update README.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190404 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[SystemZ] Add TM and TMY
Richard Sandiford [Tue, 10 Sep 2013 10:20:32 +0000 (10:20 +0000)]
[SystemZ] Add TM and TMY

The main complication here is that TM and TMY (the memory forms) set
CC differently from the register forms.  When the tested bits contain
some 0s and some 1s, the register forms set CC to 1 or 2 based on the
value the uppermost bit.  The memory forms instead set CC to 1
regardless of the uppermost bit.

Until now, I've tried to make it so that a branch never tests for an
impossible CC value.  E.g. NR only sets CC to 0 or 1, so branches on the
result will only test for 0 or 1.  Originally I'd tried to do the same
thing for TM and TMY by using custom matching code in ISelDAGToDAG.
That ended up being very ugly though, and would have meant duplicating
some of the chain checks that the common isel code does.

I've therefore gone for the simpler alternative of adding an extra
operand to the TM DAG opcode to say whether a memory form would be OK.
This means that the inverse of a "TM;JE" is "TM;JNE" rather than the
more precise "TM;JNLE", just like the inverse of "TMLL;JE" is "TMLL;JNE".
I suppose that's arguably less confusing though...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190400 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[mips][msa] Removed unsupported dot product instructions (dotp_[su].b)
Daniel Sanders [Tue, 10 Sep 2013 09:51:43 +0000 (09:51 +0000)]
[mips][msa] Removed unsupported dot product instructions (dotp_[su].b)

The dotp_[su].b instructions never existed in any revision of the MSA spec.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190398 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoAdd test cases for Mips mthc1/mfhc1 instructions. Add check for odd value of register...
Vladimir Medic [Tue, 10 Sep 2013 09:50:01 +0000 (09:50 +0000)]
Add test cases for Mips mthc1/mfhc1 instructions. Add check for odd value of register when PFU is 32 bit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190397 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRemove obsolete code from MipsAsmParser.cpp.
Vladimir Medic [Tue, 10 Sep 2013 09:39:55 +0000 (09:39 +0000)]
Remove obsolete code from MipsAsmParser.cpp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190396 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python-bindings] Changed test_memory_buffer_create_from_file to just use the generic...
Michael Gottesman [Tue, 10 Sep 2013 06:58:00 +0000 (06:58 +0000)]
[python-bindings] Changed test_memory_buffer_create_from_file to just use the generic provided test_file instead of a binary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190389 91177308-0d34-0410-b5e6-96231b3b80d8

11 years ago[python bindings] Added code to get the length of a memory buffer. Tests are included.
Michael Gottesman [Tue, 10 Sep 2013 06:57:57 +0000 (06:57 +0000)]
[python bindings] Added code to get the length of a memory buffer. Tests are included.

This is a part of a series of patches that have been sitting fallow on a
personal branch that I have been messing with for a bit.

The patches start to flesh out the python llvm-c wrapper to the point where you can:

1. Load Modules from Bitcode/Dump/Print them.
2. Iterate over Functions from those modules/get their names/dump them.
3. Iterate over the BasicBlocks from said function/get the BB's name/dump it.
4. Iterate over the Instructions in said BasicBlocks/get the instructions
   name/dump the instruction.

My main interest in developing this was to be able to gather statistics about
LLVM IR using python scripts to speed up statistical profiling of different IR
level transformations (hence the focus on printing/dumping/getting names).

This is a gift from me to the LLVM community = ).

I am going to be committing the patches slowly over the next bit as I have time
to prepare the patches.

The overall organization follows the c-api like the bindings that are already
implemented.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190388 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoMCObjectFileInfo.cpp: Fix a couple of possible typo(s), s/DwarfPub/DwarfGnuPub/,...
NAKAMURA Takumi [Tue, 10 Sep 2013 06:01:56 +0000 (06:01 +0000)]
MCObjectFileInfo.cpp: Fix a couple of possible typo(s), s/DwarfPub/DwarfGnuPub/, in r190344.

XFAIL can be removed. (in r190374)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190386 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDon't use getSetCCResultType for creating a vselect
Matt Arsenault [Tue, 10 Sep 2013 00:41:56 +0000 (00:41 +0000)]
Don't use getSetCCResultType for creating a vselect

The vselect mask isn't a setcc.

This breaks in the case when the result of getSetCCResultType
is larger than the vector operands

e.g. %tmp = select i1 %cmp <2 x i8> %a, <2 x i8> %b
when getSetCCResultType returns <2 x i32>, the assertion
that the (MaskTy.getSizeInBits() == Op1.getValueType().getSizeInBits())
is hit.

No test since I don't think I can hit this with any of the current
targets. The R600/SI implementation would break, since it returns a
vector of i1 for this, but it doesn't reach ExpandSELECT for other
reasons.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190376 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse StringRef::npos for StringRef instead of std::string one
Matt Arsenault [Tue, 10 Sep 2013 00:41:53 +0000 (00:41 +0000)]
Use StringRef::npos for StringRef instead of std::string one

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190375 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agollvm/test/DebugInfo/dwarf-public-names.ll: Mark it as XFAIL for PECOFF targets, for...
NAKAMURA Takumi [Tue, 10 Sep 2013 00:28:23 +0000 (00:28 +0000)]
llvm/test/DebugInfo/dwarf-public-names.ll: Mark it as XFAIL for PECOFF targets, for now. Investigating.

FIXME: llvm-dwarfdump doesn't handle PECOFF.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190374 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoRevert r190366. It was breaking build bots.
Bill Wendling [Tue, 10 Sep 2013 00:20:27 +0000 (00:20 +0000)]
Revert r190366. It was breaking build bots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190373 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoEnable -misched-cyclicpath by default.
Andrew Trick [Mon, 9 Sep 2013 23:31:14 +0000 (23:31 +0000)]
Enable -misched-cyclicpath by default.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190367 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoUse a default value for the prologue's debug location.
Bill Wendling [Mon, 9 Sep 2013 23:28:15 +0000 (23:28 +0000)]
Use a default value for the prologue's debug location.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190366 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug Info: In DIBuilder, the context fields of a TAG_inheritance and a
Manman Ren [Mon, 9 Sep 2013 23:07:58 +0000 (23:07 +0000)]
Debug Info: In DIBuilder, the context fields of a TAG_inheritance and a
TAG_friend are updated to use scope reference.

Added testing cases to verify that class with inheritance can be uniqued.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190364 91177308-0d34-0410-b5e6-96231b3b80d8

11 years agoDebug Info: move DIScope::getContext back from DwarfDebug.
Manman Ren [Mon, 9 Sep 2013 22:35:23 +0000 (22:35 +0000)]
Debug Info: move DIScope::getContext back from DwarfDebug.

This partially reverts r190330. DIScope::getContext now returns DIScopeRef
instead of DIScope. We construct a DIScopeRef from DIScope when we are
dealing with subprogram, lexical block or name space.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@190362 91177308-0d34-0410-b5e6-96231b3b80d8