oota-llvm.git
9 years agoRegisterPressure: Simplify close{Top|Bottom}()
Matthias Braun [Mon, 14 Sep 2015 18:24:15 +0000 (18:24 +0000)]
RegisterPressure: Simplify close{Top|Bottom}()

- There are no duplicate registers in LiveRegs list we are copying from
  and so we do not need to sort the registers.
- Simply use SmallVector::apend instead of a loop between begin() and end()
  with push_back().

Differential Revision: http://reviews.llvm.org/D12813

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247588 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing arguments...
Chen Li [Mon, 14 Sep 2015 18:10:43 +0000 (18:10 +0000)]
[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing arguments at callsite

Summary: This patch replaces isKnownNonNull() with isKnownNonNullAt() when checking nullness of passing arguments at callsite. In this way it can handle cases where the argument does not have nonnull attribute but has a dominating null check from the CFG. It also adds assertions in isKnownNonNull() and isKnownNonNullFromDominatingCondition() to make sure the value checked is pointer type (as defined in LLVM document). These assertions might trip failures in things which are not  covered under llvm/test, but fixes should be pretty obvious.

Reviewers: reames

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D12779

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247587 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert "[opaque pointer type] Pass GlobalAlias the actual pointer type rather than...
David Blaikie [Mon, 14 Sep 2015 18:01:59 +0000 (18:01 +0000)]
Revert "[opaque pointer type] Pass GlobalAlias the actual pointer type rather than decomposing it into pointee type + address space"

This was a flawed change - it just caused the getElementType call to be
deferred until later, when we really need to remove it. Now that the IR
for GlobalAliases has been updated, the root cause is addressed that way
instead and this change is no longer needed (and in fact gets in the way
- because we want to pass the pointee type directly down further).

Follow up patches to push this through GlobalValue, bitcode format, etc,
will come along soon.

This reverts commit 236160.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247585 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[llvm-mc] Better error handling in ENOENT case + test.
Davide Italiano [Mon, 14 Sep 2015 17:10:01 +0000 (17:10 +0000)]
[llvm-mc] Better error handling in ENOENT case + test.

This is a follow up to r247518.
As a general note, I think we could do a much better job testing for
error conditions in tools. I already anticipated in a previous mail,
but while implementing this I noticed that the code coverage we have
for error checking is pretty low. I can arbitrarily remove checks from
several tools and the suite still passes.

Differential Revision:  http://reviews.llvm.org/D12846

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247582 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agodsymutil: Factor out the DIE cloning into a DIECloner class. (NFC)
Adrian Prantl [Mon, 14 Sep 2015 16:46:10 +0000 (16:46 +0000)]
dsymutil: Factor out the DIE cloning into a DIECloner class. (NFC)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247577 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoImprove ISel using across lane min/max reduction
Jun Bum Lim [Mon, 14 Sep 2015 16:19:52 +0000 (16:19 +0000)]
Improve ISel using across lane min/max reduction

In vectorized integer min/max reduction code, the final "reduce" step
is sub-optimal. In AArch64, this change wll combine :
  %svn0 = vector_shuffle %0, undef<2,3,u,u>
  %smax0 = smax %0, svn0
  %svn3 = vector_shuffle %smax0, undef<1,u,u,u>
  %sc = setcc %smax0, %svn3, gt
  %n0 = extract_vector_elt %sc, #0
  %n1 = extract_vector_elt %smax0, #0
  %n2 = extract_vector_elt $smax0, #1
  %result = select %n0, %n1, n2
becomes :
  %1 = smaxv %0
  %result = extract_vector_elt %1, 0

This change extends r246790.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247575 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Unified the MipsMemSimm9GPRAsmOperand and MipsMemSimm9AsmOperand operands...
Daniel Sanders [Mon, 14 Sep 2015 15:57:24 +0000 (15:57 +0000)]
[mips] Unified the MipsMemSimm9GPRAsmOperand and MipsMemSimm9AsmOperand operands, NFC.

Summary:
These operands had the same purpose, however the MipsMemSimm9GPRAsmOperand
operand was only for micromips32r6 and the MipsMemSimm9AsmOperand did not
have a ParserMatchClass.

Patch by Scott Egerton

Reviewers: vkalintiris, dsanders

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D12730

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247573 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MergeFuncs] Fix bug in merging GetElementPointers
JF Bastien [Mon, 14 Sep 2015 15:37:48 +0000 (15:37 +0000)]
[MergeFuncs] Fix bug in merging GetElementPointers

GetElementPointers must have the first argument's type compared
for structural equivalence. Previously the code erroneously compared the
pointer's type, but this code was dead because all pointer types (of the
same address space) are the same. The pointee must be compared instead
(using the type stored in the GEP, not from the pointer type which will
be erased anyway).

Author: jrkoenig
Reviewers: dschuff, nlewycky, jfb
Subscribers: nlewycky, llvm-commits
Differential revision: http://reviews.llvm.org/D12820

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247570 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[ARM] Extract shifts out of multiply-by-constant
John Brawn [Mon, 14 Sep 2015 15:19:41 +0000 (15:19 +0000)]
[ARM] Extract shifts out of multiply-by-constant

Turning (op x (mul y k)) into (op x (lsl (mul y k>>n) n)) is beneficial when
we can do the lsl as a shifted operand and the resulting multiply constant is
simpler to generate.

Do this by doing the transformation when trying to select a shifted operand,
as that ensures that it actually turns out better (the alternative would be to
do it in PreprocessISelDAG, but we don't know for sure there if extracting the
shift would allow a shifted operand to be used).

Differential Revision: http://reviews.llvm.org/D12196

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247569 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReformat partially.
NAKAMURA Takumi [Mon, 14 Sep 2015 12:51:53 +0000 (12:51 +0000)]
Reformat partially.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247557 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReformat blank lines.
NAKAMURA Takumi [Mon, 14 Sep 2015 12:51:47 +0000 (12:51 +0000)]
Reformat blank lines.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247556 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert part of r247553, "[CMake] Reformat CLANG_TEST_DEPS." It was accidental commit.
NAKAMURA Takumi [Mon, 14 Sep 2015 12:51:01 +0000 (12:51 +0000)]
Revert part of r247553, "[CMake] Reformat CLANG_TEST_DEPS." It was accidental commit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247555 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CMake] Reformat CLANG_TEST_DEPS.
NAKAMURA Takumi [Mon, 14 Sep 2015 12:41:53 +0000 (12:41 +0000)]
[CMake] Reformat CLANG_TEST_DEPS.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247553 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Remove redundant nested-name-specifier. NFC
Simon Atanasyan [Mon, 14 Sep 2015 11:18:22 +0000 (11:18 +0000)]
[mips] Remove redundant nested-name-specifier. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247547 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Save a copy of MipsABIInfo in the MipsTargetStreamer to escape a dangling...
Simon Atanasyan [Mon, 14 Sep 2015 11:18:03 +0000 (11:18 +0000)]
[mips] Save a copy of MipsABIInfo in the MipsTargetStreamer to escape a dangling pointer

The MipsTargetELFStreamer can receive ABI info from many sources. For example,
from the MipsAsmParser instance. Lifetime of the MipsAsmParser can be shorter
than MipsTargetELFStreamer's lifetime. In that case we get a dangling pointer
to MipsABIInfo.

Differential Revision: http://reviews.llvm.org/D12805

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247546 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoWhitespace.
NAKAMURA Takumi [Mon, 14 Sep 2015 11:14:39 +0000 (11:14 +0000)]
Whitespace.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247543 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoGlobalsAAResult: Try to fix crash.
NAKAMURA Takumi [Mon, 14 Sep 2015 06:16:44 +0000 (06:16 +0000)]
GlobalsAAResult: Try to fix crash.

DeletionCallbackHandle holds GAR in its creation. It assumes;

  - It is registered as CallbackVH. It should not be moved in its life.
  - Its parent, GAR, may be moved.

To move list<DeletionCallbackHandle> GlobalsAAResult::Handles,
GAR must be updated with the destination in GlobalsAAResult(&&).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247534 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Use range-based for loops and make a helper function static. NFC
Craig Topper [Sun, 13 Sep 2015 18:01:25 +0000 (18:01 +0000)]
[TableGen] Use range-based for loops and make a helper function static. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247529 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Simplify some code by using StringRef::find instead of std::find. NFC
Craig Topper [Sun, 13 Sep 2015 18:01:20 +0000 (18:01 +0000)]
[TableGen] Simplify some code by using StringRef::find instead of std::find. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247528 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[TableGen] Use 'size_t' instead of 'unsigned' to better match the argument types...
Craig Topper [Sun, 13 Sep 2015 18:01:09 +0000 (18:01 +0000)]
[TableGen] Use 'size_t' instead of 'unsigned' to better match the argument types of addAsmOperand. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247527 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86][MMX] Added shuffle decodes for MMX/3DNow! shuffles.
Simon Pilgrim [Sun, 13 Sep 2015 11:28:45 +0000 (11:28 +0000)]
[X86][MMX] Added shuffle decodes for MMX/3DNow! shuffles.

Added shuffle decodes for MMX PUNPCK + PSHUFW shuffles.
Added shuffle decodes for 3DNow! PSWAPD shuffles.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247526 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[FunctionAttrs] Move the malloc-like test to a static helper function
Chandler Carruth [Sun, 13 Sep 2015 08:23:27 +0000 (08:23 +0000)]
[FunctionAttrs] Move the malloc-like test to a static helper function
that could be used from a new pass manager. This one makes particular
sense as a static helper as it doesn't even need TLI.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247525 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[FunctionAttrs] Factor the logic to test for a known non-null return out
Chandler Carruth [Sun, 13 Sep 2015 08:17:14 +0000 (08:17 +0000)]
[FunctionAttrs] Factor the logic to test for a known non-null return out
of a method and into a re-usable static helper. We can potentially use
this function from the implementation of a new pass manager oriented
version of the pass. Also add some better documentation of exactly what
the semantic model of this routine is (it isn't trivial) and use a more
modern naming convention for it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247524 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAVX-512: Fixed a bug in OR/XOR operations for 512-bit FP values on KNL.
Elena Demikhovsky [Sun, 13 Sep 2015 08:15:15 +0000 (08:15 +0000)]
AVX-512: Fixed a bug in OR/XOR operations for 512-bit FP values on KNL.
KNL does not have VXORPS, VORPS for 512-bit values.
I use integer VPXOR, VPOR that actually do the same.

X86ISD::FXOR/FOR are generated as a result of FSUB combining.

Differential Revision: http://reviews.llvm.org/D12753

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247523 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[FunctionAttrs] Make the per-function attribute inference a boring
Chandler Carruth [Sun, 13 Sep 2015 08:03:23 +0000 (08:03 +0000)]
[FunctionAttrs] Make the per-function attribute inference a boring
static function rather than a method. It just needed access to
TargetLibraryInfo, and this way it can be easily reused between the
current FunctionAttrs implementation and any port for the new pass
manager.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247522 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[FunctionAttrs] Collect utility functions as static helpers rather than
Chandler Carruth [Sun, 13 Sep 2015 07:50:43 +0000 (07:50 +0000)]
[FunctionAttrs] Collect utility functions as static helpers rather than
methods. They don't need anything from the class anyways.

Also, collect the declarations into the private section of the pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247521 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoClean up doxygen comments in FunctionAttrs, promoting some non-doxygen
Chandler Carruth [Sun, 13 Sep 2015 06:57:25 +0000 (06:57 +0000)]
Clean up doxygen comments in FunctionAttrs, promoting some non-doxygen
comments, deleting duplicate comments, moving comments to consistently
live on the definition since these are all really internal routines,
etc. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247520 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoDo some spring cleaning on FunctionAttrs.cpp with clang-format prior to
Chandler Carruth [Sun, 13 Sep 2015 06:47:20 +0000 (06:47 +0000)]
Do some spring cleaning on FunctionAttrs.cpp with clang-format prior to
other refactorings and cleanups here.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247519 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[llvm-mc] More meaningful error if input file doesn't exist.
Davide Italiano [Sun, 13 Sep 2015 04:09:40 +0000 (04:09 +0000)]
[llvm-mc] More meaningful error if input file doesn't exist.

Before we just printed on stderr the program name. Now at least we
print the name of the file that doesn't exist. There's probably room
for improvement of error handling in llvm-mc (and the tools in general),
but still this is a step forward.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247518 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[x86] enable machine combiner reassociations for 128-bit vector logical integer insts...
Sanjay Patel [Sat, 12 Sep 2015 19:47:50 +0000 (19:47 +0000)]
[x86] enable machine combiner reassociations for 128-bit vector logical integer insts (2nd try)

The changes in:
test/CodeGen/X86/machine-cp.ll
are just due to scheduling differences after some logic instructions were reassociated.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247516 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CodeGen] Fix AtomicExpand invalidation issue caused by r247429.
Ahmed Bougacha [Sat, 12 Sep 2015 18:51:23 +0000 (18:51 +0000)]
[CodeGen] Fix AtomicExpand invalidation issue caused by r247429.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247514 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Renamed lowerVectorShuffleAsUnpack NFCI.
Simon Pilgrim [Sat, 12 Sep 2015 18:26:47 +0000 (18:26 +0000)]
[X86] Renamed lowerVectorShuffleAsUnpack NFCI.

Renamed to lowerVectorShuffleAsPermuteAndUnpack to make it clear that it lowers to more than just a UNPCK instruction.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247513 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoDon't use bashism/kshism of test ==. From Kamil Rytarowski.
Joerg Sonnenberger [Sat, 12 Sep 2015 16:30:32 +0000 (16:30 +0000)]
Don't use bashism/kshism of test ==. From Kamil Rytarowski.

Differential Revision: http://reviews.llvm.org/D12642

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247512 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Moved lowerVectorShuffleWithUNPCK earlier to make reuse easier. NFCI.
Simon Pilgrim [Sat, 12 Sep 2015 16:03:06 +0000 (16:03 +0000)]
[X86] Moved lowerVectorShuffleWithUNPCK earlier to make reuse easier. NFCI.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247511 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Added i1 vector sextload tests
Simon Pilgrim [Sat, 12 Sep 2015 15:36:41 +0000 (15:36 +0000)]
[X86] Added i1 vector sextload tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247509 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86][FMA] Refreshed fma tests
Simon Pilgrim [Sat, 12 Sep 2015 15:33:05 +0000 (15:33 +0000)]
[X86][FMA] Refreshed fma tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247508 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agorevert r247506; need to verify changes in existing tests
Sanjay Patel [Sat, 12 Sep 2015 15:27:31 +0000 (15:27 +0000)]
revert r247506; need to verify changes in existing tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247507 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[x86] enable machine combiner reassociations for 128-bit vector logical integer insts
Sanjay Patel [Sat, 12 Sep 2015 14:58:04 +0000 (14:58 +0000)]
[x86] enable machine combiner reassociations for 128-bit vector logical integer insts

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247506 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFixed unused variable warning.
Simon Pilgrim [Sat, 12 Sep 2015 14:00:17 +0000 (14:00 +0000)]
Fixed unused variable warning.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247505 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombine] CVTPH2PS Vector Demanded Elements + Constant Folding
Simon Pilgrim [Sat, 12 Sep 2015 13:39:53 +0000 (13:39 +0000)]
[InstCombine] CVTPH2PS Vector Demanded Elements + Constant Folding

Improved InstCombine support for CVTPH2PS (F16C half 2 float conversion):

<4 x float> @llvm.x86.vcvtph2ps.128(<8 x i16>) - only uses the bottom 4 i16 elements for the conversion.

Added constant folding support.

Differential Revision: http://reviews.llvm.org/D12731

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247504 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86][SSE] Use general sext IR for (v)pmovsx stack folding tests
Simon Pilgrim [Sat, 12 Sep 2015 11:45:24 +0000 (11:45 +0000)]
[X86][SSE] Use general sext IR for (v)pmovsx stack folding tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247502 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Port SROA to the new pass manager.
Chandler Carruth [Sat, 12 Sep 2015 09:09:14 +0000 (09:09 +0000)]
[PM] Port SROA to the new pass manager.

In some ways this is a very boring port to the new pass manager as there
are no interesting analyses or dependencies or other oddities.

However, this does introduce the first good example of a transformation
pass with non-trivial state porting to the new pass manager. I've tried
to carve out patterns here to replicate elsewhere, and would appreciate
comments on whether folks like these patterns:

- A common need in the new pass manager is to effectively lift the pass
  class and some of its state into a public header file. Prior to this,
  LLVM used anonymous namespaces to provide "module private" types and
  utilities, but that doesn't scale to cases where a public header file
  is needed and the new pass manager will exacerbate that. The pattern
  I've adopted here is to use the namespace-cased-name of the core pass
  (what would be a module if we had them) as a module-private namespace.
  Then utility and other code can be declared and defined in this
  namespace. At some point in the future, we could even have
  (conditionally compiled) code that used modules features when
  available to do the same basic thing.

- I've split the actual pass run method in two in order to expose
  a private method usable by the old pass manager to wrap the new class
  with a minimum of duplicated code. I actually looked at a bunch of
  ways to automate or generate these, but they are all quite terrible
  IMO. The fundamental need is to extract the set of analyses which need
  to cross this interface boundary, and that will end up being too
  unpredictable to effectively encapsulate IMO. This is also
  a relatively small amount of boiler plate that will live a relatively
  short time, so I'm not too worried about the fact that it is boiler
  plate.

The rest of the patch is totally boring but results in a massive diff
(sorry). It just moves code around and removes or adds qualifiers to
reflect the new name and nesting structure.

Differential Revision: http://reviews.llvm.org/D12773

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247501 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoClean up: Refactoring the hardcoded value of 6 for FindAvailableLoadedValue()'s param...
Larisse Voufo [Sat, 12 Sep 2015 01:41:55 +0000 (01:41 +0000)]
Clean up: Refactoring the hardcoded value of 6 for FindAvailableLoadedValue()'s parameter MaxInstsToScan.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247497 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix typos.
Bruce Mitchener [Sat, 12 Sep 2015 01:17:08 +0000 (01:17 +0000)]
Fix typos.

Summary: This fixes a variety of typos in docs, code and headers.

Subscribers: jholewinski, sanjoy, arsenm, llvm-commits

Differential Revision: http://reviews.llvm.org/D12626

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247495 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMove asserts from PHINode::addIncoming to PHINode::setIncoming*
Hal Finkel [Sat, 12 Sep 2015 00:08:21 +0000 (00:08 +0000)]
Move asserts from PHINode::addIncoming to PHINode::setIncoming*

We had asserts in PHINode::addIncoming to check that the value types matched
the type of the PHI, and that the associated BB was not null. These did not
catch, however, later uses of setIncomingValue and setIncomingBlock (which are
called by addIncoming as well). Moving the asserts to PHINode::setIncoming*
provides better coverage. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247492 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agodsymutil: Factor out the relocation handling into a RelocationManager (NFC)
Adrian Prantl [Fri, 11 Sep 2015 23:45:30 +0000 (23:45 +0000)]
dsymutil: Factor out the relocation handling into a RelocationManager (NFC)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247490 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MC] Fix style bugs introduced in r247471. Reported by Rafael Espindola.
Davide Italiano [Fri, 11 Sep 2015 22:04:21 +0000 (22:04 +0000)]
[MC] Fix style bugs introduced in r247471. Reported by Rafael Espindola.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247483 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[MC] Don't crash on division by zero.
Davide Italiano [Fri, 11 Sep 2015 20:47:35 +0000 (20:47 +0000)]
[MC] Don't crash on division by zero.

Differential Revision: http://reviews.llvm.org/D12776

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247471 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd a non-exiting diagnostic handler for LTO.
Yunzhong Gao [Fri, 11 Sep 2015 20:01:53 +0000 (20:01 +0000)]
Add a non-exiting diagnostic handler for LTO.
This is in order to give LTO clients a chance to do some clean-up before
terminating the process.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247461 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agotypo; NFC
Sanjay Patel [Fri, 11 Sep 2015 19:29:18 +0000 (19:29 +0000)]
typo; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247454 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoUse function attribute "stackrealign" to decide whether stack
Akira Hatanaka [Fri, 11 Sep 2015 18:54:38 +0000 (18:54 +0000)]
Use function attribute "stackrealign" to decide whether stack
realignment should be forced.

With this commit, we can now force stack realignment when doing LTO and
do so on a per-function basis. Also, add a new cl::opt option
"stackrealign" to CommandFlags.h which is used to force stack
realignment via llc's command line.

Out-of-tree projects currently using -force-align-stack to force stack
realignment should make changes to attach the attribute to the functions
in the IR.

Differential Revision: http://reviews.llvm.org/D11814

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247450 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CMake] [Darwin] Need to set lto_library on CMAKE_MODULE_LINKER_FLAGS as well
Chris Bieneman [Fri, 11 Sep 2015 18:39:19 +0000 (18:39 +0000)]
[CMake] [Darwin] Need to set lto_library on CMAKE_MODULE_LINKER_FLAGS as well

This is a follow-on to r247308.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247443 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Make sure startproc/endproc are paired
David Majnemer [Fri, 11 Sep 2015 17:34:34 +0000 (17:34 +0000)]
[X86] Make sure startproc/endproc are paired

We used different conditions to determine if we should emit startproc vs
endproc.  Use the same condition to ensure that they will always be
paired.

This fixes PR24374.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247435 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[IR] Print the label operands of a catchpad like an invoke
Reid Kleckner [Fri, 11 Sep 2015 17:27:52 +0000 (17:27 +0000)]
[IR] Print the label operands of a catchpad like an invoke

The rest of the EH pads are fine, since they have at most one label and
take fewer operands for the personality.

Old catchpad vs. new:
  %5 = catchpad [i8* bitcast (i32 ()* @"\01?filt$0@0@main@@" to i8*)] to label %__except.ret.10 unwind label %catchendblock.9
-----
  %5 = catchpad [i8* bitcast (i32 ()* @"\01?filt$0@0@main@@" to i8*)]
          to label %__except.ret.10 unwind label %catchendblock.9

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247433 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CodeGen] Refactor TLI/AtomicExpand interface to make LLSC explicit.
Ahmed Bougacha [Fri, 11 Sep 2015 17:08:28 +0000 (17:08 +0000)]
[CodeGen] Refactor TLI/AtomicExpand interface to make LLSC explicit.

We used to have this magic "hasLoadLinkedStoreConditional()" callback,
which really meant two things:
- expand cmpxchg (to ll/sc).
- expand atomic loads using ll/sc (rather than cmpxchg).

Remove it, and, instead, introduce explicit callbacks:
- bool shouldExpandAtomicCmpXchgInIR(inst)
- AtomicExpansionKind shouldExpandAtomicLoadInIR(inst)

Differential Revision: http://reviews.llvm.org/D12557

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247429 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CodeGen] Rename AtomicRMWExpansionKind to AtomicExpansionKind.
Ahmed Bougacha [Fri, 11 Sep 2015 17:08:17 +0000 (17:08 +0000)]
[CodeGen] Rename AtomicRMWExpansionKind to AtomicExpansionKind.

This lets us generalize its usage to the other atomic instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247428 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PR24785] Appease MSC18 to tweak optimizations.
NAKAMURA Takumi [Fri, 11 Sep 2015 17:08:02 +0000 (17:08 +0000)]
[PR24785] Appease MSC18 to tweak optimizations.

This brings a warning.

  cl : Command line warning D9035: option 'Og-' has been deprecated and will be removed in a future release

We should resolve PR11951 to remove this tweak.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247427 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[libFuzzer] mention more trophies
Kostya Serebryany [Fri, 11 Sep 2015 16:34:14 +0000 (16:34 +0000)]
[libFuzzer] mention more trophies

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247425 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing disassembler tests for MIPS64-MIPS64R5.
Daniel Sanders [Fri, 11 Sep 2015 16:24:11 +0000 (16:24 +0000)]
[mips] Add missing disassembler tests for MIPS64-MIPS64R5.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247422 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing MIPS32 - MIPS32R5 disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 15:28:19 +0000 (15:28 +0000)]
[mips] Add missing MIPS32 - MIPS32R5 disassembler tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247420 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Attempt to fix llvm-s390x-linux1
Daniel Sanders [Fri, 11 Sep 2015 14:57:54 +0000 (14:57 +0000)]
[mips] Attempt to fix llvm-s390x-linux1

It doesn't seem to like the '|&' in the test command.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247418 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing MIPS-IV disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 14:54:58 +0000 (14:54 +0000)]
[mips] Add missing MIPS-IV disassembler tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247417 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing MIPS-III disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 14:48:46 +0000 (14:48 +0000)]
[mips] Add missing MIPS-III disassembler tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247416 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoTweak 2 x86 gold tests so they can run on non-x86 platforms
Arnaud A. de Grandmaison [Fri, 11 Sep 2015 14:45:34 +0000 (14:45 +0000)]
Tweak 2 x86 gold tests so they can run on non-x86 platforms

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247415 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing MIPS-II disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 14:34:41 +0000 (14:34 +0000)]
[mips] Add missing MIPS-II disassembler tests.

These tests were found by llvm-mc-fuzzer (see http://reviews.llvm.org/D12723)
and were verified by checking the disassembler output is accepted by GAS.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247414 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd #include llvm-config.h to Locale.cpp which depends on LLVM_ON_WIN32.
Yaron Keren [Fri, 11 Sep 2015 13:22:47 +0000 (13:22 +0000)]
Add #include llvm-config.h to Locale.cpp which depends on LLVM_ON_WIN32.
Source code was assuming that llvm-config.h would be included somehow but
up to r247253 that added #include "llvm/Support/Compiler.h" to StringRef.h
the config file was not actually included. The inclusion of llvm-config.h
caused a change of behaviour in tools/clang/test/Frontend/source-col-map.c:
previously it would output the original UTF-8 but now it outputs <U+03B1>.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247409 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRe-commit r247405: [mips] Add missing MIPS-I disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 12:59:03 +0000 (12:59 +0000)]
Re-commit r247405: [mips] Add missing MIPS-I disassembler tests.

These tests were found by llvm-mc-fuzzer (see http://reviews.llvm.org/D12723)
and verified by checking the disassembler output is accepted by GAS.

The problematic tests from the previous commit have been moved to
valid-xfail.txt for now.

Also, give invalid instructions some coverage. invalid-xfail.txt contains
instructions that should be invalid but successfully disassemble.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247407 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert r247405: [mips] Add missing MIPS-I disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 12:42:38 +0000 (12:42 +0000)]
Revert r247405: [mips] Add missing MIPS-I disassembler tests.

A small number of the added tests have operands that change on each round trip.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247406 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[mips] Add missing MIPS-I disassembler tests.
Daniel Sanders [Fri, 11 Sep 2015 12:24:06 +0000 (12:24 +0000)]
[mips] Add missing MIPS-I disassembler tests.

These tests were found by llvm-mc-fuzzer (see http://reviews.llvm.org/D12723)
and verified by checking the disassembler output is accepted by GAS.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247405 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoPPCFrameLowering::emitEpilogue(): Avoid manipulating MBBI on iterator end.
NAKAMURA Takumi [Fri, 11 Sep 2015 08:20:56 +0000 (08:20 +0000)]
PPCFrameLowering::emitEpilogue(): Avoid manipulating MBBI on iterator end.

It caused crash in MachineInstr::hasPropertyInBundle() since r247237.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247395 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix llvm/test/tools/gold/X86/bad-alias.ll.
NAKAMURA Takumi [Fri, 11 Sep 2015 08:03:17 +0000 (08:03 +0000)]
Fix llvm/test/tools/gold/X86/bad-alias.ll.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247391 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Discard useless location attributes.
Frederic Riss [Fri, 11 Sep 2015 04:17:30 +0000 (04:17 +0000)]
[dsymutil] Discard useless location attributes.

When cloning the debug info for a function that hasn't been linked,
strip the DIEs from all location attributes that wouldn't contain any
meaningful information anyway.

This kind of situation can happen when a function got discarded by the
linker, but its debug information is still wanted in the final link
because it was marked as required as some other DIE dependency. The easiest
way to get into that situation is to have using directives. They get
linked unconditionally, but their targets might not always be present.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247386 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[dsymutil] Rename some variables NFC.
Frederic Riss [Fri, 11 Sep 2015 04:17:25 +0000 (04:17 +0000)]
[dsymutil] Rename some variables NFC.

lldb doesn't like having variables named as an existing type. In order to
ease debugging, rename those variables to avoid that conflict.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247385 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix the gold test cases after alias changes
David Blaikie [Fri, 11 Sep 2015 03:28:37 +0000 (03:28 +0000)]
Fix the gold test cases after alias changes

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247381 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[opaque pointer type] Add textual IR support for explicit type parameter for global...
David Blaikie [Fri, 11 Sep 2015 03:22:04 +0000 (03:22 +0000)]
[opaque pointer type] Add textual IR support for explicit type parameter for global aliases

update.py:
import fileinput
import sys
import re

alias_match_prefix = r"(.*(?:=|:|^)\s*(?:external |)(?:(?:private|internal|linkonce|linkonce_odr|weak|weak_odr|common|appending|extern_weak|available_externally) )?(?:default |hidden |protected )?(?:dllimport |dllexport )?(?:unnamed_addr |)(?:thread_local(?:\([a-z]*\))? )?alias"
plain = re.compile(alias_match_prefix + r" (.*?))(| addrspace\(\d+\) *)\*($| *(?:%|@|null|undef|blockaddress|addrspacecast|\[\[[a-zA-Z]|\{\{).*$)")
cast  = re.compile(alias_match_prefix + r") ((?:bitcast|inttoptr|addrspacecast)\s*\(.* to (.*?)(| addrspace\(\d+\) *)\*\)\s*(?:;.*)?$)")
gep   = re.compile(alias_match_prefix + r") ((?:getelementptr)\s*(?:inbounds)?\s*\((?P<type>.*), (?P=type)(?:\s*addrspace\(\d+\)\s*)?\* .*\)\s*(?:;.*)?$)")

def conv(line):
  m = re.match(cast, line)
  if m:
    return m.group(1) + " " + m.group(3) + ", " + m.group(2)
  m = re.match(gep, line)
  if m:
    return m.group(1) + " " + m.group(3) + ", " + m.group(2)
  m = re.match(plain, line)
  if m:
    return m.group(1) + ", " + m.group(2) + m.group(3) + "*" + m.group(4) + "\n"
  return line

for line in sys.stdin:
  sys.stdout.write(conv(line))

apply.sh:
for name in "$@"
do
  python3 `dirname "$0"`/update.py < "$name" > "$name.tmp" && mv "$name.tmp" "$name"
  rm -f "$name.tmp"
done

The actual commands:
From llvm/src:
find test/ -name *.ll | xargs ./apply.sh
From llvm/src/tools/clang:
find test/ -name *.mm -o -name *.m -o -name *.cpp -o -name *.c | xargs -I '{}' ../../apply.sh "{}"
From llvm/src/tools/polly:
find test/ -name *.ll | xargs ./apply.sh

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247378 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[modules] Move ConvertUTF.h to a separate submodule that doesn't require C++.
Richard Smith [Fri, 11 Sep 2015 03:14:00 +0000 (03:14 +0000)]
[modules] Move ConvertUTF.h to a separate submodule that doesn't require C++.
The former setup once resulted in us ignoring the module for C compilations,
but Clang now errors on this if the header is included from C code (which it is).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247377 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFixed a bug that BranchProbability is not defined in BlockFrequency.cpp. NFC.
Cong Hou [Fri, 11 Sep 2015 02:47:30 +0000 (02:47 +0000)]
Fixed a bug that BranchProbability is not defined in BlockFrequency.cpp. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247376 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAsmWriter: Avoid O(N^2) processing of metadata
Duncan P. N. Exon Smith [Fri, 11 Sep 2015 01:34:59 +0000 (01:34 +0000)]
AsmWriter: Avoid O(N^2) processing of metadata

Fix embarrassing bugs I introduced to the `SlotTracker` in or around
r235785.  I had us iterating through every instruction in a function
(and hitting a map in the LLVMContext) for every basic block in the
function.

While there, completely avoid the call to
`SlotTracker::processFunctionMetadata()` from
`SlotTracker::processFunction()` if we've speculatively done this
already in `SlotTracker::processModule()` by checking
`ShouldInitializeAllMetadata` (this wasn't an algorithmic problem, but
it's touching the same line of code).

Fixes PR24699.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247372 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert "[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing argume...
Mehdi Amini [Fri, 11 Sep 2015 01:33:48 +0000 (01:33 +0000)]
Revert "[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing arguments at callsite"

This reverts commit r247356.

Breaks test/Transforms/InstCombine/pr8547.ll with:

Wrong types for attribute: byval inalloca nest noalias nocapture nonnull readnone readonly sret dereferenceable(1) dereferenceable_or_null(1)
  %call = call i32 (i8*, ...) @printf(i8* getelementptr inbounds ([10 x i8], [10 x i8]* @.str, i64 0, i64 0), i32 nonnull %conv2) #0
LLVM ERROR: Broken function found, compilation aborted!

From: Mehdi Amini <mehdi.amini@apple.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247371 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[libFuzzer] perform fewer crossover operations compared to plain mutations
Kostya Serebryany [Fri, 11 Sep 2015 00:20:58 +0000 (00:20 +0000)]
[libFuzzer] perform fewer crossover operations compared to plain mutations

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247364 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd .exe check to Execute to fix clang-modernize tests broken in r247358
Reid Kleckner [Thu, 10 Sep 2015 23:59:45 +0000 (23:59 +0000)]
Add .exe check to Execute to fix clang-modernize tests broken in r247358

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247361 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[modules] Add another .def file to our list of textual headers.
Richard Smith [Thu, 10 Sep 2015 23:30:27 +0000 (23:30 +0000)]
[modules] Add another .def file to our list of textual headers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247359 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoScanDirForExecutable on Windows fails to find executables with the "exe" extension...
Reid Kleckner [Thu, 10 Sep 2015 23:28:06 +0000 (23:28 +0000)]
ScanDirForExecutable on Windows fails to find executables with the "exe" extension in name

When the driver tries to locate a program by its name, e.g. a linker, it
scans the paths provided by the toolchain using the ScanDirForExecutable
function. If the lookup fails, the driver uses
llvm::sys::findProgramByName. Unlike llvm::sys::findProgramByName,
ScanDirForExecutable is not aware of file extensions. If the program has
the "exe" extension in its name, which is very common on Windows,
ScanDirForExecutable won't find it under the toolchain-provided paths.

This patch changes the Windows version of the "`can_execute`" function
called by ScanDirForExecutable to respect file extensions, similarly to
llvm::sys::findProgramByName.

Patch by Oleg Ranevskyy

Reviewers: rnk

Differential Revision: http://reviews.llvm.org/D12711

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247358 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoPass BranchProbability/BlockMass by value instead of const& as they are small. NFC.
Cong Hou [Thu, 10 Sep 2015 23:10:42 +0000 (23:10 +0000)]
Pass BranchProbability/BlockMass by value instead of const& as they are small. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247357 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing arguments...
Chen Li [Thu, 10 Sep 2015 23:04:49 +0000 (23:04 +0000)]
[InstCombineCalls] Use isKnownNonNullAt() to check nullness of passing arguments at callsite

Summary: This patch replaces isKnownNonNull() with isKnownNonNullAt() when checking nullness of passing arguments at callsite. In this way it can handle cases where the argument does not have nonnull attribute but has a dominating null check from the CFG.

Reviewers: reames

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D12779

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247356 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMark another method const. Sorry for missing this one the first time.
Rafael Espindola [Thu, 10 Sep 2015 22:46:38 +0000 (22:46 +0000)]
Mark another method const. Sorry for missing this one the first time.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247354 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[InstCombineCalls] Use isKnownNonNullAt() to check nullness of gc.relocate return...
Chen Li [Thu, 10 Sep 2015 22:35:41 +0000 (22:35 +0000)]
[InstCombineCalls] Use isKnownNonNullAt() to check nullness of gc.relocate return value

Summary: This patch replaces isKnownNonNull() with isKnownNonNullAt() when checking nullness of gc.relocate return value. In this way it can handle cases where the relocated value does not have nonnull attribute but has a dominating null check from the CFG.

Reviewers: reames

Subscribers: llvm-commits, sanjoy

Differential Revision: http://reviews.llvm.org/D12772

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247353 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove gcc warning when comparing an unsigned var for >= 0
Filipe Cabecinhas [Thu, 10 Sep 2015 22:34:39 +0000 (22:34 +0000)]
Remove gcc warning when comparing an unsigned var for >= 0

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247352 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[WinEH] Push and pop EBP for 32-bit funclets
Reid Kleckner [Thu, 10 Sep 2015 22:00:02 +0000 (22:00 +0000)]
[WinEH] Push and pop EBP for 32-bit funclets

The Win32 EH runtime caller does not preserve EBP, even though it does
preserve the CSRs (EBX, ESI, EDI) for us. The result was that each
finally funclet call would leave the frame pointer off by 12 bytes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247348 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAMDGPU: Simplify debug printing
Matt Arsenault [Thu, 10 Sep 2015 21:51:19 +0000 (21:51 +0000)]
AMDGPU: Simplify debug printing

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247345 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAMDGPU: Use StringRef value
Matt Arsenault [Thu, 10 Sep 2015 21:51:15 +0000 (21:51 +0000)]
AMDGPU: Use StringRef value

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247344 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[SPARC] Switch to the Machine Scheduler.
James Y Knight [Thu, 10 Sep 2015 21:49:06 +0000 (21:49 +0000)]
[SPARC] Switch to the Machine Scheduler.

The (mostly-deprecated) SelectionDAG-based ILPListDAGScheduler scheduler
was making poor scheduling decisions, causing high register pressure and
extraneous register spills.

Switching to the newer machine scheduler generates better code -- even
without there being a machine model defined for SPARC yet.

(Actually committing the test changes too, this time, unlike r247315)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247343 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoMark two methods const.
Rafael Espindola [Thu, 10 Sep 2015 21:48:36 +0000 (21:48 +0000)]
Mark two methods const.

While at it, optimize getOffset a bit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247342 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix SEH state numbering algorithm to handle cleanupendpads
Reid Kleckner [Thu, 10 Sep 2015 21:46:36 +0000 (21:46 +0000)]
Fix SEH state numbering algorithm to handle cleanupendpads

WinEHPrepare's new coloring algorithm really expects to see
cleanupendpads now, so Clang will start emitting them soon.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247341 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LV] Relax Small Size Reduction Type Requirement
Matthew Simpson [Thu, 10 Sep 2015 21:12:57 +0000 (21:12 +0000)]
[LV] Relax Small Size Reduction Type Requirement

This patch enables small size reductions in which the source types are smaller
than the reduction type (e.g., computing an i16 sum from the values in an i8
array). The previous behavior was to only allow small size reductions if the
source types and reduction type were the same. The change accounts for the fact
that the existing sign- and zero-extend instructions in these cases should
still be included in the cost model.

Differential Revision: http://reviews.llvm.org/D12770

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247337 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[RuntimeDyld] Support non-zero addends for the MachO X86_64 SUBTRACTOR reloc.
Lang Hames [Thu, 10 Sep 2015 21:05:58 +0000 (21:05 +0000)]
[RuntimeDyld] Support non-zero addends for the MachO X86_64 SUBTRACTOR reloc.

This functionality was accidentally left out of r247119.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247336 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[RuntimeDyld] Fix a bug in debugging output: all sections should be dumped
Lang Hames [Thu, 10 Sep 2015 20:44:36 +0000 (20:44 +0000)]
[RuntimeDyld] Fix a bug in debugging output: all sections should be dumped
before any relocations have been applied, and again after all relocations have
been applied.

Previously each section was dumped before and after relocations targetting it
were applied, but this only shows the impact of relocations that point to other
symbols in the same section.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247335 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd an explicit 'inline' specifier to these static functions. GCC is
Chandler Carruth [Thu, 10 Sep 2015 20:34:57 +0000 (20:34 +0000)]
Add an explicit 'inline' specifier to these static functions. GCC is
warning on them having always_inline attribute for reasons I don't fully
understand -- static functions are just as inlinable as inline
functions in terms of linkage.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247334 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoCorrect the return type of CreateCleanupEndPad
Reid Kleckner [Thu, 10 Sep 2015 20:07:25 +0000 (20:07 +0000)]
Correct the return type of CreateCleanupEndPad

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247331 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert "[SPARC] Switch to the Machine Scheduler."
James Y Knight [Thu, 10 Sep 2015 19:42:03 +0000 (19:42 +0000)]
Revert "[SPARC] Switch to the Machine Scheduler."

This reverts commit r247315.

Accidentally omitted test changes; will resubmit full change shortly.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@247328 91177308-0d34-0410-b5e6-96231b3b80d8