oota-llvm.git
10 years agoAssert that MCSymbolizer is constructed with a valid (or at least non-null)
Lang Hames [Thu, 27 Mar 2014 02:49:18 +0000 (02:49 +0000)]
Assert that MCSymbolizer is constructed with a valid (or at least non-null)
RelocationInfo argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204893 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoMove MCSymbolizer's constructor into header. It's trivial - there's no need for
Lang Hames [Thu, 27 Mar 2014 02:42:52 +0000 (02:42 +0000)]
Move MCSymbolizer's constructor into header. It's trivial - there's no need for
it to be out-of-line.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204892 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoUpdate MCSymbolizer and its subclasses' constructors to reflect the fact that
Lang Hames [Thu, 27 Mar 2014 02:39:01 +0000 (02:39 +0000)]
Update MCSymbolizer and its subclasses' constructors to reflect the fact that
they take ownership of the RelocationInfo they're constructed with.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204891 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoinalloca: *Really* fix the docs
Reid Kleckner [Thu, 27 Mar 2014 01:38:48 +0000 (01:38 +0000)]
inalloca: *Really* fix the docs

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204890 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoRemove unneeded stale type.
Reid Kleckner [Thu, 27 Mar 2014 01:34:51 +0000 (01:34 +0000)]
Remove unneeded stale type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204889 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoinalloca: Fix incorrect example IR and remove LangRef warning
Reid Kleckner [Thu, 27 Mar 2014 01:32:22 +0000 (01:32 +0000)]
inalloca: Fix incorrect example IR and remove LangRef warning

The LangRef warning wasn't formatting the way I intended it to anyway.

Surprisingly inalloca appears to work, even when optimizations are
enabled.  We generate very bad code for it, but we can self-host and run
lots of big tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204888 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoRemove forward declaration for Target class - Target is already defined here.
Lang Hames [Thu, 27 Mar 2014 01:05:49 +0000 (01:05 +0000)]
Remove forward declaration for Target class - Target is already defined here.

No functional change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204885 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[X86][Vectorizer Cost Model] Correct vectorization cost model for v2i64->v2f64
Quentin Colombet [Thu, 27 Mar 2014 00:52:16 +0000 (00:52 +0000)]
[X86][Vectorizer Cost Model] Correct vectorization cost model for v2i64->v2f64
and v4i64->v4f64.

The new costs match what we did for SSE2 and reflect the reality of our codegen.

<rdar://problem/16381225>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204884 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoCorrectly propagates st_size.
Rafael Espindola [Thu, 27 Mar 2014 00:28:24 +0000 (00:28 +0000)]
Correctly propagates st_size.

This also finally removes a bogus call to AliasedSymbol.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204883 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoadd 'requires asserts' to test that needs it
Jim Grosbach [Thu, 27 Mar 2014 00:20:42 +0000 (00:20 +0000)]
add 'requires asserts' to test that needs it

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204882 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agollvm-cov: When reading strings in gcov data, skip leading zeros
Justin Bogner [Thu, 27 Mar 2014 00:06:36 +0000 (00:06 +0000)]
llvm-cov: When reading strings in gcov data, skip leading zeros

It seems that gcov, when faced with a string that is apparently zero
length, just keeps reading words until it finds a length it likes
better. I'm not really sure why this is, but it's simple enough to
make llvm-cov follow suit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204881 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoX86: Correct vectorization cost model for v8f32->v8i8.
Jim Grosbach [Thu, 27 Mar 2014 00:04:11 +0000 (00:04 +0000)]
X86: Correct vectorization cost model for v8f32->v8i8.

Fix the cost model to reflect the reality of our codegen.

rdar://16370633

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204880 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoTreat lifetime.start'd memory like we treat freshly alloca'd memory. Patch by Björn...
Nick Lewycky [Wed, 26 Mar 2014 23:45:15 +0000 (23:45 +0000)]
Treat lifetime.start'd memory like we treat freshly alloca'd memory. Patch by Björn Steinbrink!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204876 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoReorder arguments on test command line to make it easier to cut and
Eric Christopher [Wed, 26 Mar 2014 23:10:28 +0000 (23:10 +0000)]
Reorder arguments on test command line to make it easier to cut and
paste.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204875 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Generate VSX permutations for v2[fi]64 vectors
Hal Finkel [Wed, 26 Mar 2014 22:58:37 +0000 (22:58 +0000)]
[PowerPC] Generate VSX permutations for v2[fi]64 vectors

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204873 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agollvm-cov: Move XFAIL after the body of the test
Justin Bogner [Wed, 26 Mar 2014 22:51:39 +0000 (22:51 +0000)]
llvm-cov: Move XFAIL after the body of the test

llvm-cov tests are sensitive to line number changes, so putting this
at the end will limit churn when we fix the XFAIL.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204871 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agollvm-cov: Disable test on big endian machines
Justin Bogner [Wed, 26 Mar 2014 22:36:48 +0000 (22:36 +0000)]
llvm-cov: Disable test on big endian machines

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204868 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoCloneFunction: Clone all attributes, including the CC
Reid Kleckner [Wed, 26 Mar 2014 22:26:35 +0000 (22:26 +0000)]
CloneFunction: Clone all attributes, including the CC

Summary:
Tested with a unit test because we don't appear to have any transforms
that use this other than ASan, I think.

Fixes PR17935.

Reviewers: nicholas

CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3194

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204866 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoThis is a fix for PR# 19051. I noticed code gen differences due to code motion when...
Ekaterina Romanova [Wed, 26 Mar 2014 22:15:28 +0000 (22:15 +0000)]
This is a fix for PR# 19051. I noticed code gen differences due to code motion when running tests with and without the debug info at O2. The problem is in branch folding. A loop wanted to skip the debug info, but actually it didn't do so.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204865 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd comments. Addressing review comments from Evan on r204690.
Manman Ren [Wed, 26 Mar 2014 22:14:09 +0000 (22:14 +0000)]
Add comments. Addressing review comments from Evan on r204690.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204864 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agollvm-cov: Handle functions with no line number
Justin Bogner [Wed, 26 Mar 2014 22:03:06 +0000 (22:03 +0000)]
llvm-cov: Handle functions with no line number

Functions may in an instrumented binary but not in the original source
when they're inserted by the compiler or the runtime. These functions
aren't meaningful to the user, so teach llvm-cov to skip over them
instead of crashing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204863 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix a problem with the ARM assembler incorrectly matching a
Kevin Enderby [Wed, 26 Mar 2014 21:54:11 +0000 (21:54 +0000)]
Fix a problem with the ARM assembler incorrectly matching a
vector list parameter that is using all lanes "{d0[], d2[]}" but can
match and instruction with a ”{d0, d2}" parameter.

I’m finishing up a fix for proper checking of the unsupported
alignments on vld/vst instructions and ran into this.  Thus I don’t
have a test case at this time.  And adding all code that will
demonstrate the bug would obscure the very simple one line fix.
So if you would indulge me on not having a test case at this
time I’ll instead offer up a detailed explanation of what is
going on in this commit message.

This instruction:

vld2.8  {d0[], d2[]}, [r4:64]

is not legal as the alignment can only be 16 when the size is 8.
Per this documentation:

A8.8.325 VLD2 (single 2-element structure to all lanes)
 <align> The alignment. It can be one of:
16 2-byte alignment, available only if <size> is 8, encoded as a = 1.
32 4-byte alignment, available only if <size> is 16, encoded as a = 1.
64 8-byte alignment, available only if <size> is 32, encoded as a = 1.
omitted Standard alignment, see Unaligned data access on page A3-108.

So when code is added to the llvm integrated assembler to not match
that instruction because of the alignment it then goes on to try to match
other instructions and comes across this:

vld2.8  {d0, d2}, [r4:64]

and and matches it. This is because of the method
ARMOperand::isVecListDPairSpaced() is missing the check of the Kind.
In this case the Kind is k_VectorListAllLanes . While the name of the method
may suggest that this is OK it really should check that the Kind is
k_VectorList.

As the method ARMOperand::isDoubleSpacedVectorAllLanes() is what was
used to match {d0[], d2[]}  and correctly checks the Kind:

  bool isDoubleSpacedVectorAllLanes() const {
    return Kind == k_VectorListAllLanes && VectorList.isDoubleSpaced;
  }

where the original ARMOperand::isVecListDPairSpaced() does not check
the Kind:

  bool isVecListDPairSpaced() const {
    if (isSingleSpacedVectorList()) return false;
    return (ARMMCRegisterClasses[ARM::DPairSpcRegClassID]
              .contains(VectorList.RegNum));
  }

Jim Grosbach has reviewed the change and said:  Yep, that sounds right. …
And by "right" I mean, "wow, that's a nasty latent bug I'm really, really
glad to see fixed." :)

rdar://16436683

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204861 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd a unit test for Invoke iteration, similar to the one for Call
Eli Bendersky [Wed, 26 Mar 2014 21:46:24 +0000 (21:46 +0000)]
Add a unit test for Invoke iteration, similar to the one for Call

The tests are refactored to use the same fixture.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204860 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoPR15967 Fix in basicaa for faulty returning no alias.
Arnold Schwaighofer [Wed, 26 Mar 2014 21:30:19 +0000 (21:30 +0000)]
PR15967 Fix in basicaa for faulty returning no alias.

This commit consist of two parts.
The first part fix the PR15967. The wrong conclusion was made when the MaxLookup
limit was reached. The fix introduce a out parameter (MaxLookupReached) to
DecomposeGEPExpression that the function aliasGEP can act upon.
The second part is introducing the constant MaxLookupSearchDepth to make sure
that DecomposeGEPExpression and GetUnderlyingObject use the same search depth.
This is a small cleanup to clarify the original algorithm.

Patch by Karl-Johan Karlsson!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204859 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoSimplify PBQP graph removeAdjEdgeId implementation.
Lang Hames [Wed, 26 Mar 2014 21:21:53 +0000 (21:21 +0000)]
Simplify PBQP graph removeAdjEdgeId implementation.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204857 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix bot breakage in InstructionsTest.
Eli Bendersky [Wed, 26 Mar 2014 21:11:34 +0000 (21:11 +0000)]
Fix bot breakage in InstructionsTest.

Makes sure the Call dies before the Function

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204856 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix problem with r204836
Eli Bendersky [Wed, 26 Mar 2014 20:41:15 +0000 (20:41 +0000)]
Fix problem with r204836

In CallInst, op_end() points at the callee, which we don't want to iterate over
when just iterating over arguments. Now take this into account when returning
a iterator_range from arg_operands. Similar reasoning for InvokeInst.

Also adds a unit test to verify this actually works as expected.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204851 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] VSX loads and stores support unaligned access
Hal Finkel [Wed, 26 Mar 2014 19:39:09 +0000 (19:39 +0000)]
[PowerPC] VSX loads and stores support unaligned access

I've not yet updated PPCTTI because I'm not sure what the actual relative cost
is compared to the aligned uses.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204848 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix the ARM VST4 (single 4-element structure from one lane)
Kevin Enderby [Wed, 26 Mar 2014 19:35:40 +0000 (19:35 +0000)]
Fix the ARM VST4 (single 4-element structure from one lane)
size 16 double-spaced registers instruction printing.

This:
vld4.16 {d17[1], d19[1], d21[1], d23[1]}, [r7]!

was being printed as:

vld4.16 {d17[1], d18[1], d19[1], d20[1]}, [r7]!

rdar://16435096

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204847 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoRemove PBQP-cost dimension sanity assertion in PBQP::Graph::addConstructedEdge.
Lang Hames [Wed, 26 Mar 2014 19:22:51 +0000 (19:22 +0000)]
Remove PBQP-cost dimension sanity assertion in PBQP::Graph::addConstructedEdge.
We're already effectively checking sanity for that in PBQP::Graph::addEdge.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204844 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Use v2f64 <-> v2i64 VSX conversion instructions
Hal Finkel [Wed, 26 Mar 2014 19:13:54 +0000 (19:13 +0000)]
[PowerPC] Use v2f64 <-> v2i64 VSX conversion instructions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204843 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoChange the PBQP graph adjacency list structure from std::set to std::vector.
Lang Hames [Wed, 26 Mar 2014 18:58:00 +0000 (18:58 +0000)]
Change the PBQP graph adjacency list structure from std::set to std::vector.

The edge data structure (EdgeEntry) now holds the indices of its entries in the
adjacency lists of the nodes it connects. This trades a little ugliness for
faster insertion/removal, which is now O(1) with a cheap constant factor. All
of this is implementation detail within the PBQP graph, the external API remains
unchanged.

Individual register allocations are likely to change, since the adjacency lists
will now be ordered differently (or rather, will now be unordered). This
shouldn't affect the average quality of allocations however.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204841 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoR600: Add a testcase for sext_in_reg I missed.
Matt Arsenault [Wed, 26 Mar 2014 18:31:06 +0000 (18:31 +0000)]
R600: Add a testcase for sext_in_reg I missed.

This sext_inreg i32 in i64 case was already handled, but not enabled.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204840 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Remove some dead VSX v4f32 store patterns
Hal Finkel [Wed, 26 Mar 2014 18:26:36 +0000 (18:26 +0000)]
[PowerPC] Remove some dead VSX v4f32 store patterns

These patterns are dead (because v4f32 stores are currently promoted to v4i32
and stored using Altivec instructions), and also are likely not correct
(because they'd store the vector elements in the opposite order from that
assumed by the rest of the Altivec code).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204839 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Use VSX vector load/stores for v2[fi]64
Hal Finkel [Wed, 26 Mar 2014 18:26:30 +0000 (18:26 +0000)]
[PowerPC] Use VSX vector load/stores for v2[fi]64

These instructions have access to the complete VSX register file. In addition,
they "swap" the order of the elements so that element 0 (the scalar part) comes
first in memory and element 1 follows at a higher address.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204838 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[MCJIT] Check if there have been errors during RuntimeDyld execution.
Juergen Ributzka [Wed, 26 Mar 2014 18:19:27 +0000 (18:19 +0000)]
[MCJIT] Check if there have been errors during RuntimeDyld execution.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204837 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoEnable range-for iteration over call/invoke arguments.
Eli Bendersky [Wed, 26 Mar 2014 18:18:02 +0000 (18:18 +0000)]
Enable range-for iteration over call/invoke arguments.

Similar to r204835

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204836 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd args() iteartor adapter to Function, for range-for loops.
Eli Bendersky [Wed, 26 Mar 2014 18:04:27 +0000 (18:04 +0000)]
Add args() iteartor adapter to Function, for range-for loops.

This patch is in similar vein to what done earlier to Module::globals/aliases
etc. It allows to iterate over function arguments like this:

  for (Argument Arg : F.args()) {
    ...
  }

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204835 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix for incorrect address sinking in the presence of potential overflows.
Jim Grosbach [Wed, 26 Mar 2014 17:27:01 +0000 (17:27 +0000)]
Fix for incorrect address sinking in the presence of potential overflows.

In some cases it is possible for CGP to attempt to reuse a base address from
another basic block. In those cases we have to be sure that all the address
math was either done at the same bit width, or that none of it overflowed
before it was extended.

Patch by Louis Gerbarg <lgg@apple.com>

rdar://16307442

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204833 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoRevert "X86 memcpy lowering: use "rep movs" even when esi is used as base pointer...
Hans Wennborg [Wed, 26 Mar 2014 16:30:54 +0000 (16:30 +0000)]
Revert "X86 memcpy lowering: use "rep movs" even when esi is used as base pointer" (r204174)

>  For functions where esi is used as base pointer, we would previously fall ba
>  from lowering memcpy with "rep movs" because that clobbers esi.
>
>  With this patch, we just store esi in another physical register, and restore
>  it afterwards. This adds a little bit of register preassure, but the more
>  efficient memcpy should be worth it.
>
>  Differential Revision: http://llvm-reviews.chandlerc.com/D2968

This didn't work. I was ending up with code like this:

  lea     edi,[esi+38h]
  mov     ecx,0Fh
  mov     edx,esi
  mov     esi,ebx
  rep movs dword ptr es:[edi],dword ptr [esi]
  lea     ecx,[esi+74h] <-- Ooops, we're now using esi before restoring it from edx.
  add     ebx,3Ch
  mov     esi,edx

I guess if we want to do this we need stronger glue or something, or doing the expansion
much later.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204829 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Add v2i64 as a legal VSX type
Hal Finkel [Wed, 26 Mar 2014 16:12:58 +0000 (16:12 +0000)]
[PowerPC] Add v2i64 as a legal VSX type

v2i64 needs to be a legal VSX type because it is the SetCC result type from
v2f64 comparisons. We need to expand all non-arithmetic v2i64 operations.

This fixes the lowering for v2f64 VSELECT.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204828 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Use TwoOperandAliasConstraint for ArithLogicR instructions.
Matheus Almeida [Wed, 26 Mar 2014 16:09:43 +0000 (16:09 +0000)]
[mips] Use TwoOperandAliasConstraint for ArithLogicR instructions.

This enables TableGen to generate an additional two operand matcher
for our ArithLogicR class of instructions (constituted by 3 register operands).
E.g.: and $1, $2 <=> and $1, $1, $2

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204826 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Add support to the '.dword' directive.
Matheus Almeida [Wed, 26 Mar 2014 15:44:18 +0000 (15:44 +0000)]
[mips] Add support to the '.dword' directive.

The '.dword' directive accepts a list of expressions and emits
them in 8-byte chunks in successive locations.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204822 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoClarify that select is only non-branching on the IR-level, it often ends
Joerg Sonnenberger [Wed, 26 Mar 2014 15:30:21 +0000 (15:30 +0000)]
Clarify that select is only non-branching on the IR-level, it often ends
up as jump table or other forms of branches on the machine level.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204819 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Rename function in MipsAsmParser.
Matheus Almeida [Wed, 26 Mar 2014 15:24:36 +0000 (15:24 +0000)]
[mips] Rename function in MipsAsmParser.

parseDirectiveWord is a generic function that parses an expression which
means there's no need for it to have such an specific name. Renaming it to
parseDataDirective so that it can also be used to handle .dword directives[1].

[1]To be added in a follow up commit.

No functional changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204818 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Add support to '.set mips64'.
Matheus Almeida [Wed, 26 Mar 2014 15:14:32 +0000 (15:14 +0000)]
[mips] Add support to '.set mips64'.

The '.set mips64' directive enables the feature Mips:FeatureMips64
from assembly. Note that it doesn't modify the ELF header as opposed
to the use of -mips64 from the command-line. The reason for this
is that we want to be as compatible as possible with existing assemblers
like GAS.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204817 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAArch64_BE Elf support for MC-JIT runtime dynamic linker
Christian Pirker [Wed, 26 Mar 2014 14:57:32 +0000 (14:57 +0000)]
AArch64_BE Elf support for MC-JIT runtime dynamic linker

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204816 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Add support to '.set mips64r2'.
Matheus Almeida [Wed, 26 Mar 2014 14:52:22 +0000 (14:52 +0000)]
[mips] Add support to '.set mips64r2'.

The '.set mips64r2' directive enables the feature Mips:FeatureMips64r2
from assembly. Note that it doesn't modify the ELF header as opposed
to the use of -mips64r2 from the command-line. The reason for this
is that we want to be as compatible as possible with existing assemblers
like GAS.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204815 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAArch64_BE function argument passing for ARM ABI
Christian Pirker [Wed, 26 Mar 2014 14:51:22 +0000 (14:51 +0000)]
AArch64_BE function argument passing for ARM ABI

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204814 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoARM: add intrinsics for the v8 ldaex/stlex
Tim Northover [Wed, 26 Mar 2014 14:39:31 +0000 (14:39 +0000)]
ARM: add intrinsics for the v8 ldaex/stlex

We've already got versions without the barriers, so this just adds IR-level
support for generating the new v8 ones.

rdar://problem/16227836

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204813 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoClarify llvm.clear_cache description.
Joerg Sonnenberger [Wed, 26 Mar 2014 14:35:21 +0000 (14:35 +0000)]
Clarify llvm.clear_cache description.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204812 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Hoist common functionality into a new function.
Matheus Almeida [Wed, 26 Mar 2014 14:26:27 +0000 (14:26 +0000)]
[mips] Hoist common functionality into a new function.

Given that we support multiple directives that enable a particular feature
(e.g. '.set mips16'), it's best to hoist that code into a new function
so that we don't repeat the same pattern w.r.t parsing and handling error cases.

No functional changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204811 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoChange @llvm.clear_cache default to call rt-lib
Renato Golin [Wed, 26 Mar 2014 14:01:32 +0000 (14:01 +0000)]
Change @llvm.clear_cache default to call rt-lib

After some discussion on IRC, emitting a call to the library function seems
like a better default, since it will move from a compiler internal error to
a linker error, that the user can work around until LLVM is fixed.

I'm also adding a note on the responsibility of the user to confirm that
the cache was cleared on platforms where nothing is done.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204806 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] The decision to use MO_GOT_PAGE and MO_GOT_OFST depends on the ABI being N32...
Daniel Sanders [Wed, 26 Mar 2014 13:59:42 +0000 (13:59 +0000)]
[mips] The decision to use MO_GOT_PAGE and MO_GOT_OFST depends on the ABI being N32 or N64 not the arch being MIPS64

Summary: No functional change (in supported use cases)

Reviewers: matheusalmeida

Reviewed By: matheusalmeida

Differential Revision: http://llvm-reviews.chandlerc.com/D3177

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204805 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix AVX512 Gather and Scatter execution domains.
Cameron McInally [Wed, 26 Mar 2014 13:50:50 +0000 (13:50 +0000)]
Fix AVX512 Gather and Scatter execution domains.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204804 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Add support for '.option pic2'.
Matheus Almeida [Wed, 26 Mar 2014 13:40:29 +0000 (13:40 +0000)]
[mips] Add support for '.option pic2'.

The directive '.option pic2' enables PIC from assembly source.
At the moment none of the macros/directives check the PIC bit
but that's going to be fixed relatively soon. For example, the
expansion of macros like 'la' depend on the relocation model.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204803 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd @llvm.clear_cache builtin
Renato Golin [Wed, 26 Mar 2014 12:52:28 +0000 (12:52 +0000)]
Add @llvm.clear_cache builtin

Implementing the LLVM part of the call to __builtin___clear_cache
which translates into an intrinsic @llvm.clear_cache and is lowered
by each target, either to a call to __clear_cache or nothing at all
incase the caches are unified.

Updating LangRef and adding some tests for the implemented architectures.
Other archs will have to implement the method in case this builtin
has to be compiled for it, since the default behaviour is to bail
unimplemented.

A Clang patch is required for the builtin to be lowered into the
llvm intrinsic. This will be done next.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204802 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Lower VSELECT using xxsel when VSX is available
Hal Finkel [Wed, 26 Mar 2014 12:49:28 +0000 (12:49 +0000)]
[PowerPC] Lower VSELECT using xxsel when VSX is available

With VSX there is a real vector select instruction, and so we should use it.
Note that VSELECT will still scalarize for v2f64 because the corresponding
SetCC result type (v2i64) is not currently a legal type.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204801 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Add tests for t0-t3 for N32/N64
Daniel Sanders [Wed, 26 Mar 2014 11:46:34 +0000 (11:46 +0000)]
[mips] Add tests for t0-t3 for N32/N64

These are aliases of t4-t7 and are provided for compatibility with both the
original ABI documentation (using t4-t7) and GNU As (using t0-t3)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204797 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] The register names depend on the ABI being N32/N64 rather than the arch being...
Daniel Sanders [Wed, 26 Mar 2014 11:39:07 +0000 (11:39 +0000)]
[mips] The register names depend on the ABI being N32/N64 rather than the arch being mips64

Summary: Added test cases for O32 and N32 on MIPS64.

Reviewers: matheusalmeida

Reviewed By: matheusalmeida

Differential Revision: http://llvm-reviews.chandlerc.com/D3175

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204796 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFollow-up to r204790: don't try to emit line tables if there are no functions with...
Timur Iskhodzhanov [Wed, 26 Mar 2014 11:24:36 +0000 (11:24 +0000)]
Follow-up to r204790: don't try to emit line tables if there are no functions with DI in the TU

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204795 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] $s8 is an alias for $fp in all ABI's, not just N32/N64.
Daniel Sanders [Wed, 26 Mar 2014 11:05:24 +0000 (11:05 +0000)]
[mips] $s8 is an alias for $fp in all ABI's, not just N32/N64.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204793 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] Move the CHECK lines in mips*-register-names.s to make it more obvious which...
Daniel Sanders [Wed, 26 Mar 2014 10:54:30 +0000 (10:54 +0000)]
[mips] Move the CHECK lines in mips*-register-names.s to make it more obvious which CHECK matches with which insn

This reveals a small mistake in mips-register-names.s ($sp is tested twice and
$s8 is not tested) which will be fixed in a follow-up commit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204792 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd tests for r204790
Timur Iskhodzhanov [Wed, 26 Mar 2014 09:51:45 +0000 (09:51 +0000)]
Add tests for r204790

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204791 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix PR19239 - Add support for generating debug info for functions without lexical...
Timur Iskhodzhanov [Wed, 26 Mar 2014 09:50:36 +0000 (09:50 +0000)]
Fix PR19239 - Add support for generating debug info for functions without lexical scopes and/or debug info at all

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204790 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoUse -LABEL checks in the COFF debug info tests
Timur Iskhodzhanov [Wed, 26 Mar 2014 08:45:02 +0000 (08:45 +0000)]
Use -LABEL checks in the COFF debug info tests

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204788 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoRevert "Prevent alias from pointing to weak aliases."
Rafael Espindola [Wed, 26 Mar 2014 06:14:40 +0000 (06:14 +0000)]
Revert "Prevent alias from pointing to weak aliases."

This reverts commit r204781.

I will follow up to with msan folks to see what is what they
were trying to do with aliases to weak aliases.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204784 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Generate logical vector VSX instructions
Hal Finkel [Wed, 26 Mar 2014 04:55:40 +0000 (04:55 +0000)]
[PowerPC] Generate logical vector VSX instructions

These instructions are essentially the same as their Altivec counterparts, but
have access to the larger VSX register file.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204782 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoPrevent alias from pointing to weak aliases.
Rafael Espindola [Wed, 26 Mar 2014 04:48:47 +0000 (04:48 +0000)]
Prevent alias from pointing to weak aliases.

Aliases are just another name for a position in a file. As such, the
regular symbol resolutions are not applied. For example, given

define void @my_func() {
  ret void
}
@my_alias = alias weak void ()* @my_func
@my_alias2 = alias void ()* @my_alias

We produce without this patch:

        .weak   my_alias
my_alias = my_func
        .globl  my_alias2
my_alias2 = my_alias

That is, in the resulting ELF file my_alias, my_func and my_alias are
just 3 names pointing to offset 0 of .text. That is *not* the
semantics of IR linking. For example, linking in a

@my_alias = alias void ()* @other_func

would require the strong my_alias to override the weak one and
my_alias2 would end up pointing to other_func.

There is no way to represent that with aliases being just another
name, so the best solution seems to be to just disallow it, converting
a miscompile into an error.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204781 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoDebugInfo: Add fission-related sections to COFF
David Blaikie [Wed, 26 Mar 2014 03:05:10 +0000 (03:05 +0000)]
DebugInfo: Add fission-related sections to COFF

Allows this test to pass on COFF platforms so we don't need to restrict
this test to a single target anymore.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204780 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoCorrectly detect if a symbol uses a reserved section index or not.
Rafael Espindola [Wed, 26 Mar 2014 00:16:43 +0000 (00:16 +0000)]
Correctly detect if a symbol uses a reserved section index or not.

The logic was incorrect for variables, causing them to end up in the wrong
section if the section had an index >= 0xff00.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204771 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[X86] Add broadcast instructions to the table used by ExeDepsFix pass.
Quentin Colombet [Wed, 26 Mar 2014 00:10:22 +0000 (00:10 +0000)]
[X86] Add broadcast instructions to the table used by ExeDepsFix pass.

Adds the different broadcast instructions to the ReplaceableInstrsAVX2 table.
That way the ExeDepsFix pass can take better decisions when AVX2 broadcasts are
across domain (int <-> float).

In particular, prior to this patch we were generating:
  vpbroadcastd  LCPI1_0(%rip), %ymm2
  vpand %ymm2, %ymm0, %ymm0
  vmaxps  %ymm1, %ymm0, %ymm0 ## <- domain change penalty

Now, we generate the following nice sequence where everything is in the float
domain:
  vbroadcastss  LCPI1_0(%rip), %ymm2
  vandps  %ymm2, %ymm0, %ymm0
  vmaxps  %ymm1, %ymm0, %ymm0

<rdar://problem/16354675>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204770 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoCreate .symtab_shndxr only when needed.
Rafael Espindola [Tue, 25 Mar 2014 23:44:25 +0000 (23:44 +0000)]
Create .symtab_shndxr only when needed.

We need .symtab_shndxr if and only if a symbol references a section with an
index >= 0xff00.

The old code was trying to figure out if the section was needed ahead of time,
making it a fairly dependent on the code actually writing the table. It was
also somewhat conservative and would create the section in cases where it was
not needed.

If I remember correctly, the old structure was there so that the sections were
created in the same order gas creates them. That was valuable when MC's support
for ELF was new and we tested with elf-dump.py.

This patch refactors the symbol table creation to another class and makes it
obvious that .symtab_shndxr is really only created when we are about to output
a reference to a section index >= 0xff00.

While here, also improve the tests to use macros. One file is one section
short of needing .symtab_shndxr, the second one has just the right number.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204769 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Select between VSX A-type and M-type FMA instructions just before RA
Hal Finkel [Tue, 25 Mar 2014 23:29:21 +0000 (23:29 +0000)]
[PowerPC] Select between VSX A-type and M-type FMA instructions just before RA

The VSX instruction set has two types of FMA instructions: A-type (where the
addend is taken from the output register) and M-type (where one of the product
operands is taken from the output register). This adds a small pass that runs
just after MI scheduling (and, thus, just before register allocation) that
mutates A-type instructions (that are created during isel) into M-type
instructions when:

 1. This will eliminate an otherwise-necessary copy of the addend

 2. One of the product operands is killed by the instruction

The "right" moment to make this decision is in between scheduling and register
allocation, because only there do we know whether or not one of the product
operands is killed by any particular instruction. Unfortunately, this also
makes the implementation somewhat complicated, because the MIs are not in SSA
form and we need to preserve the LiveIntervals analysis.

As a simple example, if we have:

%vreg5<def> = COPY %vreg9; VSLRC:%vreg5,%vreg9
%vreg5<def,tied1> = XSMADDADP %vreg5<tied0>, %vreg17, %vreg16,
                        %RM<imp-use>; VSLRC:%vreg5,%vreg17,%vreg16
  ...
  %vreg9<def,tied1> = XSMADDADP %vreg9<tied0>, %vreg17, %vreg19,
                        %RM<imp-use>; VSLRC:%vreg9,%vreg17,%vreg19
  ...

We can eliminate the copy by changing from the A-type to the
M-type instruction. This means:

  %vreg5<def,tied1> = XSMADDADP %vreg5<tied0>, %vreg17, %vreg16,
                        %RM<imp-use>; VSLRC:%vreg5,%vreg17,%vreg16

is replaced by:

  %vreg16<def,tied1> = XSMADDMDP %vreg16<tied0>, %vreg18, %vreg9,
                        %RM<imp-use>; VSLRC:%vreg16,%vreg18,%vreg9

and we remove: %vreg5<def> = COPY %vreg9; VSLRC:%vreg5,%vreg9

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204768 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agollvm/test/DebugInfo/empty.ll: Suppress crash for targeting pecoff while investigating.
NAKAMURA Takumi [Tue, 25 Mar 2014 23:16:44 +0000 (23:16 +0000)]
llvm/test/DebugInfo/empty.ll: Suppress crash for targeting pecoff while investigating.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204766 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoUse Endian.h to simplify this code a bit.
Rafael Espindola [Tue, 25 Mar 2014 22:43:53 +0000 (22:43 +0000)]
Use Endian.h to simplify this code a bit.

While at it, factor some logic into FragmentWriter. This will allow more code
to be factored out of the fairly large ELFObjectWriter.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204765 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[configure/make] Propagate names of build host tools when making BuildTools
Meador Inge [Tue, 25 Mar 2014 21:45:41 +0000 (21:45 +0000)]
[configure/make] Propagate names of build host tools when making BuildTools

When cross-compiling LLVM itself the configure/make scripts get confused when
creating the needed build host tools.  For example, building and configuring
like:

  CC_FOR_BUILD='i686-pc-linux-gnu-gcc' CXX_FOR_BUILD='i686-pc-linux-gnu-g++'
  CXX='i686-mingw32-g++' CC='i686-mingw32-gcc' LD='i686-mingw32-ld' /scratch
  /meadori/llvm-trunk/src/trunk/configure --host=i686-mingw32

  CC_FOR_BUILD='i686-pc-linux-gnu-gcc' CXX_FOR_BUILD='i686-pc-linux-gnu-g++'
  CXX='i686-mingw32-g++' CC='i686-mingw32-gcc' LD='i686-mingw32-ld' make

causes the following build break:

  checking whether the C compiler works... configure: error: cannot run C
  compiled programs.
  If you meant to cross compile, use `--host'.
  See `config.log' for more details.

The 'config.log' shows that i686-mingw32-gcc is being used to create
executables for the build host.

This patch fixes the problem by propogating the names of the build host
tools via BUILD_* when configuring/making BuildTools.

Original patch by Ekaterina Sanina.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204760 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[Constant Hoisting] Make the constant candidate map local to the collectConstantCandi...
Juergen Ributzka [Tue, 25 Mar 2014 21:21:10 +0000 (21:21 +0000)]
[Constant Hoisting] Make the constant candidate map local to the collectConstantCandidates method.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204758 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Correct commutable indices for VSX FMA instructions
Hal Finkel [Tue, 25 Mar 2014 19:26:43 +0000 (19:26 +0000)]
[PowerPC] Correct commutable indices for VSX FMA instructions

Although the first two operands are the ones that can be swapped, the tied
input operand is listed before them, so we need to adjust for that.

I have a test case for this, but it goes along with an upcoming commit (so it
will come soon).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204748 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[PowerPC] Add a TableGen relation for A-type and M-type VSX FMA instructions
Hal Finkel [Tue, 25 Mar 2014 18:55:11 +0000 (18:55 +0000)]
[PowerPC] Add a TableGen relation for A-type and M-type VSX FMA instructions

TableGen will create a lookup table for the A-type FMA instructions providing
their corresponding M-form opcodes. This will be used by upcoming commits.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204746 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoR600: Move computeMaskedBitsForTargetNode out of AMDILISelLowering.cpp
Matt Arsenault [Tue, 25 Mar 2014 18:18:27 +0000 (18:18 +0000)]
R600: Move computeMaskedBitsForTargetNode out of AMDILISelLowering.cpp

Remove handling of select_cc, since it makes no sense to be there. This
now does nothing, but I'll be adding some handling of other target nodes
soon.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204743 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoblockfreq: Implement Pass::releaseMemory()
Duncan P. N. Exon Smith [Tue, 25 Mar 2014 18:01:38 +0000 (18:01 +0000)]
blockfreq: Implement Pass::releaseMemory()

Implement Pass::releaseMemory() in BlockFrequencyInfo and
MachineBlockFrequencyInfo.  Just delete the private implementation when
not in use.  Switch to a std::unique_ptr to make the logic more clear.

<rdar://problem/14292693>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204741 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoblockfreq: Use const in MachineBlockFrequencyInfo
Duncan P. N. Exon Smith [Tue, 25 Mar 2014 18:01:32 +0000 (18:01 +0000)]
blockfreq: Use const in MachineBlockFrequencyInfo

<rdar://problem/14292693>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204740 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[X86TTI] Make constant base pointers for getElementPtr opaque.
Juergen Ributzka [Tue, 25 Mar 2014 18:01:25 +0000 (18:01 +0000)]
[X86TTI] Make constant base pointers for getElementPtr opaque.

If getElementPtr uses a constant as base pointer, then make the constant opaque.
This prevents constant folding it with the offset. The offset can usually be
encoded in the load/store instruction itself and the base address doesn't have
to be rematerialized several times.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204739 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[Stackmaps][X86TTI] Fix think-o in getIntImmCost calculation.
Juergen Ributzka [Tue, 25 Mar 2014 18:01:23 +0000 (18:01 +0000)]
[Stackmaps][X86TTI] Fix think-o in getIntImmCost calculation.

The cost for the first four stackmap operands was always TCC_Free.
This is only true for the first two operands. All other operands
are TCC_Free if they are within 64bit.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204738 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[DAG] Keep the opaque constant flag when performing unary constant folding operations.
Juergen Ributzka [Tue, 25 Mar 2014 18:01:20 +0000 (18:01 +0000)]
[DAG] Keep the opaque constant flag when performing unary constant folding operations.

Usually opaque constants shouldn't be folded, unless they are simple unary
operations that don't create new constants. Although this shouldn't drop the
opaque constant flag. This commit fixes this.

Related to <rdar://problem/14774662>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204737 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[X86] Generate VPSHUFB for in-place v16i16 shuffles
Adam Nemet [Tue, 25 Mar 2014 17:47:06 +0000 (17:47 +0000)]
[X86] Generate VPSHUFB for in-place v16i16 shuffles

This used to resort to splitting the 256-bit operation into two 128-bit
shuffles and then recombining the results.

Fixes <rdar://problem/16167303>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204735 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[X86] Factor out new helper getPSHUFB
Adam Nemet [Tue, 25 Mar 2014 17:47:03 +0000 (17:47 +0000)]
[X86] Factor out new helper getPSHUFB

I found three implementations of this.  This splits it out into a new function
and uses it from the three places.

My plan is to add a fourth use when lowering a vector_shuffle:v16i16.

Compared the assembly output of test/CodeGen/X86 before and after.

The only change is due to how the first PSHUFB was generated in
LowerVECTOR_SHUFFLEv8i16.  If the shuffle mask specified undef (i.e. -1), the
old implementation would write -1 * 2 and -1 * 2 + 1 (254 and 255) in the
control mask.  Now we write 0x80.  These are of course interchangeable since
bit 7 decides if a constant zero is written in the result byte.  The other
instances of this code use 0x80 consistently.

Related to <rdar://problem/16167303>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204734 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[InstCombine] Don't fold bitcast into store if it would need addrspacecast
Richard Osborne [Tue, 25 Mar 2014 17:21:41 +0000 (17:21 +0000)]
[InstCombine] Don't fold bitcast into store if it would need addrspacecast

Summary:
Previously the code didn't check if the before and after types for the
store were pointers to different address spaces. This resulted in
instcombine using a bitcast to convert between pointers to different
address spaces, causing an assertion due to the invalid cast.

It is not be appropriate to use addrspacecast this case because it is
not guaranteed to be a no-op cast. Instead bail out and do not do the
transformation.

CC: llvm-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3117

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204733 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoReuse earlier variables to make it clear the types involved in the cast.
Richard Osborne [Tue, 25 Mar 2014 17:21:35 +0000 (17:21 +0000)]
Reuse earlier variables to make it clear the types involved in the cast.

No functionality change.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204732 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoAdd missing slash to make the doxygen output less confusing.
Benjamin Kramer [Tue, 25 Mar 2014 17:20:28 +0000 (17:20 +0000)]
Add missing slash to make the doxygen output less confusing.

PR19187.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204731 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoR600: Add failing testcase for <3 x i32> stores.
Matt Arsenault [Tue, 25 Mar 2014 16:50:55 +0000 (16:50 +0000)]
R600: Add failing testcase for <3 x i32> stores.

This is supposed to have the same store size and alignment as <4 x i32>,
but currently is split into a 64-bit and 32-bit store.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204729 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoScalarEvolution: Compute exit counts for loops with a power-of-2 step.
Benjamin Kramer [Tue, 25 Mar 2014 16:25:12 +0000 (16:25 +0000)]
ScalarEvolution: Compute exit counts for loops with a power-of-2 step.

If we have a loop of the form
for (unsigned n = 0; n != (k & -32); n += 32) {}
then we know that n is always divisible by 32 and the loop must
terminate. Even if we have a condition where the loop counter will
overflow it'll always hold this invariant.

PR19183. Our loop vectorizer creates this pattern and it's also
occasionally formed by loop counters derived from pointers.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204728 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix creating illegal setcc cond codes.
Matt Arsenault [Tue, 25 Mar 2014 16:09:21 +0000 (16:09 +0000)]
Fix creating illegal setcc cond codes.

If GT/UGT or LT/ULT were set to expand, a comparison
with a constant would replace it with the illegal
cond code.

There are several more places later in this function that
will have the same basic problem.

Theoretically R600 should hit this problem for a test,
but for some reason it doesn't.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204727 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[msan] Relax the test some more.
Evgeniy Stepanov [Tue, 25 Mar 2014 14:32:05 +0000 (14:32 +0000)]
[msan] Relax the test some more.

This may or may not fix the bots. R204720 did not.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204721 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[msan] Make some tests less strict.
Evgeniy Stepanov [Tue, 25 Mar 2014 14:15:14 +0000 (14:15 +0000)]
[msan] Make some tests less strict.

This may or may not fix the bots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204720 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix these tests on windows.
Rafael Espindola [Tue, 25 Mar 2014 13:19:03 +0000 (13:19 +0000)]
Fix these tests on windows.

It is impossible to create a hard link to a non existing file, so create a
dummy file, create the link an delete the dummy file.

On windows one cannot remove the current directory, so chdir first.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204719 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[msan] More precise instrumentation of select IR.
Evgeniy Stepanov [Tue, 25 Mar 2014 13:08:34 +0000 (13:08 +0000)]
[msan] More precise instrumentation of select IR.

Some bits of select result may be initialized even if select condition
is not.

https://code.google.com/p/memory-sanitizer/issues/detail?id=50

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204716 91177308-0d34-0410-b5e6-96231b3b80d8

10 years ago[mips] '.set at=$0' should be equivalent to '.set noat'
Daniel Sanders [Tue, 25 Mar 2014 13:01:06 +0000 (13:01 +0000)]
[mips] '.set at=$0' should be equivalent to '.set noat'

Differential Revision: http://llvm-reviews.chandlerc.com/D3171

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204714 91177308-0d34-0410-b5e6-96231b3b80d8

10 years agoFix AVX2 Gather execution domains.
Cameron McInally [Tue, 25 Mar 2014 12:36:38 +0000 (12:36 +0000)]
Fix AVX2 Gather execution domains.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@204713 91177308-0d34-0410-b5e6-96231b3b80d8