oota-llvm.git
9 years ago[PowerPC] Put PPCVSXCopy into its own source file
Hal Finkel [Sun, 1 Feb 2015 22:01:29 +0000 (22:01 +0000)]
[PowerPC] Put PPCVSXCopy into its own source file

PPCInstrInfo.cpp has ended up containing several small MI-level passes, and
this is making the file harder to read than necessary. Split out
PPCVSXCopy into its own source file. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227771 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] Put PPCVSXFMAMutate into its own source file
Hal Finkel [Sun, 1 Feb 2015 21:51:22 +0000 (21:51 +0000)]
[PowerPC] Put PPCVSXFMAMutate into its own source file

PPCInstrInfo.cpp has ended up containing several small MI-level passes, and
this is making the file harder to read than necessary. Split out
PPCVSXFMAMutate into its own source file. NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227770 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] Remove the PPCVSXCopyCleanup pass
Hal Finkel [Sun, 1 Feb 2015 21:20:58 +0000 (21:20 +0000)]
[PowerPC] Remove the PPCVSXCopyCleanup pass

This MI-level pass was necessary when VSX support was first being developed,
specifically, before the ABI code had been updated to use VSX registers for
arguments (the register assignments did not change, in a physical sense, but
the VSX super-registers are now used). Unfortunately, I never went back and
removed this pass after that was done. I believe this code is now effectively
dead.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227767 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoLoopVectorize: Remove initializer list that blocks MSVC.
Benjamin Kramer [Sun, 1 Feb 2015 21:13:26 +0000 (21:13 +0000)]
LoopVectorize: Remove initializer list that blocks MSVC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227766 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] Add implicit ops to conditional returns in PPCEarlyReturn
Hal Finkel [Sun, 1 Feb 2015 20:16:10 +0000 (20:16 +0000)]
[PowerPC] Add implicit ops to conditional returns in PPCEarlyReturn

When PPCEarlyReturn, it should really copy implicit ops from the old return
instruction to the new one. This currently does not matter much, because we run
PPCEarlyReturn very late in the pipeline (there is nothing to do DCE on
definitions of those registers). However, for completeness, we should do it
anyway.

Noticed by inspection (and there should be no functional change); thus, no
test case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227763 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFoldingSetVectorIterator is just a subset of pointee_iterator, remove it.
Benjamin Kramer [Sun, 1 Feb 2015 19:26:05 +0000 (19:26 +0000)]
FoldingSetVectorIterator is just a subset of pointee_iterator, remove it.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227761 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoValueTracking: Make isSafeToSpeculativelyExecute a little cleaner
David Majnemer [Sun, 1 Feb 2015 19:10:19 +0000 (19:10 +0000)]
ValueTracking: Make isSafeToSpeculativelyExecute a little cleaner

No functional change intended.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227760 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] VSX stores don't also read
Hal Finkel [Sun, 1 Feb 2015 19:07:41 +0000 (19:07 +0000)]
[PowerPC] VSX stores don't also read

The VSX store instructions were also picking up an implicit "may read" from the
default pattern, which was an intrinsic (and we don't currently have a way of
specifying write-only intrinsics).

This was causing MI verification to fail for VSX spill restores.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227759 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] Better scheduling for isel on P7/P8
Hal Finkel [Sun, 1 Feb 2015 17:52:16 +0000 (17:52 +0000)]
[PowerPC] Better scheduling for isel on P7/P8

isel is actually a cracked instruction on the P7/P8, and must start a dispatch
group. The scheduling model should reflect this so that we don't bunch too many
of them together when possible.

Thanks to Bill Schmidt and Pat Haugen for helping to sort this out.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227758 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoInclude cstddef in EquivalenceClasses.h
Adam Nemet [Sun, 1 Feb 2015 17:21:06 +0000 (17:21 +0000)]
Include cstddef in EquivalenceClasses.h

This is to try to appease bots complaining that ptrdiff_t is undefined in
LoopAccessAnalysis.cpp.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227757 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Move LoopAccessAnalysis to its own module
Adam Nemet [Sun, 1 Feb 2015 16:56:15 +0000 (16:56 +0000)]
[LoopVectorize] Move LoopAccessAnalysis to its own module

Other than moving code and adding the boilerplate for the new files, the code
being moved is unchanged.

There are a few global functions that are shared with the rest of the
LoopVectorizer.  I moved these to the new module as well (emitLoopAnalysis,
stripIntegerCast, replaceSymbolicStrideSCEV) along with the Report class used
by emitLoopAnalysis.  There is probably room for further improvement in this
area.

I kept DEBUG_TYPE "loop-vectorize" because it's used as the PassName with
emitOptimizationRemarkAnalysis.  This will obviously have to change.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227756 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Move RuntimePointerCheck under LoopAccessAnalysis
Adam Nemet [Sun, 1 Feb 2015 16:56:11 +0000 (16:56 +0000)]
[LoopVectorize] Move RuntimePointerCheck under LoopAccessAnalysis

This class needs to remain public because it's used by
LoopVectorizationLegality::addRuntimeCheck.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227755 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Pass parameters explicitly to MemoryDepChecker
Adam Nemet [Sun, 1 Feb 2015 16:56:09 +0000 (16:56 +0000)]
[LoopVectorize] Pass parameters explicitly to MemoryDepChecker

Rather than using globals use a structure to pass parameters from the
vectorizer.  This prepares the class to be moved outside the LoopVectorizer.

It's not great how all this is passed through in LoopAccessAnalysis but this
is all expected to change once the class start servicing the Loop Distribution
pass as well where some of these parameters make no sense.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227754 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Make hasVectorInstrinsicScalarOpd inline
Adam Nemet [Sun, 1 Feb 2015 16:56:05 +0000 (16:56 +0000)]
[LoopVectorize] Make hasVectorInstrinsicScalarOpd inline

VectorUtils.h needs to be included in LoopAccessAnalysis.cpp for
getIntrinsicIDForCall but hasVectorInstrinsicScalarOpd is not used by this
module.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227753 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Convert esp-relative movs of function arguments to pushes, step 2
Michael Kuperstein [Sun, 1 Feb 2015 16:56:04 +0000 (16:56 +0000)]
[X86] Convert esp-relative movs of function arguments to pushes, step 2

This moves the transformation introduced in r223757 into a separate MI pass.
This allows it to cover many more cases (not only cases where there must be a
reserved call frame), and perform rudimentary call folding. It still doesn't
have a heuristic, so it is enabled only for optsize/minsize, with stack
alignment <= 8, where it ought to be a fairly clear win.

(Re-commit of r227728)

Differential Revision: http://reviews.llvm.org/D6789

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227752 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Split out LoopAccessAnalysis from LoopVectorizationLegality
Adam Nemet [Sun, 1 Feb 2015 16:56:04 +0000 (16:56 +0000)]
[LoopVectorize] Split out LoopAccessAnalysis from LoopVectorizationLegality

Move the canVectorizeMemory functionality from LoopVectorizationLegality to a
new class LoopAccessAnalysis and forward users.

Currently the collection of the symbolic stride information is kept with
LoopVectorizationLegality and it becomes an input to LoopAccessAnalysis.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227751 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Add accessors for Num{Stores,Loads,PredStores} in AccessAnalysis
Adam Nemet [Sun, 1 Feb 2015 16:56:02 +0000 (16:56 +0000)]
[LoopVectorize] Add accessors for Num{Stores,Loads,PredStores} in AccessAnalysis

These members are moving to LoopAccessAnalysis.  The accessors help to hide
this.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227750 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Rename the Report class to VectorizationReport
Adam Nemet [Sun, 1 Feb 2015 16:56:00 +0000 (16:56 +0000)]
[LoopVectorize] Rename the Report class to VectorizationReport

This class will become public in the new LoopAccessAnalysis header so the name
needs to be more global.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227749 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Factor out duplicated code into Report::emitAnalysis
Adam Nemet [Sun, 1 Feb 2015 16:55:58 +0000 (16:55 +0000)]
[LoopVectorize] Factor out duplicated code into Report::emitAnalysis

The logic in emitAnalysis is duplicated across multiple functions.  This
splits it into a function.  Another use will be added by the patchset.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227748 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[LoopVectorize] Split out RuntimePointerCheck from LoopVectorizationLegality
Adam Nemet [Sun, 1 Feb 2015 16:55:56 +0000 (16:55 +0000)]
[LoopVectorize] Split out RuntimePointerCheck from LoopVectorizationLegality

RuntimePointerCheck will be used through LoopAccessAnalysis in
LoopVectorizationLegality.  Later in the patchset it will become a local class
of LoopAccessAnalysis.

NFC.  This is part of the patchset that splits out the memory dependence logic
from LoopVectorizationLegality into a new class LoopAccessAnalysis.
LoopAccessAnalysis will be used by the new Loop Distribution pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227747 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRevert r227728 due to bad line endings.
Michael Kuperstein [Sun, 1 Feb 2015 16:15:07 +0000 (16:15 +0000)]
Revert r227728 due to bad line endings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227746 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PowerPC] Make r2 allocatable on PPC64/ELF for some leaf functions
Hal Finkel [Sun, 1 Feb 2015 15:03:28 +0000 (15:03 +0000)]
[PowerPC] Make r2 allocatable on PPC64/ELF for some leaf functions

The TOC base pointer is passed in r2, and we normally reserve this register so
that we can depend on it being there. However, for leaf functions, and
specifically those leaf functions that don't do any TOC access of their own
(which is generally due to accessing the constant pool, using TLS, etc.),
we can treat r2 as an ordinary callee-saved register (it must be callee-saved
because, for local direct calls, the linker will not insert any save/restore
code).

The allocation order has been changed slightly for PPC64/ELF systems to put r2
at the end of the list (while leaving it near the beginning for Darwin systems
to prevent unnecessary output changes). While r2 is allocatable, using it still
requires spill/restore traffic, and thus comes at the end of the list.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227745 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix some bashims. More information on https://wiki.ubuntu.com/DashAsBinSh. Reported...
Sylvestre Ledru [Sun, 1 Feb 2015 14:55:43 +0000 (14:55 +0000)]
Fix some bashims. More information on https://wiki.ubuntu.com/DashAsBinSh. Reported initially on https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=772302 & https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=772301

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227744 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Kill FunctionTargetTransformInfo, TTI itself is now
Chandler Carruth [Sun, 1 Feb 2015 14:37:03 +0000 (14:37 +0000)]
[multiversion] Kill FunctionTargetTransformInfo, TTI itself is now
per-function and supports the exact desired interface.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227743 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Remove the function parameter from the unrolling
Chandler Carruth [Sun, 1 Feb 2015 14:31:23 +0000 (14:31 +0000)]
[multiversion] Remove the function parameter from the unrolling
preferences interface on TTI now that all of TTI is per-function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227741 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Switch the TTI queries from TargetMachine to Subtarget
Chandler Carruth [Sun, 1 Feb 2015 14:22:17 +0000 (14:22 +0000)]
[multiversion] Switch the TTI queries from TargetMachine to Subtarget
now that we have a correct and cached subtarget specific to the
function.

Also, finish providing a cached per-function subtarget in the core
LLVMTargetMachine -- that layer hadn't switched over yet.

The only use of the TargetMachine was to re-lookup a subtarget for
a particular function to work around the fact that TTI was immutable.
Now that it is per-function and we haved a cached subtarget, use it.

This still leaves a few interfaces with real warts on them where we were
passing Function objects through the TTI interface. I'll remove these
and clean their usage up in subsequent commits now that this isn't
necessary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227738 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Remove the cached TargetMachine pointer from the
Chandler Carruth [Sun, 1 Feb 2015 14:01:15 +0000 (14:01 +0000)]
[multiversion] Remove the cached TargetMachine pointer from the
intermediate TTI implementation template and instead query up to the
derived class for both the TargetMachine and the TargetLowering.

Most of the derived types had a TLI cached already and there is no need
to store a less precisely typed target machine pointer.

This will in turn make it much cleaner to look up the TLI via
a per-function subtarget instead of the generic subtarget, and it will
pave the way toward pulling the subtarget used for unroll preferences
into the same form once we are *always* using the function to look up
the correct subtarget.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227737 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Remove another place we were "handling" nullptr even
Chandler Carruth [Sun, 1 Feb 2015 13:21:04 +0000 (13:21 +0000)]
[multiversion] Remove another place we were "handling" nullptr even
though it was never a reasonable input.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227736 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Switch all of the targets over to use the
Chandler Carruth [Sun, 1 Feb 2015 13:20:00 +0000 (13:20 +0000)]
[multiversion] Switch all of the targets over to use the
TargetIRAnalysis access path directly rather than implementing getTTI.

This even removes getTTI from the interface. It's more efficient for
each target to just register a precise callback that creates their
specific TTI.

As part of this, all of the targets which are building their subtargets
individually per-function now build their TTI instance with the function
and thus look up the correct subtarget and cache it. NVPTX, R600, and
XCore currently don't leverage this functionality, but its trivial for
them to add it now.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227735 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Remove a false freedom to leave the TargetMachine pointer
Chandler Carruth [Sun, 1 Feb 2015 12:38:24 +0000 (12:38 +0000)]
[multiversion] Remove a false freedom to leave the TargetMachine pointer
null.

For some reason some of the original TTI code supported a null target
machine. This seems to have been legacy, and I made matters worse when
refactoring this code by spreading that pattern further through the
various targets.

The TargetMachine can't actually be null, and it doesn't make sense to
support that use case. I've now consistently removed it and removed all
of the code trying to cope with that situation. This is probably good,
as several targets *didn't* cope with it being null despite the null
default argument in their constructors. =]

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227734 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoEarlyCSE: Replace custom hash mixing with Hashing.h
Benjamin Kramer [Sun, 1 Feb 2015 12:30:59 +0000 (12:30 +0000)]
EarlyCSE: Replace custom hash mixing with Hashing.h

Brings it in line with the other hashes in EarlyCSE.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227733 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Implement the old pass manager's TTI wrapper pass in
Chandler Carruth [Sun, 1 Feb 2015 12:26:09 +0000 (12:26 +0000)]
[multiversion] Implement the old pass manager's TTI wrapper pass in
terms of the new pass manager's TargetIRAnalysis.

Yep, this is one of the nicer bits of the new pass manager's design.
Passes can in many cases operate in a vacuum and so we can just nest
things when convenient. This is particularly convenient here as I can
now consolidate all of the TargetMachine logic on this analysis.

The most important change here is that this pushes the function we need
TTI for all the way into the TargetMachine, and re-creates the TTI
object for each function rather than re-using it for each function.
We're now prepared to teach the targets to produce function-specific TTI
objects with specific subtargets cached, etc.

One piece of feedback I'd love here is whether its worth renaming any of
this stuff. None of the names really seem that awesome to me at this
point, but TargetTransformInfoWrapperPass is particularly ... odd.
TargetIRAnalysisWrapper might make more sense. I would want to do that
rename separately anyways, but let me know what you think.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227731 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[multiversion] Thread a function argument through all the callers of the
Chandler Carruth [Sun, 1 Feb 2015 12:01:35 +0000 (12:01 +0000)]
[multiversion] Thread a function argument through all the callers of the
getTTI method used to get an actual TTI object.

No functionality changed. This just threads the argument and ensures
code like the inliner can correctly look up the callee's TTI rather than
using a fixed one.

The next change will use this to implement per-function subtarget usage
by TTI. The changes after that should eliminate the need for FTTI as that
will have become the default.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227730 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Convert esp-relative movs of function arguments to pushes, step 2
Michael Kuperstein [Sun, 1 Feb 2015 11:44:44 +0000 (11:44 +0000)]
[X86] Convert esp-relative movs of function arguments to pushes, step 2

This moves the transformation introduced in r223757 into a separate MI pass.
This allows it to cover many more cases (not only cases where there must be a
reserved call frame), and perform rudimentary call folding. It still doesn't
have a heuristic, so it is enabled only for optsize/minsize, with stack
alignment <= 8, where it ought to be a fairly clear win.

Differential Revision: http://reviews.llvm.org/D6789

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227728 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Clean up a stale comment that came from a differnt pass when
Chandler Carruth [Sun, 1 Feb 2015 11:35:56 +0000 (11:35 +0000)]
[PM] Clean up a stale comment that came from a differnt pass when
I created this header.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227727 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Port SimplifyCFG to the new pass manager.
Chandler Carruth [Sun, 1 Feb 2015 11:34:21 +0000 (11:34 +0000)]
[PM] Port SimplifyCFG to the new pass manager.

This should be sufficient to replace the initial (minor) function pass
pipeline in Clang with the new pass manager. I'll probably add an (off
by default) flag to do that just to ensure we can get extra testing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227726 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Port EarlyCSE to the new pass manager.
Chandler Carruth [Sun, 1 Feb 2015 10:51:23 +0000 (10:51 +0000)]
[PM] Port EarlyCSE to the new pass manager.

I've added RUN lines both to the basic test for EarlyCSE and the
target-specific test, as this serves as a nice test that the TTI layer
in the new pass manager is in fact working well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227725 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Teach the module-to-function adaptor to not run function passes
Chandler Carruth [Sun, 1 Feb 2015 10:47:25 +0000 (10:47 +0000)]
[PM] Teach the module-to-function adaptor to not run function passes
over declarations.

This is both quite unproductive and causes things to crash, for example
domtree would just assert.

I've added a declaration and a domtree run to the basic high-level tests
for the new pass manager.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227724 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Switch to a ranged based for loop. NFC
Chandler Carruth [Sun, 1 Feb 2015 10:40:21 +0000 (10:40 +0000)]
[PM] Switch to a ranged based for loop. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227723 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Port TTI to the new pass manager, introducing a TargetIRAnalysis to
Chandler Carruth [Sun, 1 Feb 2015 10:11:22 +0000 (10:11 +0000)]
[PM] Port TTI to the new pass manager, introducing a TargetIRAnalysis to
produce it.

This adds a function to the TargetMachine that produces this analysis
via a callback for each function. This in turn faves the way to produce
a *different* TTI per-function with the correct subtarget cached.

I've also done the necessary wiring in the opt tool to thread the target
machine down and make it available to the pass registry so that we can
construct this analysis from a target machine when available.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227721 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Add a few target specific nodes to 'getTargetNodeName'
Craig Topper [Sun, 1 Feb 2015 10:00:37 +0000 (10:00 +0000)]
[X86] Add a few target specific nodes to 'getTargetNodeName'

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227720 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAVX2: Added 2 more tests for gather intrinsics.
Elena Demikhovsky [Sun, 1 Feb 2015 08:52:15 +0000 (08:52 +0000)]
AVX2: Added 2 more tests for gather intrinsics.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227718 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemoved assert that doesn't typecheck and breaks debug MSVC build.
Michael Kuperstein [Sun, 1 Feb 2015 08:46:20 +0000 (08:46 +0000)]
Removed assert that doesn't typecheck and breaks debug MSVC build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227717 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Refactor the analysis registration and pass pipeline parsing to
Chandler Carruth [Sun, 1 Feb 2015 07:40:05 +0000 (07:40 +0000)]
[PM] Refactor the analysis registration and pass pipeline parsing to
live in a class.

While this isn't really significant right now, I need to expose some
state to the pass construction expressions, and making them get
evaluated within a class context is a nice way to collect members that
they may need to access.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227715 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[SeparateConstOffsetFromGEP] skip optnone functions
Jingyue Wu [Sun, 1 Feb 2015 02:34:41 +0000 (02:34 +0000)]
[SeparateConstOffsetFromGEP] skip optnone functions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227705 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[SeparateConstOffsetFromGEP] set PreservesCFG flag
Jingyue Wu [Sun, 1 Feb 2015 02:33:02 +0000 (02:33 +0000)]
[SeparateConstOffsetFromGEP] set PreservesCFG flag

SeparateConstOffsetFromGEP does not change the shape of the control flow graph.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227704 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[NVPTX] Emit .pragma "nounroll" for loops marked with nounroll
Jingyue Wu [Sun, 1 Feb 2015 02:27:45 +0000 (02:27 +0000)]
[NVPTX] Emit .pragma "nounroll" for loops marked with nounroll

Summary:
CUDA driver can unroll loops when jit-compiling PTX. To prevent CUDA
driver from unrolling a loop marked with llvm.loop.unroll.disable is not
unrolled by CUDA driver, we need to emit .pragma "nounroll" at the
header of that loop.

This patch also extracts getting unroll metadata from loop ID metadata
into a shared helper function.

Test Plan: test/CodeGen/NVPTX/nounroll.ll

Reviewers: eliben, meheff, jholewinski

Reviewed By: jholewinski

Subscribers: jholewinski, llvm-commits

Differential Revision: http://reviews.llvm.org/D7041

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227703 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix PR22393. When recursively replacing an aggregate with a smaller
Adrian Prantl [Sun, 1 Feb 2015 00:58:04 +0000 (00:58 +0000)]
Fix PR22393. When recursively replacing an aggregate with a smaller
aggregate or scalar, the debug info needs to refer to the absolute offset
(relative to the entire variable) instead of storing the offset inside
the smaller aggregate.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227702 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd missing tags.
Adrian Prantl [Sun, 1 Feb 2015 00:57:31 +0000 (00:57 +0000)]
Add missing tags.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227701 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CMake] LLVMLTO requires Intrinsics.gen since r227685 introduced llvm/Analysis/Target...
NAKAMURA Takumi [Sun, 1 Feb 2015 00:55:43 +0000 (00:55 +0000)]
[CMake] LLVMLTO requires Intrinsics.gen since r227685 introduced llvm/Analysis/TargetTransformInfo.h.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227700 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[CMake] LLVMTarget requires Intrinsics.gen since r227669 introduced llvm/Analysis...
NAKAMURA Takumi [Sun, 1 Feb 2015 00:55:32 +0000 (00:55 +0000)]
[CMake] LLVMTarget requires Intrinsics.gen since r227669 introduced llvm/Analysis/TargetTransformInfo.h.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227699 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Remove a bunch of stale TTI creation method declarations. I nuked
Chandler Carruth [Sun, 1 Feb 2015 00:22:15 +0000 (00:22 +0000)]
[PM] Remove a bunch of stale TTI creation method declarations. I nuked
their definitions, but forgot to clean up all the declarations which are
in different files.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227698 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix typo
Matt Arsenault [Sat, 31 Jan 2015 23:37:27 +0000 (23:37 +0000)]
Fix typo

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227697 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoR600/SI: Only select cvt_flr/cvt_rpi with no NaNs.
Matt Arsenault [Sat, 31 Jan 2015 21:28:13 +0000 (21:28 +0000)]
R600/SI: Only select cvt_flr/cvt_rpi with no NaNs.

These have different behavior from cvt_i32_f32 on NaN.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227693 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoX86: silence a GCC warning
Saleem Abdulrasool [Sat, 31 Jan 2015 17:56:11 +0000 (17:56 +0000)]
X86: silence a GCC warning

GCC 4.9 gives the following warning:
  warning: enumeral and non-enumeral type in conditional expression
Cast the enumeral value to an integer within the ternary operation.  NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227692 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove unused variable.
Diego Novillo [Sat, 31 Jan 2015 17:17:33 +0000 (17:17 +0000)]
Remove unused variable.

Summary:
This variable is only used inside an assert. This breaks builds with
asserts disabled.

OK for trunk?

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7314

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227691 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemoved a spurious semicolon; NFC
Aaron Ballman [Sat, 31 Jan 2015 15:18:47 +0000 (15:18 +0000)]
Removed a spurious semicolon; NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227690 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemoved SSE lane blend findCommutedOpIndices overrides. NFCI.
Simon Pilgrim [Sat, 31 Jan 2015 15:16:30 +0000 (15:16 +0000)]
Removed SSE lane blend findCommutedOpIndices overrides. NFCI.

The default op indices frmo TargetInstrInfo::findCommutedOpIndices are being commuted so we don't need to do this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227689 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86][SSE] Shuffle mask decode support for zero extend, scalar float/double moves...
Simon Pilgrim [Sat, 31 Jan 2015 14:09:36 +0000 (14:09 +0000)]
[X86][SSE] Shuffle mask decode support for zero extend, scalar float/double moves and integer load instructions

This patch adds shuffle mask decodes for integer zero extends (pmovzx** and movq xmm,xmm) and scalar float/double loads/moves (movss/movsd).

Also adds shuffle mask decodes for integer loads (movd/movq).

Differential Revision: http://reviews.llvm.org/D7228

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227688 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Switch the TargetMachine interface from accepting a pass manager
Chandler Carruth [Sat, 31 Jan 2015 11:17:59 +0000 (11:17 +0000)]
[PM] Switch the TargetMachine interface from accepting a pass manager
base which it adds a single analysis pass to, to instead return the type
erased TargetTransformInfo object constructed for that TargetMachine.

This removes all of the pass variants for TTI. There is now a single TTI
*pass* in the Analysis layer. All of the Analysis <-> Target
communication is through the TTI's type erased interface itself. While
the diff is large here, it is nothing more that code motion to make
types available in a header file for use in a different source file
within each target.

I've tried to keep all the doxygen comments and file boilerplate in line
with this move, but let me know if I missed anything.

With this in place, the next step to making TTI work with the new pass
manager is to introduce a really simple new-style analysis that produces
a TTI object via a callback into this routine on the target machine.
Once we have that, we'll have the building blocks necessary to accept
a function argument as well.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227685 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[asan][mips] Fix MIPS64 Asan mapping
Kumar Sukhani [Sat, 31 Jan 2015 10:43:18 +0000 (10:43 +0000)]
[asan][mips] Fix MIPS64 Asan mapping

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227684 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReplace another std::set in the core of CodeGenRegister, this time with sorted arrays.
Owen Anderson [Sat, 31 Jan 2015 09:13:36 +0000 (09:13 +0000)]
Replace another std::set in the core of CodeGenRegister, this time with sorted arrays.

The hot path through this region of code does lots of batch inserts into sets. By storing them as sorted arrays, we can defer the sorting to the end of the batch, which is dramatically more efficient. This reduces tblgen runtime by 25% on my worst-case target.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227682 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoChange more of the guts of CodeGenRegister's RegUnit tracking to be based on bit...
Owen Anderson [Sat, 31 Jan 2015 07:49:41 +0000 (07:49 +0000)]
Change more of the guts of CodeGenRegister's RegUnit tracking to be based on bit vectors.

This is a continuation of my prior work to move some of the inner workings for CodeGenRegister to use bit vectors when computing about register units. This is highly beneficial to TableGen runtime on targets with large, dense register files. This patch represents a ~40% runtime reduction over and above my earlier improvement on a stress test of this case.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227678 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agollvm-readobj: add a test case for ARM_MOV32(T) base relocation
Saleem Abdulrasool [Sat, 31 Jan 2015 04:46:50 +0000 (04:46 +0000)]
llvm-readobj: add a test case for ARM_MOV32(T) base relocation

Add a trivial binary (int main() { return 0; }) built for Windows on ARM to
ensure that we can correctly identify ARM_MOV32(T) base relocations.  Addresses
post-commit review comments.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227673 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoARM: make a table more readable (NFC)
Saleem Abdulrasool [Sat, 31 Jan 2015 04:12:06 +0000 (04:12 +0000)]
ARM: make a table more readable (NFC)

This adds some comments and splits the flag calculation on type boundaries to
make the table more readable.  Addresses some post-commit review comments to SVN
r227603.  NFC.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227670 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PM] Change the core design of the TTI analysis to use a polymorphic
Chandler Carruth [Sat, 31 Jan 2015 03:43:40 +0000 (03:43 +0000)]
[PM] Change the core design of the TTI analysis to use a polymorphic
type erased interface and a single analysis pass rather than an
extremely complex analysis group.

The end result is that the TTI analysis can contain a type erased
implementation that supports the polymorphic TTI interface. We can build
one from a target-specific implementation or from a dummy one in the IR.

I've also factored all of the code into "mix-in"-able base classes,
including CRTP base classes to facilitate calling back up to the most
specialized form when delegating horizontally across the surface. These
aren't as clean as I would like and I'm planning to work on cleaning
some of this up, but I wanted to start by putting into the right form.

There are a number of reasons for this change, and this particular
design. The first and foremost reason is that an analysis group is
complete overkill, and the chaining delegation strategy was so opaque,
confusing, and high overhead that TTI was suffering greatly for it.
Several of the TTI functions had failed to be implemented in all places
because of the chaining-based delegation making there be no checking of
this. A few other functions were implemented with incorrect delegation.
The message to me was very clear working on this -- the delegation and
analysis group structure was too confusing to be useful here.

The other reason of course is that this is *much* more natural fit for
the new pass manager. This will lay the ground work for a type-erased
per-function info object that can look up the correct subtarget and even
cache it.

Yet another benefit is that this will significantly simplify the
interaction of the pass managers and the TargetMachine. See the future
work below.

The downside of this change is that it is very, very verbose. I'm going
to work to improve that, but it is somewhat an implementation necessity
in C++ to do type erasure. =/ I discussed this design really extensively
with Eric and Hal prior to going down this path, and afterward showed
them the result. No one was really thrilled with it, but there doesn't
seem to be a substantially better alternative. Using a base class and
virtual method dispatch would make the code much shorter, but as
discussed in the update to the programmer's manual and elsewhere,
a polymorphic interface feels like the more principled approach even if
this is perhaps the least compelling example of it. ;]

Ultimately, there is still a lot more to be done here, but this was the
huge chunk that I couldn't really split things out of because this was
the interface change to TTI. I've tried to minimize all the other parts
of this. The follow up work should include at least:

1) Improving the TargetMachine interface by having it directly return
   a TTI object. Because we have a non-pass object with value semantics
   and an internal type erasure mechanism, we can narrow the interface
   of the TargetMachine to *just* do what we need: build and return
   a TTI object that we can then insert into the pass pipeline.
2) Make the TTI object be fully specialized for a particular function.
   This will include splitting off a minimal form of it which is
   sufficient for the inliner and the old pass manager.
3) Add a new pass manager analysis which produces TTI objects from the
   target machine for each function. This may actually be done as part
   of #2 in order to use the new analysis to implement #2.
4) Work on narrowing the API between TTI and the targets so that it is
   easier to understand and less verbose to type erase.
5) Work on narrowing the API between TTI and its clients so that it is
   easier to understand and less verbose to forward.
6) Try to improve the CRTP-based delegation. I feel like this code is
   just a bit messy and exacerbating the complexity of implementing
   the TTI in each target.

Many thanks to Eric and Hal for their help here. I ended up blocked on
this somewhat more abruptly than I expected, and so I appreciate getting
it sorted out very quickly.

Differential Revision: http://reviews.llvm.org/D7293

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227669 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoARM: support stack probe size on Windows on ARM
Saleem Abdulrasool [Sat, 31 Jan 2015 02:26:37 +0000 (02:26 +0000)]
ARM: support stack probe size on Windows on ARM

Now that -mstack-probe-size is piped through to the backend via the function
attribute as on Windows x86, honour the value to permit handling of non-default
values for stack probes.  This is needed /Gs with the clang-cl driver or
-mstack-probe-size with the clang driver when targeting Windows on ARM.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227667 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[fuzzer] add flags to run fuzzer in multiple parallel processes
Kostya Serebryany [Sat, 31 Jan 2015 01:14:40 +0000 (01:14 +0000)]
[fuzzer] add flags to run fuzzer in multiple parallel processes

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227664 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd the -section option to llvm-objdump used with -macho that takes the argument
Kevin Enderby [Sat, 31 Jan 2015 00:37:11 +0000 (00:37 +0000)]
Add the -section option to llvm-objdump used with -macho that takes the argument
segname,sectname to specify a Mach-O section to print.  The printing is based on
the section type or section attributes.

The printing of the module initialization and termination section types is printed
with this change.  Printing of other section types will be added next.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227649 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove the last vestiges of resetOperationActions.
Eric Christopher [Sat, 31 Jan 2015 00:21:17 +0000 (00:21 +0000)]
Remove the last vestiges of resetOperationActions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227648 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReuse a bunch of cached subtargets and remove getSubtarget calls
Eric Christopher [Sat, 31 Jan 2015 00:06:45 +0000 (00:06 +0000)]
Reuse a bunch of cached subtargets and remove getSubtarget calls
without a Function argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227647 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd PPC test for r227481, but XFAIL because this is actually more work than it appear...
David Blaikie [Fri, 30 Jan 2015 23:52:19 +0000 (23:52 +0000)]
Add PPC test for r227481, but XFAIL because this is actually more work than it appeared to be.

Same sort of bug as on ARM where the cmp+branch are lowered to br_cc
(choosing the branch's debugloc for the br_cc's debugloc) then expanded
out to a cmp and a br, but both using the debug loc of the br_cc, thus
losing fidelity.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227645 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReuse a bunch of cached subtargets and remove getSubtarget calls
Eric Christopher [Fri, 30 Jan 2015 23:46:43 +0000 (23:46 +0000)]
Reuse a bunch of cached subtargets and remove getSubtarget calls
without a Function argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227644 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAvoid using the cast and use the templated accessor function.
Eric Christopher [Fri, 30 Jan 2015 23:46:40 +0000 (23:46 +0000)]
Avoid using the cast and use the templated accessor function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227643 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[AArch64] Add a few more DUP testcases. NFC.
Ahmed Bougacha [Fri, 30 Jan 2015 23:41:15 +0000 (23:41 +0000)]
[AArch64] Add a few more DUP testcases.  NFC.

Also, don't lie about testing index 0.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227642 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFactor out statepoint verification into separate function. (NFC)
Philip Reames [Fri, 30 Jan 2015 23:28:05 +0000 (23:28 +0000)]
Factor out statepoint verification into separate function. (NFC)

Patch by: Igor Laevsky

"Simple refactoring. This is done in preparation to support verification of invokable statepoints."

Differential Revision: http://reviews.llvm.org/D7276

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227640 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[fuzzer] Add a gtest-style test
Kostya Serebryany [Fri, 30 Jan 2015 23:26:57 +0000 (23:26 +0000)]
[fuzzer] Add a gtest-style test

Summary: Add one gtest-style test.

Test Plan: run on bot

Reviewers: samsonov

Reviewed By: samsonov

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7287

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227639 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoReuse a bunch of cached subtargets and remove getSubtarget calls
Eric Christopher [Fri, 30 Jan 2015 23:24:40 +0000 (23:24 +0000)]
Reuse a bunch of cached subtargets and remove getSubtarget calls
without a Function argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227638 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix statepoint verifier tests to actually test verifier.
Philip Reames [Fri, 30 Jan 2015 23:18:42 +0000 (23:18 +0000)]
Fix statepoint verifier tests to actually test verifier.

Patch by: Igor Laevsky

"Statepoint verifier tests were using wrong names for the statepoint and gc.relocate intrinsics. This change renames them to use correct names and fixes all uncovered issues."

Differential Revision: http://reviews.llvm.org/D7266

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227636 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[AArch64] Robustize neon-scalar-copy.ll tests. NFC.
Ahmed Bougacha [Fri, 30 Jan 2015 23:13:57 +0000 (23:13 +0000)]
[AArch64] Robustize neon-scalar-copy.ll tests. NFC.

Some of those didn't even have run lines: they were removed
inadvertently during the Great Merge of 2014.

They used to check for DUPs, but now we go through W-regs?
Filed PR22418 for that potential regression.

For now, just make the tests explicit, so we now where we stand.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227635 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd ARM test for r227489, but XFAIL because this is actually more work than it appear...
David Blaikie [Fri, 30 Jan 2015 23:04:39 +0000 (23:04 +0000)]
Add ARM test for r227489, but XFAIL because this is actually more work than it appeared to be.

Also revert r227489 since it didn't actually fix the thing I thought I
was fixing (since the test case was targeting the wrong architecture
initially). The change might be correct & demonstrated by other test
cases, but it's not a priority for me to find those test cases right
now.

Filed PR22417 for the failure.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227632 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[PBQP] Fix transposed worst row/column check in handleAdd/RemoveNode in the PBQP
Lang Hames [Fri, 30 Jan 2015 22:28:49 +0000 (22:28 +0000)]
[PBQP] Fix transposed worst row/column check in handleAdd/RemoveNode in the PBQP
allocator.

Patch by Jonas Paulsson. Thanks Jonas!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227628 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoNFC. Making printOptionValues an API on the parser class.
Chris Bieneman [Fri, 30 Jan 2015 22:16:01 +0000 (22:16 +0000)]
NFC. Making printOptionValues an API on the parser class.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227626 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoFix memory leak in WinEHPrepare introduced in r227405.
Alexey Samsonov [Fri, 30 Jan 2015 22:07:05 +0000 (22:07 +0000)]
Fix memory leak in WinEHPrepare introduced in r227405.

This leak was detected by ASan bootstrap of LLVM.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227625 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove unused function.
Eric Christopher [Fri, 30 Jan 2015 22:02:36 +0000 (22:02 +0000)]
Remove unused function.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227624 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove extraneous forward declaration.
Eric Christopher [Fri, 30 Jan 2015 22:02:34 +0000 (22:02 +0000)]
Remove extraneous forward declaration.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227623 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoUse the cached subtargets and remove calls to getSubtarget/getSubtargetImpl
Eric Christopher [Fri, 30 Jan 2015 22:02:31 +0000 (22:02 +0000)]
Use the cached subtargets and remove calls to getSubtarget/getSubtargetImpl
without a Function argument.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227622 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd a similar templated cast for getSubtarget off of the MachineFunction
Eric Christopher [Fri, 30 Jan 2015 22:02:19 +0000 (22:02 +0000)]
Add a similar templated cast for getSubtarget off of the MachineFunction
to save typing a lot of static_casts.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227621 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoAdd one more vim swap file pattern
Michael Liao [Fri, 30 Jan 2015 21:59:28 +0000 (21:59 +0000)]
Add one more vim swap file pattern

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227620 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding vector shift instructions and tests.
Colin LeMahieu [Fri, 30 Jan 2015 21:58:46 +0000 (21:58 +0000)]
[Hexagon] Adding vector shift instructions and tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227619 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoR600/SI: Handle SI_SPILL_V96_RESTORE in SIRegisterInfo::eliminateFrameIndex()
Tom Stellard [Fri, 30 Jan 2015 21:51:51 +0000 (21:51 +0000)]
R600/SI: Handle SI_SPILL_V96_RESTORE in SIRegisterInfo::eliminateFrameIndex()

This fixes a crash in Unigine Heaven.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227618 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[X86] Cleanup tabs in test vector-zext.ll. NFC.
Ahmed Bougacha [Fri, 30 Jan 2015 21:41:28 +0000 (21:41 +0000)]
[X86] Cleanup tabs in test vector-zext.ll. NFC.

Some tests have tabs, some don't.
In vector-[sz]ext.ll, space wins (well duh!).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227615 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoSilence "not all paths return a value" warning in MSVC
Reid Kleckner [Fri, 30 Jan 2015 21:30:57 +0000 (21:30 +0000)]
Silence "not all paths return a value" warning in MSVC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227614 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding vector predicate instructions.
Colin LeMahieu [Fri, 30 Jan 2015 21:24:06 +0000 (21:24 +0000)]
[Hexagon] Adding vector predicate instructions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227613 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding vector permutation instructions and tests.
Colin LeMahieu [Fri, 30 Jan 2015 21:14:00 +0000 (21:14 +0000)]
[Hexagon] Adding vector permutation instructions and tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227612 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoWin64: Put a REX_W prefix on all TAILJMP* instructions
Reid Kleckner [Fri, 30 Jan 2015 21:03:31 +0000 (21:03 +0000)]
Win64: Put a REX_W prefix on all TAILJMP* instructions

MSDN's x64 software conventions page says that this is one of the fixed
list of legal epilogues:
https://msdn.microsoft.com/en-us/library/tawsa7cb.aspx

Presumably this is how the unwinder distinguishes epilogue jumps from
in-function control flow.

Also normalize the way we place "## TAILCALL" comments on such jumps.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227611 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding vector multiplies. Cleaning up tests.
Colin LeMahieu [Fri, 30 Jan 2015 20:56:54 +0000 (20:56 +0000)]
[Hexagon] Adding vector multiplies.  Cleaning up tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227609 91177308-0d34-0410-b5e6-96231b3b80d8

9 years agoRemove the preverify pass from the documentation now that it has been removed
Yunzhong Gao [Fri, 30 Jan 2015 20:51:09 +0000 (20:51 +0000)]
Remove the preverify pass from the documentation now that it has been removed
since r199487.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227608 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[Hexagon] Adding XTYPE/COMPLEX instructions and cleaning up tests.
Colin LeMahieu [Fri, 30 Jan 2015 20:08:37 +0000 (20:08 +0000)]
[Hexagon] Adding XTYPE/COMPLEX instructions and cleaning up tests.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227607 91177308-0d34-0410-b5e6-96231b3b80d8

9 years ago[AArch64] Make AArch64A57FPLoadBalancing output stable.
Chad Rosier [Fri, 30 Jan 2015 19:55:40 +0000 (19:55 +0000)]
[AArch64] Make AArch64A57FPLoadBalancing output stable.

Add tie breaker to colorChainSet() sort so that processing order doesn't
depend on std::set order, which depends on pointer order, which is
unstable from run to run.

No test case as this is nearly impossible to reproduce.

Phabricator Review: http://reviews.llvm.org/D7265
Patch by Geoff Berry <gberry@codeaurora.org>!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@227606 91177308-0d34-0410-b5e6-96231b3b80d8