Craig Topper [Thu, 7 Jan 2016 06:37:55 +0000 (06:37 +0000)]
[X86] Had hasSideEffects=0 to VBROADCASTI128.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257034
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 7 Jan 2016 06:37:52 +0000 (06:37 +0000)]
[X86] Add OpSize32 to MOVSX32_NOREX instructions to match their other versions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257033
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 7 Jan 2016 05:57:39 +0000 (05:57 +0000)]
[X86] Add hasSideEffects=0 and mayLoad=1 to MOVZX64* instructions. While there remove a superfluous _Q from the instruction names.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257032
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Thu, 7 Jan 2016 05:40:37 +0000 (05:40 +0000)]
llvm/test/CodeGen/X86/statepoint-vector.ll REQUIRES asserts due to a debug option.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257031
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 7 Jan 2016 05:18:49 +0000 (05:18 +0000)]
[X86] STOSQ without a rep prefix doesn't read or write RCX.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257030
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 7 Jan 2016 04:31:35 +0000 (04:31 +0000)]
Undo spurious change made in r256965
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257028
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Thu, 7 Jan 2016 04:20:52 +0000 (04:20 +0000)]
One more attempt at stablizing a test on all platforms.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257026
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Thu, 7 Jan 2016 04:15:31 +0000 (04:15 +0000)]
[Statepoints] Add test cases around vectors and stablize test
Unlike my comment in 257022 said, it turns out we do handle constant vectors in the statepoint lowering, but only because SelectionDAG doesn't actually produce constants for them. Add a couple of tests which show this working.
Also, add a triple to the same test file to hopefully fix a failing bot.
It turns out we do han
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257025
91177308-0d34-0410-b5e6-
96231b3b80d8
Haicheng Wu [Thu, 7 Jan 2016 04:01:02 +0000 (04:01 +0000)]
[AArch64 MachineCombine] Enhance/Add support for general reassociation to reduce the critical path
Allow fadd/fmul to be reassociated in aarch64.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257024
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Thu, 7 Jan 2016 03:32:11 +0000 (03:32 +0000)]
[Statepoints] Initial support for relocating vectors of pointers
Currently, we try to split vectors of pointers back into their component pointer elements during rewrite-statepoints-for-gc. This is less than ideal since presumably the vectorizer chose to vectorize for a reason. :) It's also been a source of bugs - in particular, the relocation logic as currently implemented was recently discovered to be wrong.
The alternate approach is to allow gc.relocates of vector-of-pointer type and update the backend to handle them. That's what this patch tries to do. This won't actually enable vector-of-pointers in practice - there are some RS4GC changes needed - but the lowering is standalone and testable so it makes sense to separate.
Note that there are some known cases around vector constants which this patch does not handle. Once this is in, I'll send another patch with individual fixes and test cases.
Differential Revision: http://reviews.llvm.org/D15632
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257022
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Thu, 7 Jan 2016 03:19:23 +0000 (03:19 +0000)]
[WebAssembly] Add -m:e to the target triple.
This enables ELF-style name mangling, which primarily means using ".L" for
private symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257020
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Thu, 7 Jan 2016 03:14:59 +0000 (03:14 +0000)]
[Linker] Also treat a DIImportedEntity scope DISubprogram as needed.
Follow-up to r257000: DIImportedEntity can reach a DISubprogram via
its entity, but also via its scope. Handle the latter case as well.
PR26037.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257019
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Thu, 7 Jan 2016 02:20:11 +0000 (02:20 +0000)]
[RS4GC] Add an option to suppress vector splitting
At the moment, this is essentially a diangostic option so that I can start collecting failing test cases, but we will eventually migrate to removing the vector splitting code entirely.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257015
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Thu, 7 Jan 2016 01:49:35 +0000 (01:49 +0000)]
[libFuzzer] add a position hint to the dictionary-based mutator
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257013
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Thu, 7 Jan 2016 01:23:49 +0000 (01:23 +0000)]
[ShrinkWrapping] Give up on irreducible CFGs.
We need to know whether or not a given basic block is in a loop for the analysis
to be correct.
Loop information may be incomplete on irreducible CFGs, therefore we may
generate incorrect code if we use it in those situations.
This fixes PR25988.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257012
91177308-0d34-0410-b5e6-
96231b3b80d8
Andrew Wilkins [Thu, 7 Jan 2016 00:18:56 +0000 (00:18 +0000)]
tools/llvm-config: improve shared library support
Summary:
r252532 added support for reporting the monolithic library
when LLVM_BUILD_LLVM_DYLIB is used. This would only be done
if the individual components were not found, and the dynamic
library is found.
This diff extends this as follows:
- If LLVM_LINK_LLVM_DYLIB is set, then prefer the shared
library, even if all component libraries exist.
- Two flags, --link-shared and --link-static are introduced
to provide explicit guidance. If --link-shared is passed
and the shared library does not exist, an error results.
Additionally, changed the expected shared library names from
(e.g.) LLVM-3.8.0 to LLVM-3.8. The former exists only in an
installation (and then only in CMake builds I think?), and not
in the build tree; this breaks usage of llvm-config during
builds, e.g. by llvm-go.
Reviewers: DiamondLovesYou, beanz
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15033
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257003
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Thu, 7 Jan 2016 00:06:27 +0000 (00:06 +0000)]
Always treat DISubprogram reached by DIImportedEntity as needed.
It is illegal to have a null entity in a DIImportedEntity, so
we must link in a DISubprogram metadata node referenced by one,
even if the associated function is not linked in or inlined anywhere.
Fixes PR26037.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@257000
91177308-0d34-0410-b5e6-
96231b3b80d8
Mehdi Amini [Wed, 6 Jan 2016 23:50:22 +0000 (23:50 +0000)]
Fix PR26051: Memcpy optimization should introduce a call to memcpy before the store destination position
This is a conservative fix, I expect Amaury to relax this.
Follow-up for r256923
From: Mehdi Amini <mehdi.amini@apple.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256999
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 23:45:05 +0000 (23:45 +0000)]
rangify; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256998
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 6 Jan 2016 23:24:40 +0000 (23:24 +0000)]
[X86] Determine if target shuffle can contain zero elements
getTargetShuffleMask may return shuffle masks with SM_SentinelZero (-2) values (currently just for PSHUFB but VPERM2X128 as well with this patch). Although some calling functions can make use of this (mainly for shuffle combining), others can not and their inclusion makes shuffle mask comparisons more difficult.
This patch adds a flag to getTargetShuffleMask to indicate if the calling function can't handle SM_SentinelZero; getTargetShuffleMask will then return false if it occurs to make handling much easier.
I've tidied up some uses of getTargetShuffleMask to better indicate what is going on - more could be done but at present I don't have test cases to demonstrate it.
Some upcoming patches will make use of this to both support more uses where SM_SentinelZero is not permitted (e.g. combineShuffleToAddSub), and also will allow us to add INSERTPS support to getTargetShuffleMask as part of better zero handling discussed in D14261.
Differential Revision: http://reviews.llvm.org/D15378
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256992
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Wed, 6 Jan 2016 23:22:38 +0000 (23:22 +0000)]
[Bitcode] Remove superflous compatibility tests
With r256990, bogner introduced comprehensive tests for constant arrays
and vectors. We no longer need the existing ones because they are
redundant.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256991
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Wed, 6 Jan 2016 23:16:37 +0000 (23:16 +0000)]
Bitcode: Move these tests into compatibility.ll
I added a couple of tests in r256982, but vedantk suggested that they
fit better into compatibility.ll, since they could catch format breaks
later on there.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256990
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Wed, 6 Jan 2016 22:55:03 +0000 (22:55 +0000)]
Recommit r256952 "Filtering IR printing for print-after-all/print-before-all"
Fix lit test fail due to outputting an extra line.
Differential Revision: http://reviews.llvm.org/D15776
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256987
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Wed, 6 Jan 2016 22:31:32 +0000 (22:31 +0000)]
Bitcode: Fix reading and writing of ConstantDataVectors of halfs
In r254991 I allowed ConstantDataVectors to contain elements of
HalfTy, but I missed updating the bitcode reader and writer to handle
this, so now we crash if we try to emit bitcode on programs that have
constant vectors of half.
This fixes the issue and adds test coverage for reading and writing
constant sequences in bitcode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256982
91177308-0d34-0410-b5e6-
96231b3b80d8
Nicolai Haehnle [Wed, 6 Jan 2016 22:01:04 +0000 (22:01 +0000)]
AMDGPU/SI: Fix crash when inline assembly is used in a graphics shader
Summary:
This is admittedly something that you could only run into by manually
playing around with shader assembly because the SITypeWriter pass is
skipped for compute.
Reviewers: arsenm, tstellarAMD
Subscribers: arsenm, llvm-commits
Differential Revision: http://reviews.llvm.org/D15902
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256980
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 20:52:21 +0000 (20:52 +0000)]
[LibCallSimplifier] less indenting; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256973
91177308-0d34-0410-b5e6-
96231b3b80d8
Chen Li [Wed, 6 Jan 2016 20:32:05 +0000 (20:32 +0000)]
[SplitLandingPadPredecessors] Create a PHINode for the original landingpad only if it has some uses
Summary: This patch adds a check in SplitLandingPadPredecessors to see if the original landingpad instruction has any uses. If not, we don't need to create a PHINode for it in the joint block since it's gonna be a dead code anyway. The motivation for this patch is that we found a bug that SplitLandingPadPredecessors created a PHINode of token type landingpad, which failed the verifier since PHINode can not be token type. However, the created PHINode will never be used in our code pattern. This patch will workaround this bug, and we might add supports in SplitLandingPadPredecessors to handle token type landingpad with uses in the future.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15835
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256972
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Wed, 6 Jan 2016 19:47:24 +0000 (19:47 +0000)]
Promote aggregate store to memset when possible
Summary: As per title. This will allow the optimizer to pick up on it.
Reviewers: craig.topper, spatel, dexonsmith, Prazek, chandlerc, joker.eph, majnemer
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15923
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256969
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Wed, 6 Jan 2016 19:45:09 +0000 (19:45 +0000)]
Remove useless DEBUG
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256968
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 19:33:12 +0000 (19:33 +0000)]
Consolidate MemRefs handling from BranchFolding and correct latent bug
Move the logic from BranchFolding to use the shared infrastructure for merging MMOs introduced in 256909. This has the effect of making BranchFolding more capable.
In the process, fix a latent bug. The existing handling for merging didn't handle the case where one of the instructions being merged had overflowed and dropped MemRefs. This was a latent bug in the places the code was commoned from, but potentially reachable in BranchFolding.
Once this is in, we're left with a single place to consider implementing MMO unique-ing as proposed in http://reviews.llvm.org/D15230.
Differential Revision: http://reviews.llvm.org/D15913
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256966
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 6 Jan 2016 19:26:30 +0000 (19:26 +0000)]
[WinEH] Remove calculateCatchReturnSuccessorColors
The functionality that calculateCatchReturnSuccessorColors provides was
once non-trivial: it was a computation layered on top of funclet
coloring.
These days, LLVM IR directly encodes what
calculateCatchReturnSuccessorColors computed, obsoleting the need for
it.
No functionality change is intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256965
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 19:23:35 +0000 (19:23 +0000)]
[LibCallSimplifier] use instruction-level fast-math-flags for tan/atan transform
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256964
91177308-0d34-0410-b5e6-
96231b3b80d8
Quentin Colombet [Wed, 6 Jan 2016 19:09:26 +0000 (19:09 +0000)]
[X86] Correctly model TLS calls w.r.t. frame requirements.
TLS calls need the stack frame to be properly set up and this
implies that such calls need ADJUSTSTACK_xxx markers.
Fixes PR25820.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256959
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Wed, 6 Jan 2016 19:05:19 +0000 (19:05 +0000)]
Make WinCOFFObjectWriter.cpp's timestamp writing not use ENABLE_TIMESTAMPS
LLVM_ENABLE_TIMESTAMPS controls if timestamps are embedded into llvm's
binaries. Turning it off is useful for deterministic builds.
r246905 made it so that the define suddenly also controls if the binaries that
the llvm binaries _create_ embed timestamps or not – but this shouldn't be a
configure-time option. r256203/r256204 added a driver option to toggle this on
and off, so this patch now passes this driver option in LLVM_ENABLE_TIMESTAMPS
builds so that if LLVM_ENABLE_TIMESTAMPS is set, the build of LLVM is
deterministic – but the built clang can still write timestamps into other
executables when requested.
This also allows removing some of the test machinery added in r292012 to work
around this problem.
See PR24740 for background.
http://reviews.llvm.org/D15783
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256958
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 18:47:09 +0000 (18:47 +0000)]
refactor divrem8 lowering; NFCI
The code duplication contributed to PR25754:
https://llvm.org/bugs/show_bug.cgi?id=25754
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256957
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Wed, 6 Jan 2016 18:40:11 +0000 (18:40 +0000)]
[ShrinkWrap] Fix FindIDom to only have one kind of failure.
FindIDom() can fail in two different ways - it can either return nullptr or the
block itself, depending on the circumstances. Some users of FindIDom() check
one error condition, while others check the other.
Change it to always return nullptr on failure.
This fixes PR26004.
Differential Revision: http://reviews.llvm.org/D15847
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256955
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Wed, 6 Jan 2016 18:31:44 +0000 (18:31 +0000)]
Revert r256952 due to lit test fails.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256954
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 6 Jan 2016 18:29:35 +0000 (18:29 +0000)]
[WebAssembly] Don't use range-based loop for a list that's being modified
The first instruction in a block is what the rend() iterator points to, so
if it moves, we need to re-evaluate rend() so that we continue to iterate
through the rest of the instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256953
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Wed, 6 Jan 2016 18:20:25 +0000 (18:20 +0000)]
Filtering IR printing for print-after-all/print-before-all
Summary:
This patch implements "-print-funcs" option to support function filtering for IR printing like -print-after-all, -print-before etc.
Examples:
-print-after-all -print-funcs=foo,bar
Reviewers: mcrosier, joker.eph
Subscribers: tejohnson, joker.eph, llvm-commits
Differential Revision: http://reviews.llvm.org/D15776
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256952
91177308-0d34-0410-b5e6-
96231b3b80d8
Weiming Zhao [Wed, 6 Jan 2016 18:18:16 +0000 (18:18 +0000)]
Fix option desc in FunctionAttrs; NFC
Summary: The example in desc should match with actual option name
Reviewers: jmolloy
Differential Revision: http://reviews.llvm.org/D15800
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256951
91177308-0d34-0410-b5e6-
96231b3b80d8
Geoff Berry [Wed, 6 Jan 2016 18:14:26 +0000 (18:14 +0000)]
ScheduleDAGInstrs: Bug fix for missed memory dependency.
Summary:
In buildSchedGraph(), when adding memory dependencies for loads, move
the call to adjustChainDeps() after the call to
addChainDependency(AliasChain) to handle the case where
addChainDependency(AliasChain) ends up not adding a dependency and
instead putting the SU on the RejectMemNodes list. The call to
adjustChainDeps() must be done after the call to addChainDependency() in
order to process the SU added to the RejectMemNodes list to create
memory dependencies for it.
Reviewers: hfinkel, atrick, jonpa, resistor
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D15927
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256950
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 18:10:35 +0000 (18:10 +0000)]
[BasicAA] Extract WriteOnly predicate on parameters [NFC]
Since writeonly is the only missing attribute and special case left for the memset/memcpy family of intrinsics, rearrange the code to make that much more clear.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256949
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Wed, 6 Jan 2016 17:08:56 +0000 (17:08 +0000)]
WebAssembly: add missing expected failures exposed by r256890
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256948
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 6 Jan 2016 16:45:05 +0000 (16:45 +0000)]
[WebAssembly] Add -asm-verbose=false to llc tests.
In general, disabling comments in the output reduces the chances of a
CHECK line accidentally matching a comment instead of its intended text.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256946
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Wed, 6 Jan 2016 16:15:51 +0000 (16:15 +0000)]
WebAssembly: add new expected failures exposed by r256890
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256945
91177308-0d34-0410-b5e6-
96231b3b80d8
Teresa Johnson [Wed, 6 Jan 2016 15:02:40 +0000 (15:02 +0000)]
Add unittest for new CanReplace flag on MDNodes
This adds a unittest for the support added in r256648 to add
a flag that can be used to prevent RAUW on temporary metadata
used as a map key.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256938
91177308-0d34-0410-b5e6-
96231b3b80d8
Krzysztof Parzyszek [Wed, 6 Jan 2016 14:22:22 +0000 (14:22 +0000)]
[Hexagon] Add system instructions for cache manipulation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256936
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Wed, 6 Jan 2016 13:23:52 +0000 (13:23 +0000)]
Revert "GlobalsAA: Take advantage of ArgMemOnly, InaccessibleMemOnly and InaccessibleMemOrArgMemOnly attributes"
Summary:
This reverts commit
5a9e526f29cf8510ab5c3d566fbdcf47ac24e1e9.
As per discussion in D15665
This also add a test case so that regression introduced by that diff are not reintroduced.
Reviewers: vaivaswatha, jmolloy, hfinkel, reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15919
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256932
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthew Simpson [Wed, 6 Jan 2016 12:50:29 +0000 (12:50 +0000)]
[LV] Avoid creating empty reduction entries (NFC)
This patch prevents us from unintentionally creating entries in the reductions
map for PHIs that are not actually reductions. This is currently not an issue
since we bail out if we encounter PHIs other than inductions or reductions.
However the behavior could become problematic as we add support for additional
recurrence types.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256930
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Wed, 6 Jan 2016 09:41:10 +0000 (09:41 +0000)]
PR25754: avoid generating UDIVREM8_ZEXT_HREG nodes with i64 result
Reviewers: spatel, srking
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15331
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256924
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Wed, 6 Jan 2016 09:30:39 +0000 (09:30 +0000)]
Improve load/store to memcpy for aggregate
Summary: It turns out that if we don't try to do it at the store location, we can do it before any operation that alias the load, as long as no operation alias the store.
Reviewers: craig.topper, spatel, dexonsmith, Prazek, chandlerc, joker.eph
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15903
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256923
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 6 Jan 2016 09:08:49 +0000 (09:08 +0000)]
[X86][SSE] There is no zmm addsubpd/addsubps instruction.
Replace the assert in combineShuffleToAddSub with an early out.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256922
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Wed, 6 Jan 2016 08:59:32 +0000 (08:59 +0000)]
[X86][SSE] An empty target shuffle mask is always a failure.
As discussed on D15378, move the mask.empty() tests to after the switch statement and consider any shuffle decode where the extracted target shuffle mask is empty as a failure.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256921
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 6 Jan 2016 06:18:41 +0000 (06:18 +0000)]
[X86] Use PS instead of TB for instructions that have PD/XS/XD variations. Use OpSize32 on an instruction that has an OpSize16 variant.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256918
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 6 Jan 2016 06:18:37 +0000 (06:18 +0000)]
[X86] Fix an incorrect usage of In32BitMode that should have been Not64BitMode.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256917
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 05:53:09 +0000 (05:53 +0000)]
Fix a warning [NFC]
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256916
91177308-0d34-0410-b5e6-
96231b3b80d8
Jordan Rose [Wed, 6 Jan 2016 05:17:12 +0000 (05:17 +0000)]
Add != to YAMLParser's basic_collection_iterator.
...and mark it as merely an input_iterator rather than a forward_iterator,
since it is destructive. And then rewrite == to take advantage of that.
Patch by Alex Denisov!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256913
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Wed, 6 Jan 2016 05:01:34 +0000 (05:01 +0000)]
[SimplifyLibCalls] Teach SimplifyLibCalls about operand bundles
If we replace one call-site with another, be sure to move over any
operand bundles that lingered on the old call-site.
This fixes PR26036.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256912
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 04:53:16 +0000 (04:53 +0000)]
[BasicAA] Remove special casing of memset_pattern16 in favor of generic attribute inference
Most of the properties of memset_pattern16 can be now covered by the generic attributes and inferred by InferFunctionAttrs. The only exceptions are:
- We don't yet have a writeonly attribute for the first argument.
- We don't have an attribute for modeling the access size facts encoded in MemoryLocation.cpp.
Differential Revision: http://reviews.llvm.org/D15879
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256911
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 04:43:03 +0000 (04:43 +0000)]
[BasicAA] Delete dead code related to memset/memcpy/memmove intrinsics [NFCI]
We only need to describe the writeonly property of one of the arguments. All of the rest of the semantics are nicely described by existing attributes in Intrinsics.td.
Differential Revision: http://reviews.llvm.org/D15880
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256910
91177308-0d34-0410-b5e6-
96231b3b80d8
Philip Reames [Wed, 6 Jan 2016 04:39:03 +0000 (04:39 +0000)]
Extract helper function to merge MemoryOperand lists [NFC]
In the discussion on http://reviews.llvm.org/D15730, Andy pointed out we had a utility function for merging MMO lists. Since it turned we actually had two copies and there's another review in progress (http://reviews.llvm.org/D15230) which needs the same, extract it into a utility function and clean up the interfaces to make it easier to use with a MachineInstBuilder.
I introduced a pair here to track size and allocation together. I think we should probably move in the direction of the MachineOperandsRef helper class, but I'm leaving that for further work. I want to get the poison state introduced before I make major changes to the interface.
Differential Revision: http://reviews.llvm.org/D15757
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256909
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Wed, 6 Jan 2016 03:53:36 +0000 (03:53 +0000)]
Delete trailing whitespace; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256908
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Wed, 6 Jan 2016 03:41:30 +0000 (03:41 +0000)]
Delete trailing whitespace; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256906
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Wed, 6 Jan 2016 03:01:10 +0000 (03:01 +0000)]
Do not define NOGDI. Mingw defines LOGFONTW type in wingdi.h and the mingw
version of shlobj.h includes shobjidl.h and the latter uses the LOGFONTW type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256904
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Wed, 6 Jan 2016 02:48:42 +0000 (02:48 +0000)]
Another attempt at fixing the i686-mingw32-RA-on-linux buildbot. I am getting
confused with what version of mingw is actually installed on the buildbot, and
for now I will just assume this is an unknown version which does not ship with
VersionHelpers.h.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256902
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Wed, 6 Jan 2016 02:32:31 +0000 (02:32 +0000)]
Another attempt at fixing the i686-mingw32-RA-on-linux buildbot.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256901
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 6 Jan 2016 02:13:04 +0000 (02:13 +0000)]
[libFuzzer] extend the dictionary mutator to optionally overwrite data with the dict entry
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256900
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Wed, 6 Jan 2016 01:36:45 +0000 (01:36 +0000)]
Hopefully fix a mingw32 buildbot (i686-mingw32-RA-on-linux) which does not have
the VersionHelpers.h header.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256896
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Wed, 6 Jan 2016 01:23:41 +0000 (01:23 +0000)]
More fix to coverage documentation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256895
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Wed, 6 Jan 2016 00:50:06 +0000 (00:50 +0000)]
Fixing PR25717: fatal IO error writing large outputs to console on Windows.
This patch is similar to the Python issue#11395. We need to cap the output
size to 32767 on Windows to work around the size limit of WriteConsole().
Reference: https://bugs.python.org/issue11395
Writing a test for this bug turns out to be harder than I thought. I am
still working on it (see phabricator review D15705).
Differential Revision: http://reviews.llvm.org/D15553
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256892
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 00:45:42 +0000 (00:45 +0000)]
rangify; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256891
91177308-0d34-0410-b5e6-
96231b3b80d8
Dan Gohman [Wed, 6 Jan 2016 00:43:06 +0000 (00:43 +0000)]
[SelectionDAGBuilder] Set NoUnsignedWrap for inbounds gep and load/store offsets.
In an inbounds getelementptr, when an index produces a constant non-negative
offset to add to the base, the add can be assumed to not have unsigned overflow.
This relies on the assumption that addresses can't occupy more than half the
address space, which isn't possible in C because it wouldn't be possible to
represent the difference between the start of the object and one-past-the-end
in a ptrdiff_t.
Setting the NoUnsignedWrap flag is theoretically useful in general, and is
specifically useful to the WebAssembly backend, since it permits stronger
constant offset folding.
Differential Revision: http://reviews.llvm.org/D15544
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256890
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 00:36:59 +0000 (00:36 +0000)]
use std::max ; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256889
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 00:32:15 +0000 (00:32 +0000)]
A (B + C) = A B + A C ; NFCI
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256884
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 00:23:12 +0000 (00:23 +0000)]
fix typo; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256883
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Aizatsky [Wed, 6 Jan 2016 00:21:22 +0000 (00:21 +0000)]
[libfuzzer] print_new_cov_pcs experimental option.
Differential Revision: http://reviews.llvm.org/D15901
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256882
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 6 Jan 2016 00:18:29 +0000 (00:18 +0000)]
fix typos; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256881
91177308-0d34-0410-b5e6-
96231b3b80d8
Kostya Serebryany [Wed, 6 Jan 2016 00:03:35 +0000 (00:03 +0000)]
[libFuzzer] make trace-based fuzzing not crash in presence of threads
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256876
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Tue, 5 Jan 2016 23:59:08 +0000 (23:59 +0000)]
[Statepoints] Check for the "gc-leaf-function" attribute on call sites as well.
Reviewers: sanjoy, reames
Subscribers: sanjoy, llvm-commits
Differential Revision: http://reviews.llvm.org/D15900
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256875
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 5 Jan 2016 20:46:19 +0000 (20:46 +0000)]
[LibCallSimplfier] use instruction-level fast-math-flags for fmin/fmax transforms
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256871
91177308-0d34-0410-b5e6-
96231b3b80d8
Nicolai Haehnle [Tue, 5 Jan 2016 20:42:49 +0000 (20:42 +0000)]
AMDGPU/SI: Do not move scratch resource register on Tonga & Iceland
Due to the SGPR init bug, every program claims to use the same number
of SGPRs anyway, so there's no point in trying to shift those registers
down from their initial spot of reservation.
Add a test that uses VGPR spilling and blocks most SGPRs from being used for
the scratch resource register. Previously, this would run into an assertion.
Differential Revision: http://reviews.llvm.org/D15724
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256870
91177308-0d34-0410-b5e6-
96231b3b80d8
Amaury Sechet [Tue, 5 Jan 2016 20:17:48 +0000 (20:17 +0000)]
Implement load to store => memcpy in MemCpyOpt for aggregates
Summary:
Most of the tool chain is able to optimize scalar and memcpy like operation effisciently while it isn't that good with aggregates. In order to improve the support of aggregate, we try to change aggregate manipulation into either scalar or memcpy like ones whenever possible without loosing informations.
This is one such opportunity.
Reviewers: craig.topper, spatel, dexonsmith, Prazek, chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D15894
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256868
91177308-0d34-0410-b5e6-
96231b3b80d8
Oleg Ranevskyy [Tue, 5 Jan 2016 19:56:12 +0000 (19:56 +0000)]
[Clang/Support/Windows/Unix] Command lines created by clang may exceed the command length limit set by the OS
Summary:
Hi Rafael,
Would you be able to review this patch, please?
(Clang part of the patch is D15832).
When clang runs an external tool, e.g. a linker, it may create a command line that exceeds the length limit.
Clang uses the llvm::sys::argumentsFitWithinSystemLimits function to check if command line length fits the OS
limitation. There are two problems in this function that may cause exceeding of the limit:
1. It ignores the length of the program path in its calculations. On the other hand, clang adds the program
path to the command line when it runs the program.
2. It assumes no space character is inserted after the last argument, which is not true for Windows. The flattenArgs function adds the trailing space for *each* argument. The result of this is that the terminating NULL character is not counted and may be placed beyond the length limit if the command line is exactly 32768 characters long. The WinAPI's CreateProcess does not find the NULL character and fails.
Reviewers: rafael, ygao, probinson
Subscribers: asl, llvm-commits
Differential Revision: http://reviews.llvm.org/D15831
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256866
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Tue, 5 Jan 2016 19:45:54 +0000 (19:45 +0000)]
Correct my last commit (revision 256860).
I forgot to save a small wording improvement before committing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256862
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Tue, 5 Jan 2016 19:40:58 +0000 (19:40 +0000)]
[PlaceSafepoints] Add a test.
Calls of functions with the "gc-leaf-function" attribute shouldn't be turned
into a safepoint.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256860
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 5 Jan 2016 19:09:47 +0000 (19:09 +0000)]
[InstCombine] insert a new shuffle before its uses (PR26015)
Although this solves the test case in PR26015:
https://llvm.org/bugs/show_bug.cgi?id=26015
And may solve PR25999:
https://llvm.org/bugs/show_bug.cgi?id=25999
...I suspect this is not the best solution. I think we want to insert the new shuffle
just ahead of the earliest ExtractElementInst that we're replacing, but I don't know
how that should be implemented.
Differential Revision: http://reviews.llvm.org/D15878
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256857
91177308-0d34-0410-b5e6-
96231b3b80d8
Manuel Jacob [Tue, 5 Jan 2016 19:08:33 +0000 (19:08 +0000)]
Add function for testing string attributes to InvokeInst and CallSite. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256856
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Tue, 5 Jan 2016 17:46:36 +0000 (17:46 +0000)]
[X86] Determine if we have an OpaqueSPAdjustment earlier
We queried hasFP before we hit ExpandISelPseudos. ExpandISelPseudos
manipulated state that hasFP relied on, potentially changing the result
after it has been queried elsewhere.
While I am not aware of any particular bug due to this state of affairs,
it seems best to avoid it entirely by changing the state during DAG
construction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256849
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Zuckerman [Tue, 5 Jan 2016 15:17:39 +0000 (15:17 +0000)]
[AVX512] add PSLLD and PSLLQ Intrinsic
Differential Revision: http://reviews.llvm.org/D15885
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256840
91177308-0d34-0410-b5e6-
96231b3b80d8
MinSeong Kim [Tue, 5 Jan 2016 14:50:15 +0000 (14:50 +0000)]
[MISched] Explanatory error message when machine model is not complete. NFC
When not all instructions have a scheduling class,
the error message now provides a possible solution.
Differential Revision: http://reviews.llvm.org/D15854
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256839
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Tue, 5 Jan 2016 14:35:01 +0000 (14:35 +0000)]
Reverting r256836; it causes a build bot failure: lab.llvm.org:8011/builders/lldb-x86-win7-msvc/builds/14050/steps/build/logs/stdio
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256837
91177308-0d34-0410-b5e6-
96231b3b80d8
Aaron Ballman [Tue, 5 Jan 2016 14:24:01 +0000 (14:24 +0000)]
Enable more strict standards conformance in MSVC for rvalue casting and string literal type conversion to non-const types. Also enables generation of intrinsics for more functions.
Patch by Alexander Riccio
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256836
91177308-0d34-0410-b5e6-
96231b3b80d8
MinSeong Kim [Tue, 5 Jan 2016 12:51:59 +0000 (12:51 +0000)]
[AArch64] Add support for Samsung Exynos-M1
Adds core tuning support for new Samsung Exynos-M1 core (ARMv8-A).
Differential Revision: http://reviews.llvm.org/D15663
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256828
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Tue, 5 Jan 2016 10:25:56 +0000 (10:25 +0000)]
(NFC) Change SubtargetFeatures::ToggleFeature and
SubtargetFeatures::ApplyFeatureFlag to be static, so that
MCSubtargetInfo doesn't need to instantiate SubtargetFeatures
for nothing. Also change the return type to void, as it
wasn't ever used.
This is a partial commit of http://reviews.llvm.org/D15746
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256823
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Tue, 5 Jan 2016 09:40:03 +0000 (09:40 +0000)]
Remove extra whitespace. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256821
91177308-0d34-0410-b5e6-
96231b3b80d8
Junmo Park [Tue, 5 Jan 2016 09:36:47 +0000 (09:36 +0000)]
Remove extra whitespace. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256820
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 5 Jan 2016 09:12:17 +0000 (09:12 +0000)]
[X86][SSE] Merge PerformBLENDICombine into PerformShuffleCombine
PBLEND/BLENDPD/BLENDPS are no different to the other target shuffles and this will make future improvements to the target shuffle combines more straightforward.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256819
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 5 Jan 2016 07:44:14 +0000 (07:44 +0000)]
[X86] Make MOV32ri64 a post-RA pseudo instead of a CodeGenOnly instruction. It was only needed for rematerialization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256818
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 5 Jan 2016 07:44:11 +0000 (07:44 +0000)]
[X86] Add OpSize32 to OR32mrLocked instruction to match the normal OR32mr instruction.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256817
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 5 Jan 2016 07:44:08 +0000 (07:44 +0000)]
[AVX512] Add hasSideEffects=0 to kunpck instructions since they lack a pattern in their instructions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@256816
91177308-0d34-0410-b5e6-
96231b3b80d8