OSDN Git Service
Matt Arsenault [Fri, 1 Jul 2016 02:16:24 +0000 (02:16 +0000)]
LoadStoreVectorizer: improvements: better pointer analysis
If OpB has an ADD NSW/NUW, we can use that to prove that adding 1
to OpA won't wrap if OpA + 1 == OpB.
Patch by Fiona Glaser
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274324
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 02:09:38 +0000 (02:09 +0000)]
LoadStoreVectorizer: Don't increase alignment with no align set
If no alignment was set on the load/stores, it would vectorize
to the new type even though this increases the default alignment.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274323
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 02:07:22 +0000 (02:07 +0000)]
LoadStoreVectorizer: Check TTI for vec reg bit width
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274322
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 01:55:52 +0000 (01:55 +0000)]
LoadStoreVectorizer: Fix assert when merging pointer ops
This needs to use inttoptr/ptrtoint if combining an int and pointer
load. If a pointer is used always do an integer load.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274321
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 01:51:40 +0000 (01:51 +0000)]
Revert "code hoisting pass based on GVN"
This reverts commit r274305, since it breaks self-hosting:
http://lab.llvm.org:8080/green/job/clang-stage1-configure-RA_build/22349/
http://lab.llvm.org:8011/builders/clang-x86_64-linux-selfhost-modules/builds/17232
Note that the blamelist on lab.llvm.org:8011 is incorrect. The previous
build was r274299, but somehow r274305 wasn't included in the blamelist:
http://lab.llvm.org:8011/builders/clang-x86_64-linux-selfhost-modules
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274320
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 01:51:32 +0000 (01:51 +0000)]
CodeGen: Use MachineInstr& in LiveVariables API, NFC
Change all the methods in LiveVariables that expect non-null
MachineInstr* to take MachineInstr& and update the call sites. This
clarifies the API, and designs away a class of iterator to pointer
implicit conversions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274319
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 01:47:46 +0000 (01:47 +0000)]
LoadStoreVectorizer: Use AA metadata
This was not passing the full instruction with metadata
to the alias query.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274318
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 01:27:19 +0000 (01:27 +0000)]
CodeGen: Remove implicit iterator conversions in PHIElimination, NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274317
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 01:18:53 +0000 (01:18 +0000)]
CodeGen: Use MachineInstr& in PostRASchedulerList, NFC
Remove another unnecessary iterator to pointer conversion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274315
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Fri, 1 Jul 2016 01:16:17 +0000 (01:16 +0000)]
Define a module map entry for DebugInfo/CodeView.
This fixes the -fmodules build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274313
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 00:56:27 +0000 (00:56 +0000)]
AMDGPU: Implement getLoadStoreVecRegBitWidth
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274312
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 00:50:29 +0000 (00:50 +0000)]
CodeGen: Use MachineInstr& in PostRAHazardRecognizer, NFC
Convert a loop to a range-based for, using MachineInstr& instead of
MachineInstr* and removing an implicit conversion from iterator to
pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274311
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 00:40:57 +0000 (00:40 +0000)]
CodeGen: Use MachineInstr& in PrologEpilogInserter, NFC
Use MachineInstr& over MachineInstr* to avoid implicit iterator to
pointer conversions. MachineInstr*-as-nullptr was being used as a flag
for whether the for loop terminated normally; I added an explicit `bool`
instead.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274310
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Fri, 1 Jul 2016 00:37:49 +0000 (00:37 +0000)]
[pdb] Avoid reporting an error when the module symbol stream is empty
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274309
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Fri, 1 Jul 2016 00:37:25 +0000 (00:37 +0000)]
[PDB] Indicate which type record failed hash validation
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274308
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 00:37:01 +0000 (00:37 +0000)]
LoadStoreVectorizer: if one element of a vector is integer, default to
integer.
Fixes issues on some architectures where we use arithmetic ops to build
vectors, which can cause bad things to happen for loads/stores of mixed
types.
Patch by Fiona Glaser
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274307
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Fri, 1 Jul 2016 00:36:54 +0000 (00:36 +0000)]
LoadStoreVectorizer: Fix crashes on sub-byte types
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274306
91177308-0d34-0410-b5e6-
96231b3b80d8
Sebastian Pop [Fri, 1 Jul 2016 00:24:31 +0000 (00:24 +0000)]
code hoisting pass based on GVN
This pass hoists duplicated computations in the program. The primary goal of
gvn-hoist is to reduce the size of functions before inline heuristics to reduce
the total cost of function inlining.
Pass written by Sebastian Pop, Aditya Kumar, Xiaoyu Hu, and Brian Rzycki.
Important algorithmic contributions by Daniel Berlin under the form of reviews.
Differential Revision: http://reviews.llvm.org/D19338
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274305
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 00:23:27 +0000 (00:23 +0000)]
Target: Remove unused arguments from overrideSchedPolicy, NFC
TargetSubtargetInfo::overrideSchedPolicy takes two MachineInstr*
arguments (begin and end) that invite implicit conversions from
MachineInstrBundleIterator. One option would be to change their type to
an iterator, but since they don't seem to have been used since the API
was added in 2010, I'm deleting the dead code.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274304
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 00:11:48 +0000 (00:11 +0000)]
CodeGen: Use MachineInstr& in MachineSink, NFC
Use MachineInstr& instead of MachineInstr* in MachineSinker to help
avoid implicit conversions from iterator to pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274303
91177308-0d34-0410-b5e6-
96231b3b80d8
Adam Nemet [Fri, 1 Jul 2016 00:09:02 +0000 (00:09 +0000)]
[LAA] Fix alphabetical sorting of headers. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274302
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Fri, 1 Jul 2016 00:05:40 +0000 (00:05 +0000)]
CodeGen: Use MachineInstr& more in MachineTraceMetrics, NFC
Push MachineInstr& through helper APIs for consistency. This doesn't
remove any more implicit conversions, but it's a nice cleanup after
r274300.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274301
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:53:20 +0000 (23:53 +0000)]
CodeGen: Use MachineInstr& in MachineTraceMetrics, NFC
This avoids an implicit conversion from iterator to pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274300
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 23:50:18 +0000 (23:50 +0000)]
LoadStoreVectorizer: Check skipFunction first.
Also add test I forgot to add to r274296.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274299
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:39:46 +0000 (23:39 +0000)]
CodeGen: Use MachineInstr& in LocalStackSlotAllocation, NFC
Avoid a number of implicit conversions from iterator to pointer by using
range-based for and MachineInstr&.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274298
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:33:35 +0000 (23:33 +0000)]
CodeGen: Use range-based for in LiveVariables, NFC
Avoid an implicit iterator to pointer conversion in
LiveVariables::runOnBlock by switching to a range-based for.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274297
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 23:30:29 +0000 (23:30 +0000)]
LoadStoreVectorizer: Skip optnone functions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274296
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:28:15 +0000 (23:28 +0000)]
CodeGen: Use MachineInstr& in HoistSpillHelper, NFC
Avoid another few implicit conversions from iterator to pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274295
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:13:38 +0000 (23:13 +0000)]
CodeGen: Use MachineInstr& in LDVImpl::handleDebugValue, NFC
Avoid another implicit conversion from iterator to pointer.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274294
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 23:11:38 +0000 (23:11 +0000)]
Add LoadStoreVectorizer pass
This was contributed by Apple, and I've been working on
minimal cleanups and generalizing it.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274293
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:09:39 +0000 (23:09 +0000)]
CodeGen: Use MachineInstr& in ExpandISelPseudos, NFC
Avoid another implicit conversion from MachineInstrBundleIterator to
MachineInstr* by using MachineInstr&.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274292
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 23:04:51 +0000 (23:04 +0000)]
CodeGen: Use MachineInstr& in IfConversion, NFC
Switch to a range-based for in IfConverter::PredicateBlock and take
MachineInstr& in MaySpeculate to avoid an implicit conversion from
MachineBasicBlock::iterator to MachineInstr*.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274290
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 30 Jun 2016 23:04:20 +0000 (23:04 +0000)]
Remove non-ASCII characters (silly smart quotes). Thanks to Hal for
noticing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274289
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Thu, 30 Jun 2016 23:04:07 +0000 (23:04 +0000)]
[TableGen] Use a SmallVector for Record::Values to avoid debug iterators
Debug iterators are valuable so we don't want to turn them off
completely. However, llvm-tblgen is critical to build speed, so we can
skip them here.
Regenerating X86GenSubtargetInfo.inc in a clang-cl self-host debug build
now takes 39s instead of 1m29s.
Helps PR28222
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274288
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 22:52:52 +0000 (22:52 +0000)]
CodeGen: Use MachineInstr& in TargetLowering, NFC
This is a mechanical change to make TargetLowering API take MachineInstr&
(instead of MachineInstr*), since the argument is expected to be a valid
MachineInstr. In one case, changed a parameter from MachineInstr* to
MachineBasicBlock::iterator, since it was used as an insertion point.
As a side effect, this removes a bunch of MachineInstr* to
MachineBasicBlock::iterator implicit conversions, a necessary step
toward fixing PR26753.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274287
91177308-0d34-0410-b5e6-
96231b3b80d8
David L Kreitzer [Thu, 30 Jun 2016 21:43:11 +0000 (21:43 +0000)]
Test commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274284
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 21:18:06 +0000 (21:18 +0000)]
Don't use unchecked dyn_cast
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274282
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 21:17:59 +0000 (21:17 +0000)]
SLPVectorizer: Move propagateMetadata to VectorUtils
This will be re-used by the LoadStoreVectorizer.
Fix handling of range metadata and testcase by Justin Lebar.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274281
91177308-0d34-0410-b5e6-
96231b3b80d8
Matt Arsenault [Thu, 30 Jun 2016 20:49:28 +0000 (20:49 +0000)]
AMDGPU: Add m0 vgpr load loop block as successor
This shows up as a verifier error when I move this
earlier, not sure why it didn't before.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274275
91177308-0d34-0410-b5e6-
96231b3b80d8
Mike Aizatsky [Thu, 30 Jun 2016 20:43:06 +0000 (20:43 +0000)]
[libFuzzer] Let user specify extra stats file.
Summary: If AFL_DRIVER_EXTRA_STATS_FILENAME is set and valid, write to it peak_rss_mb and slowest_unit_time_sec. These are both stats that libFuzzer can print but afl cannot.
Reviewers: kcc, aizatsky, metzman
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D21742
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274273
91177308-0d34-0410-b5e6-
96231b3b80d8
Vassil Vassilev [Thu, 30 Jun 2016 20:35:56 +0000 (20:35 +0000)]
[CMake] Module builds depend on target intrinsics_gen to be built first.
When compiling with modules, header A and B can be in the same module M.
B depends on intrinsics_gen and A doesn't. Compiling a source file #include-ing
header A, we implicitly request module M to be built. It puts header A and B in
the same TU and tries to build them. Since B depends on intrinsics_gen (which
might not be built yet) we run into build failures.
This should fix our modules buildbot.
Patch reviewed by Chris Bieneman.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274270
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 30 Jun 2016 20:27:54 +0000 (20:27 +0000)]
Introduce a *draft* of a code of conduct for the LLVM community and the
associated reporting guide.
I want to emphasize that at this point these are just drafts!
This is the result of very extended discussion on the mailing lists on
several different threads:
http://lists.llvm.org/pipermail/llvm-dev/2015-October/091218.html
http://lists.llvm.org/pipermail/llvm-dev/2016-May/099120.html
http://lists.llvm.org/pipermail/llvm-commits/Week-of-Mon-
20151019/307070.html
The reporting guide in particular I anticipate will be shaped somewhat
by the advisory committee when they are selected. But hopefully this
serves as a good starting point and good guidance while the advisory
committee is being sorted out.
I'd like to thank all the folks who contributed to this. Many, *many*
people worked to help with drafting, wording, suggestions, and edits.
Also, this is based on widely used existing codes of coduct as mentioned
in the text, and the original authors of those deserve many thanks as
well.
Differential Revision: http://reviews.llvm.org/D13741
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274268
91177308-0d34-0410-b5e6-
96231b3b80d8
Yunzhong Gao [Thu, 30 Jun 2016 18:49:04 +0000 (18:49 +0000)]
Add an artificial line-0 debug location when the compiler emits a call to
__stack_chk_fail(). This avoids a compiler crash.
Differential Revision: http://reviews.llvm.org/D21818
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274263
91177308-0d34-0410-b5e6-
96231b3b80d8
Wei Mi [Thu, 30 Jun 2016 18:42:56 +0000 (18:42 +0000)]
Refine the set of UniformAfterVectorization instructions.
Except the seed uniform instructions (conditional branch and consecutive ptr
instructions), dependencies to be added into uniform set should only be used
by existing uniform instructions or intructions outside of current loop.
Differential Revision: http://reviews.llvm.org/D21755
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274262
91177308-0d34-0410-b5e6-
96231b3b80d8
Justin Bogner [Thu, 30 Jun 2016 18:32:12 +0000 (18:32 +0000)]
CodeGen: Add the other BuildMI overload for MachineInstr&
The change in r274193 missed this variant.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274259
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 30 Jun 2016 18:25:11 +0000 (18:25 +0000)]
Delete MCCodeGenInfo.
MC doesn't really care about CodeGen stuff, so this was just
complicating target initialization.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274258
91177308-0d34-0410-b5e6-
96231b3b80d8
Etienne Bergeron [Thu, 30 Jun 2016 17:52:24 +0000 (17:52 +0000)]
revert http://reviews.llvm.org/D21101
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274251
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Thu, 30 Jun 2016 17:43:06 +0000 (17:43 +0000)]
[Support] Fix a bug in ErrorList::join / joinErrors.
When concatenating two error lists the ErrorList::join method (which is called
by joinErrors) was failing to set the checked bit on the second error, leading
to a 'failure to check error' assertion.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274249
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Thu, 30 Jun 2016 17:43:00 +0000 (17:43 +0000)]
[pdb] Re-add code to write PDB files.
Somehow all the functionality to write PDB files got removed,
probably accidentally when uploading the patch perhaps the wrong
one got uploaded. This re-adds all the code, as well as the
corresponding test.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274248
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Thu, 30 Jun 2016 17:42:48 +0000 (17:42 +0000)]
Update llvm-pdbdump to use subcommands.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274247
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 30 Jun 2016 17:15:44 +0000 (17:15 +0000)]
[CMake] Add an LLVM_ENABLE_MODULE_DEBUGGING flag for building with -gmodules.
This flag is only effective in builds with debug info and modules.
The default is On for Darwin only.
rdar://problem/
27019000
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274244
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 30 Jun 2016 15:58:36 +0000 (15:58 +0000)]
Revert "[CMake] Move the -Xclang option before -fmodules-cache-path"
This reverts commit
3db82f646a0890eb7664d0351b5a3c79622e8bef.
Vassil already fixed this and I mechanically undid his fix without looking
too close at what I'm actually doing. Need more coffee.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274242
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 30 Jun 2016 15:44:35 +0000 (15:44 +0000)]
[CMake] Move the -Xclang option before -fmodules-cache-path
This fixes a typo introduced in r274196.
Thanks to Vassil Vassilev for noticing!
http://reviews.llvm.org/D21827
rdar://problem/
27019000
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274240
91177308-0d34-0410-b5e6-
96231b3b80d8
Etienne Bergeron [Thu, 30 Jun 2016 15:36:59 +0000 (15:36 +0000)]
[exceptions] Upgrade exception handlers when stack protector is used
Summary:
MSVC provide exception handlers with enhanced information to deal with security buffer feature (/GS).
To be more secure, the security cookies (GS and SEH) are validated when unwinding the stack.
The following code:
```
void f() {}
void foo() {
__try {
f();
} __except(1) {
f();
}
}
```
Reviewers: majnemer, rnk
Subscribers: thakis, llvm-commits, chrisha
Differential Revision: http://reviews.llvm.org/D21101
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274239
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 30 Jun 2016 15:32:45 +0000 (15:32 +0000)]
fix formatting, add TODO; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274238
91177308-0d34-0410-b5e6-
96231b3b80d8
Jun Bum Lim [Thu, 30 Jun 2016 15:32:20 +0000 (15:32 +0000)]
[DSE] Fix bug in partial overwrite tracking
Summary:
Found cases where DSE incorrectly add partially-overwritten intervals.
Please see the test case for details.
Reviewers: mcrosier, eeckstein, hfinkel
Subscribers: mcrosier, llvm-commits
Differential Revision: http://reviews.llvm.org/D21859
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274237
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 30 Jun 2016 14:51:21 +0000 (14:51 +0000)]
[InstCombine] shrink switch conditions better (PR24766)
https://llvm.org/bugs/show_bug.cgi?id=24766#c2
This removes a hack that was added for the benefit of x86 codegen.
It prevented shrinking the switch condition even to smaller legal (DataLayout) types.
We have a safety mechanism in CGP after:
http://reviews.llvm.org/rL251857
...so we're free to use the optimal (smallest) IR type now.
Differential Revision: http://reviews.llvm.org/D12965
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274233
91177308-0d34-0410-b5e6-
96231b3b80d8
Elliot Colp [Thu, 30 Jun 2016 14:42:47 +0000 (14:42 +0000)]
Test commit
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274232
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 30 Jun 2016 14:27:41 +0000 (14:27 +0000)]
[InstCombine] use ConstantExpr::getBitCast() instead of creating useless instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274229
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 30 Jun 2016 14:18:18 +0000 (14:18 +0000)]
[InstCombine] extend matchSelectFromAndOr() to work with i1 scalar types
If the incoming types are i1, then we don't have to pattern match any sext ops.
Differential Revision: http://reviews.llvm.org/D21740
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274228
91177308-0d34-0410-b5e6-
96231b3b80d8
Vassil Vassilev [Thu, 30 Jun 2016 14:17:42 +0000 (14:17 +0000)]
[CMake] -fmodules-local-submodule-visibility is a cc1-only option.
This should fix modules builds on platforms other than Darwin after r274196.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274227
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 30 Jun 2016 12:44:52 +0000 (12:44 +0000)]
Don't repeat names in comments. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274226
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Thu, 30 Jun 2016 12:19:16 +0000 (12:19 +0000)]
Delete unused includes. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274225
91177308-0d34-0410-b5e6-
96231b3b80d8
Jonas Paulsson [Thu, 30 Jun 2016 07:13:56 +0000 (07:13 +0000)]
[SystemZ] Let z13 also support FeatureMiscellaneousExtensions.
This processor feature had been left out by mistake from the z13
ProcessorModel.
This time with updated test case. Thanks, Hans.
Reviewed by Ulrich Weigand.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274216
91177308-0d34-0410-b5e6-
96231b3b80d8
Pankaj Gode [Thu, 30 Jun 2016 06:42:31 +0000 (06:42 +0000)]
[AArch64] Add Broadcom Vulcan scheduling model.
Adding scheduling model for new Broadcom Vulcan core (ARMv8.1A).
Differential Revision: http://reviews.llvm.org/D21728
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274213
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Thu, 30 Jun 2016 04:38:51 +0000 (04:38 +0000)]
Use ShuffleVectorSDNode::isSplat member method instead of static method isSplatMask where the mask came directly from getMask() on a shuffle node.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274208
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Thu, 30 Jun 2016 03:00:20 +0000 (03:00 +0000)]
[CodeView] Implement support for bitfields in LLVM
CodeView need to know the offset of the storage allocation for a
bitfield. Encode this via the "extraData" field in DIDerivedType and
introduced a new flag, DIFlagBitField, to indicate whether or not a
member is a bitfield.
This fixes PR28162.
Differential Revision: http://reviews.llvm.org/D21782
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274200
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Thu, 30 Jun 2016 02:47:28 +0000 (02:47 +0000)]
[SCEV] Compute max be count from shift operator only if all else fails
In particular, check to see if we can compute a precise trip count by
exhaustively simulating the loop first.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274199
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Thu, 30 Jun 2016 02:32:20 +0000 (02:32 +0000)]
[ADT] Add a new data structure for managing a priority worklist where
re-insertion of entries into the worklist moves them to the end.
This is fairly similar to a SetVector, but helps in the case where in
addition to not inserting duplicates you want to adjust the sequence of
a pop-off-the-back worklist.
I'm not at all attached to the name of this data structure if others
have better suggestions, but this is one that David Majnemer brought up
in IRC discussions that seems plausible.
I've trimmed the interface down somewhat from SetVector's interface
because several things make less sense here IMO: iteration primarily.
I'd prefer to add these back as we have users that need them. My use
case doesn't even need all of what is provided here. =]
I've also included a basic unittest to make sure this functions
reasonably.
Differential Revision: http://reviews.llvm.org/D21866
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274198
91177308-0d34-0410-b5e6-
96231b3b80d8
George Burgess IV [Thu, 30 Jun 2016 02:11:26 +0000 (02:11 +0000)]
[CFLAA] Add support for ModRef queries.
This patch makes CFLAA answer some ModRef queries. Because we don't
distinguish between reading/writing when making StratifiedSets, we're
unable to offer any of the readonly-related answers.
Patch by Jia Chen.
Differential Revision: http://reviews.llvm.org/D21858
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274197
91177308-0d34-0410-b5e6-
96231b3b80d8
Adrian Prantl [Thu, 30 Jun 2016 01:46:49 +0000 (01:46 +0000)]
[CMake] Introduce a LLVM_ENABLE_LOCAL_SUBMODULE_VISIBILITY flag.
On Darwin it is currently impossible to build LLVM with modules
because the Darwin system module map is not compatible with
-fmodules-local-submodule-visibility at this point in time. This
patch makes the flag optional and off by default on Darwin so it
becomes possible to build LLVM with modules again.
http://reviews.llvm.org/D21827
rdar://problem/
27019000
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274196
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Thu, 30 Jun 2016 00:23:54 +0000 (00:23 +0000)]
RegisterScavenging: Code cleanup; NFC
- Use range based for loops
- No need for some !Reg checks: isPhysicalRegister() reports false for
NoRegister anyway
- Do not repeat function name in documentation comment.
- Do not repeat documentation comment in implementation when we already
have one at the declaration.
- Factor some common subexpressions out.
- Change file comments to use doxygen syntax.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274194
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 00:10:17 +0000 (00:10 +0000)]
CodeGen: Add an explicit BuildMI overload for MachineInstr&
Add an explicit overload to BuildMI for MachineInstr& to deal with
insertions inside of instruction bundles.
- Use it to re-implement MachineInstr* to give it coverage.
- Document how the overload for MachineBasicBlock::instr_iterator
differs from that for MachineBasicBlock::iterator (the previous
(implicit) overload for MachineInstr&).
- Add a comment explaining why the MachineInstr& and MachineInstr*
overloads don't universally forward to the
MachineBasicBlock::instr_iterator overload.
Thanks to Justin for noticing the API quirk. While this doesn't fix any
known bugs -- all uses of BuildMI with a MachineInstr& were previously
using MachineBasicBlock::iterator -- it protects against future bugs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274193
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Thu, 30 Jun 2016 00:09:13 +0000 (00:09 +0000)]
add vector tests to show missing transform
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274192
91177308-0d34-0410-b5e6-
96231b3b80d8
Marcin Koscielnicki [Thu, 30 Jun 2016 00:08:54 +0000 (00:08 +0000)]
[SystemZ] Split up PerformDAGCombine. [NFC]
This function is already a bit too long, and I'm about to make it worse.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274191
91177308-0d34-0410-b5e6-
96231b3b80d8
Duncan P. N. Exon Smith [Thu, 30 Jun 2016 00:01:54 +0000 (00:01 +0000)]
CodeGen: Use MachineInstr& in TargetInstrInfo, NFC
This is mostly a mechanical change to make TargetInstrInfo API take
MachineInstr& (instead of MachineInstr* or MachineBasicBlock::iterator)
when the argument is expected to be a valid MachineInstr. This is a
general API improvement.
Although it would be possible to do this one function at a time, that
would demand a quadratic amount of churn since many of these functions
call each other. Instead I've done everything as a block and just
updated what was necessary.
This is mostly mechanical fixes: adding and removing `*` and `&`
operators. The only non-mechanical change is to split
ARMBaseInstrInfo::getOperandLatencyImpl out from
ARMBaseInstrInfo::getOperandLatency. Previously, the latter took a
`MachineInstr*` which it updated to the instruction bundle leader; now,
the latter calls the former either with the same `MachineInstr&` or the
bundle leader.
As a side effect, this removes a bunch of MachineInstr* to
MachineBasicBlock::iterator implicit conversions, a necessary step
toward fixing PR26753.
Note: I updated WebAssembly, Lanai, and AVR (despite being
off-by-default) since it turned out to be easy. I couldn't run tests
for AVR since llc doesn't link with it turned on.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274189
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Wed, 29 Jun 2016 23:58:39 +0000 (23:58 +0000)]
regenerate checks
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274188
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Wed, 29 Jun 2016 23:54:42 +0000 (23:54 +0000)]
PrologEpilogInserter: Some code cleanup; NFC
- Use range based for
- Use the more common variable names MBB and MF for
MachineBasicBlock/MachineFunction variables.
- Add a few const modifiers
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274187
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Wed, 29 Jun 2016 23:54:10 +0000 (23:54 +0000)]
Add move constructor and move assignment to fix MSVC build.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274186
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Wed, 29 Jun 2016 22:27:42 +0000 (22:27 +0000)]
Object: Replace NewArchiveIterator with a simpler NewArchiveMember class. NFCI.
The NewArchiveIterator class has a problem: it requires too much context. Any
memory buffers added to the archive must be stored within an Archive::Member,
which must have an associated Archive. This makes it harder than necessary
to create new archive members (or new archives entirely) from scratch using
memory buffers.
This patch replaces NewArchiveIterator with a NewArchiveMember class that
stores just the memory buffer and the information that goes into the archive
member header.
Differential Revision: http://reviews.llvm.org/D21721
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274183
91177308-0d34-0410-b5e6-
96231b3b80d8
Adam Nemet [Wed, 29 Jun 2016 22:04:10 +0000 (22:04 +0000)]
[LV] Improve accuracy and formatting of function comment
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274182
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 29 Jun 2016 21:58:15 +0000 (21:58 +0000)]
[LLVMConfig.cmake] LLVM_CONFIG_EXPORTED_TARGETS => LLVM_EXPORTED_TARGETS.
This matches more closely the rest of the variables in LLVMConfig.cmake which
shed the _CONFIG_ part of their names.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274175
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Wed, 29 Jun 2016 21:55:46 +0000 (21:55 +0000)]
[llvm-cov] Use relative paths to file reports in -output-dir mode
This makes it possible to e.g copy a report to another filesystem.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274173
91177308-0d34-0410-b5e6-
96231b3b80d8
Zachary Turner [Wed, 29 Jun 2016 21:48:26 +0000 (21:48 +0000)]
Resubmit "Update llvm command line parser to support subcommands."
This fixes an issue where occurrence counts would be unexpectedly
reset when parsing different parts of a command line multiple
times.
**ORIGINAL COMMIT MESSAGE**
This allows command line tools to use syntaxes like the following:
llvm-foo.exe command1 -o1 -o2
llvm-foo.exe command2 -p1 -p2
Where command1 and command2 contain completely different sets of
valid options. This is backwards compatible with previous uses
of llvm cl which did not support subcommands, as any option
which specifies no optional subcommand (e.g. all existing
code) goes into a special "top level" subcommand that expects
dashed options to appear immediately after the program name.
For example, code which is subcommand unaware would generate
a command line such as the following, where no subcommand
is specified:
llvm-foo.exe -q1 -q2
The top level subcommand can co-exist with actual subcommands,
as it is implemented as an actual subcommand which is searched
if no explicit subcommand is specified. So llvm-foo.exe as
specified above could be written so as to support all three
aforementioned command lines simultaneously.
There is one additional "special" subcommand called AllSubCommands,
which can be used to inject an option into every subcommand.
This is useful to support things like help, so that commands
such as:
llvm-foo.exe --help
llvm-foo.exe command1 --help
llvm-foo.exe command2 --help
All work and display the help for the selected subcommand
without having to explicitly go and write code to handle each
one separately.
This patch is submitted without an example of anything actually
using subcommands, but a followup patch will convert the
llvm-pdbdump tool to use subcommands.
Reviewed By: beanz
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274171
91177308-0d34-0410-b5e6-
96231b3b80d8
Artem Belevich [Wed, 29 Jun 2016 20:51:15 +0000 (20:51 +0000)]
Revert r273313 "[NVPTX] Improve lowering of byval args of device functions."
The change causes llvm crash in some unoptimized builds.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274163
91177308-0d34-0410-b5e6-
96231b3b80d8
Evgeniy Stepanov [Wed, 29 Jun 2016 20:37:43 +0000 (20:37 +0000)]
StackColoring for SafeStack.
This is a fix for PR27842.
An IR-level implementation of stack coloring tailored to work with
SafeStack. It is a bit weaker than the MI implementation in that it
does not the "lifetime start at first access" logic. This can be
improved in the future.
This patch also replaces the naive implementation of stack frame
layout with a greedy algorithm that can split existing stack slots
and even fit small objects inside the alignment padding of other
objects.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274162
91177308-0d34-0410-b5e6-
96231b3b80d8
Kevin Enderby [Wed, 29 Jun 2016 20:35:44 +0000 (20:35 +0000)]
Change Archive::create() from ErrorOr<...> to Expected<...> and update
its clients.
This commit will break the next lld builds. I’ll be committing the matching
change for lld next.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274160
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Shen [Wed, 29 Jun 2016 20:10:17 +0000 (20:10 +0000)]
[InstCombine] Simplify and correct folding fcmps with the same children
Summary: Take advantage of FCmpInst::Predicate's bit pattern and handle (fcmp *, x, y) | (fcmp *, x, y) and (fcmp *, x, y) & (fcmp *, x, y) more consistently. Also fold more FCmpInst::FCMP_FALSE and FCmpInst::FCMP_TRUE to constants.
Currently InstCombine wrongly folds (fcmp ogt, x, y) | (fcmp ord, x, y) to (fcmp ogt, x, y); this patch also fixes that.
Reviewers: spatel
Subscribers: llvm-commits, iteratee, echristo
Differential Revision: http://reviews.llvm.org/D21775
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274156
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Shen [Wed, 29 Jun 2016 20:10:13 +0000 (20:10 +0000)]
[InstCombine, NFC] Change the generated variable names by creating new instructions
This removes some noise for D21775's test changes.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274155
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Wed, 29 Jun 2016 20:01:39 +0000 (20:01 +0000)]
[Triple] Reimplement isLittleEndian(). Now it works for arm too.
Differential Revision: http://reviews.llvm.org/D21846
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274154
91177308-0d34-0410-b5e6-
96231b3b80d8
Nirav Dave [Wed, 29 Jun 2016 19:54:27 +0000 (19:54 +0000)]
Permit memory operands in ins/outs instructions
[x86] (PR15455) While (ins|outs)[bwld] instructions do not take %dx as a
memory operand, various unofficial references do and objdump
disassembles to this format. Extend special treatment of
similar (in|out)[bwld] operations.
Reviewers: craig.topper, rnk, ab
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D18837
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274152
91177308-0d34-0410-b5e6-
96231b3b80d8
Rafael Espindola [Wed, 29 Jun 2016 18:31:48 +0000 (18:31 +0000)]
Don't verify inputs to the Linker if ODR merging.
This fixes pr28072.
The point, as Duncan pointed out, is that the file is already
partially linked by just reading it.
Long term I think the solution is to make metadata owned by the module
and then the linker will lazily read it and be in charge of all the
linking. Running a verifier in each input will defeat the lazy
loading, but will be legal.
Right now we are at the unfortunate position that to support odr
merging we cannot verify the inputs, which mildly annoying (see test
update).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274148
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Shen [Wed, 29 Jun 2016 17:55:11 +0000 (17:55 +0000)]
[InstCombine] Add full tests for FoldAndOfFCmps and FoldOrOfFCmps
Summary: This adds tests for covering all cases that FoldAndOfFCmps and FoldOrOfFCmps handle.
Reviewers: spatel
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D21844
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274144
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Wed, 29 Jun 2016 17:47:08 +0000 (17:47 +0000)]
[llvm-cov] Change some FileCheck prefixes to make tests reusable (NFC)
I'm planning on extending these two tests with checks that validate
html coverage reports. Make it easier to extend them by not using a
prefix called "CHECK".
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274143
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Wed, 29 Jun 2016 17:34:31 +0000 (17:34 +0000)]
Add a regression test for PR28348.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274142
91177308-0d34-0410-b5e6-
96231b3b80d8
Nico Weber [Wed, 29 Jun 2016 17:33:41 +0000 (17:33 +0000)]
Revert r272251, it caused PR28348.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274141
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Gottesman [Wed, 29 Jun 2016 17:32:37 +0000 (17:32 +0000)]
[LLVMConfig.cmake.in] Create a new variable LLVM_CONFIG_EXPORTED_TARGETS that contains the list of all targets exported via LLVMExports.cmake.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274140
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Wed, 29 Jun 2016 16:56:46 +0000 (16:56 +0000)]
Use LLVM_ATTRIBUTE_UNUSED instead of void casts; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274139
91177308-0d34-0410-b5e6-
96231b3b80d8
Ahmed Bougacha [Wed, 29 Jun 2016 16:56:09 +0000 (16:56 +0000)]
[X86] Lower blended PACKUSes using appropriate types.
When lowering two blended PACKUS, we used to disregard the types
of the PACKUS inputs, indiscriminately generating a v16i8 PACKUS.
This leads to non-selectable things like:
(v16i8 (PACKUS (v4i32 v0), (v4i32 v1)))
Instead, check that the PACKUSes have the same type, and use that
as the final result type.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274138
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Wed, 29 Jun 2016 16:34:57 +0000 (16:34 +0000)]
[llvm-cov] Disable PGO name compression in a test file
Some bots do not configure llvm with zlib enabled. Should fix:
http://lab.llvm.org:8011/builders/llvm-clang-lld-x86_64-scei-ps4-ubuntu-fast/builds/15571
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@274137
91177308-0d34-0410-b5e6-
96231b3b80d8