Sanjoy Das [Wed, 28 Oct 2015 21:27:14 +0000 (21:27 +0000)]
[SCEV] Compute max backedge count for loops with "shift ivs"
This teaches SCEV to compute //max// backedge taken counts for loops
like
for (int i = k; i != 0; i >>>= 1)
whatever();
SCEV yet cannot represent the exact backedge count for these loops, and
this patch does not change that. This is really geared towards teaching
SCEV that loops like the above are *not* infinite.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251558
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 28 Oct 2015 21:27:08 +0000 (21:27 +0000)]
[JumpThreading] Use dominating conditions to prove implications
Summary:
If P branches to Q conditional on C and Q branches to R conditional on
C' and C => C' then the branch conditional on C' can be folded to an
unconditional branch.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D13972
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251557
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Wed, 28 Oct 2015 20:08:51 +0000 (20:08 +0000)]
[Orc] Require target support for host before running execution unit tests.
Orc unit tests that execute code shouldn't run if the compiler doesn't have
target support for the host machine.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251551
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Wed, 28 Oct 2015 19:58:02 +0000 (19:58 +0000)]
[PowerPC] Fix CodeGen/PowerPC/crbit-asm.ll test for -O1
Add the crbits processor feature so that the test can be run at -O1, etc.
regardless of the default crbits setting.
Fixes PR23778.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251548
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Wed, 28 Oct 2015 19:34:04 +0000 (19:34 +0000)]
[PGO] RawProf Reader code cleanup
Add a couple of helper methods to make the primary
raw profile reader interface's implementation more
readable. It also hides more format details. This
patch has no functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251546
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Bieneman [Wed, 28 Oct 2015 18:36:56 +0000 (18:36 +0000)]
[CMake] Disable adding the test suite as a projects subdirectory
This will never work as an add_subdirectory call, so we should just make sure it doesn't happen. To do this properly we'll need to add it under clang similar to the external compiler-rt.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251543
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Wed, 28 Oct 2015 18:15:46 +0000 (18:15 +0000)]
[X86] A small fix in X86/X86TargetTransformInfo.cpp: check a value type is simple before calling getSimpleVT().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251538
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Laevsky [Wed, 28 Oct 2015 17:54:48 +0000 (17:54 +0000)]
[AliasAnalysis] Take into account readnone attribute for the function arguments
Differential Revision: http://reviews.llvm.org/D13992
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251535
91177308-0d34-0410-b5e6-
96231b3b80d8
JF Bastien [Wed, 28 Oct 2015 17:50:23 +0000 (17:50 +0000)]
WebAssembly: disable some loop-idiom recognition
memset/memcpy aren't fully supported yet. We should invert this test
once they are supported.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251534
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Wed, 28 Oct 2015 17:40:22 +0000 (17:40 +0000)]
SamplePGO - Clear per-function data after applying a profile.
The pass was keeping around a lot of per-function data (visited blocks,
edges, dominance, etc) that is just taking up memory for no reason. In
fact, from function to function it could potentially confuse the
propagator since some maps are indexed by line offsets which can be
common between functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251531
91177308-0d34-0410-b5e6-
96231b3b80d8
Igor Laevsky [Wed, 28 Oct 2015 16:42:00 +0000 (16:42 +0000)]
[AliasAnalysis] Take into account readonly attribute for the function arguments
In getArgModRefInfo we consider all arguments as having MRI_ModRef.
However for arguments marked with readonly attribute we can return
more precise answer - MRI_Ref.
Differential Revision: http://reviews.llvm.org/D13992
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251525
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 28 Oct 2015 15:08:33 +0000 (15:08 +0000)]
Typo.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251521
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 28 Oct 2015 14:38:49 +0000 (14:38 +0000)]
Reapply: [LIR] Add support for creating memsets from loops with a negative stride.
The simple fix is to prevent forming memcpy from loops with a negative stride.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251518
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Wed, 28 Oct 2015 14:30:53 +0000 (14:30 +0000)]
[GlobalOpt] Add newlines to DEBUG messages
I think these were affected by a change way back when to stop printing newlines in Value::dump() by default. This change simply allows the debug output to be readable.
NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251517
91177308-0d34-0410-b5e6-
96231b3b80d8
Artyom Skrobov [Wed, 28 Oct 2015 13:58:36 +0000 (13:58 +0000)]
[ARM] Allow SP in rGPR, starting from ARMv8
Summary:
This patch handles assembly and disassembly, but not codegen, as of yet.
Additionally, it fixes a bug whereby SP and PC as shifted-reg operands
were treated as predictable in ARMv7 Thumb; and it enables the tests
for invalid and unpredictable instructions to run on both ARMv7 and ARMv8.
Reviewers: jmolloy, rengolin
Subscribers: aemerson, rengolin, llvm-commits
Differential Revision: http://reviews.llvm.org/D14141
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251516
91177308-0d34-0410-b5e6-
96231b3b80d8
Benjamin Kramer [Wed, 28 Oct 2015 13:54:36 +0000 (13:54 +0000)]
Put global classes into the appropriate namespace.
Most of the cases belong into an anonymous namespace. No
functionality change intended.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251515
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 28 Oct 2015 13:54:09 +0000 (13:54 +0000)]
Revert "[LIR] Add support for creating memsets from loops with a negative stride."
This reverts commit r251512. This is causing LNT/chomp to fail.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251513
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 28 Oct 2015 12:55:34 +0000 (12:55 +0000)]
[LIR] Add support for creating memsets from loops with a negative stride.
http://reviews.llvm.org/D14125
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251512
91177308-0d34-0410-b5e6-
96231b3b80d8
Chad Rosier [Wed, 28 Oct 2015 12:30:08 +0000 (12:30 +0000)]
Add newline to test. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251511
91177308-0d34-0410-b5e6-
96231b3b80d8
Hrvoje Varga [Wed, 28 Oct 2015 11:04:29 +0000 (11:04 +0000)]
[mips][microMIPS] Implement PAUSE, RDHWR, RDPGPR, SDBBP, SSNOP, SYNC, SYNCI and WAIT instructions
Differential Revision: http://reviews.llvm.org/D12628
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251510
91177308-0d34-0410-b5e6-
96231b3b80d8
Vasileios Kalintiris [Wed, 28 Oct 2015 11:02:01 +0000 (11:02 +0000)]
[Orc] Remove unnecessary semicolon. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251509
91177308-0d34-0410-b5e6-
96231b3b80d8
James Molloy [Wed, 28 Oct 2015 10:41:29 +0000 (10:41 +0000)]
[GlobalsAA] An indirect global that is initialized is not fair game
When checking if an indirect global (a global with pointer type) is only assigned by allocation functions, first check if the global is itself initialized. If it is, it's not only assigned by allocation functions.
This fixes PR25309. Thanks to David Majnemer for reducing the test case!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251508
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 28 Oct 2015 09:59:50 +0000 (09:59 +0000)]
OrcJITTests: Prune unused libdeps.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251506
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Wed, 28 Oct 2015 09:37:09 +0000 (09:37 +0000)]
OrcJITTests: Update libdeps.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251504
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 28 Oct 2015 05:48:32 +0000 (05:48 +0000)]
[X86] Make some for loops over MVTs more explicit (and shorter) by just mentioning all the relevant types in an initializer list. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251500
91177308-0d34-0410-b5e6-
96231b3b80d8
Chen Li [Wed, 28 Oct 2015 05:15:51 +0000 (05:15 +0000)]
Revert r251492 "[IndVarSimplify] Rewrite loop exit values with their
initial values from loop preheader", because it broke some bots.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251498
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 28 Oct 2015 04:54:46 +0000 (04:54 +0000)]
Fix a -Wpessimizing-move warning.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251495
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 28 Oct 2015 04:53:27 +0000 (04:53 +0000)]
Use range-based for loops and use initializer list to remove a small static array. NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251494
91177308-0d34-0410-b5e6-
96231b3b80d8
Chen Li [Wed, 28 Oct 2015 04:45:47 +0000 (04:45 +0000)]
[IndVarSimplify] Rewrite loop exit values with their initial values from loop preheader
Summary:
This patch adds support to check if a loop has loop invariant conditions which lead to loop exits. If so, we know that if the exit path is taken, it is at the first loop iteration. If there is an induction variable used in that exit path whose value has not been updated, it will keep its initial value passing from loop preheader. We can therefore rewrite the exit value with
its initial value. This will help remove phis created by LCSSA and enable other optimizations like loop unswitch.
Reviewers: sanjoy
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D13974
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251492
91177308-0d34-0410-b5e6-
96231b3b80d8
Xinliang David Li [Wed, 28 Oct 2015 04:20:31 +0000 (04:20 +0000)]
[PGO] Indexed Prof Reader refactoring (NFC)
Change InstrProfReaderIndex from typedef into a wrapper
class with helper methods. This makes the index profile
reader code more readable. It also hides the implementation
detail of the index format and make it more flexible to allow
support different (or more than one) format in the future.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251491
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Wed, 28 Oct 2015 04:02:12 +0000 (04:02 +0000)]
Remove templates from CostTableLookup functions. All instantiations had the same type.
This also lets us remove the versions of the functions that took a statically sized array as we can rely on ArrayRef implicit conversion now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251490
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Wed, 28 Oct 2015 03:26:45 +0000 (03:26 +0000)]
[PowerPC] Replace cntlz[.] with cntlzw[.]
cntlz is the old POWER mnemonic. cntlzw is the PowerPC mnemonic.
This change fixes an issue when -no-integrated-as: The opcode cntlz is
unrecognized by gas
Alias the POWER mnemonic cntlz[.] to the PowerPC mnemonic cntlzw[.]
This is done for because the POWER cntlz mnemonic has be used by LLVM for
a very long time. We need to make sure that assembly programs
that are using the cntlz[.] do not break with this change.
Change PowerPC tests to reflect the insn change from cntlz to cntlzw.
Add assembly test to verify cntlz[.] is encoded correctly.
Patch by Tom Rix!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251489
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 28 Oct 2015 03:20:19 +0000 (03:20 +0000)]
[ValueTracking] Expose `implies` via ValueTracking, NFC
Summary: This will allow a later patch to `JumpThreading` use this functionality.
Reviewers: reames
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D13971
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251488
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 28 Oct 2015 03:20:15 +0000 (03:20 +0000)]
[ValueTracking] Use !range metadata more aggressively in KnownBits
Summary:
Teach `computeKnownBitsFromRangeMetadata` to use `!range` metadata more
aggressively.
Reviewers: majnemer, nlewycky, jingyue
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D14100
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251487
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 28 Oct 2015 03:20:10 +0000 (03:20 +0000)]
[SelectionDAG] Don't inspect !range metadata for extended loads
Summary:
Don't call `computeKnownBitsFromRangeMetadata` for extended loads --
this can cause a mismatch between the width of the !range metadata and
the width of the APInt's accumulating `KnownZero` (and `KnownOne` in the
future). This isn't a problem now, but will be after a future change.
Note: this can be made more aggressive in the future.
Reviewers: nlewycky
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D14107
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251486
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Wed, 28 Oct 2015 03:20:05 +0000 (03:20 +0000)]
[GVN] Make a test case more robust
The singleton !range metadata gets simplified more aggressively after a
later change, so change the !range metadata to contain more than one
element.
While at it, turn some `; CHECK` s to `; CHECK-LABEL:` s.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251485
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Wed, 28 Oct 2015 03:12:51 +0000 (03:12 +0000)]
[Orc] Disable Orc C API unit tests on non-Darwin while I investigate more
builder failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251484
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Wed, 28 Oct 2015 02:40:04 +0000 (02:40 +0000)]
[Orc] Re-add C bindings for the Orc APIs, with a fix to remove the union that
was causing builder failures.
The bindings were originally added in r251472, and reverted in r251473 due to
the builder failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251482
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Wed, 28 Oct 2015 02:36:45 +0000 (02:36 +0000)]
lit/TestRunner.py: Factor variable subsitution into an own function; NFCI
This is a clearer separation of concerns and makes it easier to reuse
the function.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251481
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Wed, 28 Oct 2015 02:36:42 +0000 (02:36 +0000)]
lit/TestRunner.py: Factor out Substitution construction; NFC
This is a clearer separation of concerns and makes it easier to reuse
the functions.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251480
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Wed, 28 Oct 2015 02:36:38 +0000 (02:36 +0000)]
lit/TestRunner.py: Get execdir from test.getExecPath() instead of passing it around; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251479
91177308-0d34-0410-b5e6-
96231b3b80d8
Matthias Braun [Wed, 28 Oct 2015 02:36:35 +0000 (02:36 +0000)]
lit/TestRunner.py: Make parseIntegratedTestScriptCommands() keyword list a parameter; NFC
This allows the function to be easily reused and also simplifies the
code as the keyword list is next to the keyword handling now.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251478
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Wed, 28 Oct 2015 01:03:09 +0000 (01:03 +0000)]
[Orc] Revert the C bindngs commit, r251472, while I debug some builder failures.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251473
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Wed, 28 Oct 2015 00:28:26 +0000 (00:28 +0000)]
[Orc] Add experimental C bindings for Orc.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251472
91177308-0d34-0410-b5e6-
96231b3b80d8
Dylan McKay [Wed, 28 Oct 2015 00:24:54 +0000 (00:24 +0000)]
Add myself as the the code owner for the AVR backend
Summary:
As I maintain the AVR backend and am currently in the process of migrating it in tree, it makes sense to add myself as the code owner.
Thoughts welcome!
Differential Revision: http://reviews.llvm.org/D14002
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251471
91177308-0d34-0410-b5e6-
96231b3b80d8
James Y Knight [Tue, 27 Oct 2015 23:09:03 +0000 (23:09 +0000)]
Make the SelectionDAG graph printer use SDNode::PersistentId labels.
r248010 changed the -debug output to use short ids, but did not
similarly modify the graph printer. Change to be consistent, for ease of
cross-reference.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251465
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Tue, 27 Oct 2015 23:01:25 +0000 (23:01 +0000)]
Bitcode: Fix more unsigned integer overflow bugs.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251464
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Tue, 27 Oct 2015 22:43:56 +0000 (22:43 +0000)]
[SimplifyCFG] Don't DCE catchret because the successor is unreachable
CatchReturnInst has side-effects: it runs a destructor. This destructor
could conceivably run forever/call exit/etc. and should not be removed.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251461
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Tue, 27 Oct 2015 22:10:17 +0000 (22:10 +0000)]
[Bitcode] Fix accidental syntax errors in compatibility tests
We used automated tools to update our IR to its current syntax in commit
21f77df7(r247378). While it correctly updated the CHECK lines in our
compatibility tests, the IR should have remained untouched. This commit
fixes the syntax errors.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251458
91177308-0d34-0410-b5e6-
96231b3b80d8
Simon Pilgrim [Tue, 27 Oct 2015 21:18:45 +0000 (21:18 +0000)]
[X86][AVX512] Test UNPCK with non-sequential scalars
Missing tests for r251297
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251453
91177308-0d34-0410-b5e6-
96231b3b80d8
Vedant Kumar [Tue, 27 Oct 2015 21:17:06 +0000 (21:17 +0000)]
[IR] Limit bits used for CallingConv::ID, update tests
Use 10 bits to represent calling convention ID's instead of 13, and
update the bitcode compatibility tests accordingly. We now error-out in
the bitcode reader when we see bad calling conv ID's.
Thanks to rnk and dexonsmith for feedback!
Differential Revision: http://reviews.llvm.org/D13826
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251452
91177308-0d34-0410-b5e6-
96231b3b80d8
Hal Finkel [Tue, 27 Oct 2015 20:37:04 +0000 (20:37 +0000)]
[AliasSetTracker] Use mod/ref information for UnknownInstr
AliasSetTracker does not need to convert the access mode to ModRefAccess if the
new visited UnknownInst has only 'REF' modrefinfo to existing pointers in the
sets.
Patch by Andrew Zhogin!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251451
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 27 Oct 2015 20:27:25 +0000 (20:27 +0000)]
Use the 'arcp' fast-math-flag when combining repeated FP divisors
This is a usage of the IR-level fast-math-flags now that they are propagated to SDNodes.
This was originally part of D8900.
Removing the global 'enable-unsafe-fp-math' checks will require auto-upgrade and
possibly other changes.
Differential Revision: http://reviews.llvm.org/D9708
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251450
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Tue, 27 Oct 2015 19:48:28 +0000 (19:48 +0000)]
[ScalarEvolutionExpander] PHI on a catchpad can be used on both edges
A PHI on a catchpad might be used by both edges out of the catchpad,
feeding back into a loop. In this case, just use the insertion point.
Anything more clever would require new basic blocks or PHI placement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251442
91177308-0d34-0410-b5e6-
96231b3b80d8
Jun Bum Lim [Tue, 27 Oct 2015 19:16:03 +0000 (19:16 +0000)]
[AArch64]Merge halfword loads into a 32-bit load
This recommits r250719, which caused a failure in SPEC2000.gcc
because of the incorrect insert point for the new wider load.
Convert two halfword loads into a single 32-bit word load with bitfield extract
instructions. For example :
ldrh w0, [x2]
ldrh w1, [x2, #2]
becomes
ldr w0, [x2]
ubfx w1, w0, #16, #16
and w0, w0, #ffff
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251438
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 27 Oct 2015 19:02:52 +0000 (19:02 +0000)]
Whitespace.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251437
91177308-0d34-0410-b5e6-
96231b3b80d8
NAKAMURA Takumi [Tue, 27 Oct 2015 19:02:36 +0000 (19:02 +0000)]
Revert r251291, "Loop Vectorizer - skipping "bitcast" before GEP"
It causes miscompilation of llvm/lib/ExecutionEngine/Interpreter/Execution.cpp.
See also PR25324.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251436
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Tue, 27 Oct 2015 18:41:46 +0000 (18:41 +0000)]
Tidy a comment. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251434
91177308-0d34-0410-b5e6-
96231b3b80d8
Cong Hou [Tue, 27 Oct 2015 17:59:36 +0000 (17:59 +0000)]
Create a new interface addSuccessorWithoutWeight(MBB*) in MBB to add successors when optimization is disabled.
When optimization is disabled, edge weights that are stored in MBB won't be used so that we don't have to store them. Currently, this is done by adding successors with default weight 0, and if all successors have default weights, the weight list will be empty. But that the weight list is empty doesn't mean disabled optimization (as is stated several times in MachineBasicBlock.cpp): it may also mean all successors just have default weights.
We should discourage using default weights when adding successors, because it is very easy for users to forget update the correct edge weights instead of using default ones (one exception is that the MBB only has one successor). In order to detect such usages, it is better to differentiate using default weights from the case when optimizations is disabled.
In this patch, a new interface addSuccessorWithoutWeight(MBB*) is created for when optimization is disabled. In this case, MBB will try to maintain an empty weight list, but it cannot guarantee this as for many uses of addSuccessor() whether optimization is disabled or not is not checked. But it can guarantee that if optimization is enabled, then the weight list always has the same size of the successor list.
Differential revision: http://reviews.llvm.org/D13963
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251429
91177308-0d34-0410-b5e6-
96231b3b80d8
Charlie Turner [Tue, 27 Oct 2015 17:59:03 +0000 (17:59 +0000)]
[SLP] Be more aggressive about reduction width selection.
Summary:
This change could be way off-piste, I'm looking for any feedback on whether it's an acceptable approach.
It never seems to be a problem to gobble up as many reduction values as can be found, and then to attempt to reduce the resulting tree. Some of the workloads I'm looking at have been aggressively unrolled by hand, and by selecting reduction widths that are not constrained by a vector register size, it becomes possible to profitably vectorize. My test case shows such an unrolling which SLP was not vectorizing (on neither ARM nor X86) before this patch, but with it does vectorize.
I measure no significant compile time impact of this change when combined with D13949 and D14063. There are also no significant performance regressions on ARM/AArch64 in SPEC or LNT.
The more principled approach I thought of was to generate several candidate tree's and use the cost model to pick the cheapest one. That seemed like quite a big design change (the algorithms seem very much one-shot), and would likely be a costly thing for compile time. This seemed to do the job at very little cost, but I'm worried I've misunderstood something!
Reviewers: nadav, jmolloy
Subscribers: mssimpso, llvm-commits, aemerson
Differential Revision: http://reviews.llvm.org/D14116
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251428
91177308-0d34-0410-b5e6-
96231b3b80d8
Charlie Turner [Tue, 27 Oct 2015 17:54:16 +0000 (17:54 +0000)]
[SLP] Try a bit harder to find reduction PHIs
Summary:
Currently, when the SLP vectorizer considers whether a phi is part of a reduction, it dismisses phi's whose incoming blocks are not the same as the block containing the phi. For the patterns I'm looking at, extending this rule to allow phis whose incoming block is a containing loop latch allows me to vectorize certain workloads.
There is no significant compile-time impact, and combined with D13949, no performance improvement measured in ARM/AArch64 in any of SPEC2000, SPEC2006 or LNT.
Reviewers: jmolloy, mcrosier, nadav
Subscribers: mssimpso, nadav, aemerson, llvm-commits
Differential Revision: http://reviews.llvm.org/D14063
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251425
91177308-0d34-0410-b5e6-
96231b3b80d8
Charlie Turner [Tue, 27 Oct 2015 17:49:11 +0000 (17:49 +0000)]
[SLP] Treat SelectInsts as reduction values.
Summary:
Certain workloads, in particular sum-of-absdiff loops, can be vectorized using SLP if it can treat select instructions as reduction values.
The test case is a bit awkward. The AArch64 cost model needs some tuning to not be so pessimistic about selects. I've had to tweak the SLP threshold here.
Reviewers: jmolloy, mzolotukhin, spatel, nadav
Subscribers: nadav, mssimpso, aemerson, llvm-commits
Differential Revision: http://reviews.llvm.org/D13949
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251424
91177308-0d34-0410-b5e6-
96231b3b80d8
Lang Hames [Tue, 27 Oct 2015 17:45:48 +0000 (17:45 +0000)]
[Orc] Fix indentation.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251423
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Tue, 27 Oct 2015 17:37:00 +0000 (17:37 +0000)]
Fix SamplePGO segfault when debug info is missing.
When emitting a remark for a conditional branch annotation, the remark
uses the line location information of the conditional branch in the
message. In some cases, that information is unavailable and the
optimization would segfaul. I'm still not sure whether this is a bug or
WAI, but the optimizer should not die because of this.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251420
91177308-0d34-0410-b5e6-
96231b3b80d8
Reid Kleckner [Tue, 27 Oct 2015 17:32:48 +0000 (17:32 +0000)]
[ms-inline-asm] Leave alignment in bytes if the native assembler uses bytes
The existing behavior was correct on Darwin, which is probably the
platform it was written for.
Before this change, we would rewrite "align 8" to ".align 3" and then
fail to make it through the integrated assembler because 3 is not a
power of 2.
Differential Revision: http://reviews.llvm.org/D14120
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251418
91177308-0d34-0410-b5e6-
96231b3b80d8
Rui Ueyama [Tue, 27 Oct 2015 16:57:50 +0000 (16:57 +0000)]
Rename qsort -> multikey_qsort. NFC.
`qsort` as a file-scope local function name was confusing.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251414
91177308-0d34-0410-b5e6-
96231b3b80d8
Ed Schouten [Tue, 27 Oct 2015 16:37:49 +0000 (16:37 +0000)]
Prefer ranlib mode over ar mode.
For CloudABI's toolchain I have a symlink that goes from <target>-ar and
<target>-ranlib to LLVM's ar binary, to mimick GNU Binutils' naming
scheme. The problem is that if we're targetting ARM64, the name of the
ranlib executable is aarch64-unknown-cloudabi-ranlib. This already
contains the string "ar".
Let's move the "ranlib" test above the "ar" test. It's not that likely
that we're going to see operating systems or harwdare architectures that
are called "ranlib".
Reviewed by: rafael
Differential Revision: http://reviews.llvm.org/D14123
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251413
91177308-0d34-0410-b5e6-
96231b3b80d8
Chris Bieneman [Tue, 27 Oct 2015 16:02:04 +0000 (16:02 +0000)]
[CMake] Get rid of LLVM_DYLIB_EXPORT_ALL, and make it the default, add libLLVM-C on darwin to cover the C API needs.
Summary:
We've had a lot of discussion in the past about the meaningful and useful default behaviors for the llvm-shlib tool. The original implementation was heavily geared toward Apple's use, and I think that was wrong. This patch seeks to correct that.
I've removed the LLVM_DYLIB_EXPORT_ALL variable and made libLLVM export everything by default.
I've also added a new target that is only built on Darwin for libLLVM-C as a library that re-exports the LLVM-C API. This library is not built on Linux because ELF doesn't support re-export libraries in the same way MachO does.
Reviewers: chapuni, resistor, bogner, axw
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D13842
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251411
91177308-0d34-0410-b5e6-
96231b3b80d8
Asaf Badouh [Tue, 27 Oct 2015 15:37:17 +0000 (15:37 +0000)]
[X86][AVX512] [X86][AVX512] add convert float to half
convert float to half with mask/maskz for the reg to reg version and mask for the reg to mem version (there is no maskz version for reg to mem).
Differential Revision: http://reviews.llvm.org/D14113
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251409
91177308-0d34-0410-b5e6-
96231b3b80d8
Charlie Turner [Tue, 27 Oct 2015 10:25:20 +0000 (10:25 +0000)]
[ARM] Expand ROTL and ROTR of vector value types
Summary: After D13851 landed, we saw backend crashes when compiling the reduced test case included in this patch. The right fix seems to be to allow these vector types for expansion in instruction selection.
Reviewers: rengolin, t.p.northover
Subscribers: RKSimon, t.p.northover, aemerson, llvm-commits, rengolin
Differential Revision: http://reviews.llvm.org/D14082
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251401
91177308-0d34-0410-b5e6-
96231b3b80d8
Mehdi Amini [Tue, 27 Oct 2015 08:12:08 +0000 (08:12 +0000)]
Do not use "else" when both branches return (NFC)
From: Mehdi Amini <mehdi.amini@apple.com>
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251398
91177308-0d34-0410-b5e6-
96231b3b80d8
David Majnemer [Tue, 27 Oct 2015 07:36:42 +0000 (07:36 +0000)]
[ScalarEvolutionExpander] Properly insert no-op casts + EH Pads
We want to insert no-op casts as close as possible to the def. This is
tricky when the cast is of a PHI node and the BasicBlocks between the
def and the use cannot hold any instructions. Iteratively walk EH pads
until we hit a non-EH pad.
This fixes PR25326.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251393
91177308-0d34-0410-b5e6-
96231b3b80d8
Michael Kuperstein [Tue, 27 Oct 2015 07:23:59 +0000 (07:23 +0000)]
[X86] Make elfiamcu an OS, not an environment.
GNU tools require elfiamcu to take up the entire OS field, so, e.g.
i?86-*-linux-elfiamcu is not considered a legal triple.
Make us compatible.
Differential Revision: http://reviews.llvm.org/D14081
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251390
91177308-0d34-0410-b5e6-
96231b3b80d8
Davide Italiano [Tue, 27 Oct 2015 04:17:51 +0000 (04:17 +0000)]
[SimplifyLibCalls] Use range-based loop. No functional change.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251383
91177308-0d34-0410-b5e6-
96231b3b80d8
Craig Topper [Tue, 27 Oct 2015 04:14:24 +0000 (04:14 +0000)]
Convert cost table lookup functions to return a pointer to the entry or nullptr instead of the index.
This avoid mentioning the table name an extra time and allows the lookup to be done directly in the ifs by relying on the bool conversion of the pointer.
While there make use of ArrayRef and std::find_if.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251382
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Tue, 27 Oct 2015 01:41:43 +0000 (01:41 +0000)]
[function-attrs] Refactor code to handle shorter code with early exits.
No functionality changed here, but the indentation is substantially
reduced and IMO the code is much easier to read. I've also added some
helpful comments.
This is just a clean-up I wrote while studying the code, and that has
been in my backlog for a while.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251381
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 27 Oct 2015 01:36:06 +0000 (01:36 +0000)]
[ValueTracking] Don't special case wrapped ConstantRanges; NFCI
Use `getUnsignedMax` directly instead of special casing a wrapped
ConstantRange.
The previous code would have been "buggy" (and this would have been a
semantic change) if LLVM allowed !range metadata to denote full
ranges. E.g. in
%val = load i1, i1* %ptr, !range !{i1 1, i1 1} ;; == full set
ValueTracking would conclude that the high bit (IOW the only bit) in
%val was zero.
Since !range metadata does not allow empty or full ranges, this change
is just a minor stylistic improvement.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251380
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Tue, 27 Oct 2015 01:28:07 +0000 (01:28 +0000)]
[x86] replace integer logic ops with packed SSE FP logic ops
If we have an operand to a bitwise logic op that's already in
an XMM register and the result is going to be sent to an XMM
register, then use an SSE logic op to avoid moves between the
integer and vector register files.
Related commits:
http://reviews.llvm.org/rL248395
http://reviews.llvm.org/rL248399
http://reviews.llvm.org/rL248404
http://reviews.llvm.org/rL248409
http://reviews.llvm.org/rL248415
This should solve PR22428:
https://llvm.org/bugs/show_bug.cgi?id=22428
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251378
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjoy Das [Tue, 27 Oct 2015 00:52:09 +0000 (00:52 +0000)]
[SCEV] Refactor out ScalarEvolution::getDataLayout; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251375
91177308-0d34-0410-b5e6-
96231b3b80d8
Steve King [Tue, 27 Oct 2015 00:14:06 +0000 (00:14 +0000)]
Fix llc crash processing S/UREM for -Oz builds caused by rL250825.
When taking the remainder of a value divided by a constant, visitREM()
attempts to convert the REM to a longer but faster sequence of instructions.
This conversion calls combine() on a speculative DIV instruction. Commit
rL250825 may cause this combine() to return a DIVREM, corrupting nearby nodes.
Flow eventually hits unreachable().
This patch adds a test case and a check to prevent visitREM() from trying
to convert the REM instruction in cases where a DIVREM is possible.
See http://reviews.llvm.org/D14035
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251373
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 26 Oct 2015 23:52:42 +0000 (23:52 +0000)]
add FP logic test cases to show current codegen (PR22428)
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251370
91177308-0d34-0410-b5e6-
96231b3b80d8
Daniel Sanders [Mon, 26 Oct 2015 23:50:00 +0000 (23:50 +0000)]
[mips][ias] Fold needsExpansion() and expandInstruction() together. NFC.
Summary:
Previously we maintained two separate switch statements that had to be kept in
sync. This patch merges them into a single switch.
Reviewers: vkalintiris
Subscribers: llvm-commits, dsanders
Differential Revision: http://reviews.llvm.org/D14012
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251369
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 26 Oct 2015 23:33:13 +0000 (23:33 +0000)]
Switch ownership of miscellaneous ARM target to myself.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251367
91177308-0d34-0410-b5e6-
96231b3b80d8
Chandler Carruth [Mon, 26 Oct 2015 22:54:53 +0000 (22:54 +0000)]
[x86] Make the vselect-minmax test 2x to 3x faster by deleting all the
instructions that aren't relevant for instruction selection of vector
min and max.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251366
91177308-0d34-0410-b5e6-
96231b3b80d8
Oleksiy Vyalov [Mon, 26 Oct 2015 22:37:36 +0000 (22:37 +0000)]
Use Twin instead of std::to_string.
http://reviews.llvm.org/D14095
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251365
91177308-0d34-0410-b5e6-
96231b3b80d8
Ivan Krasin [Mon, 26 Oct 2015 22:35:40 +0000 (22:35 +0000)]
Fix indents. It's a follow up to r251353.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251364
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 26 Oct 2015 22:34:56 +0000 (22:34 +0000)]
[LLVMSymbolize] Don't use LLVMSymbolizer::Options in ModuleInfo. NFC.
LLVMSymbolizer::Options is mostly used in LLVMSymbolizer class anyway.
Let's keep their usage restricted to that class, especially given that
it's worth to move ModuleInfo to a different header, independent from
the symbolizer class.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251363
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 26 Oct 2015 21:54:14 +0000 (21:54 +0000)]
reorganize logic; NFCI (retry r251349)
This is a preliminary step before adding another optimization
to PerformBITCASTCombine().
..and I really hope it's NFC this time!
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251357
91177308-0d34-0410-b5e6-
96231b3b80d8
Ivan Krasin [Mon, 26 Oct 2015 21:36:35 +0000 (21:36 +0000)]
Move imported entities into DwarfCompilationUnit to speed up LTO linking.
Summary:
In particular, this CL speeds up the official Chrome linking with LTO by
1.8x.
See more details in https://crbug.com/542426
Reviewers: dblaikie
Subscribers: jevinskie
Differential Revision: http://reviews.llvm.org/D13918
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251353
91177308-0d34-0410-b5e6-
96231b3b80d8
Tim Northover [Mon, 26 Oct 2015 21:32:53 +0000 (21:32 +0000)]
ARM: make sure VFP loads and stores are properly aligned.
Both VLDRS and VLDRD fault if the memory is not 4 byte aligned, which wasn't
really being checked before, leading to faults at runtime.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251352
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 26 Oct 2015 21:28:02 +0000 (21:28 +0000)]
revert r251349; it included code for a functional change
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251350
91177308-0d34-0410-b5e6-
96231b3b80d8
Sanjay Patel [Mon, 26 Oct 2015 21:24:09 +0000 (21:24 +0000)]
reorganize logic; NFCI
This is a preliminary step before adding another optimization
to PerformBITCASTCombine().
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251349
91177308-0d34-0410-b5e6-
96231b3b80d8
Keno Fischer [Mon, 26 Oct 2015 21:22:58 +0000 (21:22 +0000)]
Initialize BasicAAWrapperPass in it's constructor
Summary: This idiom is used elsewhere in LLVM, but was overlooked here.
Reviewers: chandlerc
Subscribers: llvm-commits
Differential Revision: http://reviews.llvm.org/D13628
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251348
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 26 Oct 2015 21:20:37 +0000 (21:20 +0000)]
Fix build failure on GCC 4.7 (old libstdc++ doesn't have std::map::emplace).
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251347
91177308-0d34-0410-b5e6-
96231b3b80d8
David Blaikie [Mon, 26 Oct 2015 21:10:36 +0000 (21:10 +0000)]
Remove use of std::map<>::emplace which is not supported on some older versions of libstdc++
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251346
91177308-0d34-0410-b5e6-
96231b3b80d8
Diego Novillo [Mon, 26 Oct 2015 20:50:26 +0000 (20:50 +0000)]
Remove unused local variable. NFC.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251344
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Mon, 26 Oct 2015 20:49:49 +0000 (20:49 +0000)]
Fix tests.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251343
91177308-0d34-0410-b5e6-
96231b3b80d8
Peter Collingbourne [Mon, 26 Oct 2015 20:46:44 +0000 (20:46 +0000)]
ARM/ELF: Restore original (pre-r251322) logic for deciding whether to use GOT.
Unbreaks linking with gold, which cannot resolve direct relocations referring
to global symbols.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251342
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 26 Oct 2015 20:12:29 +0000 (20:12 +0000)]
[LLVMSymbolize] Use symbol table only if function linkage name was requested.
Now it's enough to just specify -functions=short without additionally
providing -use-symbol-table=false.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251339
91177308-0d34-0410-b5e6-
96231b3b80d8
Alexey Samsonov [Mon, 26 Oct 2015 20:12:27 +0000 (20:12 +0000)]
Fix build error by fully qualifying llvm::make_unique.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@251338
91177308-0d34-0410-b5e6-
96231b3b80d8