Commit Graph

114572 Commits

Author SHA1 Message Date
Craig Topper
256cb3b576 [TableGen] Implement at least some support for multiple explicit results in an instruction pattern. No functional change to existing patterns.
This should help with the AVX512 masked gather changes Elena is working on. This patch is derived from some of the changes Elena made to tablegen, but modified by me to support arbitrary number of results.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231357 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 07:11:36 +00:00
Craig Topper
8ad519fd3c [TableGen] Add support constraining a vector type in a pattern to have a specific element type and for constraining a vector type to have the same number of elements as another vector type. This is useful for AVX512 mask operations so we relate the mask type to the type of the other arguments.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231356 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 07:11:34 +00:00
Craig Topper
62eaac6087 [X86] Use vmovss to handle inserting an element into index 0 of a v8f32 vector of zeros.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231354 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 06:38:42 +00:00
Frederic Riss
1fffc4d8a0 Remove useless break after return.
Pointed out by Paul Robinson.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231353 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 06:13:39 +00:00
Philip Reames
599a7dc9ae Add a few more performance tips
These came from my own experience and may not apply equally to all use cases.  Any alternate perspective anyone has should be used to refine these.  

As always, grammar and spelling adjustments are more than welcome.  Please just directly commit a fix if you see something problematic.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231352 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 05:55:55 +00:00
Frederic Riss
497ca634e1 Revert "[dsymutil] MSVC does generate move constructors, but it should accept to default them"
This reverts commit r231350.

It turns out MSVC doesn't generate implicit move constructors and also doesn't accept to default them...
See for example http://lab.llvm.org:8011/builders/lldb-x86-windows-msvc/builds/2786

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231351 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 05:29:05 +00:00
Frederic Riss
0ff2c1e37f [dsymutil] MSVC does generate move constructors, but it should accept to default them
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231350 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 05:17:06 +00:00
Philip Reames
315492ba12 Add a link to the new PerformanceTips docs from the 3.7 release notes
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231349 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 05:11:05 +00:00
Hans Wennborg
f082c5b27f Revert r231324 "Remove the conditional addition of the execution dependency fixing"
See PR22799.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231348 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 03:24:49 +00:00
Chandler Carruth
bb535bcc20 [MBP] Use range based for-loops throughout this code. Several had
already been added and the inconsistency made choosing names and
changing code more annoying. Plus, wow are they better for this code!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231347 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 03:19:05 +00:00
Chandler Carruth
35742dd20c [MBP] NFC, run clang-format over this code and tweak things to make the
result reasonable.

This code predated clang-format and so there was a reasonable amount of
crufty formatting that had accumulated. This should ensure that neither
myself nor others end up with formatting-only changes sneaking into
other fixes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231341 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 02:35:31 +00:00
Chandler Carruth
559a329625 [MBP] This is no longer 'block-placement2'. ;] The old variants are long
gone, update this code to reflect that.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231340 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 02:28:25 +00:00
Rafael Espindola
304fe62b74 Use the existing begin and end symbol for debug info.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231338 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 02:05:42 +00:00
NAKAMURA Takumi
9e112cc561 Reformat.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231336 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:25:19 +00:00
NAKAMURA Takumi
71fa4016bb Revert r231103, "FullDependenceAnalysis: Avoid using the (deprecated in C++11) copy ctor"
It is miscompiled on msc18.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231335 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:25:12 +00:00
NAKAMURA Takumi
c1899b151e Revert r231104, "unique_ptrify FullDependenceAnalysis::DV", to appease msc18 C2280.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231334 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:25:06 +00:00
Kostya Serebryany
c2f4077b88 [sanitizer] add nosanitize metadata to more coverage instrumentation instructions
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231333 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:20:05 +00:00
Chandler Carruth
4197c13062 [MBP] Revert r231238 which attempted to fix a nasty bug where MBP is
just arbitrarily interleaving unrelated control flows once they get
moved "out-of-line" (both outside of natural CFG ordering and with
diamonds that cannot be fully laid out by chaining fallthrough edges).

This easy solution doesn't work in practice, and it isn't just a small
bug. It looks like a very different strategy will be required. I'm
working on that now, and it'll again go behind some flag so that
everyone can experiment and make sure it is working well for them.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231332 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:07:03 +00:00
NAKAMURA Takumi
d12ce78ca9 ScalarEvolution.cpp: Appease g++-4.7. He missed implicit "this" in lambda.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231331 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 01:02:45 +00:00
Eric Christopher
061096482f Remove the conditional addition of the execution dependency fixing
pass from the ARM backend as the pass itself will detect any use
of the appropriate register class.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231324 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 00:28:55 +00:00
Eric Christopher
b56d7b0316 Cleanup and remove a chunk of getARMSubtarget calls in the
ARM TargetMachine pass pipeline construction by pushing them down
into the appropriate pass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231323 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 00:23:40 +00:00
Paul Robinson
948b2db8a7 Turn off .debug_pubnames/pubtypes for PS4.
Differential Revision: http://reviews.llvm.org/D8067


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231322 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-05 00:08:27 +00:00
Aaron Ballman
5451090e4c Initializer lists are supported in MSVC 2013. Since that's our minimum required version, we can move that to the list of acceptable C++11 features.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231313 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 23:17:31 +00:00
Argyrios Kyrtzidis
05b7eaffdd [Support] Increase timeout for the LockFileManager back to 5 mins.
Waiting for just 1 min may not be enough for some contexts.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231309 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:54:38 +00:00
Matthias Braun
29aeaf5408 Improve test robustness
Improve test robustness in preparation of coming commits:
- Avoid undefs which may get propagated too much.
- Remove several pointless add 0, instructions

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231307 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:31:18 +00:00
Sanjoy Das
295be8492e [IndVarSimplify] use the "canonical" way to infer no-wrap.
Summary:
rL225282 introduced an ad-hoc way to promote some additions to nuw or
nsw.  Since then SCEV has become smarter in directly proving no-wrap;
and using the canonical "ext(A op B) == ext(A) op ext(B)" method of
proving no-wrap is just as powerful now.  Rip out the existing
complexity in favor of getting SCEV to do all the heaving lifting
internally.

This change does not add any unit tests because it is supposed to be a
non-functional change.  Tests added in rL225282 and rL226075 are valid
tests for this change.

Reviewers: atrick, majnemer

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7981

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231306 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:24:23 +00:00
Sanjoy Das
12aa70b7e9 [SCEV] make SCEV smarter about proving no-wrap.
Summary:
Teach SCEV to prove no overflow for an add recurrence by proving
something about the range of another add recurrence a loop-invariant
distance away from it.

Reviewers: atrick, hfinkel

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D7980

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231305 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:24:17 +00:00
David Blaikie
274eae84d8 Provide an explicit move ctor because MSVC can't synthesize one
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231303 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:20:52 +00:00
Frederic Riss
d0d92e7d30 [dsymutil] Add minimal code to emit DIE trees.
This commit adds code to emit DIE trees that have been pruned from the
parts that haven't been marked as kept in the previous pass.

It works by 'cloning' the input DIE tree (as read by libDebugInfoDwarf)
into a tree of DIE objects. Cloning the DIEs means essentially cloning
their attributes. The code in this commit does only handle scalar and
block attributes (scalar because they are trivial, blocks because they
can't be easily replaced by a scalr placeholder), all the other ones
are replaced by placeholder zero values and will be handled in
further commits.

The added tests mostly check that the DIE tree has the correct layout and
also verify that a few chosen scalar and block attributes correctly make
their way into the output.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231300 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:07:44 +00:00
Frederic Riss
17b5682094 DWARFFormValue: Add getAsSignedConstant method.
The implementation accepts explicitely signed forms (DW_FORM_sdata),
but also unsigned forms as long as they fit in an int64_t.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231299 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:07:41 +00:00
Frederic Riss
13d5de32ac Teach DIEInteger to emit FORM_strp and FORM_ref_addr attributes.
To be used/tested by llvm-dsymutil. (llvm-dsymutil does a 'static' link,
no need for relocations for most things, so it'll just emit raw integers
for most attributes)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231298 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:07:36 +00:00
Frederic Riss
c7e4df3846 Make the DWARFAbbreviationDeclaration::AttributeSpec type public.
It was already exposed through the iterators anyway.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231297 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:07:30 +00:00
David Blaikie
0effae854a Update LangRef for explicit type changes to 'load' instruction
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231296 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:06:14 +00:00
Rafael Espindola
236aa85873 Expand variables when evaluating absolute expressions.
This allows for variables to be used in .size.
This matches gnu AS functionality.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231295 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:03:21 +00:00
David Blaikie
af9251f66b Update LangRef for getelementptr explicit type changes
Here's a rough/first draft - it at least hits the actual textual IR
examples and some of the phrasing. It's probably worth a full pass over,
but I'm not sure how much these docs should reflect the strange
intermediate state we're in anyway.

Totally open to lots of review/feedback/suggestions.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231294 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 22:02:58 +00:00
Sanjay Patel
198cef7d7a don't repeat class / function / variable names in comments; NFC
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231292 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 21:49:03 +00:00
Paul Robinson
4ceab42509 Support standard DWARF TLS opcode; Darwin and PS4 use it.
Differential Revision: http://reviews.llvm.org/D8018


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231286 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 20:55:11 +00:00
Nemanja Ivanovic
b69d556c37 Add LLVM support for PPC cryptography builtins
Review: http://reviews.llvm.org/D7955


git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231285 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 20:44:33 +00:00
Reid Kleckner
c56226c6d1 Try to satisfy sanitizer lint check
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231284 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 20:38:59 +00:00
David Blaikie
5109ec06a6 Add a FIXME for PR22796, broken ordering of ClassInfo in TableGen
As discussed (at length) in code review of r222935, with Duncan.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231282 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 19:56:44 +00:00
Rafael Espindola
671944ff8f Fix the build of the gold-plugin and examples.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231279 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 19:15:29 +00:00
Reid Kleckner
b34736c630 Add missing <atomic> include to PassRegistry.h
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231277 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 19:06:17 +00:00
Erik Eckstein
953be88190 Add a lock() function in PassRegistry to speed up multi-thread synchronization.
When calling lock() after all passes are registered, the PassRegistry doesn't need a mutex anymore to look up passes.
This speeds up multithreaded llvm execution by ~5% (tested with 4 threads).
In an asserts build of llvm this has an even bigger impact.

Note that it's not required to use the lock function.



git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231276 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 18:57:11 +00:00
David Blaikie
be61f2b51c Recommit r231221: "Devirtualize ~parser<T> by making it protected in base classes and making derived classes final"
Reverted in r231254 due to a self-hosting crash of Clang (see Clang
PR22793). Workaround the crash by using {} instead of = default to
define a dtor.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231274 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 18:52:32 +00:00
Rafael Espindola
c90e7f79ca Bring r231132 back with a fix.
The issue was that we were always printing the remarks. Fix that and add a test
showing that it prints nothing if -pass-remarks is not given.

Original message:
Correctly handle -pass-remarks in the gold plugin.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231273 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 18:51:45 +00:00
Mehdi Amini
c94da20917 Make DataLayout Non-Optional in the Module
Summary:
DataLayout keeps the string used for its creation.

As a side effect it is no longer needed in the Module.
This is "almost" NFC, the string is no longer
canonicalized, you can't rely on two "equals" DataLayout
having the same string returned by getStringRepresentation().

Get rid of DataLayoutPass: the DataLayout is in the Module

The DataLayout is "per-module", let's enforce this by not
duplicating it more than necessary.
One more step toward non-optionality of the DataLayout in the
module.

Make DataLayout Non-Optional in the Module

Module->getDataLayout() will never returns nullptr anymore.

Reviewers: echristo

Subscribers: resistor, llvm-commits, jholewinski

Differential Revision: http://reviews.llvm.org/D7992

From: Mehdi Amini <mehdi.amini@apple.com>

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231270 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 18:43:29 +00:00
Reid Kleckner
2337c1fd83 Revert "unique_ptrify ValID::ConstantStructElts"
This reverts r231200 and r231204. The second one added an explicit move
ctor for MSVC.

This change broke the clang-cl self-host due to weirdness in MSVC's
implementation of std::map::insert. Somehow we lost our rvalue ref-ness
when going through variadic placement new:

  template <class _Objty, class... _Types>
  void construct(_Objty *_Ptr,
                 _Types &&... _Args) { // construct _Objty(_Types...) at _Ptr
    ::new ((void *)_Ptr) _Objty(_STD forward<_Types>(_Args)...);
  }

For some reason, Clang decided to call the deleted std::pair copy
constructor at this point. Needs further investigation, once I can
build.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231269 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 18:31:10 +00:00
Wei Mi
a2b8275694 Revert the test commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231264 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 17:44:22 +00:00
Wei Mi
f1bc8c34cf Test commit. It will be reverted in the next commit.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231262 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 17:41:17 +00:00
Adrian Prantl
2e74ddea3a Update the out-of-date dwarf expressions in these testcases.
git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@231261 91177308-0d34-0410-b5e6-96231b3b80d8
2015-03-04 17:39:59 +00:00