OSDN Git Service

android-x86/external-llvm.git
8 years agoSample profile summary cleanup
Easwaran Raman [Mon, 28 Mar 2016 23:14:29 +0000 (23:14 +0000)]
Sample profile summary cleanup

Replace references to MaxHeadSamples with MaxFunctionCount

Differential Revision: http://reviews.llvm.org/D18522

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264686 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[WebAssembly] Remove duplicate disabling of passes
Derek Schuff [Mon, 28 Mar 2016 22:52:20 +0000 (22:52 +0000)]
[WebAssembly] Remove duplicate disabling of passes

Also put all the disabled passes together

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264684 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PowerPC] Clarify a comment in PPCTTI about vector loads
Hal Finkel [Mon, 28 Mar 2016 22:39:35 +0000 (22:39 +0000)]
[PowerPC] Clarify a comment in PPCTTI about vector loads

This should say that we could do unaligned vector loads on the P7 using VSX
instructions, not that we should.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264683 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[docs] Corrections w.r.t V2 of the coverage mapping format
Vedant Kumar [Mon, 28 Mar 2016 22:16:01 +0000 (22:16 +0000)]
[docs] Corrections w.r.t V2 of the coverage mapping format

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264679 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PGO] More comments how function pointers for indirect calls are mapped
Adam Nemet [Mon, 28 Mar 2016 22:15:15 +0000 (22:15 +0000)]
[PGO] More comments how function pointers for indirect calls are mapped
to function names

Summary:
Hopefully this will make it easier for the next person to figure all
this out...

Reviewers: bogner, davidxl

Subscribers: davidxl, cfe-commits

Differential Revision: http://reviews.llvm.org/D18489

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264678 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoregenerate checks
Sanjay Patel [Mon, 28 Mar 2016 22:12:21 +0000 (22:12 +0000)]
regenerate checks

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264677 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix checks: *_DAG -> *-DAG
Sanjay Patel [Mon, 28 Mar 2016 22:11:06 +0000 (22:11 +0000)]
fix checks: *_DAG -> *-DAG

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264676 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Coverage] Fix the expected counts in instrprof-comdat.h
Vedant Kumar [Mon, 28 Mar 2016 22:10:40 +0000 (22:10 +0000)]
[Coverage] Fix the expected counts in instrprof-comdat.h

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264675 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix CHECK_NEXT -> CHECK-NEXT
Sanjay Patel [Mon, 28 Mar 2016 22:03:07 +0000 (22:03 +0000)]
fix CHECK_NEXT -> CHECK-NEXT

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264674 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix CHECK_DAG -> CHECK-DAG
Sanjay Patel [Mon, 28 Mar 2016 22:00:38 +0000 (22:00 +0000)]
fix CHECK_DAG -> CHECK-DAG

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264673 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix CHECK_NEXT -> CHECK-NEXT
Sanjay Patel [Mon, 28 Mar 2016 21:58:27 +0000 (21:58 +0000)]
fix CHECK_NEXT -> CHECK-NEXT

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264672 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix CHECK_LABEL -> CHECK-LABEL
Sanjay Patel [Mon, 28 Mar 2016 21:56:48 +0000 (21:56 +0000)]
fix CHECK_LABEL -> CHECK-LABEL

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264671 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agotrailing whitespace
Sanjay Patel [Mon, 28 Mar 2016 21:52:53 +0000 (21:52 +0000)]
trailing whitespace

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264670 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoMove use of scoped enum in MachineFunctionProperties after its declaration
Derek Schuff [Mon, 28 Mar 2016 21:38:35 +0000 (21:38 +0000)]
Move use of scoped enum in MachineFunctionProperties after its declaration

This should fix the build with MSVC 2013

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264668 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRemove personality for declarations in CloneModule.
Evgeniy Stepanov [Mon, 28 Mar 2016 21:37:02 +0000 (21:37 +0000)]
Remove personality for declarations in CloneModule.

Personality is copied as part of copyFunctionAttributes, but it is
invalid on a declaration. Remove the personality attribute it the
function body is not cloned.

Also add a verifier run over output modules in the llvm-split tool.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264667 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[X86][SSE] Vectorize a bit (AND/XOR/OR) op if a BUILD_VECTOR has the same op for...
Simon Pilgrim [Mon, 28 Mar 2016 21:33:52 +0000 (21:33 +0000)]
[X86][SSE] Vectorize a bit (AND/XOR/OR) op if a BUILD_VECTOR has the same op for all their scalar elements.

If all a BUILD_VECTOR's source elements are the same bit (AND/XOR/OR) operation type and each has one constant operand, lower to a pair of BUILD_VECTOR and just apply the bit operation to the vectors.

The constant operands will form a constant vector meaning that we still only have a single BUILD_VECTOR to lower and we will have replaced all the scalarized operations with a single SSE equivalent.

Its not in our interest to start make a general purpose vectorizer from this, but I'm seeing enough of these scalar bit operations from the later legalization/scalarization stages to support them at least.

Differential Revision: http://reviews.llvm.org/D18492

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264666 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agodocs: Fix footnote after r260042.
Nico Weber [Mon, 28 Mar 2016 21:24:46 +0000 (21:24 +0000)]
docs: Fix footnote after r260042.

r260042 removed a footnote referring to autoconf, but it left
around one item still referring to that footnote (libtool), and
it didn't renumber the later footnote reference.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264663 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agofix CHECK_NEXT -> CHECK-NEXT
Sanjay Patel [Mon, 28 Mar 2016 21:14:24 +0000 (21:14 +0000)]
fix CHECK_NEXT -> CHECK-NEXT

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264661 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoReapply (2x) "[PGO] Fix name encoding for ObjC-like functions"
Vedant Kumar [Mon, 28 Mar 2016 21:06:42 +0000 (21:06 +0000)]
Reapply (2x) "[PGO] Fix name encoding for ObjC-like functions"

Function names in ObjC can have spaces in them. This interacts poorly
with name compression, which uses spaces to separate PGO names. Fix the
issue by using a different separator and update a test.

I chose "\01" as the separator because 1) it's non-printable, 2) we
strip it from PGO names, and 3) it's the next natural choice once "\00"
is discarded (that one's overloaded).

What's changed since the original commit?

- I fixed up the covmap-V2 binary format tests using a linux VM.
- I weakened the CHECK lines in instrprof-comdat.h to account for the
  fact that there have been bugfixes to clang coverage. These will be
  fixed up in a follow-up.
- I added an assert to make sure we don't get bitten by this again.
- I constructed the c-general.profraw file without name compression
  enabled to appease some bots.

Differential Revision: http://reviews.llvm.org/D18516

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264658 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoAdd an IR Verifier check for orphaned DICompileUnits.
Adrian Prantl [Mon, 28 Mar 2016 21:06:26 +0000 (21:06 +0000)]
Add an IR Verifier check for orphaned DICompileUnits.
A DICompileUnit that is not listed in llvm.dbg.cu will cause assertion
failures and/or crashes in the backend. The Verifier should reject this.

rdar://problem/25369499

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264657 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[LVers] Change CHECK_LABEL to CHECK-LABEL (underscore->dash)
Adam Nemet [Mon, 28 Mar 2016 21:04:13 +0000 (21:04 +0000)]
[LVers] Change CHECK_LABEL to CHECK-LABEL (underscore->dash)

Thanks to Sanjoy for catching this.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264656 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agodocs: Try to remove weird linebreak from generated html.
Nico Weber [Mon, 28 Mar 2016 21:00:21 +0000 (21:00 +0000)]
docs: Try to remove weird linebreak from generated html.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264654 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[asan] Fix testcase for r264645
Ryan Govostes [Mon, 28 Mar 2016 20:42:56 +0000 (20:42 +0000)]
[asan] Fix testcase for r264645

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264652 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoExplicitly test BitstreamReader::setArtificialByteLimit, NFC
Duncan P. N. Exon Smith [Mon, 28 Mar 2016 20:39:41 +0000 (20:39 +0000)]
Explicitly test BitstreamReader::setArtificialByteLimit, NFC

Explicitly check that artificial byte limit is rounded correctly by
exposing BitstreamReader::Size through a new accessor, getSizeIfKnown.

The original code for rounding (from r264547) wasn't obviously correct,
and even though r264623 cleaned it up (by calling llvm::alignTo) I think
it's worth testing.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264650 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoHandle section vs global name conflict.
Evgeniy Stepanov [Mon, 28 Mar 2016 20:36:28 +0000 (20:36 +0000)]
Handle section vs global name conflict.

This is a fix for PR26941.

When there is both a section and a global definition with the same
name, the global wins.

Section symbols are not added to the symbol table; section references
are left undefined and fixed up in the object writer unless they've
been satisfied by some other definition.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264649 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agodocs: Update Ninja link, also fix link syntax.
Nico Weber [Mon, 28 Mar 2016 20:32:38 +0000 (20:32 +0000)]
docs: Update Ninja link, also fix link syntax.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264648 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRemove accidentally duplicated test
Duncan P. N. Exon Smith [Mon, 28 Mar 2016 20:30:15 +0000 (20:30 +0000)]
Remove accidentally duplicated test

This was identical to setArtificialByteLimitNotWordBoundary.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264646 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[asan] Support dead code stripping on Mach-O platforms
Ryan Govostes [Mon, 28 Mar 2016 20:28:57 +0000 (20:28 +0000)]
[asan] Support dead code stripping on Mach-O platforms

On OS X El Capitan and iOS 9, the linker supports a new section
attribute, live_support, which allows dead stripping to remove dead
globals along with the ASAN metadata about them.

With this change __asan_global structures are emitted in a new
__DATA,__asan_globals section on Darwin.

Additionally, there is a __DATA,__asan_liveness section with the
live_support attribute. Each entry in this section is simply a tuple
that binds together the liveness of a global variable and its ASAN
metadata structure. Thus the metadata structure will be alive if and
only if the global it references is also alive.

Review: http://reviews.llvm.org/D16737

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264645 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRevert "Reapply "[PGO] Fix name encoding for ObjC-like functions""
Vedant Kumar [Mon, 28 Mar 2016 20:20:40 +0000 (20:20 +0000)]
Revert "Reapply "[PGO] Fix name encoding for ObjC-like functions""

This reverts commit r264641 to investigate why c-general.test is failing
on the bots.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264643 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoReapply "[PGO] Fix name encoding for ObjC-like functions"
Vedant Kumar [Mon, 28 Mar 2016 20:12:07 +0000 (20:12 +0000)]
Reapply "[PGO] Fix name encoding for ObjC-like functions"

Function names in ObjC can have spaces in them. This interacts poorly
with name compression, which uses spaces to separate PGO names. Fix the
issue by using a different separator and update a test.

I chose "\01" as the separator because 1) it's non-printable, 2) we
strip it from PGO names, and 3) it's the next natural choice once "\00"
is discarded (that one's overloaded).

This reverts the revert commit beaf3d18. What's changed?

- I fixed up the covmap-V2 binary format tests using a linux VM.
- I updated the expected counts in instrprof-comdat.h to account for
  the fact that there have been bugfixes to clang coverage.
- I added an assert to make sure we don't get bitten by this again.

Differential Revision: http://reviews.llvm.org/D18516

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264641 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoSimplify how we represent relocation iterators.
Rafael Espindola [Mon, 28 Mar 2016 19:23:51 +0000 (19:23 +0000)]
Simplify how we represent relocation iterators.

Instead of using a bit to detect if they are "dynamic", just look at
sh_link.

This is a simplification on its own, and will help with using
llvm-objdump in dynamic objects.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264624 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoUse 'alignTo' instead of re-inventing it. NFC
Douglas Katzman [Mon, 28 Mar 2016 19:23:23 +0000 (19:23 +0000)]
Use 'alignTo' instead of re-inventing it. NFC

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264623 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoSubmitted new file with wrong line endings. Correcting...
Mike Spertus [Mon, 28 Mar 2016 19:06:17 +0000 (19:06 +0000)]
Submitted new file with wrong line endings. Correcting...

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264620 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoProfile summary cleanup.
Easwaran Raman [Mon, 28 Mar 2016 18:58:05 +0000 (18:58 +0000)]
Profile summary cleanup.

Differential Revision: http://reviews.llvm.org/D18468

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264619 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoFix builds on Windows broken in r264598.
Eugene Zelenko [Mon, 28 Mar 2016 18:54:58 +0000 (18:54 +0000)]
Fix builds on Windows broken in r264598.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264618 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PGO] Update r264611 to use C-style comment
Adam Nemet [Mon, 28 Mar 2016 18:45:30 +0000 (18:45 +0000)]
[PGO] Update r264611 to use C-style comment

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264614 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PGO] Comment how function pointers for indirect calls are mapped to function names
Adam Nemet [Mon, 28 Mar 2016 18:27:44 +0000 (18:27 +0000)]
[PGO] Comment how function pointers for indirect calls are mapped to function names

Summary:
Hopefully this will make it easier for the next person to figure all
this out...

Reviewers: bogner, davidxl

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D18490

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264611 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agodos2unix CMakeLists.txt
Reid Kleckner [Mon, 28 Mar 2016 18:19:32 +0000 (18:19 +0000)]
dos2unix CMakeLists.txt

The sign now reads "0 days since the last line ending confusion"

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264609 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoMIRParser: Add %subreg.xxx syntax for subregister index operands
Matthias Braun [Mon, 28 Mar 2016 18:18:46 +0000 (18:18 +0000)]
MIRParser: Add %subreg.xxx syntax for subregister index operands

Differential Revision: http://reviews.llvm.org/D18279

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264608 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoCodeGen: Correct specification of PHI nodes
Matthias Braun [Mon, 28 Mar 2016 18:18:41 +0000 (18:18 +0000)]
CodeGen: Correct specification of PHI nodes

They do have a def machine operand.

Fixing the definition is necessary for an upcoming patch.

Differential Revision: http://reviews.llvm.org/D18384

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264607 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[AArch64] Do not lower scalar sdiv/udiv to a shifts + mul sequence when optimizing...
Haicheng Wu [Mon, 28 Mar 2016 18:17:07 +0000 (18:17 +0000)]
[AArch64] Do not lower scalar sdiv/udiv to a shifts + mul sequence when optimizing for minsize

Mimic what x86 does when optimizing sdiv/udiv for minsize.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264606 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRevert "[SimlifyCFG] Prevent passes from destroying canonical loop structure, especia...
Reid Kleckner [Mon, 28 Mar 2016 18:07:40 +0000 (18:07 +0000)]
Revert "[SimlifyCFG] Prevent passes from destroying canonical loop structure, especially for nested loops"

This reverts commit r264596.

It does not compile.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264604 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoForgot to commit this file in revision 264601
Mike Spertus [Mon, 28 Mar 2016 17:59:40 +0000 (17:59 +0000)]
Forgot to commit this file in revision 264601

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264602 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoUse VS2015 Project Support for Natvis to eliminate the need to manually install natvi...
Mike Spertus [Mon, 28 Mar 2016 17:58:38 +0000 (17:58 +0000)]
Use VS2015 Project Support for Natvis to eliminate the need to manually install natvis files

When using Visual Studio 2015, cmake now puts the native visualizers in llvm.sln, so the developer automatically sees custom visualizations.
Much thanks to ariccio who provided extensive help on this change. (manual installation still needed on VS2013)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264601 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PowerPC] On the A2, popcnt[dw] are very slow
Hal Finkel [Mon, 28 Mar 2016 17:52:08 +0000 (17:52 +0000)]
[PowerPC] On the A2, popcnt[dw] are very slow

The A2 cores support the popcntw/popcntd instructions, but they're microcoded,
and slower than our default software emulation. Specifically, popcnt[dw] take
approximately 74 cycles, whereas our software emulation takes only 24-28
cycles.

I've added a new target feature to indicate a slow popcnt[dw], instead of just
removing the existing target feature from the a2/a2q processor models, because:
  1. This allows us to return more accurate information via the TTI interface
     (I recognize that this currently makes no practical difference)
  2. Is hopefully easier to understand (it allows the core's features to match
     its manual while still having the desired effect).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264600 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRemove else after return
David Blaikie [Mon, 28 Mar 2016 17:45:48 +0000 (17:45 +0000)]
Remove else after return

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264599 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoFix Clang-tidy modernize-deprecated-headers warnings in some files; other minor fixes.
Eugene Zelenko [Mon, 28 Mar 2016 17:40:08 +0000 (17:40 +0000)]
Fix Clang-tidy modernize-deprecated-headers warnings in some files; other minor fixes.

Differential revision: http://reviews.llvm.org/D18469

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264598 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[SimlifyCFG] Prevent passes from destroying canonical loop structure, especially...
Hyojin Sung [Mon, 28 Mar 2016 17:22:25 +0000 (17:22 +0000)]
[SimlifyCFG] Prevent passes from destroying canonical loop structure, especially for nested loops

When eliminating or merging almost empty basic blocks, the existence of non-trivial PHI nodes
is currently used to recognize potential loops of which the block is the header and keep the block.
However, the current algorithm fails if the loops' exit condition is evaluated only with volatile
values hence no PHI nodes in the header. Especially when such a loop is an outer loop of a nested
loop, the loop is collapsed into a single loop which prevent later optimizations from being
applied (e.g., transforming nested loops into simplified forms and loop vectorization).

The patch augments the existing PHI node-based check by adding a pre-test if the BB actually
belongs to a set of loop headers and not eliminating it if yes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264596 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[llvm-readobj] NFC Replace case by macros for PT_* enums
Hemant Kulkarni [Mon, 28 Mar 2016 17:20:23 +0000 (17:20 +0000)]
[llvm-readobj] NFC Replace case by macros for PT_* enums

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264595 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PGO] Don't set the function hotness attribute when populating counters
Rong Xu [Mon, 28 Mar 2016 17:08:56 +0000 (17:08 +0000)]
[PGO] Don't set the function hotness attribute when populating counters

Don't set the function hotness attribute on the fly. This changes the CFG
branch probability of the caller function, which leads to inconsistent BB
ordering. This patch moves the attribute setting to a separated loop after
 the counts in all functions are populated.

Fixes PR27024 - PGO instrumentation profile data is not reflected in correct
basic blocks.

Differential Revision: http://reviews.llvm.org/D18491

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264594 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoIntroduce MachineFunctionProperties and the AllVRegsAllocated property
Derek Schuff [Mon, 28 Mar 2016 17:05:30 +0000 (17:05 +0000)]
Introduce MachineFunctionProperties and the AllVRegsAllocated property

MachineFunctionProperties represents a set of properties that a MachineFunction
can have at particular points in time. Existing examples of this idea are
MachineRegisterInfo::isSSA() and MachineRegisterInfo::tracksLiveness() which
will eventually be switched to use this mechanism.
This change introduces the AllVRegsAllocated property; i.e. the property that
all virtual registers have been allocated and there are no VReg operands
left.

With this mechanism, passes can declare that they require a particular property
to be set, or that they set or clear properties by implementing e.g.
MachineFunctionPass::getRequiredProperties(). The MachineFunctionPass base class
verifies that the requirements are met, and handles the setting and clearing
based on the delcarations. Passes can also directly query and update the current
properties of the MF if they want to have conditional behavior.

This change annotates the target-independent post-regalloc passes; future
changes will also annotate target-specific ones.

Reviewers: qcolombet, hfinkel

Differential Revision: http://reviews.llvm.org/D18421

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264593 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[llvm-size] Implement --common option
Hemant Kulkarni [Mon, 28 Mar 2016 16:48:10 +0000 (16:48 +0000)]
[llvm-size] Implement --common option

Differential Revision: http://reviews.llvm.org/D16820

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264591 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRevert "[PGO] Fix name encoding for ObjC-like functions"
Vedant Kumar [Mon, 28 Mar 2016 16:14:07 +0000 (16:14 +0000)]
Revert "[PGO] Fix name encoding for ObjC-like functions"

This reverts commit r264587. Reverting to investigate 6 unexpected
failures on the ppc bot:

http://lab.llvm.org:8011/builders/clang-ppc64be-linux/builds/2822

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264590 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoAMDGPU/SI: Limit load clustering to 16 bytes instead of 4 instructions
Tom Stellard [Mon, 28 Mar 2016 16:10:13 +0000 (16:10 +0000)]
AMDGPU/SI: Limit load clustering to 16 bytes instead of 4 instructions

Summary:
This helps prevent load clustering from drastically increasing register
pressure by trying to cluster 4 SMRDx8 loads together.  The limit of 16
bytes was chosen, because it seems like that was the original intent
of setting the limit to 4 instructions, but more analysis could show
that a different limit is better.

This fixes yields small decreases in register usage with shader-db, but
also helps avoid a large increase in register usage when lane mask
tracking is enabled in the machine scheduler, because lane mask tracking
enables more opportunities for load clustering.

shader-db stats:

2379 shaders in 477 tests
Totals:
SGPRS: 49744 -> 48600 (-2.30 %)
VGPRS: 34120 -> 34076 (-0.13 %)
Code Size: 1282888 -> 1283184 (0.02 %) bytes
LDS: 28 -> 28 (0.00 %) blocks
Scratch: 495616 -> 492544 (-0.62 %) bytes per wave
Max Waves: 6843 -> 6853 (0.15 %)
Wait states: 0 -> 0 (0.00 %)

Reviewers: nhaehnle, arsenm

Subscribers: arsenm, llvm-commits

Differential Revision: http://reviews.llvm.org/D18451

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264589 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[SimplifyLibCalls] Transform printf("%s", "a") -> putchar('a').
Davide Italiano [Mon, 28 Mar 2016 15:54:01 +0000 (15:54 +0000)]
[SimplifyLibCalls] Transform printf("%s", "a") -> putchar('a').

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264588 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PGO] Fix name encoding for ObjC-like functions
Vedant Kumar [Mon, 28 Mar 2016 15:52:08 +0000 (15:52 +0000)]
[PGO] Fix name encoding for ObjC-like functions

Function names in ObjC can have spaces in them. This interacts poorly
with name compression, which uses spaces to separate PGO names. Fix the
issue by using a different separator and update a test.

I chose "\01" as the separator because 1) it's non-printable, 2) we
strip it from PGO names, and 3) it's the next natural choice once "\00"
is discarded (that one's overloaded).

Differential Revision: http://reviews.llvm.org/D18516

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264587 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Coverage] Strip <unknown> from PGO names if no filenames are available
Vedant Kumar [Mon, 28 Mar 2016 15:49:08 +0000 (15:49 +0000)]
[Coverage] Strip <unknown> from PGO names if no filenames are available

Patch suggested by David Li!

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264586 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Hexagon] Improve handling of unaligned vector loads and stores
Krzysztof Parzyszek [Mon, 28 Mar 2016 15:43:03 +0000 (15:43 +0000)]
[Hexagon] Improve handling of unaligned vector loads and stores

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264584 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoNFC: skip FenceInst up-front in AtomicExpandPass.
James Y Knight [Mon, 28 Mar 2016 15:05:30 +0000 (15:05 +0000)]
NFC: skip FenceInst up-front in AtomicExpandPass.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264583 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Hexagon] Only use restore functions for single register at -Oz
Krzysztof Parzyszek [Mon, 28 Mar 2016 14:52:21 +0000 (14:52 +0000)]
[Hexagon] Only use restore functions for single register at -Oz

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264581 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Hexagon] Speed up frame lowering when no optimizations are enabled
Krzysztof Parzyszek [Mon, 28 Mar 2016 14:42:03 +0000 (14:42 +0000)]
[Hexagon] Speed up frame lowering when no optimizations are enabled

- Do not optimize stack slots in optnone functions.
- Get aligned-base register from HexagonMachineFunctionInfo instead of
  looking for ALIGNA instruction in the function's body.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264580 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoSparc: silently ignore .proc assembler directive
Douglas Katzman [Mon, 28 Mar 2016 14:00:11 +0000 (14:00 +0000)]
Sparc: silently ignore .proc assembler directive

Differential Revision: http://reviews.llvm.org/D18463

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264579 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[lanai] Add Lanai backend.
Jacques Pienaar [Mon, 28 Mar 2016 13:09:54 +0000 (13:09 +0000)]
[lanai] Add Lanai backend.

Add the Lanai backend to lib/Target.

General Lanai backend discussion on llvm-dev thread "[RFC] Lanai backend" (http://lists.llvm.org/pipermail/llvm-dev/2016-February/095118.html).

Differential Revision: http://reviews.llvm.org/D17011

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264578 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[SROA] Fix typo in comment
Hal Finkel [Mon, 28 Mar 2016 11:23:21 +0000 (11:23 +0000)]
[SROA] Fix typo in comment

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264573 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoC++11 is required, remove some preprocessor checks for it
Hal Finkel [Mon, 28 Mar 2016 11:13:03 +0000 (11:13 +0000)]
C++11 is required, remove some preprocessor checks for it

We require C++11 to build, so remove a few remaining preprocessor checks for
'__cplusplus >= 201103L'. This should always be true.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264572 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Power9] Implement new altivec instructions: bcd* series
Chuang-Yu Cheng [Mon, 28 Mar 2016 09:04:23 +0000 (09:04 +0000)]
[Power9] Implement new altivec instructions: bcd* series

This patch implements the following altivec instructions:

- Decimal Convert From/to National/Zoned/Signed-QWord:
    bcdcfn. bcdcfz. bcdctn. bcdctz. bcdcfsq. bcdctsq.

- Decimal Copy-Sign/Set-Sign:
    bcdcpsgn. bcdsetsgn.

- Decimal Shift/Unsigned-Shift/Shift-and-Round:
    bcds. bcdus. bcdsr.

- Decimal (Unsigned) Truncate:
    bcdtrunc. bcdutrunc.

Total 13 instructions

Thanks Amehsan's advice! Thanks Kit's great help!
Reviewers: hal, nemanja, kbarton, tjablin, amehsan

http://reviews.llvm.org/D17838

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264568 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Power9] Implement new vsx instructions: insert, extract, test data class, min/max...
Chuang-Yu Cheng [Mon, 28 Mar 2016 08:34:28 +0000 (08:34 +0000)]
[Power9] Implement new vsx instructions: insert, extract, test data class, min/max, reverse, permute, splat

This change implements the following vsx instructions:

- Scalar Insert/Extract
    xsiexpdp xsiexpqp xsxexpdp xsxsigdp xsxexpqp xsxsigqp

- Vector Insert/Extract
    xviexpdp xviexpsp xvxexpdp xvxexpsp xvxsigdp xvxsigsp
    xxextractuw xxinsertw

- Scalar/Vector Test Data Class
    xststdcdp xststdcsp xststdcqp
    xvtstdcdp xvtstdcsp

- Maximum/Minimum
    xsmaxcdp xsmaxjdp
    xsmincdp xsminjdp

- Vector Byte-Reverse/Permute/Splat
    xxbrd xxbrh xxbrq xxbrw
    xxperm xxpermr
    xxspltib

30 instructions

Thanks Nemanja for invaluable discussion! Thanks Kit's great help!
Reviewers: hal, nemanja, kbarton, tjablin, amehsan

http://reviews.llvm.org/D16842

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264567 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoAVX-512: Fixed ICMP instruction selection for i1 operands
Elena Demikhovsky [Mon, 28 Mar 2016 07:47:58 +0000 (07:47 +0000)]
AVX-512: Fixed ICMP instruction selection for i1 operands

ICMP instruction selection fails on SKX and KNL for i1 operand.
I use XOR to resolve:
(A == B) is equivalent to (A xor B) == 0

Differential Revision: http://reviews.llvm.org/D18511

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264566 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Power9] Implement new vsx instructions: quad-precision move, fp-arithmetic
Chuang-Yu Cheng [Mon, 28 Mar 2016 07:38:01 +0000 (07:38 +0000)]
[Power9] Implement new vsx instructions: quad-precision move, fp-arithmetic

This change implements the following vsx instructions:

- quad-precision move
    xscpsgnqp, xsabsqp, xsnegqp, xsnabsqp

- quad-precision fp-arithmetic
    xsaddqp(o) xsdivqp(o) xsmulqp(o) xssqrtqp(o) xssubqp(o)
    xsmaddqp(o) xsmsubqp(o) xsnmaddqp(o) xsnmsubqp(o)

22 instructions

Thanks Nemanja and Kit for careful review and invaluable discussion!
Reviewers: hal, nemanja, kbarton, tjablin, amehsan

http://reviews.llvm.org/D16110

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264565 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agollvm/test/Transforms/FunctionImport/funcimport.ll: -stats REQUIRES +Asserts.
NAKAMURA Takumi [Mon, 28 Mar 2016 02:14:49 +0000 (02:14 +0000)]
llvm/test/Transforms/FunctionImport/funcimport.ll: -stats REQUIRES +Asserts.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264561 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Coverage] Fix the way we load "<unknown>:func" records
Vedant Kumar [Mon, 28 Mar 2016 01:16:12 +0000 (01:16 +0000)]
[Coverage] Fix the way we load "<unknown>:func" records

When emitting coverage mappings for functions with local linkage and an
unknown filename, we use "<unknown>:func" for the PGO function name. The
problem is that we don't strip "<unknown>" from the name when loading
coverage data, like we do for other file names. Fix that and add a test.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264559 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Replace dead code with an assertion, NFC
Duncan P. N. Exon Smith [Mon, 28 Mar 2016 00:03:12 +0000 (00:03 +0000)]
BitcodeWriter: Replace dead code with an assertion, NFC

The caller of ValueEnumerator::EnumerateOperandType never sends in
metadata.  Assert that, and remove the unnecessary logic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264558 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Reuse writeMetadataRecords, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:59:32 +0000 (23:59 +0000)]
BitcodeWriter: Reuse writeMetadataRecords, NFC

Change writeFunctionMetadata to call writeMetadataRecords.  For now
there's no functionality change, but makes it easy to serialize other
types of metadata in the function block in the future.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264557 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Rename some functions for consistency, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:56:04 +0000 (23:56 +0000)]
BitcodeWriter: Rename some functions for consistency, NFC

To match writeMetadataRecords, writeNamedMetadata and
writeMetadataStrings, change:

    WriteModuleMetadata        => writeModuleMetadata
    WriteFunctionLocalMetadata => writeFunctionMetadata
    Write##CLASS               => write##CLASS

The only major change is "FunctionLocal" => "Function".  The point is to
be less specific, in preparation for emitting normal metadata records
inside function metadata blocks (currently we only emit
`LocalAsMetadata` there).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264556 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Split out writeMetadataRecords, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:53:30 +0000 (23:53 +0000)]
BitcodeWriter: Split out writeMetadataRecords, NFC

Besides being a nice cleanup, this is preparation for reusing the code
in function metadata blocks.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264555 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Restructure WriteFunctionLocalMetadata, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:38:36 +0000 (23:38 +0000)]
BitcodeWriter: Restructure WriteFunctionLocalMetadata, NFC

Use an early return to simplify logic.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264554 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcode: Fix MSVC bot failure from r264549
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:36:55 +0000 (23:36 +0000)]
Bitcode: Fix MSVC bot failure from r264549

make_unique => llvm::make_unique

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264553 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Simplify tracking of function-local metadata, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:22:31 +0000 (23:22 +0000)]
BitcodeWriter: Simplify tracking of function-local metadata, NFC

We don't really need a separate vector here; instead, point at a range
inside the main MDs array.  This matches how r264551 references the
ranges of strings and non-strings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264552 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoReapply ~"Bitcode: Collect all MDString records into a single blob"
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:17:54 +0000 (23:17 +0000)]
Reapply ~"Bitcode: Collect all MDString records into a single blob"

Spiritually reapply commit r264409 (reverted in r264410), albeit with a
bit of a redesign.

Firstly, avoid splitting the big blob into multiple chunks of strings.

r264409 imposed an arbitrary limit to avoid a massive allocation on the
shared 'Record' SmallVector.  The bug with that commit only reproduced
when there were more than "chunk-size" strings.  A test for this would
have been useless long-term, since we're liable to adjust the chunk-size
in the future.

Thus, eliminate the motivation for chunk-ing by storing the string sizes
in the blob.  Here's the layout:

    vbr6: # of strings
    vbr6: offset-to-blob
    blob:
       [vbr6]: string lengths
       [char]: concatenated strings

Secondly, make the output of llvm-bcanalyzer readable.

I noticed when debugging r264409 that llvm-bcanalyzer was outputting a
massive blob all in one line.  Past a small number, the strings were
impossible to split in my head, and the lines were way too long.  This
version adds support in llvm-bcanalyzer for pretty-printing.

    <STRINGS abbrevid=4 op0=3 op1=9/> num-strings = 3 {
      'abc'
      'def'
      'ghi'
    }

From the original commit:

Inspired by Mehdi's similar patch, http://reviews.llvm.org/D18342, this
should (a) slightly reduce bitcode size, since there is less record
overhead, and (b) greatly improve reading speed, since blobs are super
cheap to deserialize.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264551 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcodeWriter: Simplify and test writing blobs, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:04:04 +0000 (23:04 +0000)]
BitcodeWriter: Simplify and test writing blobs, NFC

Split helper out of EmitRecordWithAbbrevImpl called emitBlob to reduce
code duplication, and add a few tests for it.

No functionality change intended.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264550 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoSupport: Implement StreamingMemoryObject::getPointer
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 23:00:59 +0000 (23:00 +0000)]
Support: Implement StreamingMemoryObject::getPointer

The implementation is fairly obvious.  This is preparation for using
some blobs in bitcode.

For clarity (and perhaps future-proofing?), I moved the call to
JumpToBit in BitstreamCursor::readRecord ahead of calling
MemoryObject::getPointer, since JumpToBit can theoretically (a) read
bytes, which (b) invalidates the blob pointer.

This isn't strictly necessary the two memory objects we have:

  - The return of RawMemoryObject::getPointer is valid until the memory
    object is destroyed.

  - StreamingMemoryObject::getPointer is valid until the next chunk is
    read from the stream.  Since the JumpToBit call is only going ahead
    to a word boundary, we'll never load another chunk.

However, reordering makes it clear by inspection that the blob returned
by BitstreamCursor::readRecord will be valid.

I added some tests for StreamingMemoryObject::getPointer and
BitstreamCursor::readRecord.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264549 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoSupport: Move StreamingMemoryObject{,Test}.cpp, NFC
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 22:55:19 +0000 (22:55 +0000)]
Support: Move StreamingMemoryObject{,Test}.cpp, NFC

Change the filename to indicate this is a test, rename the tests, move
them into an anonymous namespace, and rename some variables.  All to
match our usual style before making further changes.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264548 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcode: Add SimpleBitstreamCursor::setArtificialByteLimit
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 22:49:32 +0000 (22:49 +0000)]
Bitcode: Add SimpleBitstreamCursor::setArtificialByteLimit

Allow users of SimpleBitstreamCursor to limit the number of bytes
available to the cursor.  This is preparation for instantiating a cursor
that isn't allowed to load more bytes from a StreamingMemoryObject (just
move around the ones already-loaded).

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264547 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcode: Add SimpleBitstreamCursor::getPointerToByte, etc.
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 22:45:25 +0000 (22:45 +0000)]
Bitcode: Add SimpleBitstreamCursor::getPointerToByte, etc.

Add API to SimpleBitstreamCursor to allow users to translate between
byte addresses and pointers.

  - jumpToPointer: move the bit position to a particular pointer.
  - getPointerToByte: get the pointer for a particular byte.
  - getPointerToBit: get the pointer for the byte of the current bit.
  - getCurrentByteNo: convenience function for assertions and tests.

Mainly adds unit tests (getPointerToBit/Byte already has a use), but
also preparation for eventually using jumpToPointer.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264546 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoBitcode: Split out SimpleBitstreamCursor
Duncan P. N. Exon Smith [Sun, 27 Mar 2016 22:40:55 +0000 (22:40 +0000)]
Bitcode: Split out SimpleBitstreamCursor

Split out SimpleBitstreamCursor from BitstreamCursor, which is a
lower-level cursor with no knowledge of bitcode blocks, abbreviations,
or records.  It just knows how to read bits and navigate the stream.

This is mainly organizational, to separate the API for manipulating raw
bits from that for bitcode concepts like Record and Block.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264545 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRevert "isPodLike: more precise"
JF Bastien [Sun, 27 Mar 2016 20:50:05 +0000 (20:50 +0000)]
Revert "isPodLike: more precise"

This reverts commit c45f2afac5d6855a4804456a0f718563dc47ada0.

Looks like it may be causing a failure, I'll revert for now.

                 from
lib/CodeGen/AsmPrinter/DwarfDebug.cpp:14:
/usr/include/c++/4.9.2/bits/stl_pair.h: In instantiation of
                 'std::pair<_T1, _T2>& std::pair<_T1,
                 _T2>::operator=(const std::pair<_T1, _T2>&) [with _T1 =
                 std::unique_ptr<llvm::DwarfTypeUnit>; _T2 = const
                 llvm::DICompositeType*]':

/usr/include/c++/4.9.2/bits/stl_pair.h:160:8: error: use of deleted
function 'std::unique_ptr<_Tp, _Dp>& std::unique_ptr<_Tp,
_Dp>::operator=(const std::unique_ptr<_Tp, _Dp>&) [with _Tp =
llvm::DwarfTypeUnit; _Dp = std::default_delete<llvm::DwarfTypeUnit>]'
  first = __p.first;
        ^

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264544 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoworkaround for an IR variable named %.
Sanjay Patel [Sun, 27 Mar 2016 20:44:35 +0000 (20:44 +0000)]
workaround for an IR variable named %.

(which SimplifyCFG can produce...)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264543 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoadd scrubber for excessive leading whitespace
Sanjay Patel [Sun, 27 Mar 2016 20:43:02 +0000 (20:43 +0000)]
add scrubber for excessive leading whitespace

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264542 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoisPodLike: more precise
JF Bastien [Sun, 27 Mar 2016 20:32:21 +0000 (20:32 +0000)]
isPodLike: more precise

I tried to use isPodLike in:
  http://reviews.llvm.org/D18483

That failed because !is_class is too strict on platforms which don't yet
have is_trivially_copyable. This update tries to make isPodLike smarter
for platforms which don't have is_trivially_copyable, and AFAICT it
Should Just Work on all of them. I'll revert if the bots disagree with
me.

I'll also rename isPodLike to isTriviallyCopyable if this all works out,
since that's what the standard calls it now and one day we'll be rid of
isPodLike.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264541 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoUse DAG check to try to appease bot
Teresa Johnson [Sun, 27 Mar 2016 15:36:43 +0000 (15:36 +0000)]
Use DAG check to try to appease bot

Try to appease
http://bb.pgr.jp/builders/cmake-llvm-x86_64-linux/builds/34772. This was
the only check that didn't use DAG and it wasn't found.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264538 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[ThinLTO] Add optional import message and statistics
Teresa Johnson [Sun, 27 Mar 2016 15:27:30 +0000 (15:27 +0000)]
[ThinLTO] Add optional import message and statistics

Summary:
Add a statistic to count the number of imported functions. Also, add a
new -print-imports option to emit a trace of imported functions, that
works even for an NDEBUG build.

Note that emitOptimizationRemark does not work for the above printing as
it expects a Function object and DebugLoc, neither of which we have
with summary-based importing.

This is part 2 of D18487, the first part was committed separately as
r264536.

Reviewers: joker.eph

Subscribers: llvm-commits, joker.eph

Differential Revision: http://reviews.llvm.org/D18487

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264537 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[ThinLTO] Don't try to import alias unless aliasee can be imported
Teresa Johnson [Sun, 27 Mar 2016 15:01:11 +0000 (15:01 +0000)]
[ThinLTO] Don't try to import alias unless aliasee can be imported

With r264503, aliases are now being added to the GlobalsToImport set
even when their aliasees can't be imported due to their linkage type.
While the importing worked correctly (the aliases imported as
declarations) due to the logic in doImportAsDefinition, there is no
point to adding them to the GlobalsToImport set.

Additionally, with D18487 it was resulting in incorrectly printing a
message indicating that the alias was imported.

To avoid this, delay adding aliases to the GlobalsToImport set until
after the linkage type of the aliasee is checked.

This patch is part of D18487.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264536 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[PowerPC] Map max/minnum intrinsics and fmax/fmin to ISD nodes for CTR-based loop...
Hal Finkel [Sun, 27 Mar 2016 05:40:56 +0000 (05:40 +0000)]
[PowerPC] Map max/minnum intrinsics and fmax/fmin to ISD nodes for CTR-based loop legality

Intrinsic::maxnum and Intrinsic::minnum, along with the associated libc
function calls (fmax[f], etc.) generally map to function calls after lowering.
For some vector types with QPX at least, however, we can legally lower these,
and we don't need to prohibit CTR-based loops on their account.

It turned out, however, that the logic that checked the opcodes associated with
intrinsics was broken (it would set the Opcode variable, but that variable was
later checked only if set for some otherwise-external function call.

This fixes the latter problem and adds the FMAX/MINNUM mappings.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264532 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[Verifier] Reject PHIs using defs from own block.
Michael Kruse [Sat, 26 Mar 2016 23:32:57 +0000 (23:32 +0000)]
[Verifier] Reject PHIs using defs from own block.

Reject the following IR as malformed (assuming that %entry, %next are
not in a loop):

    next:
      %y = phi i32 [ 0, %entry ]
      %x = phi i32 [ %y, %entry ]

Such PHI nodes came up in PR26718. While there was no consensus on
whether or not this is valid IR, most opinions on that bug and in a
discussion on the llvm-dev mailing list tended towards a
"strict interpretation" (term by Joseph Tremoulet) of PHI node uses.
Also, the language reference explicitly states that "the use of each
incoming value is deemed to occur on the edge from the corresponding
predecessor block to the current block" and
`DominatorTree::dominates(Instruction*, Use&)` uses this definition as
well.

For the code mentioned in PR15384, clang does not compile to such PHIs
(anymore?). The test case still hangs when replacing `%tmp6` with `%tmp`
in revisions before r176366 (where PR15384 has been fixed). The
occurrence of %tmp6 therefore was probably unintentional. Its value is
not used except in other PHIs.

Reviewers: majnemer, reames, JosephTremoulet, bkramer, grosser, jdoerfert, kparzysz, sanjoy

Differential Revision: http://reviews.llvm.org/D18443

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264528 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[SimplifyCFG] propagate branch metadata when creating select (PR26636)
Sanjay Patel [Sat, 26 Mar 2016 23:30:50 +0000 (23:30 +0000)]
[SimplifyCFG] propagate branch metadata when creating select (PR26636)

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264527 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agominimize test cases
Sanjay Patel [Sat, 26 Mar 2016 23:09:25 +0000 (23:09 +0000)]
minimize test cases

These are tests for store transforms.
The loads, adds, and geps were irrelevant.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264526 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agollvm-dwp: Include the dwo name (if available) when diagnosing duplicate CU IDs from...
David Blaikie [Sat, 26 Mar 2016 20:32:14 +0000 (20:32 +0000)]
llvm-dwp: Include the dwo name (if available) when diagnosing duplicate CU IDs from dwp input files

If you're building dwps from other dwps, it can be hard to track down a
duplicate CU ID if it comes from two compilations of the same file in
different modes, etc. By including the .dwo path (which is hopefully
more unique than the file path) it can help track down where the
duplicates came from.

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264520 91177308-0d34-0410-b5e6-96231b3b80d8

8 years ago[X86][AVX] Enabled SMUL_LOHI/UMUL_LOHI v8i32 vectors on AVX1 targets
Simon Pilgrim [Sat, 26 Mar 2016 18:32:13 +0000 (18:32 +0000)]
[X86][AVX] Enabled SMUL_LOHI/UMUL_LOHI v8i32 vectors on AVX1 targets

Correct splitting of v8i32 vectors into v4i32 vectors to prevent scalarization

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264517 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoRevert "NFC: static_assert instead of comment"
JF Bastien [Sat, 26 Mar 2016 18:20:02 +0000 (18:20 +0000)]
Revert "NFC: static_assert instead of comment"

This reverts commit fa36fcff16c7d4f78204d6296bf96c3558a4a672.

Causes the following Windows failure:

  C:\Buildbot\Slave\llvm-clang-lld-x86_64-scei-ps4-windows10pro-fast\llvm.src\lib\CodeGen\MachineInstr.cpp(762):
  error C2338: must be trivially copyable to memmove

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264516 91177308-0d34-0410-b5e6-96231b3b80d8

8 years agoNFC: static_assert instead of comment
JF Bastien [Sat, 26 Mar 2016 18:14:27 +0000 (18:14 +0000)]
NFC: static_assert instead of comment

Summary: isPodLike is as close as we have for is_trivially_copyable.

Subscribers: llvm-commits

Differential Revision: http://reviews.llvm.org/D18483

git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@264515 91177308-0d34-0410-b5e6-96231b3b80d8