intel/llvm
1
0
Fork 0
mirror of https://github.com/intel/llvm.git synced 2026-01-25 10:55:58 +08:00
Code Issues Packages Projects Releases Wiki Activity
262,582 Commits 76 Branches 2,120 Tags
a6465bbb726badf86e520d15d29b6a65a2d1007d
Commit Graph

262582 Commits

This Branch
This Branch
All Branches
Author SHA1 Message Date
Rafael Espindola
a6465bbb72 Alternative way to detemplate GotSection. 
GetSection is a template because write calls relocate.

relocate has two parts. The non alloc code really has to be a
template, as it is looking a raw input file data.

The alloc part is only a template because of getSize.

This patch folds the value of getSize early, detemplates
getRelocTargetVA and splits relocate into a templated non alloc case
and a regular function for the alloc case. This has the nice advantage
of making sure we collect all the information we need for relocations
before getting to InputSection::relocateNonAlloc.

Since we know got is alloc, it can just call the function directly and
avoid the template.

llvm-svn: 303355
 2017-05-18 16:45:36 +00:00
Rafael Espindola
246c1c47ea Fix flag to start with 1 << 0. NFC. 
Thanks to Andrew Ng for noticing it.

llvm-svn: 303354
 2017-05-18 16:20:12 +00:00
Krasimir Georgiev
9f5608a862 [clang-format] Fix MatchingOpeningBlockLineIndex computation 
Summary:
Computed line index must be relative to the current 'parent' node, and
thus use CurrentLines instead of Lines.

Without this, a child line's MatchingOpeningBlockLineIndex is out of
range of the parent's list of line, which can cause crash or unexpected
behavior if this field is used in childs.

Contributed by @Typz!

Reviewers: krasimir, djasper

Reviewed By: krasimir

Subscribers: cfe-commits, klimek

Differential Revision: https://reviews.llvm.org/D32524

llvm-svn: 303353
 2017-05-18 15:16:24 +00:00
Zachary Turner
8a7508970a [COFF] Fix interaction between /DEBUG and /PDB 
When /DEBUG is not specified, /PDB should be ignored.  When
/DEBUG is specified, a PDB should be output regardless of
whether or not /PDB is specified.  /PDB just overrides the
default name.

This patch implements this behavior, and adds some tests, while
also removing a dead option /DEBUGPDB which was unused in any
code.

Differential Revision: https://reviews.llvm.org/D33302

llvm-svn: 303352
 2017-05-18 15:15:10 +00:00
Zachary Turner
5a83fb153f Fix some minor issues in PDB parsing library. 
1) Until now I'd never seen a valid PDB where the DBI stream and
   the PDB Stream disagreed on the "Age" field.  Because of that,
   we had code to assert that they matched.  Recently though I was
   given a PDB where they disagreed, so this assumption has proven
   to be incorrect.  Remove this check.

2) We were walking the entire list of hash values for types up front
   and then throwing away the values.  For large PDBs this was a
   significant slow down.  Remove this.

With this patch, I can dump the list of all compilands from a
1.5GB PDB file in just a few seconds.

llvm-svn: 303351
 2017-05-18 15:14:44 +00:00
Simon Dardis
657899bad1 [compiler-rt][cmake] Build unit tests conditionally with _FILE_OFFSET_BITS=64 and _LARGEFILE_SOURCE 
The sanitizer library unit tests for libc can get a different definition
of 'struct stat' to what the sanitizer library is built with for certain
targets.

For MIPS the size element of 'struct stat' is after a macro guarded
explicit padding element.

This patch resolves any possible inconsistency by adding the same
_FILE_OFFSET_BITS=64 and _LARGE_SOURCE with the same
conditions as the sanitizer library to the build flags for the unit tests.

This resolves a recurring build failure on the MIPS buildbots due to
'struct stat' defintion differences.

Reviewers: slthakur

Differential Revision: https://reviews.llvm.org/D33131

llvm-svn: 303350
 2017-05-18 13:19:35 +00:00
Anna Thomas
7bca59152a [JumpThreading] Dont RAUW condition incorrectly 
Summary:
We have a bug when RAUWing the condition if experimental.guard or assumes is a use of that
condition. This is because LazyValueInfo may have used the guards/assumes to identify the
value of the condition at the end of the block. RAUW replaces the uses
at the guard/assume as well as uses before the guard/assume. Both of
these are incorrect.
For now, disable RAUW for conditions and fix the logic as a next
step: https://reviews.llvm.org/D33257

Reviewers: sanjoy, reames, trentxintong

Subscribers: llvm-commits

Differential Revision: https://reviews.llvm.org/D33279

llvm-svn: 303349
 2017-05-18 13:12:18 +00:00
Pavel Labath
a24a3a30d0 Add Status -- llvm::Error glue 
Summary:
This adds functions to convert between llvm::Error and Status classes.
Posix errors in Status are represented as llvm::ECError, and the rest as
llvm::StringError.

For the conversion from Error to Status, ECError is again represented as
a posix error in Status, while other errors are stored as generic errors
and only the string value is preserved.

Reviewers: zturner, jingham

Subscribers: lldb-commits

Differential Revision: https://reviews.llvm.org/D33241

llvm-svn: 303348
 2017-05-18 12:46:50 +00:00
Sam Kolton
ebfdaf7394 [AMDGPU] SDWA operands should not intersect with potential MIs 
Summary:
There should be no intesection between SDWA operands and potential MIs. E.g.:
```
v_and_b32 v0, 0xff, v1 -> src:v1 sel:BYTE_0
v_and_b32 v2, 0xff, v0 -> src:v0 sel:BYTE_0
v_add_u32 v3, v4, v2
```
In that example it is possible that we would fold 2nd instruction into 3rd (v_add_u32_sdwa) and then try to fold 1st instruction into 2nd (that was already destroyed). So if SDWAOperand is also a potential MI then do not apply it.

Reviewers: vpykhtin, arsenm

Subscribers: kzhuravl, wdng, nhaehnle, yaxunl, dstuttard, tpr, t-tye

Differential Revision: https://reviews.llvm.org/D32804

llvm-svn: 303347
 2017-05-18 12:12:03 +00:00
Guy Blank
d19632fa16 [MVT] add v1i1 MVT 
Adds the v1i1 MVT as a preparation for another commit (https://reviews.llvm.org/D32273)

Differential Revision: https://reviews.llvm.org/D32540

llvm-svn: 303346
 2017-05-18 11:29:41 +00:00
Igor Breger
842b5b36ba [GlobalISel][X86] G_ADD/G_SUB vector legalizer/selector support. 
Summary: G_ADD/G_SUB vector legalizer/selector support.

Reviewers: zvi, guyblank

Reviewed By: guyblank

Subscribers: rovka, llvm-commits, kristof.beyls

Differential Revision: https://reviews.llvm.org/D33232

llvm-svn: 303345
 2017-05-18 11:10:56 +00:00
Simon Pilgrim
8d81c22164 Fix 'not all control paths return a value' warning on windows buildbots. 
llvm-svn: 303344
 2017-05-18 10:48:23 +00:00
Alex Lorenz
e1e566d4df [index] Record references to class receivers used in property references 
rdar://32250025

llvm-svn: 303343
 2017-05-18 10:43:11 +00:00
Simon Pilgrim
6bba6068be [X86][AVX512] Add 512-bit vector ctpop costs + tests 
llvm-svn: 303342
 2017-05-18 10:42:34 +00:00
Daniel Sanders
89e9308623 Re-commit: [globalisel][tablegen] Import rules containing intrinsic_wo_chain. 
Summary:
As of this patch, 1018 out of 3938 rules are currently imported.

Depends on D32275

Reviewers: qcolombet, kristof.beyls, rovka, t.p.northover, ab, aditya_nandakumar

Reviewed By: qcolombet

Subscribers: dberris, igorb, llvm-commits

Differential Revision: https://reviews.llvm.org/D32278

The previous commit failed on test-suite/Bitcode/simd_ops/AArch64_halide_runtime.bc
because isImmOperandEqual() assumed MO was a register operand and that's not
always true.

llvm-svn: 303341
 2017-05-18 10:33:36 +00:00
Zvi Rackover
d17d13d2a9 [X86] Add explicit triple to test invocation 
llvm-svn: 303340
 2017-05-18 09:32:56 +00:00
Daniel Jasper
f97310fb7a Revert r302781 and subsequent attempts to disable part of it. 
The Msan unit tests are still broken and by this point, I think we
should start over.

llvm-svn: 303339
 2017-05-18 09:31:37 +00:00
Diana Picus
44514860f4 Fixup r303324 - temporary disable stndup interceptor, due to r302781 being buggy 
r303324 missed one of the tests added by r302781. This commit applies
the same fix as r303324 to the missed test (strndup.cc).

llvm-svn: 303338
 2017-05-18 09:12:27 +00:00
Peter Smith
d54f368e6a [ELF] Support R_ARM_SBREL32 Relocation 
This change adds support for the R_ARM_SBREL32 relocation. The relocation
is a base relative relocation that is produced by clang/llvm when -frwpi
is used. The use case for the -frwpi option is position independent data
for embedded systems that do not have a GOT. With -frwpi all data is
accessed via an offset from a base register (usually r9), where r9 is set
at run time to where the data has been loaded. The base of the data is
known as the static base.

The ARM ABI defines the static base as:
B(S) is the addressing origin of the output segment defining the symbol S.
The origin is not required to be the base address of the segment. For
simplicity we choose to use the base address of the segment.

The ARM procedure call standard only defines a read write variant using
R_ARM_SBREL32 relocations. The read-only data is accessed via pc-relative
offsets from the code, this is implemented in clang as -fropi.

Fixes PR32924

Differential Revision: https://reviews.llvm.org/D33280

llvm-svn: 303337
 2017-05-18 09:12:21 +00:00
Max Kazantsev
627ad0fec3 [SCEV][NFC] Remove duplication of isLoopInvariant code 
Replace two places that duplicate the code of isLoopInvariant method with
the invocation of this method.

Differential Revision: https://reviews.llvm.org/D33313

llvm-svn: 303336
 2017-05-18 08:26:41 +00:00
George Rimar
47f84b1a3c [DWARF] - Simplify RelocVisitor implementation. 
We do not need to store relocation width field.
Patch removes relative code, that simplifies implementation.

Differential revision: https://reviews.llvm.org/D33274

llvm-svn: 303335
 2017-05-18 08:25:11 +00:00
George Rimar
d2aaf5bd37 [lib/Object] - Fix build bot after r303331 "[lib/Object] - Minor API update for llvm::Decompressor". 
Error was:
Decompressor.h:33:28: error: extra qualification ‘llvm::object::Decompressor::’ on member ‘resizeAndDecompress’ [-fpermissive]
   template <class T> Error Decompressor::resizeAndDecompress(T &Out) {

llvm-svn: 303334
 2017-05-18 08:14:43 +00:00
Lama Saba
2ea271b54a [X86] Replace slow LEA instructions in X86 
According to Intel's Optimization Reference Manual for SNB+:
  " For LEA instructions with three source operands and some specific situations, instruction latency has increased to 3 cycles, and must
    dispatch via port 1:
  - LEA that has all three source operands: base, index, and offset
  - LEA that uses base and index registers where the base is EBP, RBP,or R13
  - LEA that uses RIP relative addressing mode
  - LEA that uses 16-bit addressing mode "
  This patch currently handles the first 2 cases only.
 
Differential Revision: https://reviews.llvm.org/D32277

llvm-svn: 303333
 2017-05-18 08:11:50 +00:00
Krasimir Georgiev
994b6c9b8e [clang-format] Make NoLineBreakFormatter respect MustBreakBefore 
Summary:
This patch makes NoLineBreakFormatter to insert a break before tokens where
MustBreakBefore is true.

Reviewers: djasper

Reviewed By: djasper

Subscribers: cfe-commits, klimek

Differential Revision: https://reviews.llvm.org/D33238

llvm-svn: 303332
 2017-05-18 08:07:52 +00:00
George Rimar
f98b9ac5da [lib/Object] - Minor API update for llvm::Decompressor. 
I revisited Decompressor API (issue with it was triggered during D32865 review)
and found it is probably provides more then we really need.

Issue was about next method's signature:

Error decompress(SmallString<32> &Out);
It is too strict. At first I wanted to change it to decompress(SmallVectorImpl<char> &Out),
but then found it is still not flexible because sticks to SmallVector.

During reviews was suggested to use templating to simplify code. Patch do that.

Differential revision: https://reviews.llvm.org/D33200

llvm-svn: 303331
 2017-05-18 08:00:01 +00:00
Krasimir Georgiev
ba6b315ea9 clang-format: fix prefix for doxygen comments after member 
Summary:
Doxygen supports putting documentation blocks after member, by adding
an additional < marker in the comment block. This patch makes sure
this marker is used in lines which are introduced by breaking the
comment.

  int foo; ///< Some very long comment.

becomes:

  int foo; ///< Some very long
           ///< comment.

Contributed by @Typz!

Reviewers: krasimir

Reviewed By: krasimir

Subscribers: djasper, klimek, cfe-commits

Differential Revision: https://reviews.llvm.org/D33282

llvm-svn: 303330
 2017-05-18 07:36:21 +00:00
Serguei Katkov
00211c3faa Fix buildbot failure after rL303327: [BPI] Reduce the probability of unreachable edge to minimal value greater than 0. 
One more test is updated to meet new branch probability for unreachable branches.

llvm-svn: 303329
 2017-05-18 07:20:52 +00:00
Zvi Rackover
c20c6d07cf [X86] Adding tests for scalar bitcasts from vsetcc. NFC. 
llvm-svn: 303328
 2017-05-18 07:04:48 +00:00
Serguei Katkov
ba831f78fd [BPI] Reduce the probability of unreachable edge to minimal value greater than 0 
The probability of edge coming to unreachable block should be as low as possible.
The change reduces the probability to minimal value greater than zero.

The bug https://bugs.llvm.org/show_bug.cgi?id=32214 show the example when
the probability of edge coming to unreachable block is greater than for edge
coming to out of the loop and it causes incorrect loop rotation.

Please note that with this change the behavior of unreachable heuristic is a bit different
than others. Specifically, before this change the sum of probabilities
coming to unreachable blocks have the same weight for all branches
(it was just split over all edges of this block coming to unreachable blocks).
With this change it might be slightly different but not to much due to probability of
taken branch to unreachable block is really small.

Reviewers: chandlerc, sanjoy, vsk, congh, junbuml, davidxl, dexonsmith
Reviewed By: chandlerc, dexonsmith
Subscribers: reames, llvm-commits
Differential Revision: https://reviews.llvm.org/D30633

llvm-svn: 303327
 2017-05-18 06:11:56 +00:00
Akira Hatanaka
b10bff1183 [ThinLTO] Do not assert when adding a module with a different but 
compatible target triple

Currently, an assertion fails in ThinLTOCodeGenerator::addModule when
the target triple of the module being added doesn't match that of the
one stored in TMBuilder. This patch relaxes the constraint and makes
changes to allow target triples that only differ in their version
numbers on Apple platforms, similarly to what r228999 did.

rdar://problem/30133904

Differential Revision: https://reviews.llvm.org/D33291

llvm-svn: 303326
 2017-05-18 03:52:29 +00:00
Alexander Kornienko
fd3e39846d Fix an assertion failure in FormatASTNodeDiagnosticArgument. 
Summary:
The test being added in this patch used to cause an assertion failure:

/build/./bin/clang -cc1 -internal-isystem /build/lib/clang/5.0.0/include -nostdsysteminc -verify -fsyntax-only -std=c++11 -Wshadow-all /src/tools/clang/test/SemaCXX/warn-shadow.cpp
--
Exit Code: 134

Command Output (stderr):
--
clang: /src/tools/clang/lib/AST/ASTDiagnostic.cpp:424: void clang::FormatASTNodeDiagnosticArgument(DiagnosticsEngine::ArgumentKind, intptr_t, llvm::StringRef, llvm::StringRef, ArrayRef<DiagnosticsEngine::ArgumentValue>, SmallVectorImpl<char> &, void *, ArrayRef<intptr_t>): Assertion `isa<NamedDecl>(DC) && "Expected a NamedDecl"' failed.
#0 0x0000000001c7a1b4 PrintStackTraceSignalHandler(void*) (/build/./bin/clang+0x1c7a1b4)
#1 0x0000000001c7a4e6 SignalHandler(int) (/build/./bin/clang+0x1c7a4e6)
#2 0x00007f30880078d0 __restore_rt (/lib/x86_64-linux-gnu/libpthread.so.0+0xf8d0)
#3 0x00007f3087054067 gsignal (/lib/x86_64-linux-gnu/libc.so.6+0x35067)
#4 0x00007f3087055448 abort (/lib/x86_64-linux-gnu/libc.so.6+0x36448)
#5 0x00007f308704d266 (/lib/x86_64-linux-gnu/libc.so.6+0x2e266)
#6 0x00007f308704d312 (/lib/x86_64-linux-gnu/libc.so.6+0x2e312)
#7 0x00000000035b7f22 clang::FormatASTNodeDiagnosticArgument(clang::DiagnosticsEngine::ArgumentKind, long, llvm::StringRef, llvm::StringRef, llvm::ArrayRef<std::pair<clang::DiagnosticsEngine::ArgumentKind, long> >, llvm::SmallVectorImpl<char>&, void*, llvm::ArrayRef<long>) (/build/
./bin/clang+0x35b7f22)
#8 0x0000000001ddbae4 clang::Diagnostic::FormatDiagnostic(char const*, char const*, llvm::SmallVectorImpl<char>&) const (/build/./bin/clang+0x1ddbae4)
#9 0x0000000001ddb323 clang::Diagnostic::FormatDiagnostic(char const*, char const*, llvm::SmallVectorImpl<char>&) const (/build/./bin/clang+0x1ddb323)
#10 0x00000000022878a4 clang::TextDiagnosticBuffer::HandleDiagnostic(clang::DiagnosticsEngine::Level, clang::Diagnostic const&) (/build/./bin/clang+0x22878a4)
#11 0x0000000001ddf387 clang::DiagnosticIDs::ProcessDiag(clang::DiagnosticsEngine&) const (/build/./bin/clang+0x1ddf387)
#12 0x0000000001dd9dea clang::DiagnosticsEngine::EmitCurrentDiagnostic(bool) (/build/./bin/clang+0x1dd9dea)
#13 0x0000000002cad00c clang::Sema::EmitCurrentDiagnostic(unsigned int) (/build/./bin/clang+0x2cad00c)
#14 0x0000000002d91cd2 clang::Sema::CheckShadow(clang::NamedDecl*, clang::NamedDecl*, clang::LookupResult const&) (/build/./bin/clang+0x2d91cd2)

Stack dump:
0.      Program arguments: /build/./bin/clang -cc1 -internal-isystem /build/lib/clang/5.0.0/include -nostdsysteminc -verify -fsyntax-only -std=c++11 -Wshadow-all /src/tools/clang/test/SemaCXX/warn-shadow.cpp
1.      /src/tools/clang/test/SemaCXX/warn-shadow.cpp:214:23: current parser token ';'
2.      /src/tools/clang/test/SemaCXX/warn-shadow.cpp:213:26: parsing function body 'handleLinkageSpec'
3.      /src/tools/clang/test/SemaCXX/warn-shadow.cpp:213:26: in compound statement ('{}')
/build/tools/clang/test/SemaCXX/Output/warn-shadow.cpp.script: line 1: 15595 Aborted                 (core dumped) /build/./bin/clang -cc1 -internal-isystem /build/lib/clang/5.0.0/include -nostdsysteminc -verify -fsyntax-only -std=c++11 -Wshadow-all /src/tools/clang/test/SemaCXX/warn-shadow.cpp

Reviewers: rsmith

Reviewed By: rsmith

Subscribers: krytarowski, cfe-commits

Differential Revision: https://reviews.llvm.org/D33207

llvm-svn: 303325
 2017-05-18 03:02:15 +00:00
Kostya Serebryany
9aa8ef8504 temporary disable stndup interceptor, due to r302781 being buggy 
llvm-svn: 303324
 2017-05-18 03:00:07 +00:00
Davide Italiano
9ae69a75ec [Target/X86] Remove unneeded return. NFCI. 
llvm-svn: 303323
 2017-05-18 02:36:42 +00:00
Richard Smith
54f0440c1f [modules] Switch from inferring owning modules based on source location to 
inferring based on the current module at the point of creation.

This should result in no functional change except when building a preprocessed
module (or more generally when using #pragma clang module begin/end to switch
module in the middle of a file), in which case it allows us to correctly track
the owning module for declarations. We can't map from FileID to module in the
preprocessed module case, since all modules would have the same FileID.

There are still a couple of remaining places that try to infer a module from a
source location; I'll clean those up in follow-up changes.

llvm-svn: 303322
 2017-05-18 02:29:20 +00:00
Alexander Kornienko
6180337fe1 [clang-tidy] Optimize GlobList::contains 
With large lists of checks and large number of warnings GlobList::contains
starts being ridiculously CPU hungry, since it runs regexp match per glob.
Caching results of glob matching in a StringMap significantly speeds up check
filtering even for small GlobLists.

/tmp/q.cc:

void f() {
  int I;
  {int I;}
  {int I;}
  {int I;}
  ... // 200k times
}

Before the patch:

GlobList with 2 entries:
  $ time clang-tidy-old -checks=-*,modernize-use-override /tmp/q.cc -- -Wshadow
  200000 warnings generated.
  Suppressed 200000 warnings (200000 with check filters).

  real    0m3.826s
  user    0m3.176s
  sys     0m0.504s

GlobList with 28 entries:
  $ time clang-tidy-old -checks=-*,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,modernize-use-override /tmp/q.cc -- -Wshadow
  200000 warnings generated.
  Suppressed 200000 warnings (200000 with check filters).

  real    0m5.000s
  user    0m4.744s
  sys     0m0.060s

GlobList with 158 entries:
  $ time clang-tidy-old -checks=-*,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,modernize-use-override /tmp/q.cc -- -Wshadow
  200000 warnings generated.
  Suppressed 200000 warnings (200000 with check filters).

  real    0m13.920s
  user    0m13.636s
  sys     0m0.104s

With the patch runtime is practically independent from the length of the GlobList:
  $ time clang-tidy-new -checks=-*,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,a,b,c,d,e,f,g,h,i,j,k,l,m,n,o,p,q,r,s,t,u,v,w,x,y,z,modernize-use-override /tmp/q.cc -- -Wshadow
  200000 warnings generated.
  Suppressed 200000 warnings (200000 with check filters).

  real    0m2.300s
  user    0m2.104s
  sys     0m0.044s

llvm-svn: 303321
 2017-05-18 01:13:51 +00:00
Craig Topper
704b4fbbf6 [Statistics] Use the new Statistic::updateMax to atomically calculate a maximum value statistic. 
llvm-svn: 303320
 2017-05-18 01:11:52 +00:00
Justin Bogner
be42c4aec8 Update three tests I missed in r302979 and r302990 
llvm-svn: 303319
 2017-05-18 00:58:06 +00:00
Craig Topper
8a950275f7 [Statistics] Add a method to atomically update a statistic that contains a maximum 
Summary:
There are several places in the codebase that try to calculate a maximum value in a Statistic object. We currently do this in one of two ways:

  MaxNumFoo = std::max(MaxNumFoo, NumFoo);

or

  MaxNumFoo = (MaxNumFoo > NumFoo) ? MaxNumFoo : NumFoo;

The first version reads from MaxNumFoo one time and uncontionally rwrites to it. The second version possibly reads it twice depending on the result of the first compare.  But we have no way of knowing if the value was changed by another thread between the reads and the writes.

This patch adds a method to the Statistic object that can ensure that we only store if our value is the max and the previous max didn't change after we read it. If it changed we'll recheck if our value should still be the max or not and try again.

This spawned from an audit I'm trying to do of all places we uses the implicit conversion to unsigned on the Statistics objects. See my previous thread on llvm-dev https://groups.google.com/forum/#!topic/llvm-dev/yfvxiorKrDQ

Reviewers: dberlin, chandlerc, hfinkel, dblaikie

Reviewed By: chandlerc

Subscribers: llvm-commits, sanjoy

Differential Revision: https://reviews.llvm.org/D33301

llvm-svn: 303318
 2017-05-18 00:51:39 +00:00
Nick Lewycky
9add1594d2 The constant expression evaluator should examine function arguments for non-constexpr function calls unless the EvalInfo says to stop. 
llvm-svn: 303317
 2017-05-17 23:56:54 +00:00
Kyle Butt
0cf5b2f88a CodeGen: BlockPlacement: Add Message strings to asserts. NFC 
Add message strings to all the unlabeled asserts in the file.

Differential Revision: https://reviews.llvm.org/D33078

llvm-svn: 303316
 2017-05-17 23:44:41 +00:00
Sanjay Patel
7f4687f164 [InstCombine] add test for xor-of-icmps; NFC 
This is another form of the problem discussed in D32143.

llvm-svn: 303315
 2017-05-17 23:22:52 +00:00
Craig Topper
48187cffe2 [Statistics] Use Statistic::operator+= instead of adding and assigning separately. 
I believe this technically fixes a multithreaded race condition in this code. But my primary concern was as part of looking at removing the ability to treat Statistics like a plain unsigned. There are many weird operations on Statistics in the codebase.

llvm-svn: 303314
 2017-05-17 23:22:10 +00:00
Quentin Colombet
a072d13e54 Revert "[globalisel][tablegen] Import rules containing intrinsic_wo_chain." 
This reverts commit r303259.

This breaks the GISel bot:
http://lab.llvm.org:8080/green/job/Compiler_Verifiers_GlobalISEL/5163/consoleFull#-134276167849ba4694-19c4-4d7e-bec5-911270d8a58c

llvm-svn: 303313
 2017-05-17 23:17:29 +00:00
Sanjay Patel
ba212c241a [InstCombine] handle icmp i1 X, C early to avoid creating an unknown pattern 
The missing optimization for xor-of-icmps still needs to be added, but by
being more efficient (not generating unnecessary logic ops with constants)
we avoid the bug.

See discussion in post-commit comments:
https://reviews.llvm.org/D32143

llvm-svn: 303312
 2017-05-17 22:29:40 +00:00
Reid Kleckner
cde4b3f4e6 Attempt to pacify ASan and UBSan reports in CrashRecovery tests 
llvm-svn: 303311
 2017-05-17 22:23:20 +00:00
Sanjay Patel
3cd38a8d4c [InstCombine] add test for missing icmp bool fold; NFC 
llvm-svn: 303310
 2017-05-17 22:20:02 +00:00
Sanjay Patel
e5747e3cbd [InstCombine] move icmp bool canonicalizations to helper; NFC 
As noted in the post-commit comments in D32143, we should be
catching the constant operand cases sooner to be more efficient
and less likely to expose a missing fold.

llvm-svn: 303309
 2017-05-17 22:15:07 +00:00
Matt Arsenault
2b1f9aa577 AMDGPU: Start defining a calling convention 
Partially implement callee-side for arguments and return values.
byval doesn't work properly, and most likely sret or other on-stack
return values most as well.

llvm-svn: 303308
 2017-05-17 21:56:25 +00:00
Kyle Butt
f6c61ef64d CodeGen: Power: Add lowering for shifts of v1i128. 
When legalizing vector operations on vNi128, they will be split to v1i128
because that is a legal type on ppc64, but then the compiler will crash in
selection dag because it fails to select for these operations. This patch fixes
shift operations. Logical shift right and left shift can be performed in the
vector unit, but algebraic shift right requires being split.

Differential Revision: https://reviews.llvm.org/D32774

llvm-svn: 303307
 2017-05-17 21:54:41 +00:00
Michael Liao
ab12984634 Fix PR33028 
- '-verify-mahcineinstrs' starts to complain allocatable live-in physical
  registers on non-entry or non-landing-pad basic blocks.
- Refactor the XBEGIN translation to define EAX on a dedicated fallback code
  path due to XABORT. Add a pseudo instruction to define EAX explicitly to
  avoid add physical register live-in.

Differential Revision: https://reviews.llvm.org/D33168

llvm-svn: 303306
 2017-05-17 21:48:00 +00:00