Commit Graph

156037 Commits

Author SHA1 Message Date
Peter Collingbourne 03c3324ccd Remove SpecialCaseList::findCategory.
It turned out that I didn't need this for DFSan.

llvm-svn: 188646
2013-08-19 00:24:20 +00:00
Craig Topper 048b883416 Revert r188644 to unbreak buildbots.
llvm-svn: 188645
2013-08-18 20:38:37 +00:00
Craig Topper 260f7e6893 Remove Stmt::operator new that takes a ASTContext*. All callers use the ASTContext& version.
llvm-svn: 188644
2013-08-18 20:03:21 +00:00
Tim Northover 55349a29c6 ARM: make sure we keep inline asm operands tied.
When patching inlineasm nodes to use GPRPair for 64-bit values, we
were dropping the information that two operands were tied, which
effectively broke the live-interval of vregs affected.

llvm-svn: 188643
2013-08-18 18:06:03 +00:00
David Blaikie 25b2085490 Revert "Revert "DebugInfo: Omit debug info for dynamic classes in TUs that do not have the vtable for that class""
This reverts commit r188600.

r188640/r188639 fixed the root cause of the crash-on-valid that r188600
originally introduced. This now appears to bootstrap debug clang
successfully to the best of my testing.

llvm-svn: 188642
2013-08-18 17:59:12 +00:00
Craig Topper 5a050016bb Remove throw() from Stmt::operator new so the compiler will omit the null check on the result since ASTContext allocator won't return null.
llvm-svn: 188641
2013-08-18 17:45:38 +00:00
David Blaikie d278589329 DebugInfo: Avoid duplicating types that may be created during the process of creating their context
A partner to r188639, this is a somewhat heavy-handed fix to the general
issue, since even after that prior change the issue does still
unavoidably arise with template parameters (see test case).

There are other ways we could consider addressing this (see FIXME).

llvm-svn: 188640
2013-08-18 17:36:19 +00:00
David Blaikie adfbf993d6 DebugInfo: Don't emit vbase 'containing types' for context chain limited types
Possible minor reduction in debug info & avoid some cases where creating
a context chain could lead to the type the context chain is being
created for, being created. (this is still possible with template
parameters - tests/fixes/improvements to follow)

llvm-svn: 188639
2013-08-18 16:55:33 +00:00
NAKAMURA Takumi c96c699a16 InitHeaderSearch.cpp: [Cygwin] Add 4.7.3.
llvm-svn: 188638
2013-08-18 15:03:03 +00:00
Elena Demikhovsky 3ce8dbbac2 AVX-512: Added VMOVD, VMOVQ, VMOVSS, VMOVSD instructions.
llvm-svn: 188637
2013-08-18 13:08:57 +00:00
Craig Topper 3793291d50 Make expression allocation methods use a 'const' reference to the ASTContext since the underlying operator new only needs a const reference.
llvm-svn: 188636
2013-08-18 10:09:15 +00:00
Chandler Carruth 6b8d17bf35 It appears that glibc 2.13 has the same thread descriptor size as 2.11,
bump up the inflection point to 2.14. If someone can tell me how to
actually figure out value for this, that would be awesome.

Anyways, this takes me to one ASan failure, one LSan failure, and three
TSan failures for 'check-all' on Linux.

llvm-svn: 188635
2013-08-18 10:04:51 +00:00
Craig Topper e6861c9ce5 Make more of the lowering helpers static. Also use MVT instead of EVT in a couple places.
llvm-svn: 188629
2013-08-18 08:53:01 +00:00
Chandler Carruth 022a3d4e88 Revert r188369 (and r188455) which breaks the build on at least some
Linux systems. I've replied on the original commit thread with details
of the problem.

llvm-svn: 188628
2013-08-18 08:47:06 +00:00
Dmitri Gribenko 8cac835723 docs: command guide: cleanups, no text changes
llvm-svn: 188627
2013-08-18 08:32:32 +00:00
Dmitri Gribenko 8b2a3d1fea Remove unused stdio.h includes
llvm-svn: 188626
2013-08-18 08:29:51 +00:00
Dmitri Gribenko 96933de7f5 libclang: cleanup unused includes in public header
llvm-svn: 188625
2013-08-18 07:57:43 +00:00
Logan Chien f166ea6715 Rename "polly-test" in the document and script.
The Makefile rule "polly-test" has been renamed to
"check-polly" in r182171.  This CL updates the document and
the automatic build script.

llvm-svn: 188624
2013-08-18 07:33:28 +00:00
Chandler Carruth 7deaae76ec Fix a use-after-free found in libclang when doing code completion. The
loop processing the candidates can cause new declerations to be added to
the context, invalidating lookup_result. To avoid that, make a copy of
the list of declarations to iterate over.

I don't have a way to check in a test case for this as it involves
a giant pile of source code and a generated PCH file used to accelerate
code completion, all of this running under ASan.

llvm-svn: 188623
2013-08-18 07:20:52 +00:00
David Blaikie 5043815a60 DebugInfo: don't require full definitions for friend classes
Fixes a crash-on-valid introduced by r188486 (which should've occurred
earlier but for a blatant bug where calling createFwdDecl from the
requireCompleteType callback was useless under -flimit-debug-info and we
were just getting lucky with other later callbacks requiring the type
anyway).

llvm-svn: 188622
2013-08-18 04:50:23 +00:00
NAKAMURA Takumi 6dd203288d {c-index-test|c-arcmt-test}/Makefile: Update USEDLIBS along corresponding to clangIndex.
llvm-svn: 188621
2013-08-18 03:48:57 +00:00
NAKAMURA Takumi ee4763c248 Makefile.rules: Avoid -fomit-frame-pointer also on cygwin due to PR14646.
llvm-svn: 188620
2013-08-18 03:38:40 +00:00
NAKAMURA Takumi 1ade5e9574 Makefile.rules: Simplify nested if(s) on OmitFramePointer.
llvm-svn: 188619
2013-08-18 02:46:21 +00:00
Chandler Carruth 46919be781 Port the detection of zlib from the main autoconf system to the sample
project's autoconf. This is the last of the missing optional checks used
by libSupport that seemed to be missing from the sample project, but
I could easily have missed some as this was done by inspection when
Craig asked me to add the terminfo support.

llvm-svn: 188618
2013-08-18 01:55:15 +00:00
Chandler Carruth 718d3f5e76 Add support for linking librt and using clock_gettime to the sample
autoconf setup.

llvm-svn: 188617
2013-08-18 01:46:34 +00:00
Chandler Carruth 03ee1a4c45 Update the sample project autoconf setup to include support for
detecting terminfo. Requested by Craig Topper, and probably should be
done much more systematically.

llvm-svn: 188616
2013-08-18 01:43:57 +00:00
Chandler Carruth 67ff8b7185 Go through the really awkward dance required to delete the memory
allocated by setupterm. Without this, some folks are seeing leaked
memory whenever this routine is called more than once. Thanks to Craig
Topper for the report.

llvm-svn: 188615
2013-08-18 01:20:32 +00:00
Hal Finkel 3f5279cc26 Fix SCEVExpander creating distinct duplicate PHI entries
This fixes SCEVExpander so that it does not create multiple distinct induction
variables for duplicate PHI entries. Specifically, given some code like this:

do.body6:                                         ; preds = %do.body6, %do.body6, %if.then5
  %end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ]
...

Note that it is legal to have multiple entries for a basic block so long as the
associated value is the same. So the above input is okay, but expanding an
AddRec in this loop could produce code like this:

do.body6:                                         ; preds = %do.body6, %do.body6, %if.then5
  %indvar = phi i64 [ %indvar.next, %do.body6 ], [ %indvar.next1, %do.body6 ], [ 0, %if.then5 ]
  %end.0 = phi i8* [ undef, %if.then5 ], [ %incdec.ptr, %do.body6 ], [ %incdec.ptr, %do.body6 ]
...
  %indvar.next = add i64 %indvar, 1
  %indvar.next1 = add i64 %indvar, 1

And this is not legal because there are two PHI entries for %do.body6 each with
a distinct value.

Unfortunately, I don't have an in-tree test case.

llvm-svn: 188614
2013-08-18 00:16:23 +00:00
Joerg Sonnenberger 9dce26947e Fix endian conditions on Solaris. From Alexander Esilevich.
llvm-svn: 188613
2013-08-17 22:32:10 +00:00
David Blaikie 949939cdbb PR16927: Don't assert (or, previously, skip) static data members of enumeration type
llvm-svn: 188612
2013-08-17 20:01:53 +00:00
Guillaume Papin 8d951fbc18 Revert "cpp11-migrate: Add a class to support include directives modifications"
This reverts commit r188610.

Issue with the absolute include paths not found in the unit tests on the Windows
bots. Needs investigation.

llvm-svn: 188611
2013-08-17 19:04:56 +00:00
Guillaume Papin 114d035521 cpp11-migrate: Add a class to support include directives modifications
The IncludeDirectives class helps with detecting and modifying #include
directives. For now it allows the users to add angled-includes in a source file.
This is a start for this class that will evolve in the future to add more
functionality.

llvm-svn: 188610
2013-08-17 18:48:31 +00:00
Juergen Ributzka 2c2dbf4542 Fix the name and the type of the argument for intrinisc
_mm256_broadcastsi128_si256 to align with the Intel documentation.

This fixes bug PR 16581 and rdar:14747994.

llvm-svn: 188609
2013-08-17 16:40:09 +00:00
Juergen Ributzka b4c2fa244a The vbroadcastsi256 intrinsic does not exactly resemble the GCC
builtin. The GCC builtin expects the arguments to be passed by val,
whereas the LLVM intrinsic expects a pointer instead.

This is related to PR 16581 and rdar:14747994.

llvm-svn: 188608
2013-08-17 16:38:37 +00:00
Anton Yartsev 9ddb5fe2d7 + make scan-build work with Strawberry Perl, ActiveState Perl, cygwin perl and msys perl ports.
llvm-svn: 188607
2013-08-17 15:43:19 +00:00
Joerg Sonnenberger 4e5a399b60 Recognize NetBSD's terminfo implementation.
llvm-svn: 188606
2013-08-17 11:06:00 +00:00
Joerg Sonnenberger 8e3050db51 PR 16899: Do not modify the basic block using the iterator, but keep the
next value. This avoids crashes due to invalidation.

Patch by Joey Gouly.

llvm-svn: 188605
2013-08-17 11:04:47 +00:00
Jason Molenda 5894a73862 Follow-up to the checkin of r188532 -- make sure that
we've read the ObjectFile's Symtab before we change the
File addresses in the Sections.

llvm-svn: 188604
2013-08-17 03:39:52 +00:00
Richard Smith 2648fab8de Fix indentation.
llvm-svn: 188603
2013-08-17 00:51:11 +00:00
Richard Smith f9b1510576 Refactor all diagnosing of TypoCorrections through a common function, in
preparation for teaching this function how to diagnose a correction that
includes importing a module.

llvm-svn: 188602
2013-08-17 00:46:16 +00:00
Argyrios Kyrtzidis 15a2fccb78 Rename libIDE to libIndex.
Per feedback from Chandler, it's better to have libraries with more specific functionality.
LibIndex will contain the indexing functionality of libclang, which includes USR generation.

llvm-svn: 188601
2013-08-17 00:40:41 +00:00
David Blaikie 7d5d7c7e90 Revert "DebugInfo: Omit debug info for dynamic classes in TUs that do not have the vtable for that class"
This reverts commit r188576.

Reverting while I investigate a selfhosting buildbot failure on Darwin.

llvm-svn: 188600
2013-08-17 00:06:55 +00:00
Tom Stellard 59ed08b238 R600: Fix possible use of an uninitialized variable
Spotted by Nick Lewycky!

llvm-svn: 188599
2013-08-17 00:06:51 +00:00
Tom Stellard b249b75726 R600: Expand vector FRINT ops
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 188598
2013-08-16 23:51:33 +00:00
Tom Stellard ad3aff246c R600: Expand vector FFLOOR ops
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 188597
2013-08-16 23:51:29 +00:00
Tom Stellard a92ff87929 R600: Expand vector float operations for both SI and R600
Reviewed-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 188596
2013-08-16 23:51:24 +00:00
Jim Grosbach d786679049 ARM: Properly constrain comparison fastisel register classes.
Ongoing 'make the verifier happy' improvements to ARM fast-isel.

rdar://12594152

llvm-svn: 188595
2013-08-16 23:37:40 +00:00
Jim Grosbach 3fa749102a ARM: Fast-isel register class constrain for extends.
Properly constrain the operand register class for instructions used
in [sz]ext expansion. Update more tests to use the verifier now that
we're getting the register classes correct.

rdar://12594152

llvm-svn: 188594
2013-08-16 23:37:36 +00:00
Jim Grosbach 06c2a68125 ARM: Fix more fast-isel verifier failures.
Teach the generic instruction selection helper functions to constrain
the register classes of their input operands. For non-physical register
references, the generic code needs to be careful not to mess that up
when replacing references to result registers. As the comment indicates
for MachineRegisterInfo::replaceRegWith(), it's important to call
constrainRegClass() first.

rdar://12594152

llvm-svn: 188593
2013-08-16 23:37:31 +00:00
Jim Grosbach d69f3ed947 ARM: Clean up fast-isel machine verifier errors.
Lots of machine verifier errors result from using a plain GPR regclass
for incoming argument copies. A more restrictive rGPR class is more
appropriate since it more accurately represents what's happening, plus
it lines up better with isel later on so the verifier is happier.
Reduces the number of ARM fast-isel tests not running with the verifier
enabled by over half.

rdar://12594152

llvm-svn: 188592
2013-08-16 23:37:23 +00:00