platform/upstream/llvm.git
12 years ago[mips] Change the order of template parameters. Move the default parameters to
Akira Hatanaka [Thu, 20 Dec 2012 03:52:08 +0000 (03:52 +0000)]
[mips] Change the order of template parameters. Move the default parameters to
the end.

llvm-svn: 170651

12 years ago[mips] Refactor shift instructions with register operands. Separate encoding
Akira Hatanaka [Thu, 20 Dec 2012 03:48:24 +0000 (03:48 +0000)]
[mips] Refactor shift instructions with register operands. Separate encoding
information from the rest.

llvm-svn: 170650

12 years ago[mips] Refactor shift immediate instructions. Separate encoding information
Akira Hatanaka [Thu, 20 Dec 2012 03:44:41 +0000 (03:44 +0000)]
[mips] Refactor shift immediate instructions. Separate encoding information
from the rest.

llvm-svn: 170649

12 years ago[mips] Refactor arithmetic and logic instructions with immediate operands.
Akira Hatanaka [Thu, 20 Dec 2012 03:40:03 +0000 (03:40 +0000)]
[mips] Refactor arithmetic and logic instructions with immediate operands.
Separate encoding information from the rest.

llvm-svn: 170648

12 years ago[mips] Refactor arithmetic and logic instructions. Separate encoding
Akira Hatanaka [Thu, 20 Dec 2012 03:34:05 +0000 (03:34 +0000)]
[mips] Refactor arithmetic and logic instructions. Separate encoding
information from the rest.

llvm-svn: 170647

12 years agodocs: Show TOC for GettingStarted.rst.
Sean Silva [Thu, 20 Dec 2012 03:32:39 +0000 (03:32 +0000)]
docs: Show TOC for GettingStarted.rst.

This is a pretty lengthy document, so put the table of contents in your
face so that it's easier to scope out the content.

This document is a mess currently and needs to be
refactored/revised/split-up.

llvm-svn: 170646

12 years agoclang/test/Misc/ast-dump-decl.c: Add triple. __thread is not supported on all targets.
NAKAMURA Takumi [Thu, 20 Dec 2012 03:30:30 +0000 (03:30 +0000)]
clang/test/Misc/ast-dump-decl.c: Add triple. __thread is not supported on all targets.

llvm-svn: 170645

12 years agoUpdated link in LanguageExtensions.rst to point to BlockLanguageSpec.rst instead...
Michael Gottesman [Thu, 20 Dec 2012 03:23:52 +0000 (03:23 +0000)]
Updated link in LanguageExtensions.rst to point to BlockLanguageSpec.rst instead of BlockLanguageSpec.txt.

Thanks to Jean-Daniel Dupas for comments!

llvm-svn: 170644

12 years agoChanged UTF-8 emdashes to --- ala Dmitri Gribenko's comments.
Michael Gottesman [Thu, 20 Dec 2012 03:23:49 +0000 (03:23 +0000)]
Changed UTF-8 emdashes to --- ala Dmitri Gribenko's comments.

llvm-svn: 170643

12 years ago[mips] Delete ArithOverflowR and ArithOverflow and use ArithLogicR and
Akira Hatanaka [Thu, 20 Dec 2012 03:00:16 +0000 (03:00 +0000)]
[mips] Delete ArithOverflowR and ArithOverflow and use ArithLogicR and
ArithLogicI as the instruction base classes.

llvm-svn: 170642

12 years agoClarify the comments in PlatformDarwin::LocateExecutableScriptingResource
Jason Molenda [Thu, 20 Dec 2012 02:53:18 +0000 (02:53 +0000)]
Clarify the comments in PlatformDarwin::LocateExecutableScriptingResource
a bit.

llvm-svn: 170641

12 years agoFix another uninitialized bool member bug found by -fsanitize=bool. This one
Richard Smith [Thu, 20 Dec 2012 02:47:01 +0000 (02:47 +0000)]
Fix another uninitialized bool member bug found by -fsanitize=bool. This one
appears to currently be benign (we happen to test the flags in the right
order, so we never depend on the uninitialized value).

llvm-svn: 170640

12 years agoRobustify the @encode code and add support for _Atomic.
John McCall [Thu, 20 Dec 2012 02:45:14 +0000 (02:45 +0000)]
Robustify the @encode code and add support for _Atomic.

Fixes PR14628.

llvm-svn: 170639

12 years agodocs: Clean up adornments.
Sean Silva [Thu, 20 Dec 2012 02:40:45 +0000 (02:40 +0000)]
docs: Clean up adornments.

For whatever reason the usage of '^^^' and '---' adornments were
reversed compared to the "canonical" style of the LLVM docs (which is
currently "the style used in SphinxQuickstartTemplate.rst"). This change
doesn't affect the document structure at all, I'm just doing it for
trivial stylistic consistency (the document content is *much* more
important---thanks Nadav for writing this up!).

Also, trim the adornments to be the same length as the section names.

llvm-svn: 170638

12 years agodocs: ASCII-fy
Sean Silva [Thu, 20 Dec 2012 02:23:25 +0000 (02:23 +0000)]
docs: ASCII-fy

llvm-svn: 170637

12 years agoFix code that attempted to produce a diagnostic with one DiagnosticEngine, then
Richard Smith [Thu, 20 Dec 2012 02:22:15 +0000 (02:22 +0000)]
Fix code that attempted to produce a diagnostic with one DiagnosticEngine, then
produce a note for that diagnostic either with a different DiagnosticEngine or
after calling DiagnosticEngine::Reset(). That didn't make any sense, and did the
wrong thing if the original diagnostic was suppressed.

llvm-svn: 170636

12 years agoProvide couple of DeclPrint tests for my last patch.
Fariborz Jahanian [Thu, 20 Dec 2012 02:20:09 +0000 (02:20 +0000)]
Provide couple of DeclPrint tests for my last patch.

llvm-svn: 170635

12 years agoImplement AST dumper for Decls.
Alexander Kornienko [Thu, 20 Dec 2012 02:09:13 +0000 (02:09 +0000)]
Implement AST dumper for Decls.
http://llvm-reviews.chandlerc.com/D52

Patch by Philip Craig!

llvm-svn: 170634

12 years ago<rdar://problem/12903081>
Enrico Granata [Thu, 20 Dec 2012 02:07:45 +0000 (02:07 +0000)]
<rdar://problem/12903081>

When looking for the Python script to load for a given module, replace dots with underscores if there are any after stripping the extension
This means that for a module named foo.devel.xyz a file named foo_devel.py will be what we try to load

llvm-svn: 170633

12 years agoLoop Vectorizer: Enable if-conversion.
Nadav Rotem [Thu, 20 Dec 2012 02:00:02 +0000 (02:00 +0000)]
Loop Vectorizer: Enable if-conversion.

llvm-svn: 170632

12 years agos/AttributesImpl/AttributeImpl/g This is going to apply to Attribute, not Attributes.
Bill Wendling [Thu, 20 Dec 2012 01:36:59 +0000 (01:36 +0000)]
s/AttributesImpl/AttributeImpl/g This is going to apply to Attribute, not Attributes.

llvm-svn: 170631

12 years agoDo not introduce vector operations in functions marked with noimplicitfloat.
Bob Wilson [Thu, 20 Dec 2012 01:36:20 +0000 (01:36 +0000)]
Do not introduce vector operations in functions marked with noimplicitfloat.

<rdar://problem/12879313>

llvm-svn: 170630

12 years agoClean up some DOxygen comments.
Jim Grosbach [Thu, 20 Dec 2012 01:14:48 +0000 (01:14 +0000)]
Clean up some DOxygen comments.

llvm-svn: 170629

12 years agoClean up some DOxygen comments.
Jim Grosbach [Thu, 20 Dec 2012 01:14:45 +0000 (01:14 +0000)]
Clean up some DOxygen comments.

llvm-svn: 170628

12 years agoFix an uninitialized member variable, found by -fsanitize=bool.
Richard Smith [Thu, 20 Dec 2012 01:05:39 +0000 (01:05 +0000)]
Fix an uninitialized member variable, found by -fsanitize=bool.

llvm-svn: 170627

12 years agowhitespace
Nadav Rotem [Thu, 20 Dec 2012 00:49:56 +0000 (00:49 +0000)]
whitespace

llvm-svn: 170626

12 years ago[analyzer] Add the pointer escaped callback.
Anna Zaks [Thu, 20 Dec 2012 00:38:25 +0000 (00:38 +0000)]
[analyzer] Add the pointer escaped callback.

Instead of using several callbacks to identify the pointer escape event,
checkers now can register for the checkPointerEscape.

Converted the Malloc checker to use the new callback.
SimpleStreamChecker will be converted next.

llvm-svn: 170625

12 years ago[ELF][Reader] Ignore SHF_GROUP when switching on flags.
Michael J. Spencer [Thu, 20 Dec 2012 00:37:24 +0000 (00:37 +0000)]
[ELF][Reader] Ignore SHF_GROUP when switching on flags.

llvm-svn: 170624

12 years agoStyle fixes.
Michael J. Spencer [Thu, 20 Dec 2012 00:37:10 +0000 (00:37 +0000)]
Style fixes.

llvm-svn: 170623

12 years agodoc: resize the image.
Nadav Rotem [Thu, 20 Dec 2012 00:29:18 +0000 (00:29 +0000)]
doc: resize the image.

llvm-svn: 170622

12 years agoTarget/R600: Update MIB according to r170588.
NAKAMURA Takumi [Thu, 20 Dec 2012 00:22:11 +0000 (00:22 +0000)]
Target/R600: Update MIB according to r170588.

llvm-svn: 170620

12 years agoSetting TestInferiorChanged.py to expectedFailureLinux because of bugzilla 14662.
Andrew Kaylor [Thu, 20 Dec 2012 00:19:27 +0000 (00:19 +0000)]
Setting TestInferiorChanged.py to expectedFailureLinux because of bugzilla 14662.

Note that the test actually fails before reaching the assertion described in that bug, but the failure is a trivial matter of adding OS-conditional expected stop reason strings (as noted by new comments in this patch).  However, with the stop reason strings changed, the test fails for the more substantial reason noted in bugzilla 14662.

llvm-svn: 170619

12 years agoDoc: update the chart.
Nadav Rotem [Thu, 20 Dec 2012 00:03:36 +0000 (00:03 +0000)]
Doc: update the chart.

llvm-svn: 170618

12 years ago[preprocessor] When "merging" macro argument tokens into one SLocEntry chunk,
Argyrios Kyrtzidis [Wed, 19 Dec 2012 23:55:44 +0000 (23:55 +0000)]
[preprocessor] When "merging" macro argument tokens into one SLocEntry chunk,
make sure they came from the same kind of FileIDs.

Thanks to Abramo Bagnara for providing the test case.

llvm-svn: 170616

12 years agoAdd a context so that once we uniquify strings we can access them easily.
Bill Wendling [Wed, 19 Dec 2012 23:55:43 +0000 (23:55 +0000)]
Add a context so that once we uniquify strings we can access them easily.

llvm-svn: 170615

12 years agoTypo.
Chad Rosier [Wed, 19 Dec 2012 23:47:16 +0000 (23:47 +0000)]
Typo.

llvm-svn: 170613

12 years ago<rdar://problem/12446222>
Enrico Granata [Wed, 19 Dec 2012 23:42:07 +0000 (23:42 +0000)]
<rdar://problem/12446222>

Implement the ability for Python commands to be interrupted by pressing CTRL+C
Also add a new Mutex subclass that attempts to be helpful for debugging by logging actions performed on it

FYI of all interested - there is a separate deadlocking issue related to how LLDB dispatches CTRL+C that might cause LLDB to deadlock upon pressing CTRL+C while in a Python command.
This is not a regression, and was just previously masked by us not even trying to bail out of Python commands, so that it would not be clear from a user perspective whether we were
deadlocked or stuck in an inconsistent state within the Python interpreter.

llvm-svn: 170612

12 years ago[driver] Have -isysroot warn on nonexistent paths.
Chad Rosier [Wed, 19 Dec 2012 23:41:50 +0000 (23:41 +0000)]
[driver] Have -isysroot warn on nonexistent paths.
rdar://12282267

llvm-svn: 170611

12 years agoMC: Add MCInstrDesc::mayAffectControlFlow() method.
Jim Grosbach [Wed, 19 Dec 2012 23:38:53 +0000 (23:38 +0000)]
MC: Add MCInstrDesc::mayAffectControlFlow() method.

MC disassembler clients (LLDB) are interested in querying if an
instruction may affect control flow other than by virtue of being
an explicit branch instruction. For example, instructions which
write directly to the PC on some architectures.

llvm-svn: 170610

12 years agoAdd isSubRegisterEq() and isSuperRegisterEq().
Jim Grosbach [Wed, 19 Dec 2012 23:38:49 +0000 (23:38 +0000)]
Add isSubRegisterEq() and isSuperRegisterEq().

isSub and isSuper return false if RegA == RegB. Add variants which also
include the identity function.

llvm-svn: 170609

12 years agoMove isSubRegister() and isSuperRegister to MCRegisterInfo.
Jim Grosbach [Wed, 19 Dec 2012 23:38:46 +0000 (23:38 +0000)]
Move isSubRegister() and isSuperRegister to MCRegisterInfo.

These were defined on TargetRegisterInfo, but they don't use any information
that's not available in MCRegisterInfo, so sink them down to be available
at the MC layer.

llvm-svn: 170608

12 years agoFix doc comment. '///' not '//'.
Jim Grosbach [Wed, 19 Dec 2012 23:38:44 +0000 (23:38 +0000)]
Fix doc comment. '///' not '//'.

llvm-svn: 170607

12 years agoAudit DeclPrinter with -ast-dump on Cocoa.h and
Fariborz Jahanian [Wed, 19 Dec 2012 23:36:00 +0000 (23:36 +0000)]
Audit DeclPrinter with -ast-dump on Cocoa.h and
fix any bad objectiveC syntax coming out of
DeclPrinter. This is on going. Also, introduce a new
PrintPolicy and use it as needed when declaration tag
is to be produced via DeclPrinter.

llvm-svn: 170606

12 years agoMark register tests as expectedFailureLinux because of bugzilla reports 14600 and...
Andrew Kaylor [Wed, 19 Dec 2012 23:31:12 +0000 (23:31 +0000)]
Mark register tests as expectedFailureLinux because of bugzilla reports 14600 and 14661.

llvm-svn: 170605

12 years agoDisable confirmation prompt in TestProcessLaunch (to avoid hang on Linux once bug...
Daniel Malea [Wed, 19 Dec 2012 23:22:11 +0000 (23:22 +0000)]
Disable confirmation prompt in TestProcessLaunch (to avoid hang on Linux once bug is fixed)

llvm-svn: 170603

12 years agoRefactor isIntrinsic() to be quicker, and change classof() (and thus, isa<IntrinsicIn...
Michael Ilseman [Wed, 19 Dec 2012 23:17:20 +0000 (23:17 +0000)]
Refactor isIntrinsic() to be quicker, and change classof() (and thus, isa<IntrinsicInst>()) to use it. This decreases the number of occurrences of the slow-path string matching performed by getIntrinsicID().

llvm-svn: 170602

12 years agoThis patch removes the SymbolFileSymtab support
Sean Callanan [Wed, 19 Dec 2012 23:05:01 +0000 (23:05 +0000)]
This patch removes the SymbolFileSymtab support
for reporting class types from Objective-C runtime
class symbols.  Instead, LLDB now queries the
Objective-C runtime for class types.

We have also added a (minimal) Objective-C runtime
type vendor for Objective-C runtime version 1, to
prevent regressions when calling class methods in
the V1 runtime.

Other components of this fix include:

- We search the Objective-C runtime in a few more
  places.

- We enable enumeration of all members of
  Objective-C classes, which Clang does in certain
  circumstances.

- SBTarget::FindFirstType and SBTarget::FindTypes
  now query the Objective-C runtime as needed.

- I fixed several test cases.

<rdar://problem/12885034>

llvm-svn: 170601

12 years agos/AttributeListImpl/AttributeSetImpl/g to match the namechange of AttributeList.
Bill Wendling [Wed, 19 Dec 2012 22:42:22 +0000 (22:42 +0000)]
s/AttributeListImpl/AttributeSetImpl/g to match the namechange of AttributeList.

llvm-svn: 170600

12 years agoAlways use addOperand(MF, MO) from MachineInstrBuilder.
Jakob Stoklund Olesen [Wed, 19 Dec 2012 22:35:46 +0000 (22:35 +0000)]
Always use addOperand(MF, MO) from MachineInstrBuilder.

The single-argument MachineInstr::addOperand(MO) will be removed soon.

llvm-svn: 170599

12 years ago[PCH/Modules] In ASTReader::completeVisibleDeclsMap, make sure to visit all
Argyrios Kyrtzidis [Wed, 19 Dec 2012 22:21:18 +0000 (22:21 +0000)]
[PCH/Modules] In ASTReader::completeVisibleDeclsMap, make sure to visit all
modules when getting the decls for a namespace or translation unit.

Otherwise the code-completion results will not be complete.

rdar://12889089

llvm-svn: 170596

12 years agoAdd a missing 'else'. Found by grep '} if'
Dmitri Gribenko [Wed, 19 Dec 2012 22:13:01 +0000 (22:13 +0000)]
Add a missing 'else'.  Found by grep '} if'

No testcase because it is apparently not so trivial to construct.

llvm-svn: 170595

12 years agoR600: Add entry in CODE_OWNERS.TXT
Tom Stellard [Wed, 19 Dec 2012 22:10:35 +0000 (22:10 +0000)]
R600: Add entry in CODE_OWNERS.TXT

llvm-svn: 170594

12 years agoR600: Remove unecessary VREG alignment.
Tom Stellard [Wed, 19 Dec 2012 22:10:34 +0000 (22:10 +0000)]
R600: Remove unecessary VREG alignment.

Unlike SGPRs VGPRs doesn't need to be aligned.

Patch by: Christian König

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
llvm-svn: 170593

12 years agoR600: control flow optimization
Tom Stellard [Wed, 19 Dec 2012 22:10:33 +0000 (22:10 +0000)]
R600: control flow optimization

Branch if we have enough instructions so that it makes sense.
Also remove branches if they don't make sense.

Patch by: Christian König

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
Signed-off-by: Christian König <deathsimple@vodafone.de>
llvm-svn: 170592

12 years agoR600: New control flow for SI v2
Tom Stellard [Wed, 19 Dec 2012 22:10:31 +0000 (22:10 +0000)]
R600: New control flow for SI v2

This patch replaces the control flow handling with a new
pass which structurize the graph before transforming it to
machine instruction. This has a couple of different advantages
and currently fixes 20 piglit tests without a single regression.

It is now a general purpose transformation that could be not
only be used for SI/R6xx, but also for other hardware
implementations that use a form of structurized control flow.

v2: further cleanup, fixes and documentation

Patch by: Christian König

Signed-off-by: Christian König <deathsimple@vodafone.de>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
Tested-by: Michel Dänzer <michel.daenzer@amd.com>
llvm-svn: 170591

12 years agoDocumentation: UsersManual.rst: some cleanups after semi-automatic conversion
Dmitri Gribenko [Wed, 19 Dec 2012 22:06:59 +0000 (22:06 +0000)]
Documentation: UsersManual.rst: some cleanups after semi-automatic conversion

llvm-svn: 170590

12 years agoSplit out abbreviations for the skeleton info from the rest of
Eric Christopher [Wed, 19 Dec 2012 22:02:53 +0000 (22:02 +0000)]
Split out abbreviations for the skeleton info from the rest of
the abbreviations. Part of implementing split dwarf.

llvm-svn: 170589

12 years agoRemove the explicit MachineInstrBuilder(MI) constructor.
Jakob Stoklund Olesen [Wed, 19 Dec 2012 21:31:56 +0000 (21:31 +0000)]
Remove the explicit MachineInstrBuilder(MI) constructor.

Use the version that also takes an MF reference instead.

It would technically be possible to extract an MF reference from the MI
as MI->getParent()->getParent(), but that would not work for MIs that
are not inserted into any basic block.

Given the reasonably small number of places this constructor was used at
all, I preferred the compile time check to a run time assertion.

llvm-svn: 170588

12 years agoFix a bug that was found by building clang with -fsanitize.
Nadav Rotem [Wed, 19 Dec 2012 20:47:04 +0000 (20:47 +0000)]
Fix a bug that was found by building clang with -fsanitize.
I introduced it in r166785. PR14291.

If TD is unavailable use getScalarSizeInBits, but don't optimize
pointers or vectors of pointers.

llvm-svn: 170586

12 years agodocs: Fix title underline warnings
Meador Inge [Wed, 19 Dec 2012 20:16:40 +0000 (20:16 +0000)]
docs: Fix title underline warnings

Building Vectorizers.rst produces a few warnings of the form:

   WARNING: Title underline too short.

Fixed by adding the extra needed dashes under the title.

llvm-svn: 170582

12 years agoLLVM sdisel normalize bit extraction of the form:
Evan Cheng [Wed, 19 Dec 2012 20:16:09 +0000 (20:16 +0000)]
LLVM sdisel normalize bit extraction of the form:
 ((x & 0xff00) >> 8) << 2
to
 (x >> 6) & 0x3fc

This is general goodness since it folds a left shift into the mask. However,
the trailing zeros in the mask prevents the ARM backend from using the bit
extraction instructions. And worse since the mask materialization may require
an addition instruction. This comes up fairly frequently when the result of
the bit twiddling is used as memory address. e.g.

 = ptr[(x & 0xFF0000) >> 16]

We want to generate:
  ubfx   r3, r1, #16, #8
  ldr.w  r3, [r0, r3, lsl #2]

vs.
  mov.w  r9, #1020
  and.w  r2, r9, r1, lsr #14
  ldr    r2, [r0, r2]

Add a late ARM specific isel optimization to
ARMDAGToDAGISel::PreprocessISelDAG(). It folds the left shift to the
'base + offset' address computation; change the mask to one which doesn't have
trailing zeros and enable the use of ubfx.

Note the optimization has to be done late since it's target specific and we
don't want to change the DAG normalization. It's also fairly restrictive
as shifter operands are not always free. It's only done for lsh 1 / 2. It's
known to be free on some cpus and they are most common for address
computation.

This is a slight win for blowfish, rijndael, etc.

rdar://12870177

llvm-svn: 170581

12 years agoRemove edis remnant.
Benjamin Kramer [Wed, 19 Dec 2012 20:11:17 +0000 (20:11 +0000)]
Remove edis remnant.

llvm-svn: 170580

12 years agoRemove edis - the enhanced disassembler. Fixes PR14654.
Roman Divacky [Wed, 19 Dec 2012 19:55:47 +0000 (19:55 +0000)]
Remove edis - the enhanced disassembler. Fixes PR14654.

llvm-svn: 170578

12 years agoTransform (x&C)>V into (x&C)!=0 where possible
Paul Redmond [Wed, 19 Dec 2012 19:47:13 +0000 (19:47 +0000)]
Transform (x&C)>V into (x&C)!=0 where possible

When the least bit of C is greater than V, (x&C) must be greater than V
if it is not zero, so the comparison can be simplified.

Although this was suggested in Target/X86/README.txt, it benefits any
architecture with a directly testable form of AND.

Patch by Kevin Schoedel

llvm-svn: 170576

12 years agoSkip libc++ data formatter tests on Linux because there is no standard location for...
Andrew Kaylor [Wed, 19 Dec 2012 19:32:40 +0000 (19:32 +0000)]
Skip libc++ data formatter tests on Linux because there is no standard location for the makefile to find libc++.

llvm-svn: 170575

12 years agoAdd an MF argument to MachineInstr::addOperand().
Jakob Stoklund Olesen [Wed, 19 Dec 2012 19:19:01 +0000 (19:19 +0000)]
Add an MF argument to MachineInstr::addOperand().

Just like for addMemOperand(), the function pointer provides a context
for allocating memory. This will make it possible to use a better memory
allocation strategy for the MI operand list, which is currently a slow
std::vector.

Most calls to addOperand() come from MachineInstrBuilder, so give that
class an MF reference as well. Code using BuildMI() won't need changing
at all since the MF reference is already required to allocate a
MachineInstr.

Future patches will fix code that calls MI::addOperand(Op) directly, as
well as code that uses the now deprecated MachineInstrBuilder(MI)
constructor.

llvm-svn: 170574

12 years agoobjective-C: Don't warn of unimplemented property of protocols in
Fariborz Jahanian [Wed, 19 Dec 2012 18:58:55 +0000 (18:58 +0000)]
objective-C: Don't warn of unimplemented property of protocols in
category, when those properties will be implemented in category's
primary class or one of its super classes. // rdar://12568064

llvm-svn: 170573

12 years agoForgot to remove trailing whitespace when I converted the BlockLanguageSpec.txt ...
Michael Gottesman [Wed, 19 Dec 2012 18:42:42 +0000 (18:42 +0000)]
Forgot to remove trailing whitespace when I converted the BlockLanguageSpec.txt => BlockLanguageSpec.rst. It has been eliminated.

llvm-svn: 170572

12 years agoConverted BlockLanguageSpec.txt => BlockLanguageSpec.rst.
Michael Gottesman [Wed, 19 Dec 2012 18:41:09 +0000 (18:41 +0000)]
Converted BlockLanguageSpec.txt => BlockLanguageSpec.rst.

llvm-svn: 170571

12 years agoChange test to dump Filecheck output to an intermediate file. This makes it easier...
Ted Kremenek [Wed, 19 Dec 2012 18:30:38 +0000 (18:30 +0000)]
Change test to dump Filecheck output to an intermediate file.  This makes it easier to update this test.

llvm-svn: 170570

12 years agoRemove superfluous brief command from getAsString.
Chad Rosier [Wed, 19 Dec 2012 18:06:44 +0000 (18:06 +0000)]
Remove superfluous brief command from getAsString.

llvm-svn: 170569

12 years agodoc: add subsections.
Nadav Rotem [Wed, 19 Dec 2012 18:04:44 +0000 (18:04 +0000)]
doc: add subsections.

llvm-svn: 170568

12 years agoDOC: document the use of O2, O3 and Os with -fvectorize.
Nadav Rotem [Wed, 19 Dec 2012 18:02:36 +0000 (18:02 +0000)]
DOC: document the use of O2, O3 and Os with -fvectorize.

llvm-svn: 170567

12 years agoFix PR14591: Windows newlines in doxygen comments cause failed assertion in
Dmitri Gribenko [Wed, 19 Dec 2012 17:34:55 +0000 (17:34 +0000)]
Fix PR14591: Windows newlines in doxygen comments cause failed assertion in
TextDiagnostic

Patch by Janusz Chorko.

llvm-svn: 170566

12 years agoAdd a missing 'else'. Found by grep '} if'
Dmitri Gribenko [Wed, 19 Dec 2012 17:29:30 +0000 (17:29 +0000)]
Add a missing 'else'.  Found by grep '} if'

No testcase because this did not affect correctness: a declaration can only be
ClassTemplateDecl or a FunctionTemplateDecl, not both.

llvm-svn: 170565

12 years agoTest suite cleanup for Linux: mark test cases expected to fail due to open bugzillas
Daniel Malea [Wed, 19 Dec 2012 17:19:28 +0000 (17:19 +0000)]
Test suite cleanup for Linux: mark test cases expected to fail due to open bugzillas
- bugzillas covered: 14323, 14600, 14541, 14437, 14540, 14541

llvm-svn: 170564

12 years agoComment parsing: add a missing 'else'. Found by inspection.
Dmitri Gribenko [Wed, 19 Dec 2012 17:17:09 +0000 (17:17 +0000)]
Comment parsing: add a missing 'else'.  Found by inspection.

No testcase because we were just building an extra AST node and eventually
throwing it away, so it did not affect correctness.

llvm-svn: 170563

12 years agoImprove on header comment suggested by James Dennet.
Fariborz Jahanian [Wed, 19 Dec 2012 17:03:46 +0000 (17:03 +0000)]
Improve on header comment suggested by James Dennet.

llvm-svn: 170555

12 years ago[ASan] explicitly link with dynamic ASan runtime when building unittests on Android
Alexey Samsonov [Wed, 19 Dec 2012 15:52:30 +0000 (15:52 +0000)]
[ASan] explicitly link with dynamic ASan runtime when building unittests on Android

llvm-svn: 170552

12 years agoPowerPC: Expand VSELECT nodes.
Benjamin Kramer [Wed, 19 Dec 2012 15:49:14 +0000 (15:49 +0000)]
PowerPC: Expand VSELECT nodes.

There's probably a better expansion for those nodes than the default for
altivec, but this is better than crashing. VSELECTs occur in loop vectorizer
output.

llvm-svn: 170551

12 years agoChange AsmOperandInfo::ConstraintVT to MVT, instead of EVT.
Patrik Hagglund [Wed, 19 Dec 2012 15:19:11 +0000 (15:19 +0000)]
Change AsmOperandInfo::ConstraintVT to MVT, instead of EVT.

Accordingly, add MVT::getVT.

llvm-svn: 170550

12 years ago[ASan] Support building both 32- and 64-bit unit tests if we can target both architec...
Alexey Samsonov [Wed, 19 Dec 2012 15:17:23 +0000 (15:17 +0000)]
[ASan] Support building both 32- and 64-bit unit tests if we can target both architectures

llvm-svn: 170549

12 years ago[asan] asan_allocator2: add mmap/munmap stats
Kostya Serebryany [Wed, 19 Dec 2012 14:56:38 +0000 (14:56 +0000)]
[asan] asan_allocator2: add mmap/munmap stats

llvm-svn: 170548

12 years agoRevert 170545 while I debug the ppc failures.
Rafael Espindola [Wed, 19 Dec 2012 14:48:05 +0000 (14:48 +0000)]
Revert 170545 while I debug the ppc failures.

llvm-svn: 170547

12 years agoMake TargetLowering::getTypeConversion more resilient against odd illegal MVTs.
Benjamin Kramer [Wed, 19 Dec 2012 14:34:28 +0000 (14:34 +0000)]
Make TargetLowering::getTypeConversion more resilient against odd illegal MVTs.

- An MVT can become an EVT when being split (e.g. v2i8 -> v1i8, the latter doesn't exist)
- Return the scalar value when an MVT is scalarized (v1i64 -> i64)

Fixes PR14639ff.

llvm-svn: 170546

12 years agoAdd r170095 back.
Rafael Espindola [Wed, 19 Dec 2012 14:15:04 +0000 (14:15 +0000)]
Add r170095 back.

I cannot reproduce it the failures locally, so I will keep an eye at the ppc
bots. This patch does add the change to the "Disassembly of section" message,
but that is not what was failing on the bots.

Original message:

Add a funciton to get the segment name of a section.

On MachO, sections also have segment names. When a tool looking at a .o file
prints a segment name, this is what they mean. In reality, a .o has only one
anonymous, segment.

This patch adds a MachO only function to fetch that segment name. I named it
getSectionFinalSegmentName since the main use for the name seems to be infor
the linker with segment this section should go to.

The patch also changes MachOObjectFile::getSectionName to return just the
section name instead of computing SegmentName,SectionName.

llvm-svn: 170545

12 years ago[msan] Add track-origins argument to the pass constructor.
Evgeniy Stepanov [Wed, 19 Dec 2012 13:55:51 +0000 (13:55 +0000)]
[msan] Add track-origins argument to the pass constructor.

llvm-svn: 170544

12 years ago[ASan] make sure Android unittests depend on ASan runtime
Alexey Samsonov [Wed, 19 Dec 2012 13:46:58 +0000 (13:46 +0000)]
[ASan] make sure Android unittests depend on ASan runtime

llvm-svn: 170543

12 years agoDocumentation: add a missing space
Dmitri Gribenko [Wed, 19 Dec 2012 12:51:48 +0000 (12:51 +0000)]
Documentation: add a missing space

llvm-svn: 170542

12 years agoSignificantly change the way we build ASan unittests in CMake
Alexey Samsonov [Wed, 19 Dec 2012 12:33:39 +0000 (12:33 +0000)]
Significantly change the way we build ASan unittests in CMake
build tree. Now just-built Clang is used to:
  1) compile instrumented sources (as before);
  2) compile non-instrumented sources;
  3) compile our own instrumented version of googletest;
  4) link it all together using -fsanitize=address flag
     (instead of trying to copy linker behavior in
      CMake build rules).

This makes ASan unittests pretty much self-consistent
and independent of other LLVM libraries.

llvm-svn: 170541

12 years agoSplit the usage of 'EVT PartVT' into 'MVT PartVT' and 'EVT PartEVT'.
Patrik Hagglund [Wed, 19 Dec 2012 12:33:30 +0000 (12:33 +0000)]
Split the usage of 'EVT PartVT' into 'MVT PartVT' and 'EVT PartEVT'.

llvm-svn: 170540

12 years agoCMake: factor out a function that returns the expected directory for unit test
Alexey Samsonov [Wed, 19 Dec 2012 12:30:33 +0000 (12:30 +0000)]
CMake: factor out a function that returns the expected directory for unit test

llvm-svn: 170539

12 years agoChange RegVT in BitTestBlock and RegsForValue, to contain MVTs,
Patrik Hagglund [Wed, 19 Dec 2012 12:23:01 +0000 (12:23 +0000)]
Change RegVT in BitTestBlock and RegsForValue, to contain MVTs,
instead of EVTs.

llvm-svn: 170538

12 years agoChange TargetLowering::getTypeForExtArgOrReturn to take and return
Patrik Hagglund [Wed, 19 Dec 2012 12:02:25 +0000 (12:02 +0000)]
Change TargetLowering::getTypeForExtArgOrReturn to take and return
MVTs, instead of EVTs.

llvm-svn: 170537

12 years agoChange a parameter of TargetLowering::getVectorTypeBreakdown to MVT,
Patrik Hagglund [Wed, 19 Dec 2012 11:53:21 +0000 (11:53 +0000)]
Change a parameter of TargetLowering::getVectorTypeBreakdown to MVT,
from EVT.

llvm-svn: 170536

12 years agoChange TargetLowering::RegisterTypeForVT to contain MVTs, instead of
Patrik Hagglund [Wed, 19 Dec 2012 11:48:16 +0000 (11:48 +0000)]
Change TargetLowering::RegisterTypeForVT to contain MVTs, instead of
EVTs.

llvm-svn: 170535

12 years agoChange TargetLowering::TransformToType to contain MVTs, instead of
Patrik Hagglund [Wed, 19 Dec 2012 11:42:00 +0000 (11:42 +0000)]
Change TargetLowering::TransformToType to contain MVTs, instead of
EVTs.

llvm-svn: 170534

12 years agoChange TargetLowering::getRepRegClassCostFor, getIndexedLoadAction,
Patrik Hagglund [Wed, 19 Dec 2012 11:37:12 +0000 (11:37 +0000)]
Change TargetLowering::getRepRegClassCostFor, getIndexedLoadAction,
getIndexedStoreAction, and addRegisterClass to take and MVT, instead
of EVT.

llvm-svn: 170533

12 years agoChange TargetLowering::findRepresentativeClass to take an MVT, instead
Patrik Hagglund [Wed, 19 Dec 2012 11:30:36 +0000 (11:30 +0000)]
Change TargetLowering::findRepresentativeClass to take an MVT, instead
of EVT.

llvm-svn: 170532

12 years ago[ASan] Change the boilerplate check in the libdispatch tests to fix them.
Alexander Potapenko [Wed, 19 Dec 2012 11:26:41 +0000 (11:26 +0000)]
[ASan] Change the boilerplate check in the libdispatch tests to fix them.

llvm-svn: 170531