Douglas Gregor [Mon, 28 Jan 2013 20:13:44 +0000 (20:13 +0000)]
Undo my re-wording of the "ARC forbids Objective-C objects in ..."
error. Jordan is right.
llvm-svn: 173713
Andrew Kaylor [Mon, 28 Jan 2013 19:52:37 +0000 (19:52 +0000)]
Add support for source and line information to IntelJITEventListener for object emitted by MCJIT.
llvm-svn: 173712
Edwin Vane [Mon, 28 Jan 2013 19:34:42 +0000 (19:34 +0000)]
Fix gcc/printf/ISO C++ warning
Remove the use of the 't' length modifier to avoid a gcc warning. Based
on usage, 32 bits of precision is good enough for printing a stack
offset for a stack trace.
't' length modifier isn't in C++03 but it *is* in C++11. Added a FIXME
to reintroduce once LLVM makes the switch to C++11.
Reviewer: gribozavr
llvm-svn: 173711
Shankar Easwaran [Mon, 28 Jan 2013 19:21:04 +0000 (19:21 +0000)]
add register section, remove contentType from sectionKey
llvm-svn: 173709
Douglas Gregor [Mon, 28 Jan 2013 19:08:09 +0000 (19:08 +0000)]
Forbid the use of objects in unions in Objective-C++ ARC. Fixes
<rdar://problem/
13098104>.
llvm-svn: 173708
Douglas Gregor [Mon, 28 Jan 2013 18:38:02 +0000 (18:38 +0000)]
Enable the global module index by default. Introduce the
-fno-modules-global-index -cc1 option to allow one to disable the
index for performance testing purposes, but with a 10% win in
-fsyntax-only time, there is no reason a user would do this.
llvm-svn: 173707
Bill Schmidt [Mon, 28 Jan 2013 18:36:58 +0000 (18:36 +0000)]
This patch addresses bug 15031.
The common code in the post-RA scheduler to break anti-dependencies on the
critical path contained a flaw. In the reported case, an anti-dependency
between the overlapping registers %X4 and %R4 exists:
%X29<def> = OR8 %X4, %X4
%R4<def>, %X3<def,dead,tied3> = LBZU 1, %X3<kill,tied1>
The unpatched code breaks the dependency by replacing %R4 and its uses
with %R3, the first register on the available list. However, %R3 and
%X3 overlap, so this creates two overlapping definitions on the same
instruction.
The fix is straightforward, preventing selection of a register that
overlaps any other defined register on the same instruction.
The test case is reduced from the bug report, and verifies that we no
longer produce "lbzu 3, 1(3)" when breaking this anti-dependency.
llvm-svn: 173706
Douglas Gregor [Mon, 28 Jan 2013 18:29:39 +0000 (18:29 +0000)]
ASTReader and profiling statistics indicate that implementing a method
pool in the global module index is not worthwhile. Update comments to
limit the scope of the global module index to identifiers.
llvm-svn: 173705
Dmitri Gribenko [Mon, 28 Jan 2013 18:01:33 +0000 (18:01 +0000)]
Tests: remove disabled llvm-gcc invocations
llvm-svn: 173703
Douglas Gregor [Mon, 28 Jan 2013 17:54:36 +0000 (17:54 +0000)]
Add some more ASTReader statistics for global method pool lookups.
llvm-svn: 173702
Jordan Rose [Mon, 28 Jan 2013 17:49:02 +0000 (17:49 +0000)]
PR15067 (again): Don't warn about UCNs in C90 if we're raw-lexing.
Fixes a crash. Thanks, Richard.
llvm-svn: 173701
Eric Christopher [Mon, 28 Jan 2013 17:39:03 +0000 (17:39 +0000)]
Fix comment.
llvm-svn: 173700
Eric Christopher [Mon, 28 Jan 2013 17:33:26 +0000 (17:33 +0000)]
Fix comment.
llvm-svn: 173698
Dmitri Gribenko [Mon, 28 Jan 2013 17:31:40 +0000 (17:31 +0000)]
FileCheck'ize and merge tests
llvm-svn: 173697
Dmitri Gribenko [Mon, 28 Jan 2013 17:30:37 +0000 (17:30 +0000)]
FileCheck'ize test
llvm-svn: 173696
Daniel Jasper [Mon, 28 Jan 2013 17:30:17 +0000 (17:30 +0000)]
Timur Iskhodzhanov [Mon, 28 Jan 2013 17:29:50 +0000 (17:29 +0000)]
Use the correct order of NOINLINE vs ret type to fix Windows build
llvm-svn: 173694
Douglas Gregor [Mon, 28 Jan 2013 17:26:34 +0000 (17:26 +0000)]
Back out r173689. It doesn't matter when we're building with Clang anyway.
llvm-svn: 173693
Douglas Gregor [Mon, 28 Jan 2013 16:46:33 +0000 (16:46 +0000)]
Eliminate memory allocation from most invocations of
ModuleManager::visit() by keeping a free list of the two data
structures used to store state (a preallocated stack and a visitation
number vector). Improves -fsyntax-only performance for my modules test
case by 2.8%. Modules has pulled ahead by almost 10% with the global
module index.
llvm-svn: 173692
Renato Golin [Mon, 28 Jan 2013 16:02:45 +0000 (16:02 +0000)]
Vectorization Factor clarification
llvm-svn: 173691
Daniel Jasper [Mon, 28 Jan 2013 15:40:20 +0000 (15:40 +0000)]
Don't put a function's return type on its own line in Google style.
This would be against the style guide:
http://google-styleguide.googlecode.com/svn/trunk/cppguide.xml?showone=Function_Declarations_and_Definitions#Function_Declarations_and_Definitions
Not sure what to do as a last resort if the function signature does not
fit onto a single line in Google style ..
llvm-svn: 173690
Douglas Gregor [Mon, 28 Jan 2013 15:39:10 +0000 (15:39 +0000)]
Note that SrcMgr::SLocEntry's are POD-like, so SmallVector can do a
better job with them. Shaves off 0.7% of -fparse-only time for a
modules test case. Sure makes you wonder...
llvm-svn: 173689
Daniel Jasper [Mon, 28 Jan 2013 15:16:31 +0000 (15:16 +0000)]
Let clang-format break after a function's return type.
Before:
TypeSpecDecl *TypeSpecDecl::Create(ASTContext &C, DeclContext *DC,
SourceLocation L, IdentifierInfo *II,
Type *T) {}
After:
TypeSpecDecl *
TypeSpecDecl::Create(ASTContext &C, DeclContext *DC, SourceLocation L,
IdentifierInfo *II, Type *T) {}
This fixes llvm.org/PR14717.
llvm-svn: 173688
Evgeniy Stepanov [Mon, 28 Jan 2013 13:52:49 +0000 (13:52 +0000)]
[msan] A runtime call to support custom allocators.
llvm-svn: 173687
Evgeniy Stepanov [Mon, 28 Jan 2013 13:45:16 +0000 (13:45 +0000)]
[msan] Remove icmp tests that require exact shadow propagation.
llvm-svn: 173686
Daniel Jasper [Mon, 28 Jan 2013 13:31:35 +0000 (13:31 +0000)]
Make continuations in constructor initializers consistent.
Before:
Constructor
: LooooooooooooooooooooongVariable(
LooooooooooooooooooooongValue) {}
After:
Constructor
: LooooooooooooooooooooongVariable(
LooooooooooooooooooooongValue) {}
llvm-svn: 173685
Daniel Jasper [Mon, 28 Jan 2013 13:21:16 +0000 (13:21 +0000)]
Initial support for formatting range-based for-loops.
Before (in good cases):
for (auto aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
:
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa) {}
for (auto
aaaaaaaaaaaaaaaaaaaa :
aaaaaaaaaaaaaaaaaaaa(
aaaaaaaaaaaaaaaaaaaa,
aaaa)) {}
After:
for (auto aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa :
aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa) {}
for (auto
aaaaaaaaaaaaaaaaaaaa :
aaaaaaaaaaaaaaaaaaaa(
aaaaaaaaaaaaaaaaaaaa, aaaa)) {}
llvm-svn: 173684
Daniel Jasper [Mon, 28 Jan 2013 12:45:14 +0000 (12:45 +0000)]
Improve formatting of conditional expressions.
Before we did not really systematically format those. Now, we format the
different cases as:
- 1 Line: a ? b : c;
- 2 Lines: short ? loooooooooong
: loooooooooong
- 2 Lines: loooooooooooooooong
? short : short
- 3 Lines: loooooooooooooooong
? loooooooooooooong
: loooooooooooooong
Not sure whether "?" and ":" should go on the new line, but it seems to
be the most consistent approach.
llvm-svn: 173683
Evgeniy Stepanov [Mon, 28 Jan 2013 11:42:28 +0000 (11:42 +0000)]
[msan] Mostly disable msan-handle-icmp-exact.
It is way too slow. Change the default option value to 0.
Always do exact shadow propagation for unsigned ICmp with constants, it is
cheap (under 1% cpu time) and required for correctness.
llvm-svn: 173682
Alexey Samsonov [Mon, 28 Jan 2013 11:24:13 +0000 (11:24 +0000)]
[ASan] fix a bug in allocator-v2 which could lead to SEGV on realloc(malloc(0), 4)
llvm-svn: 173681
Daniel Jasper [Mon, 28 Jan 2013 09:35:24 +0000 (09:35 +0000)]
Several small changes in formatting decisions.
1. Use a hanging ident for function calls nested in binary expressions.
E.g.:
int aaaaa =
aaaaaaaaa &&
aaaaaaaaaa(
aaaaaaaaaa);
2. Slightly improve heuristic for builder type expressions and reduce
penalty for breaking before "." and "->" in those.
3. Remove mostly obsolete metric of decreasing indent level. This
fixes: llvm.org/PR14931.
Changes #1 and #2 were necessary to keep tests passing after #3.
llvm-svn: 173680
Evgeniy Stepanov [Mon, 28 Jan 2013 09:18:40 +0000 (09:18 +0000)]
Revert r173678.
Broken tests.
llvm-svn: 173679
Evgeniy Stepanov [Mon, 28 Jan 2013 09:15:15 +0000 (09:15 +0000)]
[msan] Make msan-handle-icmp-exact=0 by default.
50% slowdown on one of the specs.
llvm-svn: 173678
Alexey Samsonov [Mon, 28 Jan 2013 09:07:30 +0000 (09:07 +0000)]
CMake: simplify build rules for compiler-rt unit tests. This fixes warnings in Ninja build tree.
llvm-svn: 173677
Kostya Serebryany [Mon, 28 Jan 2013 08:05:47 +0000 (08:05 +0000)]
[asan] fix a crash in asan stats printing (initialize the allocator in __asan_init)
llvm-svn: 173676
Daniel Jasper [Mon, 28 Jan 2013 07:43:15 +0000 (07:43 +0000)]
Improve indentation after breaking at nested name specifiers.
These always represent a continuation and we should increase the ident.
Before:
aaaaaaaaa(
aaaaaaaaaaaaaaaaaaaaa::
aaaaaaaaaaaaaaaaaaaa);
After:
aaaaaaaaa(
aaaaaaaaaaaaaaaaaaaaa::
aaaaaaaaaaaaaaaaaaaa);
llvm-svn: 173675
Craig Topper [Mon, 28 Jan 2013 07:41:18 +0000 (07:41 +0000)]
Fix 256-bit PALIGNR comment decoding to understand that it works on independent 256-bit lanes.
llvm-svn: 173674
Daniel Jasper [Mon, 28 Jan 2013 07:35:34 +0000 (07:35 +0000)]
Avoid confusing identations for multi-parameter functions.
Before:
aaaaaaaa(
aaaaaaaaa(
aaaaaaaaaa(),
aaaaaaaaa);
After:
aaaaaaaa(
aaaaaaaaa(
aaaaaaaaaa(),
aaaaaaaaa);
llvm-svn: 173673
Nadav Rotem [Mon, 28 Jan 2013 07:35:33 +0000 (07:35 +0000)]
A bugfix for tblgen, in the function ‘emitSourceFileHeader’.
When the first parameter (‘Desc’) is more than 80 characters long, it will result the header line that contains the description to be more
Than (4GB!) long. Not only it takes forever to produce, the output file cannot be open, since its ginormous.
Patch by Elior Malul.
llvm-svn: 173672
Kostya Serebryany [Mon, 28 Jan 2013 07:34:22 +0000 (07:34 +0000)]
[asan] two more internal flags for asan-rt: print_stats (0) and print_legend (1)
llvm-svn: 173671
Alexey Samsonov [Mon, 28 Jan 2013 07:20:44 +0000 (07:20 +0000)]
Print warning instead of error if optional ASan features are enabled w/o specifying -fsanitize=address
llvm-svn: 173670
Craig Topper [Mon, 28 Jan 2013 07:19:11 +0000 (07:19 +0000)]
Add missing break in 256-bit palignr comment printing. No test case yet because the comment itself is still wrong.
llvm-svn: 173669
Alexey Samsonov [Mon, 28 Jan 2013 07:16:22 +0000 (07:16 +0000)]
[CMake] Fix compiler-rt tests after r173617
llvm-svn: 173668
Craig Topper [Mon, 28 Jan 2013 06:48:25 +0000 (06:48 +0000)]
Fix inconsistent usage of PALIGN and PALIGNR when referring to the same instruction.
llvm-svn: 173667
Michael Gottesman [Mon, 28 Jan 2013 06:39:31 +0000 (06:39 +0000)]
Created ObjCARCUtil.cpp for functions which in my humble opinion are too large to static inline and place in a header file such as ObjCARC.h.
llvm-svn: 173666
Michael J. Spencer [Mon, 28 Jan 2013 06:11:39 +0000 (06:11 +0000)]
[ELF] Use correct regex in test.
llvm-svn: 173665
Craig Topper [Mon, 28 Jan 2013 06:09:24 +0000 (06:09 +0000)]
Remove addToNoHelperNeeded function that was left unused after r173649. Fixes a -Wunused warning.
llvm-svn: 173664
Michael Gottesman [Mon, 28 Jan 2013 05:51:58 +0000 (05:51 +0000)]
Cleaned up includes in various ObjCARC files and removed some whitespace violations.
llvm-svn: 173663
Michael Gottesman [Mon, 28 Jan 2013 05:51:54 +0000 (05:51 +0000)]
Refactor ObjCARCAliasAnalysis into its own file.
llvm-svn: 173662
Bill Wendling [Mon, 28 Jan 2013 05:51:40 +0000 (05:51 +0000)]
Mid-air collision. reapply r173656.
llvm-svn: 173661
Bill Wendling [Mon, 28 Jan 2013 05:44:14 +0000 (05:44 +0000)]
Rewrite the removeAttr() method.
This now uses the AttributeSet object instead of the Attribute /
AttributeWithIndex objects. It's fairly simple now. It goes through all of the
subsets before the one we're modifying, adds them to the new set. It then adds
the modified subset (with the requested attributes removed). And then adds the
rest of the subsets.
llvm-svn: 173660
Bill Wendling [Mon, 28 Jan 2013 05:23:28 +0000 (05:23 +0000)]
Rewrite the addAttr() method.
This now uses the AttributeSet object instead of the Attribute /
AttributeWithIndex objects. It's fairly simple now. It goes through all of the
subsets before the one we're modifying, adds them to the new set. It then adds
the modified subset. And then adds the rest of the subsets.
llvm-svn: 173659
Hal Finkel [Mon, 28 Jan 2013 04:37:38 +0000 (04:37 +0000)]
Fix a typo in the name of a tooling unit test
llvm-svn: 173658
Hal Finkel [Mon, 28 Jan 2013 04:37:37 +0000 (04:37 +0000)]
Fix the indentation of the first line of preprocessor output
The -E output from clang did not produce the correct indentation on the first line.
This is because MoveToLine returned false, and when this happens,
the regular process for producing initial indentation is skipped.
Thanks to Eli for suggesting a way to simplify this to a one-line change.
llvm-svn: 173657
NAKAMURA Takumi [Mon, 28 Jan 2013 04:29:01 +0000 (04:29 +0000)]
Revert r173646, "Use proper type for the index."
Unfortunately, msvc miscompiles it. Investigating.
llvm-svn: 173656
Michael J. Spencer [Mon, 28 Jan 2013 04:15:44 +0000 (04:15 +0000)]
[ELF] Use entry point from LinkerOptions.
Patch by Ahmed Bougacha!
llvm-svn: 173655
Michael Gottesman [Mon, 28 Jan 2013 04:12:07 +0000 (04:12 +0000)]
Refactored out pass ObjCARCAPElim from ObjCARCOpts.cpp => ObjCARCAPElim.cpp.
llvm-svn: 173654
Michael Gottesman [Mon, 28 Jan 2013 03:35:20 +0000 (03:35 +0000)]
Fixed case insensitive issue.
llvm-svn: 173653
Michael Gottesman [Mon, 28 Jan 2013 03:30:34 +0000 (03:30 +0000)]
Removed extraneous doxygen end module statement.
llvm-svn: 173652
Michael Gottesman [Mon, 28 Jan 2013 03:28:38 +0000 (03:28 +0000)]
Extracted pass ObjCARCExpand from ObjCARC.cpp => ObjCARCExpand.cpp.
I also added the local header ObjCARC.h for common functions used by the
various passes.
llvm-svn: 173651
Edwin Vane [Mon, 28 Jan 2013 03:19:57 +0000 (03:19 +0000)]
Improved svn repo searching for 'make update'
Use a simple recursive bash function to search for svn repos for the 'make
update' target thus including projects like clang-tools-extra.
Reviewers: bkramer, echristo
llvm-svn: 173650
Reed Kotler [Mon, 28 Jan 2013 02:46:49 +0000 (02:46 +0000)]
Make some code a little simpler.
llvm-svn: 173649
Michael Gottesman [Mon, 28 Jan 2013 01:36:00 +0000 (01:36 +0000)]
Since ObjCARC has been refactored into its own library with its own declaration header, we need to include the declaration header alongside Scalar.h in BackendUtil.
llvm-svn: 173648
Michael Gottesman [Mon, 28 Jan 2013 01:35:51 +0000 (01:35 +0000)]
Extracted ObjCARC.cpp into its own library libLLVMObjCARCOpts in preparation for refactoring the ARC Optimizer.
llvm-svn: 173647
Bill Wendling [Mon, 28 Jan 2013 01:30:29 +0000 (01:30 +0000)]
Use proper type for the index.
llvm-svn: 173646
Richard Smith [Mon, 28 Jan 2013 01:15:38 +0000 (01:15 +0000)]
Tests and a minor bugfix for [dcl.attr.depend]p1 (C++11 [[carries_dependency]]
attribute).
llvm-svn: 173645
Bill Wendling [Mon, 28 Jan 2013 01:11:42 +0000 (01:11 +0000)]
Remove another use of AttributeWithIndex, using the AttributeSetImpl accessors instead.
llvm-svn: 173644
Richard Smith [Mon, 28 Jan 2013 00:54:05 +0000 (00:54 +0000)]
Add a -pedantic warning: an anonymous union within an anonymous union is not
permitted in standard C++, despite being silently accepted by many (all?) major
C++ implementations.
llvm-svn: 173643
Bill Wendling [Mon, 28 Jan 2013 00:46:02 +0000 (00:46 +0000)]
Remove another use of AttributeWithIndex, using the AttributeSetImpl accessors instead.
llvm-svn: 173642
Bill Wendling [Mon, 28 Jan 2013 00:21:34 +0000 (00:21 +0000)]
Remove a use of AttributeWithIndex.
We want to remove AttributeWithIndex because it provides a non-encapsulated view
of the AttributeSetImpl object. Instead, use accessor methods and iterators.
Eventually, this code can be simplified because the Attribute object will hold
only one attribute instead of multiple attributes.
llvm-svn: 173641
Bill Wendling [Sun, 27 Jan 2013 23:53:56 +0000 (23:53 +0000)]
Improve the debug output a bit.
llvm-svn: 173640
Bill Wendling [Sun, 27 Jan 2013 23:50:44 +0000 (23:50 +0000)]
Use proper return type for attribute index.
llvm-svn: 173639
Bill Wendling [Sun, 27 Jan 2013 23:49:44 +0000 (23:49 +0000)]
Use proper return type for attribute index.
llvm-svn: 173638
Bill Wendling [Sun, 27 Jan 2013 23:41:29 +0000 (23:41 +0000)]
Push the calculation of the 'Raw' attribute mask down into the implementation. It in turn uses the correct list for calculating the 'Raw' value.
llvm-svn: 173637
Bill Wendling [Sun, 27 Jan 2013 22:46:17 +0000 (22:46 +0000)]
Don't erase these methods. They're used during testing.
llvm-svn: 173636
Bill Wendling [Sun, 27 Jan 2013 22:43:04 +0000 (22:43 +0000)]
Add special 'get' methods to create an Attribute with an alignment. Also do some random cleanup. No functionality change.
llvm-svn: 173635
Richard Osborne [Sun, 27 Jan 2013 22:28:30 +0000 (22:28 +0000)]
[XCore] Add missing l2rus instructions.
These instructions are not targeted by the compiler but they are
needed for the MC layer.
llvm-svn: 173634
Dmitri Gribenko [Sun, 27 Jan 2013 22:11:38 +0000 (22:11 +0000)]
Comment parsing: fix test in r173626 for -fms-extensions -fms-compatibility
llvm-svn: 173633
Bill Wendling [Sun, 27 Jan 2013 21:38:03 +0000 (21:38 +0000)]
Privitize some the copy c'tor and assignment operator of uniquified objects.
llvm-svn: 173632
Bill Wendling [Sun, 27 Jan 2013 21:32:11 +0000 (21:32 +0000)]
Add some helpful comments.
llvm-svn: 173631
Dmitri Gribenko [Sun, 27 Jan 2013 21:28:24 +0000 (21:28 +0000)]
Decl printer: fix CXXConstructExpr with implicit default argument
Patch by Will Wilson.
llvm-svn: 173630
Richard Osborne [Sun, 27 Jan 2013 21:26:02 +0000 (21:26 +0000)]
[XCore] Add missing l2r instructions.
These instructions are not targeted by the compiler but they are
needed for the MC layer.
llvm-svn: 173629
Bill Wendling [Sun, 27 Jan 2013 21:23:46 +0000 (21:23 +0000)]
s/AttrList/pImpl/g in AttributeSet. No functionality change.
llvm-svn: 173628
Bill Wendling [Sun, 27 Jan 2013 21:20:06 +0000 (21:20 +0000)]
Start using more of the AttrNode in the AttributeSetImpl class.
Also add some asserts.
llvm-svn: 173627
Dmitri Gribenko [Sun, 27 Jan 2013 21:18:39 +0000 (21:18 +0000)]
Comment parsing: attach any tag type's documentation to its typedef if latter
does not have one of its own. // rdar://
13067629
Original patch (r173586 and r173587) by Fariborz Jahanian, modified by me.
llvm-svn: 173626
Benjamin Kramer [Sun, 27 Jan 2013 21:02:52 +0000 (21:02 +0000)]
Legalizer: Reword comment again, per Duncan's suggestion.
llvm-svn: 173625
Richard Osborne [Sun, 27 Jan 2013 20:46:21 +0000 (20:46 +0000)]
[XCore] Add missing 1r instructions.
These instructions are not targeted by the compiler but they are
needed for the MC layer.
llvm-svn: 173624
Richard Osborne [Sun, 27 Jan 2013 20:42:57 +0000 (20:42 +0000)]
[XCore] Add missing 0r instructions.
These instructions are not targeted by the compiler but they are
needed for the MC layer.
llvm-svn: 173623
Jordan Rose [Sun, 27 Jan 2013 20:12:04 +0000 (20:12 +0000)]
PR15067: Don't assert when a UCN appears in a C90 file.
Unfortunately, we can't accept the UCN as an extension because we're
required to treat it as two tokens for preprocessing purposes.
llvm-svn: 173622
Hal Finkel [Sun, 27 Jan 2013 20:07:01 +0000 (20:07 +0000)]
BBVectorize: Better use of TTI->getShuffleCost
When flipping the pair of subvectors that form a vector, if the
vector length is 2, we can use the SK_Reverse shuffle kind to get
more-accurate cost information. Also we can use the SK_ExtractSubvector
shuffle kind to get accurate subvector extraction costs.
The current cost model implementations don't yet seem complex enough
for this to make a difference (thus, there are no test cases with this
commit), but it should help in future.
Depending on how the various targets optimize and combine shuffles in
practice, we might be able to get more-accurate costs by combining the
costs of multiple shuffle kinds. For example, the cost of flipping the
subvector pairs could be modeled as two extractions and two subvector
insertions. These changes, however, should probably be motivated
by specific test cases.
llvm-svn: 173621
Benjamin Kramer [Sun, 27 Jan 2013 15:04:43 +0000 (15:04 +0000)]
Legalizer: Add an assert and tweak a comment to clarify the assumptions this code makes.
llvm-svn: 173620
NAKAMURA Takumi [Sun, 27 Jan 2013 14:12:25 +0000 (14:12 +0000)]
AddCompilerRT.cmake: Try to unbreak since r173617.
llvm-svn: 173619
Bill Wendling [Sun, 27 Jan 2013 12:50:02 +0000 (12:50 +0000)]
In the AttributeSetImpl c'tor, fill in the AttrNodes data structure with the attributes being passed in.
llvm-svn: 173618
NAKAMURA Takumi [Sun, 27 Jan 2013 12:46:53 +0000 (12:46 +0000)]
[CMake][Lit][unittests] Cleanup in AddLLVM.cmake.
llvm-svn: 173617
NAKAMURA Takumi [Sun, 27 Jan 2013 12:20:50 +0000 (12:20 +0000)]
[CMake][Lit][unittests] Deprecate CMAKE_BUILD_TYPE in each build directory for unittests.
For example,
cur) unittests/ADT/Release/ADTTests
new) unittests/ADT/ADTTests
RUNTIME_BUILD_MODE can be substituted to CMAKE_CFG_INTDIR.
With Make and Ninja, the tree is not built with multiple configurations.
Then, including the build type in target directory doesn't make sense.
See also "How can I build multiple modes without switching?"
http://www.cmake.org/Wiki/CMake_FAQ
CMAKE_CFG_INTDIR is set to "."
With multiple-configuration-aware build system, like Visual Studio, each unittest is built on appropriate directory, for example,
unittests/ADT/Release/ADTTests.exe
CMAKE_CFG_INTDIR is set to build system's variable, like "$(Configuration)" or "$(OutDir)".
Thus, "--param build_config" is also deprecated.
llvm-svn: 173616
Benjamin Kramer [Sun, 27 Jan 2013 11:19:11 +0000 (11:19 +0000)]
When the legalizer is splitting vector shifts, the result may not have the right shift amount type.
Fix that by adding a cast to the shift expander. This came up with vector shifts
on sse-less X86 CPUs.
<2 x i64> = shl <2 x i64> <2 x i64>
-> i64,i64 = shl i64 i64; shl i64 i64
-> i32,i32,i32,i32 = shl_parts i32 i32 i64; shl_parts i32 i32 i64
Now we cast the last two i64s to the right type. Fixes the crash in PR14668.
llvm-svn: 173615
NAKAMURA Takumi [Sun, 27 Jan 2013 11:09:41 +0000 (11:09 +0000)]
Revert r173586 (and r173587) , "Attach any tag type's documentation to its typedef if"
It caused unexpected warnings with @tparam.
llvm-svn: 173614
Bill Wendling [Sun, 27 Jan 2013 10:28:39 +0000 (10:28 +0000)]
Fix miscompile. Add back the use of the ArrayRef version of the ::get method.
llvm-svn: 173613
Bill Wendling [Sun, 27 Jan 2013 10:00:13 +0000 (10:00 +0000)]
Rearrange some deckchairs. Moving the class before it's use.
llvm-svn: 173612
Bill Wendling [Sun, 27 Jan 2013 09:55:44 +0000 (09:55 +0000)]
Remove dead methods.
llvm-svn: 173611