David Blaikie [Tue, 13 Nov 2012 18:51:45 +0000 (18:51 +0000)]
Simplify function try/catch scope handling.
Based on post-commit review feedback for r167766 by Richard Smith.
llvm-svn: 167856
Ulrich Weigand [Tue, 13 Nov 2012 18:40:58 +0000 (18:40 +0000)]
Do not consider a machine instruction that uses and defines the same
physical register as candidate for common subexpression elimination
in MachineCSE.
This fixes a bug on PowerPC in MultiSource/Applications/oggenc/oggenc
caused by MachineCSE invalidly merging two separate DYNALLOC insns.
llvm-svn: 167855
Shankar Easwaran [Tue, 13 Nov 2012 18:39:10 +0000 (18:39 +0000)]
Adding support to resolve symbols with archive libraries in lld
llvm-svn: 167854
Shankar Easwaran [Tue, 13 Nov 2012 18:38:42 +0000 (18:38 +0000)]
Adding changes to support GNU style archive library reading
llvm-svn: 167853
Shankar Easwaran [Tue, 13 Nov 2012 18:26:22 +0000 (18:26 +0000)]
Adding convenience function to ELF Header
llvm-svn: 167852
Andrew Trick [Tue, 13 Nov 2012 17:37:46 +0000 (17:37 +0000)]
Fix -join-splitedges: my previous "cleanup" broke it.
Working on reducing unit tests.
This won't be enabled unless a subtarget enables misched.
llvm-svn: 167851
Daniel Jasper [Tue, 13 Nov 2012 17:14:11 +0000 (17:14 +0000)]
Fix AST-matcher descendant visiting for Types, TypeLocs and NestedNamespecifierLocs.
The RecursiveASTVisitor assumes that any given Traverse-method can be called with a NULL-node. So the subclass needs to handle these appropriately.
llvm-svn: 167850
Chad Rosier [Tue, 13 Nov 2012 16:42:19 +0000 (16:42 +0000)]
Revert 167755/167760. We don't want to emit crash diagnostics on command-line syntax errors.
llvm-svn: 167849
Howard Hinnant [Tue, 13 Nov 2012 15:57:18 +0000 (15:57 +0000)]
Update CREDIT.TXT
llvm-svn: 167847
Benjamin Kramer [Tue, 13 Nov 2012 15:32:35 +0000 (15:32 +0000)]
This patch makes the behavior of clang consistent with the behavior of gcc 4.6 in cases where both -fPIC and -fPIE is used.
- Separately check if -fPIE was specified in the command line and define both __PIC__ and __PIE__ when -fPIE is used. We need to check this separately because -fPIE will infer -fPIC even if its not explicitly used.
- Fixed existing tests.
- Added new tests for cases where both -fPIC and -fPIE is used.
Author: Tareq A. Siraj <tareq.a.siraj@intel.com>
Fixes: PR13221
Review: http://llvm-reviews.chandlerc.com/D94
llvm-svn: 167846
NAKAMURA Takumi [Tue, 13 Nov 2012 15:03:33 +0000 (15:03 +0000)]
llvm/test/Other/close-stderr.ll: Mark it as XFAIL:mingw32 for now.
On MSYS, 70 is not seen, but 1.
r127726 should be reworked. Candidate options are;
1) Use not exit(70), but _exit(70), in report_fatal_error().
2) Return with _exit(70) in ~raw_ostream().
llvm-svn: 167836
Dmitry Vyukov [Tue, 13 Nov 2012 14:13:21 +0000 (14:13 +0000)]
tsan: remove unused const
llvm-svn: 167835
Dmitry Vyukov [Tue, 13 Nov 2012 14:05:58 +0000 (14:05 +0000)]
tsan: better function names
llvm-svn: 167834
Alexey Samsonov [Tue, 13 Nov 2012 14:05:02 +0000 (14:05 +0000)]
[TSan] Add output test for write under reader lock
llvm-svn: 167833
Dmitry Vyukov [Tue, 13 Nov 2012 13:53:43 +0000 (13:53 +0000)]
tsan: fix stats collection
llvm-svn: 167832
Duncan Sands [Tue, 13 Nov 2012 13:01:58 +0000 (13:01 +0000)]
Codegen support for arbitrary vector getelementptrs.
llvm-svn: 167830
Duncan Sands [Tue, 13 Nov 2012 13:01:00 +0000 (13:01 +0000)]
Fix the instcombine GEP index widening transform to work correctly for vector
getelementptrs.
llvm-svn: 167829
Duncan Sands [Tue, 13 Nov 2012 12:59:33 +0000 (12:59 +0000)]
Relax the restrictions on vector of pointer types, and vector getelementptr.
Previously in a vector of pointers, the pointer couldn't be any pointer type,
it had to be a pointer to an integer or floating point type. This is a hassle
for dragonegg because the GCC vectorizer happily produces vectors of pointers
where the pointer is a pointer to a struct or whatever. Vector getelementptr
was restricted to just one index, but now that vectors of pointers can have
any pointer type it is more natural to allow arbitrary vector getelementptrs.
There is however the issue of struct GEPs, where if each lane chose different
struct fields then from that point on each lane will be working down into
unrelated types. This seems like too much pain for too little gain, so when
you have a vector struct index all the elements are required to be the same.
llvm-svn: 167828
Benjamin Kramer [Tue, 13 Nov 2012 12:12:02 +0000 (12:12 +0000)]
DependenceAnalysis: Print all dependency pairs when dumping. Update all testcases.
Part of a patch by Preston Briggs.
llvm-svn: 167827
Andrew Trick [Tue, 13 Nov 2012 08:47:29 +0000 (08:47 +0000)]
misched: Allow subtargets to enable misched and dependent options.
This allows me to begin enabling (or backing out) misched by default
for one subtarget at a time. To run misched we typically want to:
- Disable SelectionDAG scheduling (use the source order scheduler)
- Enable more aggressive coalescing (until we decide to always run the coalescer this way)
- Enable MachineScheduler pass itself.
Disabling PostRA sched may follow for some subtargets.
llvm-svn: 167826
Andrew Trick [Tue, 13 Nov 2012 08:47:25 +0000 (08:47 +0000)]
Added RegisterCoalescer support for joining global copies first.
This adds the -join-globalcopies option which can be enabled by
default once misched is also enabled.
Ideally, the register coalescer would be able to split local live
ranges in a way that produces copies that can be easily resolved by
the scheduler. Until then, this heuristic should be good enough to at
least allow the scheduler to run after coalescing.
llvm-svn: 167825
Alexey Samsonov [Tue, 13 Nov 2012 07:55:16 +0000 (07:55 +0000)]
[ASan] now we don't need llvm_obj_root to access llvm-symbolizer tool
llvm-svn: 167823
Alexey Samsonov [Tue, 13 Nov 2012 07:15:32 +0000 (07:15 +0000)]
Figure out <size> argument of llvm.lifetime intrinsics at the moment they are created (during function inlining)
llvm-svn: 167821
Jyotsna Verma [Tue, 13 Nov 2012 06:31:55 +0000 (06:31 +0000)]
Test commit.
Add a blank line.
llvm-svn: 167819
Nico Weber [Tue, 13 Nov 2012 06:25:15 +0000 (06:25 +0000)]
UCNs in char literals are done (in LiteralSupport), remove FIXME. Expand UCN FIXME in LexNumericConstant.
llvm-svn: 167818
Hal Finkel [Tue, 13 Nov 2012 05:54:54 +0000 (05:54 +0000)]
BBVectorize: Remove temporary assert used for debugging
llvm-svn: 167817
Argyrios Kyrtzidis [Tue, 13 Nov 2012 05:07:23 +0000 (05:07 +0000)]
Copy the decls returned by DeclContext::lookup_result to a
new container so we can safely iterate over them.
The container holding the lookup decls can under certain conditions
be changed while iterating (e.g. because of deserialization).
llvm-svn: 167816
Meador Inge [Tue, 13 Nov 2012 04:16:17 +0000 (04:16 +0000)]
instcombine: Migrate math library call simplifications
This patch migrates the math library call simplifications from the
simplify-libcalls pass into the instcombine library call simplifier.
I have typically migrated just one simplifier at a time, but the math
simplifiers are interdependent because:
1. CosOpt, PowOpt, and Exp2Opt all depend on UnaryDoubleFPOpt.
2. CosOpt, PowOpt, Exp2Opt, and UnaryDoubleFPOpt all depend on
the option -enable-double-float-shrink.
These two factors made migrating each of these simplifiers individually
more of a pain than it would be worth. So, I migrated them all together.
llvm-svn: 167815
Anna Zaks [Tue, 13 Nov 2012 03:34:49 +0000 (03:34 +0000)]
Add a test that shows that reporting a leak after failure to free is
tricky.
llvm-svn: 167814
Anna Zaks [Tue, 13 Nov 2012 03:18:01 +0000 (03:18 +0000)]
Fix a Malloc Checker FP by tracking return values from initWithCharacter
and other functions.
When these functions return null, the pointer is not freed by
them/ownership is not transfered. So we should allow the user to free
the pointer by calling another function when the return value is NULL.
llvm-svn: 167813
Evan Cheng [Tue, 13 Nov 2012 03:14:16 +0000 (03:14 +0000)]
Remove virtual keyword for two routines that should never be overridden.
llvm-svn: 167812
Hal Finkel [Tue, 13 Nov 2012 03:12:40 +0000 (03:12 +0000)]
BBVectorize: Don't vectorize vector-manipulation chains
Don't choose a vectorization plan containing only shuffles and
vector inserts/extracts. Due to inperfections in the cost model,
these can lead to infinite recusion.
llvm-svn: 167811
Enrico Granata [Tue, 13 Nov 2012 02:57:43 +0000 (02:57 +0000)]
Giving at least some error information when a Python exception happens during command script import
llvm-svn: 167810
Evan Cheng [Tue, 13 Nov 2012 02:56:38 +0000 (02:56 +0000)]
Revert r167759. Ben is right this isn't likely to help much.
llvm-svn: 167809
Andrew Trick [Tue, 13 Nov 2012 02:35:06 +0000 (02:35 +0000)]
misched: Don't consider artificial edges weak edges.
For now be more conservative in case other out-of-tree schedulers rely
on the old behavior of artificial edges.
llvm-svn: 167808
Bill Wendling [Tue, 13 Nov 2012 02:31:58 +0000 (02:31 +0000)]
Use the 'count' attribute instead of the 'upper_bound' attribute.
If we have a type 'int a[1]' and a type 'int b[0]', the generated DWARF is the
same for both of them because we use the 'upper_bound' attribute. Instead use
the 'count' attrbute, which gives the correct number of elements in the array.
<rdar://problem/
12566646>
llvm-svn: 167807
Bill Wendling [Tue, 13 Nov 2012 02:31:47 +0000 (02:31 +0000)]
Use the 'count' attribute instead of the 'upper_bound' attribute.
If we have a type 'int a[1]' and a type 'int b[0]', the generated DWARF is the
same for both of them because we use the 'upper_bound' attribute. Instead use
the 'count' attrbute, which gives the correct number of elements in the array.
<rdar://problem/
12566646>
llvm-svn: 167806
Enrico Granata [Tue, 13 Nov 2012 02:05:45 +0000 (02:05 +0000)]
<rdar://problem/
11814875>
If a ValueObjectDynamic has no formatter, try using its static type to figure one out
llvm-svn: 167803
Eli Friedman [Tue, 13 Nov 2012 02:05:15 +0000 (02:05 +0000)]
Fix IR generation for bool on PPC (and any other target where bool is not 8 bits in memory).
PR11777.
llvm-svn: 167802
Argyrios Kyrtzidis [Tue, 13 Nov 2012 01:03:15 +0000 (01:03 +0000)]
[preprocessor] When #including something that contributes no tokens at all,
don't recursively continue lexing.
This avoids a stack overflow with a sequence of many empty #includes.
rdar://
11988695
llvm-svn: 167801
Argyrios Kyrtzidis [Tue, 13 Nov 2012 01:02:40 +0000 (01:02 +0000)]
In Lexer::LexTokenInternal, avoid code duplication; no functionality change.
llvm-svn: 167800
Bill Wendling [Tue, 13 Nov 2012 00:54:24 +0000 (00:54 +0000)]
Disable accelerator tables when compiling with LTO.
LTO doesn't generate correct accelerator tables. This is due to the general lack
correct of debug info for LTO. Disable it when using LTO.
<rdar://problem/
12401423>
llvm-svn: 167799
Richard Smith [Tue, 13 Nov 2012 00:54:12 +0000 (00:54 +0000)]
Fix some wrong-code bugs in implicitly-defined assignment operators:
- In C++11, perform overload resolution over all assignment operators, rather than just looking for copy/move assignment operators.
- Clean up after temporaries produced by operator= immediately, rather than accumulating them until the end of the function.
llvm-svn: 167798
Andrew Trick [Tue, 13 Nov 2012 00:34:44 +0000 (00:34 +0000)]
Cleanup the main RegisterCoalescer loop.
Block priorities still apply outside loops.
llvm-svn: 167793
Kaelyn Uhrain [Tue, 13 Nov 2012 00:18:47 +0000 (00:18 +0000)]
A couple of small fixes to r167783
llvm-svn: 167791
Anna Zaks [Tue, 13 Nov 2012 00:13:44 +0000 (00:13 +0000)]
[analyzer] Address Jordan's feedback for r167780.
llvm-svn: 167790
Ted Kremenek [Tue, 13 Nov 2012 00:12:13 +0000 (00:12 +0000)]
Fix bad CFG construction bug when handling C++ 'try' statements.
This code assigned the last created CFGBlock* to the variable 'Block',
which is a scratch variable which is null'ed out after a block is
completed. By assigning the last created block to 'Block', we start
editing a completed block, inserting CFGStmts that should be in
another block. This was the case with 'try'. The test case that
showed this had a while loop inside a 'try', and the logic before
the while loop was being included as part of the "condition block"
for the loop. This showed up as a bogus dead store, but could
have lots of implications.
Turns out this bug was replicated a few times within CFG.cpp, so
I went and fixed up those as well.
llvm-svn: 167788
Shuxin Yang [Tue, 13 Nov 2012 00:08:49 +0000 (00:08 +0000)]
revert r167740
llvm-svn: 167787
Nick Lewycky [Tue, 13 Nov 2012 00:08:34 +0000 (00:08 +0000)]
When filtering the list of associated namespaces so that we don't suggest people
add functions to namespace 'std', also filter out namespaces with '__' anywhere
in the name.
llvm-svn: 167786
Andrew Trick [Mon, 12 Nov 2012 23:59:48 +0000 (23:59 +0000)]
Cleanup -join-splitedges. Make the loop more obvious.
llvm-svn: 167785
Hal Finkel [Mon, 12 Nov 2012 23:55:36 +0000 (23:55 +0000)]
BBVectorize: Only some insert element operand pairs are free.
This fixes another infinite recursion case when using target costs.
We can only replace insert element input chains that are pure (end
with inserting into an undef).
llvm-svn: 167784
Kaelyn Uhrain [Mon, 12 Nov 2012 23:48:05 +0000 (23:48 +0000)]
Enable C++11 attribute syntax for warn_unused_result and allow it to be
applied to CXXRecordDecls, where functions with that return type will
inherit the warn_unused_result attribute.
Also includes a tiny fix (with no discernable behavior change for
existing code) to re-sync AttributeDeclKind enum and
err_attribute_wrong_decl_type with warn_attribute_wrong_decl_type since
the enum is used with both diagnostic messages to chose the correct
description.
llvm-svn: 167783
Jean-Daniel Dupas [Mon, 12 Nov 2012 23:43:34 +0000 (23:43 +0000)]
Update link to external document
llvm-svn: 167781
Anna Zaks [Mon, 12 Nov 2012 23:40:29 +0000 (23:40 +0000)]
[analyzer] Follow up to r167762 - precisely determine the adjustment
conditions.
The adjustment is needed only in case of dynamic dispatch performed by
the analyzer - when the runtime declaration is different from the static
one.
Document this explicitly in the code (by adding a helper). Also, use
canonical Decls to avoid matching against the case where the definition
is different from found declaration.
This fix suppresses the testcase I added in r167762, so add another
testcase to make sure we do test commit r167762.
llvm-svn: 167780
Richard Smith [Mon, 12 Nov 2012 23:33:00 +0000 (23:33 +0000)]
Factor duplicated implicit memcpy call generation code out of copy/move
assignment generation. This incidentally avoids reusing the same Expr* across
multiple statements in the same object; that was generating slightly broken
ASTs, but I couldn't trigger any observable bad behavior, so no test.
llvm-svn: 167779
Greg Clayton [Mon, 12 Nov 2012 23:15:22 +0000 (23:15 +0000)]
Patch from Andrew Kaylor that fixes a race condition in the Listener.cpp.
llvm-svn: 167778
Eric Christopher [Mon, 12 Nov 2012 23:13:34 +0000 (23:13 +0000)]
Finish reverting r167761, it's causing test failures.
llvm-svn: 167777
Greg Clayton [Mon, 12 Nov 2012 22:54:26 +0000 (22:54 +0000)]
New compiler warnings caught issues with the m_encoding_uid field that should have been a lldb::user_id_t type, but was a uint32_t a long time ago and never got updated.
llvm-svn: 167774
Greg Clayton [Mon, 12 Nov 2012 22:53:16 +0000 (22:53 +0000)]
Fixed an error in the ELF parser that was comparing a bool to 4 causing 32 bit ELF relocations to get parsed incorrectly.
llvm-svn: 167773
Daniel Malea [Mon, 12 Nov 2012 22:43:13 +0000 (22:43 +0000)]
Fix typo (dwarf/dsym mismatch) in testcase causing early failure on Linux
llvm-svn: 167771
Daniel Malea [Mon, 12 Nov 2012 22:34:24 +0000 (22:34 +0000)]
Fix minor typo on build page.
llvm-svn: 167770
Michael Liao [Mon, 12 Nov 2012 22:33:18 +0000 (22:33 +0000)]
Fix test case added in patch fixing PR14314
llvm-svn: 167769
Daniel Malea [Mon, 12 Nov 2012 22:32:00 +0000 (22:32 +0000)]
Update platform section to mention Linux/FreeBSD local debugging support, and also mention that LLDB builds on Linux with clang and libstdc++/libc++.
llvm-svn: 167768
Rafael Espindola [Mon, 12 Nov 2012 22:29:43 +0000 (22:29 +0000)]
Remove duplicated line.
llvm-svn: 167767
David Blaikie [Mon, 12 Nov 2012 22:25:41 +0000 (22:25 +0000)]
Fix more try scoping bugs introduced by r167650.
Introduces more clear scoping flags & flag combinations which should hopefully
be more understandable.
llvm-svn: 167766
Eric Christopher [Mon, 12 Nov 2012 22:22:20 +0000 (22:22 +0000)]
Add an option to enable prototype "fission" capabilities and debug changes.
llvm-svn: 167765
Daniel Malea [Mon, 12 Nov 2012 22:16:27 +0000 (22:16 +0000)]
Replace const_iterator with iterator to build against libstdc++
- libstdc++ defines vector::erase(iterator) but not vector::erase(const_iterator)
llvm-svn: 167764
Bill Wendling [Mon, 12 Nov 2012 22:14:27 +0000 (22:14 +0000)]
That's causing an error.
llvm-svn: 167763
Anna Zaks [Mon, 12 Nov 2012 22:06:24 +0000 (22:06 +0000)]
[analyzer] Fix a regression (from r 165079): compare canonical types.
Suppresses a leak false positive (radar://
12663777).
In addition, we'll need to rewrite the adjustReturnValue() method not to
return UnknownVal by default, but rather assert in cases we cannot
handle. To make it possible, we need to correctly handle some of the
edge cases we already know about.
llvm-svn: 167762
Bill Wendling [Mon, 12 Nov 2012 22:01:56 +0000 (22:01 +0000)]
Don't test for incomplete types.
llvm-svn: 167761
Chad Rosier [Mon, 12 Nov 2012 21:51:08 +0000 (21:51 +0000)]
Update test case for r167754/r167755.
llvm-svn: 167760
Evan Cheng [Mon, 12 Nov 2012 21:42:53 +0000 (21:42 +0000)]
Cache size of PassVector to speed up getNumContainedPasses().
getNumContainedPasses() used to compute the size of the vector on demand. It is
called repeated in loops (such as runOnFunction()) and it can be updated while
inside the loop.
llvm-svn: 167759
Andrew Trick [Mon, 12 Nov 2012 21:42:40 +0000 (21:42 +0000)]
Added a temporary option to avoid critical edges splitting.
This teaches the register coalescer to be less prone to split critical
edges. I am currently benchmarking this with the new (post-coalescer)
scheduler. I plan to enable this by default and remove the option as
soon as misched is enabled.
llvm-svn: 167758
Eric Christopher [Mon, 12 Nov 2012 21:40:38 +0000 (21:40 +0000)]
Rewrite DIContext interface to take an object. Update all callers.
llvm-svn: 167757
Richard Smith [Mon, 12 Nov 2012 21:38:00 +0000 (21:38 +0000)]
Remove abuse of hasTrivial*, and fix miscompile wherein global arrays with
internal linkage, no uses, trivial construction, and nontrivial destruction
were not emitted.
llvm-svn: 167756
Chad Rosier [Mon, 12 Nov 2012 21:32:44 +0000 (21:32 +0000)]
Revert r167620; this can be implemented using an existing CL option.
llvm-svn: 167755
Chad Rosier [Mon, 12 Nov 2012 21:32:24 +0000 (21:32 +0000)]
If we encounter a fatal error, exit with status 70. For BSD systems this is
defined as an internal software error. This notifies the driver to report
diagnostics information.
rdar://
11951540
llvm-svn: 167754
Andrew Trick [Mon, 12 Nov 2012 21:28:10 +0000 (21:28 +0000)]
misched: rename interfaceto avoid gcc warnings
llvm-svn: 167753
Greg Clayton [Mon, 12 Nov 2012 21:27:20 +0000 (21:27 +0000)]
<rdar://problem/
12153915>
When uniquing classes against one another we can't depend on any or all of the artificial functions (default ctor, dtor, copy ctor, move ctor, etc) being in each definition. Now we treat those separately and handle those to the best of our ability.
llvm-svn: 167752
Filipe Cabecinhas [Mon, 12 Nov 2012 21:26:32 +0000 (21:26 +0000)]
s/BCPLComment/LineComment/ (llvm r167690)
llvm-svn: 167751
Hal Finkel [Mon, 12 Nov 2012 21:21:02 +0000 (21:21 +0000)]
BBVectorize: Use a more sophisticated check for input cost
The old checking code, which assumed that input shuffles and insert-elements
could always be folded (and thus were free) is too simple.
This can only happen in special circumstances.
Using the simple check caused infinite recursion.
llvm-svn: 167750
Ted Kremenek [Mon, 12 Nov 2012 21:20:48 +0000 (21:20 +0000)]
Per discussion on cfe-dev, re-enable suppression of -Wimplicit-fallthrough on C, but also include dialects of C++ earlier than C++11.
There was enough consensus that we *can* get a good language solution
to have an annotation outside of C++11, and without this annotation
this warning doesn't quite mean's completeness criteria for this
kind of warning. For now, restrict this warning to C++11 (where an
annotation exists), and make this the behavior for the LLVM 3.2 release.
Afterwards, we will hammer out a language solution that we are all
happy with.
llvm-svn: 167749
Bill Wendling [Mon, 12 Nov 2012 21:13:35 +0000 (21:13 +0000)]
Update testcase to show that we don't emit an error for sizes <= 32-bits.
llvm-svn: 167748
Daniel Malea [Mon, 12 Nov 2012 21:02:14 +0000 (21:02 +0000)]
Fix libstdc++ build
- Add missing operator= definition for DelayedAddObjCClassProperty
- needed to be compatible with libstdc++ vector implementation
llvm-svn: 167747
Andrew Trick [Mon, 12 Nov 2012 19:52:20 +0000 (19:52 +0000)]
misched: Target-independent support for MacroFusion.
Uses the infrastructure from r167742 to support clustering instructure
that the target processor can "fuse". e.g. cmp+jmp.
Next step: target hook implementations with test cases, and enable.
llvm-svn: 167744
Hal Finkel [Mon, 12 Nov 2012 19:41:38 +0000 (19:41 +0000)]
BBVectorize: Check the types of compare instructions
The pass would previously assert when trying to compute the cost of
compare instructions with illegal vector types (like struct pointers).
llvm-svn: 167743
Andrew Trick [Mon, 12 Nov 2012 19:40:10 +0000 (19:40 +0000)]
misched: Target-independent support for load/store clustering.
This infrastructure is generally useful for any target that wants to
strongly prefer two instructions to be adjacent after scheduling.
A following checkin will add target-specific hooks with unit
tests. Then this feature will be enabled by default with misched.
llvm-svn: 167742
Chad Rosier [Mon, 12 Nov 2012 19:39:37 +0000 (19:39 +0000)]
If we encounter a fatal error, call the interrupt handler to ensure any
temporary files are removed.
rdar://
12282296
llvm-svn: 167741
Shuxin Yang [Mon, 12 Nov 2012 19:34:11 +0000 (19:34 +0000)]
This change is to fix rdar://
12571717 which is about assertion in Reassociate pass.
The assertion is trigged when the Reassociater tries to transform expression
... + 2 * n * 3 + 2 * m + ...
into:
... + 2 * (n*3 + m).
In the process of the transformation, a helper routine folds the constant 2*3 into 6,
confusing optimizer which is trying the to eliminate the common factor 2, and cannot
find 2 any more.
Review is pending. But I'd like commit first in order to help those who are waiting
for this fix.
llvm-svn: 167740
David Blaikie [Mon, 12 Nov 2012 19:32:32 +0000 (19:32 +0000)]
80 cols & unnecessary braces from r167736.
Post-commit review feedback by Eli Friedman.
llvm-svn: 167739
Andrew Trick [Mon, 12 Nov 2012 19:28:57 +0000 (19:28 +0000)]
misched: Infrastructure for weak DAG edges.
This adds support for weak DAG edges to the general scheduling
infrastructure in preparation for MachineScheduler support for
heuristics based on weak edges.
llvm-svn: 167738
Ulrich Weigand [Mon, 12 Nov 2012 19:13:24 +0000 (19:13 +0000)]
Make TOC order deterministic by using MapVector instead of DenseMap.
llvm-svn: 167737
David Blaikie [Mon, 12 Nov 2012 19:12:12 +0000 (19:12 +0000)]
Correct printing of nested anonymous type member accesses.
Patch by Florent Bruneau!
llvm-svn: 167736
Bill Wendling [Mon, 12 Nov 2012 18:52:32 +0000 (18:52 +0000)]
The register constraint could mean a 16- or 8-bit register.
llvm-svn: 167735
Nadav Rotem [Mon, 12 Nov 2012 18:45:12 +0000 (18:45 +0000)]
fix a spelling mistake
llvm-svn: 167734
Howard Hinnant [Mon, 12 Nov 2012 18:19:15 +0000 (18:19 +0000)]
Wen-Han Gu: Fix for llvm.org/bugs/show_bug.cgi?id=14312 Exception Table out-of-range but still keep walking.
llvm-svn: 167733
Sean Callanan [Mon, 12 Nov 2012 18:00:29 +0000 (18:00 +0000)]
Fixed a potential crash in the Darwin platform.
llvm-svn: 167732
Hal Finkel [Mon, 12 Nov 2012 14:50:59 +0000 (14:50 +0000)]
BBVectorize: Check the input types of shuffles for legality
This fixes a bug where shuffles were being fused such that the
resulting input types were not legal on the target. This would
occur only when both inputs and dependencies were also foldable
operations (such as other shuffles) and there were other connected
pairs in the same block.
llvm-svn: 167731
Alexander Potapenko [Mon, 12 Nov 2012 14:49:58 +0000 (14:49 +0000)]
Don't use __cxa_demangle under MSVC (which doesn't have it)
llvm-svn: 167730
Alexey Samsonov [Mon, 12 Nov 2012 14:47:00 +0000 (14:47 +0000)]
[ASan] fixup for r167725: Don't fetch name of StructType if it is literal
llvm-svn: 167729