platform/upstream/llvm.git
12 years agoobjc++: more atomic property api code for
Fariborz Jahanian [Sun, 8 Jan 2012 19:13:23 +0000 (19:13 +0000)]
objc++: more atomic property api code for
c++ object properties. wip.

llvm-svn: 147750

12 years agoInstCombine: If we have a bit test and a sign test anded/ored together, merge the...
Benjamin Kramer [Sun, 8 Jan 2012 18:32:24 +0000 (18:32 +0000)]
InstCombine: If we have a bit test and a sign test anded/ored together, merge the sign bit into the bit test.

This is common in bit field code, e.g. checking if the first or the last bit of a bit field is set.

llvm-svn: 147749

12 years agoReverted commit #147601 upon Evan's request.
Victor Umansky [Sun, 8 Jan 2012 17:20:33 +0000 (17:20 +0000)]
Reverted commit #147601 upon Evan's request.

llvm-svn: 147748

12 years ago<rdar://problem/10645694>
Greg Clayton [Sun, 8 Jan 2012 05:54:35 +0000 (05:54 +0000)]
<rdar://problem/10645694>

Fixed an ARM backtracing issue where if the previous frame was a thumb
function and it was a tail call so that the current frame returned to
an address that would fall into the next function, we would use the
next function as the basis for how we unwound the previous frame's
registers and of course get things wrong. We now fix the PC code
address using the current ABI plug-in, and the ARM ABI plug-in has
been modified to correctly fix the code address. So when we do the
symbol context lookup, instead of taking an address like 0x1001 and
decrementing 1, and looking up the symbol context for a frame, we
now correctly fix 0x1001 to 0x1000, then decrement that by 1 to
get the correct symbol context.

I added a bunch more logging to "log enable lldb uwnind" to help
us in the future. We now log the PC, FP and SP (if they are available),
and we also dump the "active_row" that we find for unwinding a frame.

llvm-svn: 147747

12 years ago<rdar://problem/10660369>
Greg Clayton [Sun, 8 Jan 2012 05:49:22 +0000 (05:49 +0000)]
<rdar://problem/10660369>

Fixed the default ARM regiter set to be able to refer to "r7" as "fp" for the apple debugserver.

llvm-svn: 147746

12 years agoRemove MCELFStreamer.h.
Rafael Espindola [Sat, 7 Jan 2012 23:18:39 +0000 (23:18 +0000)]
Remove MCELFStreamer.h.

llvm-svn: 147745

12 years agoRemove unused variable.
Rafael Espindola [Sat, 7 Jan 2012 22:52:07 +0000 (22:52 +0000)]
Remove unused variable.

llvm-svn: 147744

12 years agoDon't print a label before .cfi_startproc when we don't need to. This makes
Rafael Espindola [Sat, 7 Jan 2012 22:42:19 +0000 (22:42 +0000)]
Don't print a label before .cfi_startproc when we don't need to. This makes
the produce assembly when using CFI just a bit more readable.

llvm-svn: 147743

12 years agoMake clever use of alignment and padding to shrink GlobalValue.
Benjamin Kramer [Sat, 7 Jan 2012 21:17:16 +0000 (21:17 +0000)]
Make clever use of alignment and padding to shrink GlobalValue.

-8 bytes on x86_64, no change on x86.

llvm-svn: 147742

12 years agoRecursive calls to ValueObject::GetSummaryAsCString() are causing crashes.
Greg Clayton [Sat, 7 Jan 2012 20:58:07 +0000 (20:58 +0000)]
Recursive calls to ValueObject::GetSummaryAsCString() are causing crashes.
The previous approach to controlling the recursion was doing it from
outside the function which is not reliable. Now it is being done inside
the function. This might not solve all of the crashes that we were seeing
since there are other functions that clear the bit that indicates that
the summary is in the process of being generated, but it might solve some.

llvm-svn: 147741

12 years agoMatch SelectionDAG logic for enabling movt.
Jakob Stoklund Olesen [Sat, 7 Jan 2012 20:49:15 +0000 (20:49 +0000)]
Match SelectionDAG logic for enabling movt.

Darwin doesn't do static, and ELF targets only support static.

llvm-svn: 147740

12 years agoFix typo in the X86 backend readme. Patch from Jaeden Amero.
Craig Topper [Sat, 7 Jan 2012 20:35:21 +0000 (20:35 +0000)]
Fix typo in the X86 backend readme. Patch from Jaeden Amero.

llvm-svn: 147739

12 years agoRemove VectorExtras. This unused helper was written for a type of API that is discour...
Benjamin Kramer [Sat, 7 Jan 2012 19:42:13 +0000 (19:42 +0000)]
Remove VectorExtras. This unused helper was written for a type of API that is discouraged now.

llvm-svn: 147738

12 years agoRemove unused include of VectorExtras.h.
Benjamin Kramer [Sat, 7 Jan 2012 19:41:22 +0000 (19:41 +0000)]
Remove unused include of VectorExtras.h.

llvm-svn: 147737

12 years agoPack UsingDecl more.
Benjamin Kramer [Sat, 7 Jan 2012 19:09:05 +0000 (19:09 +0000)]
Pack UsingDecl more.

88 -> 80 bytes on x86_64.

llvm-svn: 147736

12 years agoobjc++: More codegen stuff for atomic properties of c++ objects
Fariborz Jahanian [Sat, 7 Jan 2012 18:56:22 +0000 (18:56 +0000)]
objc++: More codegen stuff for atomic properties of c++ objects
with non-trivial copies. // rdar://6137845

llvm-svn: 147735

12 years agoRemove unnecessary check of hasAVX(). It's already included in hasXMM().
Craig Topper [Sat, 7 Jan 2012 18:48:43 +0000 (18:48 +0000)]
Remove unnecessary check of hasAVX(). It's already included in hasXMM().

llvm-svn: 147734

12 years agoReplace some uses of hasNUsesOfValue(0, X) with !hasAnyUseOfValue(X)
Craig Topper [Sat, 7 Jan 2012 18:31:09 +0000 (18:31 +0000)]
Replace some uses of hasNUsesOfValue(0, X) with !hasAnyUseOfValue(X)

llvm-svn: 147733

12 years ago[analyzer] Fix use-after-free in HandleTranslationUnit.
Anna Zaks [Sat, 7 Jan 2012 16:49:46 +0000 (16:49 +0000)]
[analyzer] Fix use-after-free in HandleTranslationUnit.

A patch by Dmitri Gribenko!

The attached patch fixes a use-after-free in AnalysisConsumer::HandleTranslationUnit.  The problem is that
BugReporter's destructor runs after AnalysisManager has been already
deleted.  The fix introduces a scope to force correct destruction
order.

A crash happens only when reports have been added in AnalysisConsumer::HandleTranslationUnit's BugReporter. We don't have such checkers in clang so no test.

llvm-svn: 147732

12 years agoPort the trick to skip the check for empty buckets from StringMap to DenseMap.
Benjamin Kramer [Sat, 7 Jan 2012 13:12:07 +0000 (13:12 +0000)]
Port the trick to skip the check for empty buckets from StringMap to DenseMap.

This should fix the odd behavior that find() is slower than lookup().

llvm-svn: 147731

12 years agoFixed TypeofExpr AST and code generation.
Abramo Bagnara [Sat, 7 Jan 2012 10:52:36 +0000 (10:52 +0000)]
Fixed TypeofExpr AST and code generation.

llvm-svn: 147730

12 years agoSwitch NamespaceDecl from its own hand-rolled redeclaration chain over
Douglas Gregor [Sat, 7 Jan 2012 09:11:48 +0000 (09:11 +0000)]
Switch NamespaceDecl from its own hand-rolled redeclaration chain over
to Redeclarable<NamespaceDecl>, so that we benefit from the improveed
redeclaration deserialization and merging logic provided by
Redeclarable<T>. Otherwise, no functionality change.

As a drive-by fix, collapse the "inline" bit into the low bit of the
original namespace/anonymous namespace, saving 8 bytes per
NamespaceDecl on x86_64.

llvm-svn: 147729

12 years agoAdd some DAG combines for SUBC/SUBE. If nothing uses the carry/borrow out of subc...
Craig Topper [Sat, 7 Jan 2012 09:06:39 +0000 (09:06 +0000)]
Add some DAG combines for SUBC/SUBE. If nothing uses the carry/borrow out of subc, turn it into a sub. Turn (subc x, x) into 0 with no borrow. Turn (subc x, 0) into x with no borrow. Turn (subc -1, x) into (xor x, -1) with no borrow. Turn sube with no borrow in into subc.

llvm-svn: 147728

12 years agoFix TableGen so that it will emit the correct signature for FastEmit_f:
Cameron Zwarich [Sat, 7 Jan 2012 08:18:37 +0000 (08:18 +0000)]
Fix TableGen so that it will emit the correct signature for FastEmit_f:

  /// FastEmit_f - This method is called by target-independent code
  /// to request that an instruction with the given type, opcode, and
  /// floating-point immediate operand be emitted.
  virtual unsigned FastEmit_f(MVT VT,
                              MVT RetVT,
                              unsigned Opcode,
                              const ConstantFP *FPImm);

Currently, it emits an accidentally overloaded version without the const on the
ConstantFP*. This doesn't affect anything in the tree, since nothing causes that
method to be autogenerated, but I have been playing with some ARM TableGen
refactorings that hit this problem.

llvm-svn: 147727

12 years agoOptimize reserved register coalescing.
Jakob Stoklund Olesen [Sat, 7 Jan 2012 07:39:50 +0000 (07:39 +0000)]
Optimize reserved register coalescing.

Reserved registers don't have proper live ranges, their LiveInterval
simply has a snippet of liveness for each def.  Virtual registers with a
single value that is a copy of a reserved register (typically %esp) can
be coalesced with the reserved register if the live range doesn't
overlap any reserved register defs.

When coalescing with a reserved register, don't modify the reserved
register live range.  Just leave it as a bunch of dead defs.  This
eliminates quadratic coalescer behavior in i386 functions with many
function calls.

PR11699

llvm-svn: 147726

12 years agoUse the 'regalloc' debug tag for most register allocator tracing.
Jakob Stoklund Olesen [Sat, 7 Jan 2012 07:39:47 +0000 (07:39 +0000)]
Use the 'regalloc' debug tag for most register allocator tracing.

llvm-svn: 147725

12 years agoEnable redundant phi elimination after LSR.
Andrew Trick [Sat, 7 Jan 2012 07:08:17 +0000 (07:08 +0000)]
Enable redundant phi elimination after LSR.

This will be more important as we extend the LSR pass in ways that don't rely on the formula solver. In particular, we need it for constructing IV chains.

llvm-svn: 147724

12 years agoMore lambda work: semantic analysis of capturing 'this'. It's a bit complicated...
Eli Friedman [Sat, 7 Jan 2012 04:59:52 +0000 (04:59 +0000)]
More lambda work: semantic analysis of capturing 'this'.  It's a bit complicated, but we have to be careful about when exactly captures are marked given PotentiallyPotentiallyEvaluated contexts.  (Actually, it's not 100% correct yet, but it's close enough for the moment.)

llvm-svn: 147723

12 years ago<rdar://problem/10658091>
Greg Clayton [Sat, 7 Jan 2012 04:21:42 +0000 (04:21 +0000)]
<rdar://problem/10658091>

Fixed dynamic types for objective C to not try and make everything dynamic including base classes.

llvm-svn: 147722

12 years agoFix dead link
Eli Bendersky [Sat, 7 Jan 2012 04:11:27 +0000 (04:11 +0000)]
Fix dead link

llvm-svn: 147721

12 years agoUse getRegForValue() to materialize the address of ARM globals.
Jakob Stoklund Olesen [Sat, 7 Jan 2012 04:07:22 +0000 (04:07 +0000)]
Use getRegForValue() to materialize the address of ARM globals.

This enables basic local CSE, giving us 20% smaller code for
consumer-typeset in -O0 builds.

<rdar://problem/10658692>

llvm-svn: 147720

12 years agoRevert part of r147716. Looks like x87 instructions kill markers are all messed
Evan Cheng [Sat, 7 Jan 2012 03:35:48 +0000 (03:35 +0000)]
Revert part of r147716. Looks like x87 instructions kill markers are all messed
up so branch folding pass can't use the scavenger. :-(  This doesn't breaks
anything currently. It just means targets which do not carefully update kill
markers cannot run post-ra scheduler (not new, it has always been the case).

We should fix this at some point since it's really hacky.

llvm-svn: 147719

12 years agoLSR: Don't optimize loops if an outer loop has no preheader.
Andrew Trick [Sat, 7 Jan 2012 03:16:50 +0000 (03:16 +0000)]
LSR: Don't optimize loops if an outer loop has no preheader.

LoopSimplify may not run on some outer loops, e.g. because of indirect
branches. SCEVExpander simply cannot handle outer loops with no preheaders.
Fixes rdar://10655343 SCEVExpander segfault.

llvm-svn: 147718

12 years agoSplit Finish into Finish and FinishImpl to have a common place to do end of
Rafael Espindola [Sat, 7 Jan 2012 03:13:18 +0000 (03:13 +0000)]
Split Finish into Finish and FinishImpl to have a common place to do end of
file error checking. Use that to error on an unfinished cfi_startproc.

The error is not nice, but is already better than a segmentation fault.

llvm-svn: 147717

12 years agoAdded a late machine instruction copy propagation pass. This catches
Evan Cheng [Sat, 7 Jan 2012 03:02:36 +0000 (03:02 +0000)]
Added a late machine instruction copy propagation pass. This catches
opportunities that only present themselves after late optimizations
such as tail duplication .e.g.
## BB#1:
        movl    %eax, %ecx
        movl    %ecx, %eax
        ret

The register allocator also leaves some of them around (due to false
dep between copies from phi-elimination, etc.)

This required some changes in codegen passes. Post-ra scheduler and the
pseudo-instruction expansion passes have been moved after branch folding
and tail merging. They were before branch folding before because it did
not always update block livein's. That's fixed now. The pass change makes
independently since we want to properly schedule instructions after
branch folding / tail duplication.

rdar://10428165
rdar://10640363

llvm-svn: 147716

12 years agoCopy implicit defs (e.g. r0) when changing tBX_RET to tPOP_RET. This bug is
Evan Cheng [Sat, 7 Jan 2012 02:55:54 +0000 (02:55 +0000)]
Copy implicit defs (e.g. r0) when changing tBX_RET to tPOP_RET. This bug is
exposed with an upcoming change will would delete the copy to return register
because there is no use! It's amazing anything works.

llvm-svn: 147715

12 years ago[analyzer] Add basic format string vulnerability checking.
Anna Zaks [Sat, 7 Jan 2012 02:33:10 +0000 (02:33 +0000)]
[analyzer] Add basic format string vulnerability checking.

We already have a more conservative check in the compiler (if the
format string is not a literal, we warn). Still adding it here for
completeness and since this check is stronger - only triggered if the
format string is tainted.

llvm-svn: 147714

12 years ago[arcmt] Disable tests in mingw, no idea why they are failing there.
Argyrios Kyrtzidis [Sat, 7 Jan 2012 02:24:23 +0000 (02:24 +0000)]
[arcmt] Disable tests in mingw, no idea why they are failing there.

llvm-svn: 147713

12 years agoUse movw+movt in ARMFastISel::ARMMaterializeGV.
Jakob Stoklund Olesen [Sat, 7 Jan 2012 01:47:05 +0000 (01:47 +0000)]
Use movw+movt in ARMFastISel::ARMMaterializeGV.

This eliminates a lot of constant pool entries for -O0 builds of code
with many global variable accesses.

This speeds up -O0 codegen of consumer-typeset by 2x because the
constant island pass no longer has to look at thousands of constant pool
entries.

<rdar://problem/10629774>

llvm-svn: 147712

12 years agoLSR: run DeleteDeadPhis before replaceCongruentPhis.
Andrew Trick [Sat, 7 Jan 2012 01:36:44 +0000 (01:36 +0000)]
LSR: run DeleteDeadPhis before replaceCongruentPhis.

llvm-svn: 147711

12 years agoRefactor.
Devang Patel [Sat, 7 Jan 2012 01:33:34 +0000 (01:33 +0000)]
Refactor.

Store AsmParser info locally. A small step towards emitting match entries for multiple asm variants.

llvm-svn: 147710

12 years agoCleanup comments and argument types related to my previous replaceCongruentPhis checkin.
Andrew Trick [Sat, 7 Jan 2012 01:29:21 +0000 (01:29 +0000)]
Cleanup comments and argument types related to my previous replaceCongruentPhis checkin.

llvm-svn: 147709

12 years agoEliminate an error check that may not work with all asm syntax variants.
Devang Patel [Sat, 7 Jan 2012 01:22:23 +0000 (01:22 +0000)]
Eliminate an error check that may not work with all asm syntax variants.

llvm-svn: 147708

12 years agoExtended replaceCongruentPhis to handle mixed phi types.
Andrew Trick [Sat, 7 Jan 2012 01:12:09 +0000 (01:12 +0000)]
Extended replaceCongruentPhis to handle mixed phi types.

llvm-svn: 147707

12 years agoLambdas: semantic analysis of explicit captures.
Eli Friedman [Sat, 7 Jan 2012 01:08:17 +0000 (01:08 +0000)]
Lambdas: semantic analysis of explicit captures.

This patch (and some of my other commits related to lambdas) is heavily based off of John Freeman's work-in-progress patches.

llvm-svn: 147706

12 years ago[analyzer] Remove CallEnterNodeBuilder and simplify ExprEngine::processCallEnter().
Ted Kremenek [Sat, 7 Jan 2012 01:03:17 +0000 (01:03 +0000)]
[analyzer] Remove CallEnterNodeBuilder and simplify ExprEngine::processCallEnter().

This removes analysis of other translation units, but that was an experimental feature anyway that we will revisit later.

llvm-svn: 147705

12 years agoMake the 'x' constraint work for AVX registers as well.
Eric Christopher [Sat, 7 Jan 2012 01:02:09 +0000 (01:02 +0000)]
Make the 'x' constraint work for AVX registers as well.

Fixes rdar://10614894

llvm-svn: 147704

12 years agoMissing raw_ostream.h breaks MSVC build.
Andrew Trick [Sat, 7 Jan 2012 00:54:28 +0000 (00:54 +0000)]
Missing raw_ostream.h breaks MSVC build.

llvm-svn: 147703

12 years agoPatch from Enrico Granata that moves SBData related functions into the SBData
Greg Clayton [Sat, 7 Jan 2012 00:45:50 +0000 (00:45 +0000)]
Patch from Enrico Granata that moves SBData related functions into the SBData
class instead of requiring a live process in order to be able to create useful
SBData objects.

llvm-svn: 147702

12 years agocomment typo
Andrew Trick [Sat, 7 Jan 2012 00:29:20 +0000 (00:29 +0000)]
comment typo

llvm-svn: 147701

12 years agoExpose isNonConstantNegative to users of ScalarEvolution.
Andrew Trick [Sat, 7 Jan 2012 00:27:31 +0000 (00:27 +0000)]
Expose isNonConstantNegative to users of ScalarEvolution.

llvm-svn: 147700

12 years agoAdd field PaddingType to ABIArgInfo which specifies the type of padding that
Akira Hatanaka [Sat, 7 Jan 2012 00:25:33 +0000 (00:25 +0000)]
Add field PaddingType to ABIArgInfo which specifies the type of padding that
is inserted before the real argument. Padding is needed to ensure the backend
reads from or writes to the correct argument slots when the original alignment
of a byval structure is unavailable due to flattening.

llvm-svn: 147699

12 years agoCorrectly enqueue successors in ExprEngine::processCallExit().
Ted Kremenek [Sat, 7 Jan 2012 00:10:49 +0000 (00:10 +0000)]
Correctly enqueue successors in ExprEngine::processCallExit().

llvm-svn: 147698

12 years ago[analyzer] Remove CallExitNodeBuilder, and have ExprEngine::processCallExit() do...
Ted Kremenek [Sat, 7 Jan 2012 00:00:59 +0000 (00:00 +0000)]
[analyzer] Remove CallExitNodeBuilder, and have ExprEngine::processCallExit() do the work manually.  This is a nice simplification.

Along the way, fix Exprengine::processCallExit() to also perform the postStmt callback for checkers for CallExprs.

llvm-svn: 147697

12 years agoAdd comment.
Chad Rosier [Fri, 6 Jan 2012 23:45:47 +0000 (23:45 +0000)]
Add comment.

llvm-svn: 147696

12 years agoAdd a comment and ensure that anyone else looking at this code doesn't start
Eric Christopher [Fri, 6 Jan 2012 23:03:37 +0000 (23:03 +0000)]
Add a comment and ensure that anyone else looking at this code doesn't start
to bleed from the eyes.

llvm-svn: 147695

12 years agoUse const vector references instead of a vector copy. Spotted by Devang.
Eric Christopher [Fri, 6 Jan 2012 23:03:34 +0000 (23:03 +0000)]
Use const vector references instead of a vector copy. Spotted by Devang.

llvm-svn: 147694

12 years agoUse -> instead of (*iter).
Eric Christopher [Fri, 6 Jan 2012 23:03:27 +0000 (23:03 +0000)]
Use -> instead of (*iter).

llvm-svn: 147693

12 years agoSuppress -Wunused-value within macros from system headers.
Matt Beaumont-Gay [Fri, 6 Jan 2012 22:43:58 +0000 (22:43 +0000)]
Suppress -Wunused-value within macros from system headers.

Along the way, move a helper function from SemaChecking.cpp to a more
accessible home in SourceManager.

llvm-svn: 147692

12 years agoobjc++: more code gen stuff for atomic property api,
Fariborz Jahanian [Fri, 6 Jan 2012 22:33:54 +0000 (22:33 +0000)]
objc++: more code gen stuff for atomic property api,
currently turned off. // rdar://6137845
Also, fixes a test case which should be nonatomic under
new API.

llvm-svn: 147691

12 years agoEnable aligned NEON spilling by default.
Jakob Stoklund Olesen [Fri, 6 Jan 2012 22:19:37 +0000 (22:19 +0000)]
Enable aligned NEON spilling by default.

Experiments show this to be a small speedup for modern ARM cores.

llvm-svn: 147689

12 years ago[analyzer] Make the entries in 'Environment' context-sensitive by making entries...
Ted Kremenek [Fri, 6 Jan 2012 22:09:28 +0000 (22:09 +0000)]
[analyzer] Make the entries in 'Environment' context-sensitive by making entries map from
(Stmt*,LocationContext*) pairs to SVals instead of Stmt* to SVals.

This is needed to support basic IPA via inlining.  Without this, we cannot tell
if a Stmt* binding is part of the current analysis scope (StackFrameContext) or
part of a parent context.

This change introduces an uglification of the use of getSVal(), and thus takes
two steps forward and one step back.  There are also potential performance implications
of enlarging the Environment.  Both can be addressed going forward by refactoring the
APIs and optimizing the internal representation of Environment.  This patch
mainly introduces the functionality upon when we want to build upon (and clean up).

llvm-svn: 147688

12 years agoDuring name lookup, use redecl_iterator to walk over the redeclaration
Douglas Gregor [Fri, 6 Jan 2012 22:05:37 +0000 (22:05 +0000)]
During name lookup, use redecl_iterator to walk over the redeclaration
chain to determine whether any declaration of the given entity is
visible, eliminating the redundant (and less efficient)
getPreviousDeclaration() implementation.

This tweak uncovered an omission in the handling of
RedeclarableTemplateDecl, where we weren't making sure to search for
additional redeclarations of a template in other module files. Things
would be cleaner if RedeclarableTemplateDecl actually used Redeclarable.

llvm-svn: 147687

12 years agoPut all IVUsers in the processed set. Allow querying IVUsers with isIVUserOrOperand.
Andrew Trick [Fri, 6 Jan 2012 21:41:55 +0000 (21:41 +0000)]
Put all IVUsers in the processed set. Allow querying IVUsers with isIVUserOrOperand.

llvm-svn: 147686

12 years agoAbort AdjustBBOffsetsAfter early when possible.
Jakob Stoklund Olesen [Fri, 6 Jan 2012 21:40:15 +0000 (21:40 +0000)]
Abort AdjustBBOffsetsAfter early when possible.

llvm-svn: 147685

12 years agolit/lit.TestRunner: Add an extra_substitutions argument for executeShTest --
Daniel Dunbar [Fri, 6 Jan 2012 21:39:06 +0000 (21:39 +0000)]
lit/lit.TestRunner: Add an extra_substitutions argument for executeShTest --
useful for test suites which want to piggyback onto the "shtest" format style.

llvm-svn: 147684

12 years agoSCEVExpander: hoistStep should check strict dominance.
Andrew Trick [Fri, 6 Jan 2012 21:23:43 +0000 (21:23 +0000)]
SCEVExpander: hoistStep should check strict dominance.

llvm-svn: 147683

12 years agoTracing to help investigate issues with SjLj spill code.
Andrew Trick [Fri, 6 Jan 2012 21:16:27 +0000 (21:16 +0000)]
Tracing to help investigate issues with SjLj spill code.

llvm-svn: 147682

12 years agoRevert r147664; it's breaking clang regression tests.
Eli Friedman [Fri, 6 Jan 2012 20:42:20 +0000 (20:42 +0000)]
Revert r147664; it's breaking clang regression tests.

llvm-svn: 147681

12 years ago__gxx_personality_v0 update. This is completely untested code. But my brain is...
Howard Hinnant [Fri, 6 Jan 2012 20:39:47 +0000 (20:39 +0000)]
__gxx_personality_v0 update.  This is completely untested code.  But my brain is scrambled and I wanted to get it checked in.  Code review from anyone who knows anything at all about personality functions would be much appreciated.  What is in here is borrowed heavily from llvm/examples/ExceptionDemo/ExceptionDemo.cpp.  There are some calls to abort() that should probably be changed to return an error code instead.  There may be encodings under readEncodedPointer that need to be implemented.  And my handling of type_info is almost a complete guess.

llvm-svn: 147680

12 years agoInitializing to false makes better sense. Thanks, David.
Chad Rosier [Fri, 6 Jan 2012 20:11:59 +0000 (20:11 +0000)]
Initializing to false makes better sense.  Thanks, David.

llvm-svn: 147679

12 years agoFix uninitialized variable warning.
Chad Rosier [Fri, 6 Jan 2012 20:05:14 +0000 (20:05 +0000)]
Fix uninitialized variable warning.

llvm-svn: 147678

12 years agoRevert r147655; it's breaking the compiler_rt build on OSX.
Eli Friedman [Fri, 6 Jan 2012 20:03:09 +0000 (20:03 +0000)]
Revert r147655; it's breaking the compiler_rt build on OSX.

llvm-svn: 147677

12 years agoFix uninitialized variable warning.
Chad Rosier [Fri, 6 Jan 2012 20:02:49 +0000 (20:02 +0000)]
Fix uninitialized variable warning.

llvm-svn: 147676

12 years agoFix uninitialized variable warning.
Chad Rosier [Fri, 6 Jan 2012 19:59:58 +0000 (19:59 +0000)]
Fix uninitialized variable warning.

llvm-svn: 147675

12 years ago[asan] do not use new/delete for the internal thread structure
Kostya Serebryany [Fri, 6 Jan 2012 19:44:11 +0000 (19:44 +0000)]
[asan] do not use new/delete for the internal thread structure

llvm-svn: 147674

12 years agoFix a leak I noticed while reviewing the accelerator table changes. Passes
Eric Christopher [Fri, 6 Jan 2012 19:35:04 +0000 (19:35 +0000)]
Fix a leak I noticed while reviewing the accelerator table changes. Passes
lldb testsuite.

rdar://10652330

llvm-svn: 147673

12 years agoThread safety analysis: added support for trylock attribute.
DeLesley Hutchins [Fri, 6 Jan 2012 19:16:50 +0000 (19:16 +0000)]
Thread safety analysis: added support for trylock attribute.

llvm-svn: 147672

12 years ago[asan] move more code into OS-specific files
Kostya Serebryany [Fri, 6 Jan 2012 19:11:09 +0000 (19:11 +0000)]
[asan] move more code into OS-specific files

llvm-svn: 147671

12 years agoAdded LocalVariableMap
DeLesley Hutchins [Fri, 6 Jan 2012 18:36:09 +0000 (18:36 +0000)]
Added LocalVariableMap

llvm-svn: 147670

12 years ago__FLT_EVAL_METHOD__ should be 1 on NetBSD/i386, since it defaults to
Joerg Sonnenberger [Fri, 6 Jan 2012 18:32:26 +0000 (18:32 +0000)]
__FLT_EVAL_METHOD__ should be 1 on NetBSD/i386, since it defaults to
"double" rounding.

llvm-svn: 147669

12 years agoFixed a bug where the DWARF location expression
Sean Callanan [Fri, 6 Jan 2012 18:24:47 +0000 (18:24 +0000)]
Fixed a bug where the DWARF location expression
parser was creating malformed resuls.  When the
location of a variable is computed by reading a
register and adding an offset, we shouldn't say
that the variable's value is located in that
register.  This was confusing the expression
parser when trying to read a variable captured
by a block.

llvm-svn: 147668

12 years ago[asan] cleanup: remove the SIGILL-related code (compiler part)
Kostya Serebryany [Fri, 6 Jan 2012 18:09:21 +0000 (18:09 +0000)]
[asan] cleanup: remove the SIGILL-related code (compiler part)

llvm-svn: 147667

12 years agoobjc++: some declarations related to atomic
Fariborz Jahanian [Fri, 6 Jan 2012 18:07:23 +0000 (18:07 +0000)]
objc++: some declarations related to atomic
properties of c++ object types with non-trivial
assignment copy. Not used yet. // rdar://6137845

llvm-svn: 147666

12 years ago[asan] cleanup: remove the SIGILL-related code (rt part)
Kostya Serebryany [Fri, 6 Jan 2012 18:02:04 +0000 (18:02 +0000)]
[asan] cleanup: remove the SIGILL-related code (rt part)

llvm-svn: 147665

12 years agoSilence GCC warnings.
Jakub Staszak [Fri, 6 Jan 2012 17:44:30 +0000 (17:44 +0000)]
Silence GCC warnings.

llvm-svn: 147664

12 years ago[asan] mention -fno-omit-frame-pointer in the docs
Kostya Serebryany [Fri, 6 Jan 2012 17:35:27 +0000 (17:35 +0000)]
[asan] mention -fno-omit-frame-pointer in the docs

llvm-svn: 147663

12 years agoWhen inferring a submodule ID during module creation, look up the
Douglas Gregor [Fri, 6 Jan 2012 17:19:32 +0000 (17:19 +0000)]
When inferring a submodule ID during module creation, look up the
include stack to find the first file that is known to be part of the
module. This copes with situations where the module map doesn't
completely specify all of the headers that are involved in the module,
which can come up when there are very strange #include_next chains
(e.g., with weird compiler/stdlib headers like stdarg.h or float.h).

llvm-svn: 147662

12 years ago[analyzer] Correct the c++-analyzer symlink
Anna Zaks [Fri, 6 Jan 2012 17:11:23 +0000 (17:11 +0000)]
[analyzer] Correct the c++-analyzer symlink

As Matt pointed out, this should be just a link to 'ccc-analyzer'.

llvm-svn: 147661

12 years agoStash Decl's TopLevelDeclInObjCContainer and ModulePrivate bits
Douglas Gregor [Fri, 6 Jan 2012 16:59:53 +0000 (16:59 +0000)]
Stash Decl's TopLevelDeclInObjCContainer and ModulePrivate bits
into the two unused lower bits of the NextDeclInContext link, dropping
the number of bits in Decl down to 32, and saving 8 bytes per
declaration on x86-64.

llvm-svn: 147660

12 years agoC++11 generalized constant expressions: implement checking and diagnostics for
Richard Smith [Fri, 6 Jan 2012 16:39:00 +0000 (16:39 +0000)]
C++11 generalized constant expressions: implement checking and diagnostics for
pointer-arithmetic-related undefined behavior and unspecified results. We
continue to fold such values, but now notice they aren't constant expressions.

llvm-svn: 147659

12 years agoIntroduce a "Hidden" bit into Decl, to track whether that declaration
Douglas Gregor [Fri, 6 Jan 2012 16:22:39 +0000 (16:22 +0000)]
Introduce a "Hidden" bit into Decl, to track whether that declaration
is hidden from name lookup. The previous hack of tweaking the
ModulePrivate bit when loading a declaration from a hidden submodule
was brittle.

Note that we now have 34 bits in Decl. I'll fix that next.

llvm-svn: 147658

12 years agoTeach DeclContext deserialization to pay attention to endianness, from
Douglas Gregor [Fri, 6 Jan 2012 16:09:53 +0000 (16:09 +0000)]
Teach DeclContext deserialization to pay attention to endianness, from
David Fang and Takumi Nakamura. Fixes many PCH failures on PowerPC.

llvm-svn: 147657

12 years agoIf we are compiling with -fno-builtin then don't do constant folding of
David Chisnall [Fri, 6 Jan 2012 12:20:19 +0000 (12:20 +0000)]
If we are compiling with -fno-builtin then don't do constant folding of
builtins.

This fixes PR11711.

llvm-svn: 147655

12 years agoFix typo in string
Eli Bendersky [Fri, 6 Jan 2012 07:49:17 +0000 (07:49 +0000)]
Fix typo in string

llvm-svn: 147654

12 years agoFix the mangling of class template arguments in a particular
John McCall [Fri, 6 Jan 2012 05:06:35 +0000 (05:06 +0000)]
Fix the mangling of class template arguments in a particular
dependent case.  Thanks to Jason Merrill for pointing this out.

llvm-svn: 147653

12 years agoImprovements to the uninitialized variable warning: Check if the constructor
Rafael Espindola [Fri, 6 Jan 2012 04:54:01 +0000 (04:54 +0000)]
Improvements to the uninitialized variable warning: Check if the constructor
call is elidable or if the constructor is trivial instead of checking if it
is user declared.

llvm-svn: 147652

12 years agoAs part of the ongoing work in finalizing the accelerator tables, extend
Eric Christopher [Fri, 6 Jan 2012 04:35:23 +0000 (04:35 +0000)]
As part of the ongoing work in finalizing the accelerator tables, extend
the debug type accelerator tables to contain the tag and a flag
stating whether or not a compound type is a complete type.

rdar://10652330

llvm-svn: 147651

12 years agoMore lambda work. Fixes a minor bug Richard pointed out, makes lookup for lambda...
Eli Friedman [Fri, 6 Jan 2012 03:05:34 +0000 (03:05 +0000)]
More lambda work.  Fixes a minor bug Richard pointed out, makes lookup for lambda parameters work correctly, recording more information into the AST.

llvm-svn: 147650

12 years agoDavid Blaikie and Chandler would like us to diagnose
Richard Smith [Fri, 6 Jan 2012 02:30:50 +0000 (02:30 +0000)]
David Blaikie and Chandler would like us to diagnose

  int f();

in function scopes under -Wvexing-parse, so now we do.

llvm-svn: 147649