Rafael Espindola [Wed, 28 May 2014 18:15:43 +0000 (18:15 +0000)]
[pr19844] Add thread local mode to aliases.
This matches gcc's behavior. It also seems natural given that aliases
contain other properties that govern how it is accessed (linkage,
visibility, dll storage).
Clang still has to be updated to expose this feature to C.
llvm-svn: 209759
Fariborz Jahanian [Wed, 28 May 2014 18:12:10 +0000 (18:12 +0000)]
Objective-C. Diagnose use of properties in functions nested in,
now deprecated, ObjC containers instead of crashing.
// rdar://
16859666
llvm-svn: 209758
Dmitry Vyukov [Wed, 28 May 2014 18:03:32 +0000 (18:03 +0000)]
tsan: refactor suppressions machinery
The refactoring makes suppressions more flexible
and allow to suppress based on arbitrary number of stacks.
In particular it fixes:
https://code.google.com/p/thread-sanitizer/issues/detail?id=64
"Make it possible to suppress deadlock reports by any stack (not just first)"
llvm-svn: 209757
Todd Fiala [Wed, 28 May 2014 17:58:29 +0000 (17:58 +0000)]
Disabled warning C4251 on MSVC builds.
See http://reviews.llvm.org/D3934 for more details.
This is only the CMakeLists.txt portion of that change.
Change by Zachary Turner.
llvm-svn: 209756
Louis Gerbarg [Wed, 28 May 2014 17:38:31 +0000 (17:38 +0000)]
Add support for combining GEPs across PHI nodes
Currently LLVM will generally merge GEPs. This allows backends to use more
complex addressing modes. In some cases this is not happening because there
is PHI inbetween the two GEPs:
GEP1--\
|-->PHI1-->GEP3
GEP2--/
This patch checks to see if GEP1 and GEP2 are similiar enough that they can be
cloned (GEP12) in GEP3's BB, allowing GEP->GEP merging (GEP123):
GEP1--\ --\ --\
|-->PHI1-->GEP3 ==> |-->PHI2->GEP12->GEP3 == > |-->PHI2->GEP123
GEP2--/ --/ --/
This also breaks certain use chains that are preventing GEP->GEP merges that the
the existing instcombine would merge otherwise.
Tests included.
llvm-svn: 209755
Eli Bendersky [Wed, 28 May 2014 17:35:14 +0000 (17:35 +0000)]
Trim trailing whitespace in cindex.py
llvm-svn: 209754
Johannes Doerfert [Wed, 28 May 2014 17:21:02 +0000 (17:21 +0000)]
Added option for n-dimensional rectangular tiling
+ CL-option --polly-tile-sizes=<int,...,int>
The i'th value is used as a tile size for dimension i, if
there is no i'th value, the value of --polly-default-tile-size is
used
+ CL-option --polly-default-tile-size=int
Used if no tile size is given for a dimension i
+ 3 Simple testcases
llvm-svn: 209753
Todd Fiala [Wed, 28 May 2014 17:06:04 +0000 (17:06 +0000)]
Cmake build changes.
Disables exception handling in LLDB, using appropriate compiler
flags depending on the platform. This is consistent with the build
of LLVM, should improve performance, and also removes a substantial number
of warnings from the Windows build.
See http://reviews.llvm.org/D3929 for more details.
Change by Zachary Turner
llvm-svn: 209752
Fariborz Jahanian [Wed, 28 May 2014 17:02:35 +0000 (17:02 +0000)]
Objective-C. Deprecate use of function definitions
in Objective-C container declarations (but not
in their definitions. // rdar://
10414277
llvm-svn: 209751
Johannes Doerfert [Wed, 28 May 2014 16:54:42 +0000 (16:54 +0000)]
Reorder cmake include folders (polly source first)
llvm-svn: 209750
Todd Fiala [Wed, 28 May 2014 16:43:26 +0000 (16:43 +0000)]
Fix Windows warnings.
This fixes a number of trivial warnings in the Windows build. This is part of a larger effort to make the Windows build warning-free.
See http://reviews.llvm.org/D3914 for more details.
Change by Zachary Turner
llvm-svn: 209749
Hal Finkel [Wed, 28 May 2014 15:33:19 +0000 (15:33 +0000)]
Revert "[DAGCombiner] Split up an indexed load if only the base pointer value is live"
This reverts r208640 (I've just XFAILed the test) because it broke ppc64/Linux
self-hosting. Because nearly every regression test triggers a segfault, I hope
this will be easy to fix.
llvm-svn: 209747
Rafael Espindola [Wed, 28 May 2014 15:30:40 +0000 (15:30 +0000)]
InstCombine: Improvement to check if signed addition overflows.
This patch implements two things:
1. If we know one number is positive and another is negative, we return true as
signed addition of two opposite signed numbers will never overflow.
2. Implemented TODO : If one of the operands only has one non-zero bit, and if
the other operand has a known-zero bit in a more significant place than it
(not including the sign bit) the ripple may go up to and fill the zero, but
won't change the sign. e.x - (x & ~4) + 1
We make sure that we are ignoring 0 at MSB.
Patch by Suyog Sarda.
llvm-svn: 209746
Hal Finkel [Wed, 28 May 2014 15:25:06 +0000 (15:25 +0000)]
Revert "[PPC] Use alias symbols in address computation."
This reverts commit r209638 because it broke self-hosting on ppc64/Linux. (the
Clang-compiled TableGen would segfault because it jumped to an invalid address
from within _ZNK4llvm17ManagedStaticBase21RegisterManagedStaticEPFPvvEPFvS1_E
(which is within the command-line parameter registration process)).
llvm-svn: 209745
Dmitry Vyukov [Wed, 28 May 2014 15:22:12 +0000 (15:22 +0000)]
tsan: do not use 64-bit atomics in allocator code
64-bit atomics make porting of asan to 32-bits platforms problematic.
llvm-svn: 209744
Alexander Kornienko [Wed, 28 May 2014 15:21:14 +0000 (15:21 +0000)]
Renamed runClangTidy argument Ranges to InputFiles, removed a TODO comment.
No functional changes.
llvm-svn: 209743
Joerg Sonnenberger [Wed, 28 May 2014 15:12:55 +0000 (15:12 +0000)]
Don't hard-code ld when extracting host linker version, use ${LD} if
it is set.
llvm-svn: 209742
Joerg Sonnenberger [Wed, 28 May 2014 15:08:05 +0000 (15:08 +0000)]
Refactor muldf3 and mulsf3.
Patch from: GuanHong Liu
Differential Revision: http://reviews.llvm.org/D3886
llvm-svn: 209741
Joerg Sonnenberger [Wed, 28 May 2014 15:06:25 +0000 (15:06 +0000)]
Refactor addsf3.c and adddf3.c.
Differential Revision: http://reviews.llvm.org/D3885
llvm-svn: 209740
Ed Maste [Wed, 28 May 2014 14:11:20 +0000 (14:11 +0000)]
Avoid passing null signal name for Log %s argument
llvm-svn: 209739
Dario Domizioli [Wed, 28 May 2014 14:06:38 +0000 (14:06 +0000)]
[DOC] Documentation for #pragma clang optimize on/off
llvm-svn: 209738
Ed Maste [Wed, 28 May 2014 14:05:43 +0000 (14:05 +0000)]
Fix whitespace / formatting
llvm-svn: 209737
Evgeniy Stepanov [Wed, 28 May 2014 13:45:18 +0000 (13:45 +0000)]
[asan] Mass-xfail 17 tests on Android.
This makes check-asan pass on Android, which should help prevent future
regressions.
https://code.google.com/p/address-sanitizer/issues/detail?id=316
llvm-svn: 209736
Evgeniy Stepanov [Wed, 28 May 2014 13:13:30 +0000 (13:13 +0000)]
[sanitizer] Replace -lpthread with -pthread in tests.
-lpthread does not work on Android.
llvm-svn: 209735
Evgeniy Stepanov [Wed, 28 May 2014 13:07:39 +0000 (13:07 +0000)]
[asan] XFAIL:android several tests.
llvm-svn: 209734
Timur Iskhodzhanov [Wed, 28 May 2014 13:06:14 +0000 (13:06 +0000)]
[ASan tests] Exclude some tests from Windows runs
Some features are not supported yet and some are not planned to be fixed soon
llvm-svn: 209733
Evgeniy Stepanov [Wed, 28 May 2014 12:53:34 +0000 (12:53 +0000)]
-pthread does not translate to -lpthread on Android.
There is no libpthread.so, and pthread interface is implemented in libc.so.
This mirrors gcc behavior.
llvm-svn: 209731
Alp Toker [Wed, 28 May 2014 12:26:32 +0000 (12:26 +0000)]
Add a PR reference for XFAILed test
llvm-svn: 209730
Alp Toker [Wed, 28 May 2014 12:20:23 +0000 (12:20 +0000)]
Test requires exceptions
It's still XFAIL, but slightly closer to passing.
llvm-svn: 209729
Alp Toker [Wed, 28 May 2014 12:20:14 +0000 (12:20 +0000)]
Consolidate some note diagnostics
These note diags have the same message and can be unified further but for now
let's just bring them together.
Incidental change: Display a source range in the final attr diagnostic.
llvm-svn: 209728
Nikola Smiljanic [Wed, 28 May 2014 11:19:43 +0000 (11:19 +0000)]
Fix line endings.
llvm-svn: 209727
NAKAMURA Takumi [Wed, 28 May 2014 10:53:06 +0000 (10:53 +0000)]
Let clang/test/CodeGen/pr19841.cpp tolerant of MS mangler.
llvm-svn: 209726
Daniel Jasper [Wed, 28 May 2014 10:09:11 +0000 (10:09 +0000)]
clang-format: Don't break before a case's colon.
Before (with just the right line length:
switch (a) {
case some_namespace::some_constant
:
return;
}
After:
switch (a) {
case some_namespace::
some_constant:
return;
}
llvm-svn: 209725
Evgeniy Stepanov [Wed, 28 May 2014 09:58:58 +0000 (09:58 +0000)]
[asan] Fix android test setup.
llvm-svn: 209724
Nick Lewycky [Wed, 28 May 2014 09:56:42 +0000 (09:56 +0000)]
Add 'nonnull' parameter or return attribute when producing an llvm pointer type in a function type where the C++ type is a reference. Update the tests.
llvm-svn: 209723
Timur Iskhodzhanov [Wed, 28 May 2014 09:38:51 +0000 (09:38 +0000)]
[ASan tests] Move one more test to POSIX-only directory
llvm-svn: 209722
Evgeniy Stepanov [Wed, 28 May 2014 09:26:46 +0000 (09:26 +0000)]
[asancov] Don't emit extra runtime calls when compiling without coverage.
llvm-svn: 209721
Daniel Jasper [Wed, 28 May 2014 09:11:53 +0000 (09:11 +0000)]
clang-format: Format array and dict literals similar to blocks.
Especially, reduce the amount of indentation if it doesn't increase
readability.
Before:
NSMutableDictionary* dictionary = [NSMutableDictionary
dictionaryWithDictionary:@{
aaaaaaaaaaaaaaaaaaaaa :
aaaaaaaaaaaaa,
bbbbbbbbbbbbbbbbbb : bbbbb,
cccccccccccccccc :
ccccccccccccccc
}];
After:
NSMutableDictionary* dictionary =
[NSMutableDictionary dictionaryWithDictionary:@{
aaaaaaaaaaaaaaaaaaaaa :
aaaaaaaaaaaaa,
bbbbbbbbbbbbbbbbbb : bbbbb,
cccccccccccccccc :
ccccccccccccccc
}];
llvm-svn: 209720
Timur Iskhodzhanov [Wed, 28 May 2014 08:38:13 +0000 (08:38 +0000)]
[ASan/Win] Use clang rather than clang-cl by default for lit tests. Make Windows-only tests explicitly use clang-cl.
Reviewed at http://reviews.llvm.org/D3893
llvm-svn: 209719
Evgeniy Stepanov [Wed, 28 May 2014 08:34:53 +0000 (08:34 +0000)]
[asan] Add a dll thunk for __sanitizer_cov_module_init.
llvm-svn: 209718
Evgeniy Stepanov [Wed, 28 May 2014 08:26:24 +0000 (08:26 +0000)]
[asan] Add a few stubs to fix windows build.
llvm-svn: 209717
Alexey Bataev [Wed, 28 May 2014 07:40:25 +0000 (07:40 +0000)]
[OPENMP] Additional checking for local vars in initial values for threadprivate vars
llvm-svn: 209716
Tilmann Scheller [Wed, 28 May 2014 06:43:00 +0000 (06:43 +0000)]
[AArch64] Add store post-index update folding regression tests for the load/store optimizer.
Add regression tests for the following transformation:
str X, [x20]
...
add x20, x20, #32
->
str X, [x20], #32
with X being either w0, x0, s0, d0 or q0.
llvm-svn: 209715
Alexey Bataev [Wed, 28 May 2014 06:15:33 +0000 (06:15 +0000)]
[OPENMP] Reformatting of parsing code for OpenMP constructs.
llvm-svn: 209714
Alexey Bataev [Wed, 28 May 2014 05:53:51 +0000 (05:53 +0000)]
[OPENMP][C++11] Formatting and make more C++11 compliant of OpenMP sema.
llvm-svn: 209713
Justin Bogner [Wed, 28 May 2014 05:45:17 +0000 (05:45 +0000)]
utils: Teach lldbDataFormatters to load automatically
Add an __lldb_init_module function so that importing the
lldbDataFormatters script automatically adds the formatters.
llvm-svn: 209712
Tilmann Scheller [Wed, 28 May 2014 05:44:14 +0000 (05:44 +0000)]
[AArch64] Add load post-index update folding regression tests for the load/store optimizer.
Add regression tests for the following transformation:
ldr X, [x20]
...
add x20, x20, #32
->
ldr X, [x20], #32
with X being either w0, x0, s0, d0 or q0.
llvm-svn: 209711
Tobias Grosser [Wed, 28 May 2014 04:34:58 +0000 (04:34 +0000)]
Fix formatting
llvm-svn: 209710
Simon Atanasyan [Wed, 28 May 2014 04:10:01 +0000 (04:10 +0000)]
[Mips] Do not mix _gp and _gp_disp symbols in relocation handling.
No functional changes.
llvm-svn: 209709
Richard Trieu [Wed, 28 May 2014 02:16:01 +0000 (02:16 +0000)]
Move the logic for testing for namespace std into one location. This check can
be performed by using Decl::isInStdNamespace or DeclContext::isStdNamespace
llvm-svn: 209708
Nick Kledzik [Wed, 28 May 2014 02:04:45 +0000 (02:04 +0000)]
[mach-o] Add support for parsing of weak-def symbols
llvm-svn: 209707
Hans Wennborg [Wed, 28 May 2014 01:52:23 +0000 (01:52 +0000)]
Don't dllimport/export destructor variants implemented by thunks.
MSVC doesn't export these functions, so trying to import them doesnt' work.
Also, don't let any dll attributes on the CXXDestructorDecl influence the
thunk's linkage -- they should always be linkonce_odr.
This takes care of the FIXME's for this in Nico's tests.
Differential Revision: http://reviews.llvm.org/D3930
llvm-svn: 209706
Nick Kledzik [Wed, 28 May 2014 01:31:04 +0000 (01:31 +0000)]
[mach-o] rename test case
llvm-svn: 209705
Nick Kledzik [Wed, 28 May 2014 01:16:35 +0000 (01:16 +0000)]
[mach-o] Support parsing of non-lazy-pointer sections
llvm-svn: 209704
Greg Clayton [Wed, 28 May 2014 00:23:43 +0000 (00:23 +0000)]
Remove unused variable.
llvm-svn: 209703
Greg Clayton [Wed, 28 May 2014 00:21:15 +0000 (00:21 +0000)]
Allow classes to be intialized using current lldb::SB objects. This can help to import/export the current process state.
llvm-svn: 209702
Nick Kledzik [Tue, 27 May 2014 23:20:52 +0000 (23:20 +0000)]
[mach-o] Add support for initializers and terminators in object files
llvm-svn: 209700
Greg Fitzgerald [Tue, 27 May 2014 23:13:52 +0000 (23:13 +0000)]
[asan] use relative line numbers in tests
llvm-svn: 209699
Alexey Samsonov [Tue, 27 May 2014 23:09:50 +0000 (23:09 +0000)]
Change representation of instruction ranges where variable is accessible.
Use more straightforward way to represent the set of instruction
ranges where the location of a user variable is defined - vector of pairs
of instructions (defining start/end of each range),
instead of a flattened vector of instructions where some instructions
are supposed to start the range, and the rest are supposed to "clobber" it.
Simplify the code which generates actual .debug_loc entries.
No functionality change.
llvm-svn: 209698
Alexey Samsonov [Tue, 27 May 2014 22:47:41 +0000 (22:47 +0000)]
Factor out looking for prologue end into a function
llvm-svn: 209697
Sebastian Pop [Tue, 27 May 2014 22:42:11 +0000 (22:42 +0000)]
delinearize does not return a remainder
llvm-svn: 209696
Sebastian Pop [Tue, 27 May 2014 22:42:09 +0000 (22:42 +0000)]
pass element size to delinearization
Instead of relying on the delinearization to infer the size of an element,
compute the element size from the base address type. This is a much more precise
way of computing the element size than before, as we would have mixed together
the size of an element with the strides of the innermost dimension.
llvm-svn: 209695
Sebastian Pop [Tue, 27 May 2014 22:42:00 +0000 (22:42 +0000)]
avoid type mismatch when building SCEVs
This is a corner case I have stumbled upon when dealing with ARM64 type
conversions. I was not able to extract a testcase for the community codebase to
fail on. The patch conservatively discards a division that would have ended up
in an ICE due to a type mismatch when building a multiply expression. I have
also added code to a place that builds add expressions and in which we should be
careful not to pass in operands of different types.
llvm-svn: 209694
Sebastian Pop [Tue, 27 May 2014 22:41:56 +0000 (22:41 +0000)]
do not use the GCD to compute the delinearization strides
We do not need to compute the GCD anymore after we removed the constant
coefficients from the terms: the terms are now all parametric expressions and
there is no need to recognize constant terms that divide only a subset of the
terms. We only rely on the size of the terms, i.e., the number of operands in
the multiply expressions, to sort the terms and recognize the parametric
dimensions.
llvm-svn: 209693
Sebastian Pop [Tue, 27 May 2014 22:41:51 +0000 (22:41 +0000)]
remove BasePointer before delinearizing
No functional change is intended: instead of relying on the delinearization to
come up with the base pointer as a remainder of the divisions in the
delinearization, we just compute it from the array access and use that value.
We substract the base pointer from the SCEV to be delinearized and that
simplifies the work of the delinearizer.
llvm-svn: 209692
Sebastian Pop [Tue, 27 May 2014 22:41:45 +0000 (22:41 +0000)]
remove constant terms
The delinearization is needed only to remove the non linearity induced by
expressions involving multiplications of parameters and induction variables.
There is no problem in dealing with constant times parameters, or constant times
an induction variable.
For this reason, the current patch discards all constant terms and multipliers
before running the delinearization algorithm on the terms. The only thing
remaining in the term expressions are parameters and multiply expressions of
parameters: these simplified term expressions are passed to the array shape
recognizer that will not recognize constant dimensions anymore: these will be
recognized as different strides in parametric subscripts.
The only important special case of a constant dimension is the size of elements.
Instead of relying on the delinearization to infer the size of an element,
compute the element size from the base address type. This is a much more precise
way of computing the element size than before, as we would have mixed together
the size of an element with the strides of the innermost dimension.
llvm-svn: 209691
Alexey Samsonov [Tue, 27 May 2014 22:35:00 +0000 (22:35 +0000)]
Don't pre-populate the set of keys in the map with variable locations history.
Current implementation of calculateDbgValueHistory already creates the
keys in the expected order (user variables are listed in order of appearance),
and should do so later by contract.
No functionality change.
llvm-svn: 209690
Arnaud A. de Grandmaison [Tue, 27 May 2014 22:03:28 +0000 (22:03 +0000)]
No need for those tests to go thru llvm-as and/or llvm-dis.
opt can handle them by itself.
llvm-svn: 209689
Arnaud A. de Grandmaison [Tue, 27 May 2014 21:35:46 +0000 (21:35 +0000)]
Factor out comparison of Instruction "special" states.
No functional change.
llvm-svn: 209688
Reid Kleckner [Tue, 27 May 2014 21:35:17 +0000 (21:35 +0000)]
Wording fix for llvm.global_dtors docs.
llvm-svn: 209687
Alexander Kornienko [Tue, 27 May 2014 21:29:22 +0000 (21:29 +0000)]
Retain isImplicit flag for local variable declarations when instantiating
templates.
Reviewers: rsmith
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3924
llvm-svn: 209686
Nick Kledzik [Tue, 27 May 2014 20:37:08 +0000 (20:37 +0000)]
[mach-o] Add checks that string literals in object files are zero terminated
llvm-svn: 209685
Nick Kledzik [Tue, 27 May 2014 20:25:06 +0000 (20:25 +0000)]
[mach-o] Add support for reading utf16 string literal sections
llvm-svn: 209684
David Blaikie [Tue, 27 May 2014 20:20:43 +0000 (20:20 +0000)]
DebugInfo: partially revert cleanup committed in r209680
I'm not sure exactly where/how we end up with an abstract DbgVariable
with a null DIE, but we do... looking into it & will add a test and/or
fix when I figure it out.
Currently shows up in selfhost or compiler-rt builds.
llvm-svn: 209683
Ben Langmuir [Tue, 27 May 2014 19:57:48 +0000 (19:57 +0000)]
Add a test that we don't store stale modtime in modules
The change from r209195 turned out to be important to avoid saving stale
modification time/expected size information in a module file when there
are 3 or more modules in a dependency chain and the bottom one is
rebuilt. So add a test for that.
rdar://problem/
17038180
llvm-svn: 209682
Nick Kledzik [Tue, 27 May 2014 19:35:41 +0000 (19:35 +0000)]
Add make_dynamic_error_code().
This is a short-term fix to allow lld Readers to return error messages
with dynamic content.
The long term fix will be to enhance ErrorOr<> to work with errors other
than error_code. Or to change the interface to Readers to pass down a
diagnostics object through which all error messages are written.
llvm-svn: 209681
David Blaikie [Tue, 27 May 2014 19:34:32 +0000 (19:34 +0000)]
DebugInfo: Simplify solution to avoid DW_AT_artificial on inlined parameters.
Originally committed in r207717, I clearly didn't look very closely at
the code to understand how existing things were working...
llvm-svn: 209680
James Dennett [Tue, 27 May 2014 19:13:04 +0000 (19:13 +0000)]
Add range accessors for captures of a LambdaExpr.
Summary:
This adds LambdaExpr::captures(), LambdaExpr::explicit_captures()
and LambdaExpr::implicit_captures() as simple wrappers over the underlying
*_begin()/*_end() functions.
Reviewers: aaron.ballman
Differential Revision: http://reviews.llvm.org/D3926
llvm-svn: 209679
Sasa Stankovic [Tue, 27 May 2014 18:53:06 +0000 (18:53 +0000)]
[mips] Optimize long branch for MIPS64 by removing %higher and %highest.
%higher and %highest can have non-zero values only for offsets greater
than 2GB, which is highly unlikely, if not impossible when compiling a
single function. This makes long branch for MIPS64 3 instructions smaller.
Differential Revision: http://llvm-reviews.chandlerc.com/D3281.diff
llvm-svn: 209678
David Blaikie [Tue, 27 May 2014 18:37:55 +0000 (18:37 +0000)]
DebugInfo: Create abstract function definitions even when concrete definitions preceed inline definitions.
After much puppetry, here's the major piece of the work to ensure that
even when a concrete definition preceeds all inline definitions, an
abstract definition is still created and referenced from both concrete
and inline definitions.
Variables are still broken in this case (see comment in
dbg-value-inlined-parameter.ll test case) and will be addressed in
follow up work.
llvm-svn: 209677
David Blaikie [Tue, 27 May 2014 18:37:51 +0000 (18:37 +0000)]
DebugInfo: Avoid an extra map lookup when finding abstract subprogram DIEs.
llvm-svn: 209676
David Blaikie [Tue, 27 May 2014 18:37:48 +0000 (18:37 +0000)]
DebugInfo: Lazily construct subprogram definition DIEs.
A further step to correctly emitting concrete out of line definitions
preceeding inlined instances of the same program.
To do this, emission of subprograms must be delayed until required since
we don't know which (abstract only (if there's no out of line
definition), concrete only (if there are no inlined instances), or both)
DIEs are required at the start of the module.
To reduce the test churn in the following commit that actually fixes the
bug, this commit introduces the lazy DIE construction and cleans up test
cases that are impacted by the changes in the resulting DIE ordering.
llvm-svn: 209675
David Blaikie [Tue, 27 May 2014 18:37:43 +0000 (18:37 +0000)]
DebugInfo: Lazily attach definition attributes to definitions.
This is a precursor to fixing inlined debug info where the concrete,
out-of-line definition may preceed any inlined usage. To cope with this,
the attributes that may appear on the concrete definition or the
abstract definition are delayed until the end of the module. Then, if an
abstract definition was created, it is referenced (and no other
attributes are added to the out-of-line definition), otherwise the
attributes are added directly to the out-of-line definition.
In a couple of cases this causes not just reordering of attributes, but
reordering of types. When the creation of the attribute is delayed, if
that creation would create a type (such as for a DW_AT_type attribute)
then other top level DIEs may've been constructed during the delay,
causing the referenced type to be created and added after those
intervening DIEs. In the extreme case, in cross-cu-inlining.ll, this
actually causes the DW_TAG_basic_type for "int" to move from one CU to
another.
llvm-svn: 209674
David Blaikie [Tue, 27 May 2014 18:37:38 +0000 (18:37 +0000)]
DebugInfo: Separate out the addition of subprogram attribute additions so that they can be added later depending on whether or not the function is inlined.
llvm-svn: 209673
Fariborz Jahanian [Tue, 27 May 2014 18:26:09 +0000 (18:26 +0000)]
Objective-C. Fixes an obscuer crash caused by multiple inclusion of
same framework after complaining about duplicate class definition.
// rdar://
17024681
llvm-svn: 209672
Jingyue Wu [Tue, 27 May 2014 18:12:55 +0000 (18:12 +0000)]
Fixed a test in r209670
The test was outdated with r209537.
llvm-svn: 209671
Jingyue Wu [Tue, 27 May 2014 18:00:00 +0000 (18:00 +0000)]
Distribute sext/zext to the operands of and/or/xor
This is an enhancement to SeparateConstOffsetFromGEP. With this patch, we can
extract a constant offset from "s/zext and/or/xor A, B".
Added a new test @ext_or to verify this enhancement.
Refactoring the code, I also extracted some common logic to function
Distributable.
llvm-svn: 209670
David Blaikie [Tue, 27 May 2014 17:57:14 +0000 (17:57 +0000)]
DebugInfo: Fix argument ordering in test by adding argument numbering.
This old test didn't have the argument numbering that's now squirelled
away in the high bits of the line number in the DW_TAG_arg_variable
metadata.
Add the numbering and update the test to ensure arguments are in-order.
llvm-svn: 209669
Renato Golin [Tue, 27 May 2014 17:01:21 +0000 (17:01 +0000)]
Fix pr19841, bb are also unnamed
llvm-svn: 209668
Filipe Cabecinhas [Tue, 27 May 2014 16:54:33 +0000 (16:54 +0000)]
Post-commit fixes for r209643
Detected by Daniel Jasper, Ilia Filippov, and Andrea Di Biagio
Fixed the argument order to select (the mask semantics to blendv* are the
inverse of select) and fixed the tests
Added parenthesis to the assert condition
Ran clang-format
llvm-svn: 209667
Renato Golin [Tue, 27 May 2014 16:51:36 +0000 (16:51 +0000)]
Fix pr19841.cpp on release mode
llvm-svn: 209666
Tim Northover [Tue, 27 May 2014 16:50:09 +0000 (16:50 +0000)]
AArch64: add test for NZCV cross-copy save.
llvm-svn: 209665
Tim Northover [Tue, 27 May 2014 16:50:03 +0000 (16:50 +0000)]
AArch64: add AArch64-specific test for 'c' and 'n'.
llvm-svn: 209664
Renato Golin [Tue, 27 May 2014 16:46:27 +0000 (16:46 +0000)]
Revert small change to EmitDeclRefLValue
That small change, although it looked harmless, it made emitting the LValue
on the PHI node without the proper cast. Reverting it fixes PR19841.
llvm-svn: 209663
Bill Schmidt [Tue, 27 May 2014 15:57:51 +0000 (15:57 +0000)]
[PATCH] Correct type used for VADD_SPLAT optimization on PowerPC
In PPCISelLowering.cpp: PPCTargetLowering::LowerBUILD_VECTOR(), there
is an optimization for certain patterns to generate one or two vector
splats followed by a vector add or subtract. This operation is
represented by a VADD_SPLAT in the selection DAG. Prior to this
patch, it was possible for the VADD_SPLAT to be assigned the wrong
data type, causing incorrect code generation. This patch corrects the
problem.
Specifically, the code previously assigned the value type of the
BUILD_VECTOR node to the newly generated VADD_SPLAT node. This is
correct much of the time, but not always. The problem is that the
call to isConstantSplat() may return a SplatBitSize that is not the
same as the number of bits in the original element vector type. The
correct type to assign is a vector type with the same element bit size
as SplatBitSize.
The included test case shows an example of this, where the
BUILD_VECTOR node has a type of v16i8. The vector to be built is {0,
16, 0, 16, 0, 16, 0, 16, 0, 16, 0, 16, 0, 16, 0, 16}. isConstantSplat
detects that we can generate a splat of 16 for type v8i16, which is
the type we must assign to the VADD_SPLAT node. If we do not, we
generate a vspltisb of 8 and a vaddubm, which generates the incorrect
result {16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16, 16,
16}. The correct code generation is a vspltish of 8 and a vadduhm.
This patch also corrected code generation for
CodeGen/PowerPC/2008-07-10-SplatMiscompile.ll, which had been marked
as an XFAIL, so we can remove the XFAIL from the test case.
llvm-svn: 209662
NAKAMURA Takumi [Tue, 27 May 2014 15:13:51 +0000 (15:13 +0000)]
ASTMatchers.h: Appease msc17.
llvm-svn: 209661
Alexander Musman [Tue, 27 May 2014 15:12:19 +0000 (15:12 +0000)]
Parsing/Sema for OMPCollapseClause.
Actual usage in Sema for collapsing loops will in some future patch.
llvm-svn: 209660
Zoran Jovanovic [Tue, 27 May 2014 14:58:51 +0000 (14:58 +0000)]
[mips][mips64r6] Add Relocations R_MIPS_PCHI16, R_MIPS_PCLO16
Differential Revision: http://reviews.llvm.org/D3860
llvm-svn: 209659
Evgeniy Stepanov [Tue, 27 May 2014 14:18:43 +0000 (14:18 +0000)]
[tsan] Fix tsango build.
llvm-svn: 209658
Evgeniy Stepanov [Tue, 27 May 2014 14:14:03 +0000 (14:14 +0000)]
[asan] Disable dlopen interceptor on Android.
It can not work, as libdl is always first in the lookup order,
even ahead of LD_PRELOAD-ed libraries.
llvm-svn: 209657