Daniel Jasper [Thu, 14 Feb 2013 09:58:41 +0000 (09:58 +0000)]
Remove the trailing whitespace of formatted lines.
So far, clang-format has always assumed the whitespace belonging to the
subsequent token. This has the negative side-effect that when
clang-format formats a line, it does not remove its trailing whitespace,
as it belongs to the next token.
Thus, this patch fixes most of llvm.org/PR15062.
We are not zapping a file's trailing whitespace so far, as this does not
belong to any token we see during formatting. We need to fix this in a
subsequent patch.
llvm-svn: 175152
Daniel Jasper [Thu, 14 Feb 2013 09:19:04 +0000 (09:19 +0000)]
Get less confused by trailing comma in Google style.
The formatter can now format:
void
aaaaaaaaaaaaaaaaaa(int level,
double *min_x,
double *max_x,
double *min_y,
double *max_y,
double *min_z,
double *max_z, ) {
}
Although this is invalid code, it frequently happens during development and
clang-format should be nicer :-).
llvm-svn: 175151
Elena Demikhovsky [Thu, 14 Feb 2013 09:07:34 +0000 (09:07 +0000)]
Moved line-info.ll to DebugInfo\X86 directory
llvm-svn: 175150
Owen Anderson [Thu, 14 Feb 2013 09:07:33 +0000 (09:07 +0000)]
Add some legality checks for SETCC before introducing it in the DAG combiner post-operand legalization.
llvm-svn: 175149
Daniel Jasper [Thu, 14 Feb 2013 08:45:14 +0000 (08:45 +0000)]
Remove accidentally committed empty test.
llvm-svn: 175148
Daniel Jasper [Thu, 14 Feb 2013 08:42:54 +0000 (08:42 +0000)]
Align superclasses for multiple inheritence.
This fixes llvm.org/PR15179.
Before:
class ColorChooserMac : public content::ColorChooser,
public content::WebContentsObserver {
};
After:
class ColorChooserMac : public content::ColorChooser,
public content::WebContentsObserver {
};
llvm-svn: 175147
Elena Demikhovsky [Thu, 14 Feb 2013 08:23:08 +0000 (08:23 +0000)]
The test failed on Windows. I've changed the platform to run to "x86_64-apple-darwin".
llvm-svn: 175146
Evgeniy Stepanov [Thu, 14 Feb 2013 08:22:06 +0000 (08:22 +0000)]
[sanitizer] Raise argv limit for ReExec.
We are still limited by /proc/*/cmdline size (4Kb max on Linux!).
llvm-svn: 175145
Elena Demikhovsky [Thu, 14 Feb 2013 08:20:26 +0000 (08:20 +0000)]
Fixed a bug in X86TargetLowering::LowerVectorIntExtend() (assertion failure).
Added a test.
llvm-svn: 175144
Bill Wendling [Thu, 14 Feb 2013 08:18:53 +0000 (08:18 +0000)]
Revert accidental commit.
llvm-svn: 175143
Kostya Serebryany [Thu, 14 Feb 2013 08:12:26 +0000 (08:12 +0000)]
[asan] use short path for sanitizer_common/tests/sanitizer_test_utils.h, add -I sanitizer_common/tests to asan/tests/CMakeLists.txt
llvm-svn: 175142
Bill Wendling [Thu, 14 Feb 2013 08:09:20 +0000 (08:09 +0000)]
Pass the target options through to code generation.
The code generation stuff is going to set attributes on the functions it
generates. To do that it needs the target options. Pass them through.
llvm-svn: 175141
Kostya Serebryany [Thu, 14 Feb 2013 08:05:55 +0000 (08:05 +0000)]
[asan] fix x32 build (H.J. Lu)
llvm-svn: 175140
Michel Danzer [Thu, 14 Feb 2013 08:00:33 +0000 (08:00 +0000)]
R600/SI: Check for empty stack in SIAnnotateControlFlow::isTopOfStack
Fixes assertion failure in newly added lit test. Might just be a bandaid that
needs to be revisited.
llvm-svn: 175139
Michel Danzer [Thu, 14 Feb 2013 07:43:51 +0000 (07:43 +0000)]
R600: Add lit tests for texture sampling instruction selection.
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 175138
Kostya Serebryany [Thu, 14 Feb 2013 06:54:51 +0000 (06:54 +0000)]
[tsan] disable a failing test until it gets fixed. fix lint
llvm-svn: 175137
Greg Clayton [Thu, 14 Feb 2013 04:15:23 +0000 (04:15 +0000)]
<rdar://problem/
13207948>
"watchpoint set expression" fails if "--" is present without any option. Made this command match exactly what "expression" does.
llvm-svn: 175136
Greg Clayton [Thu, 14 Feb 2013 03:54:39 +0000 (03:54 +0000)]
<rdar://problem/
13200878>
When launching in the shell, make sure if you specify a relative path, and if the current working directory has a space in it, that we don't hose the shell invocation.
Currently if we launch with a relative path, we prepend the current working directory to the PATH using:
PATH=`cwd`:$PATH a.out ...
We needed to add quotes around the value for PATH to make sure if any paths in PATH contained spaces, that we don't hose the shell command. Now we do a:
PATH="`cwd`:$PATH" a.out ...
llvm-svn: 175135
Andrew Trick [Thu, 14 Feb 2013 03:45:08 +0000 (03:45 +0000)]
Reapply "s/grep/FileCheck/ in some tests"
This reverts commit
fd1335e982bbf93c5f450ed4fd29f9f787435c85.
Use a triple this time.
llvm-svn: 175134
Rafael Espindola [Thu, 14 Feb 2013 03:33:34 +0000 (03:33 +0000)]
Revert r175120 and r175121. Clang is producing the expected asm names again.
llvm-svn: 175133
Rafael Espindola [Thu, 14 Feb 2013 03:31:26 +0000 (03:31 +0000)]
Partially revert r175117 so that we don't break assumptions about how
static functions in extern "C" contexts are mangled. Should fix the
bootstrap.
llvm-svn: 175132
Greg Clayton [Thu, 14 Feb 2013 03:26:35 +0000 (03:26 +0000)]
<rdar://problem/
13198767>
When dumping instructions, resolve the address specified as a file address if the target doesn't have anything loaded.
llvm-svn: 175131
Cameron Zwarich [Thu, 14 Feb 2013 03:25:24 +0000 (03:25 +0000)]
RegisterCoalescer::reMaterializeTrivialDef() can constrain the destination
register class to match the defining instruction.
llvm-svn: 175130
Francois Pichet [Thu, 14 Feb 2013 03:24:14 +0000 (03:24 +0000)]
Add missing typename to unbreak the MSVC 2012 build.
llvm-svn: 175129
Nick Lewycky [Thu, 14 Feb 2013 03:23:37 +0000 (03:23 +0000)]
Teach the DataLayout aware constant folder to be much more aggressive towards
'and' instructions. This is a pattern that shows up a lot in ubsan binaries.
llvm-svn: 175128
Jim Ingham [Thu, 14 Feb 2013 03:05:42 +0000 (03:05 +0000)]
Add a test for handling a function call that throws an exception, and make it work.
<rdar://problem/
13183944>
llvm-svn: 175127
Reed Kotler [Thu, 14 Feb 2013 03:05:25 +0000 (03:05 +0000)]
Remove the form field from Mips16 instruction formats and set things
up so that we can apply the direct object emitter patch. This patch
should be a nop right now and it's test is to not break what is already
there.
llvm-svn: 175126
Jim Ingham [Thu, 14 Feb 2013 03:04:50 +0000 (03:04 +0000)]
Probably should return that value we took the trouble to compute.
llvm-svn: 175125
Andrew Trick [Thu, 14 Feb 2013 03:00:57 +0000 (03:00 +0000)]
Revert "s/grep/FileCheck/ in some tests"
This reverts commit
8b75e6bc35fb3f9c1e788dbd05084c0f4a60a0f3.
The FileCheck tests are not equivalent:
test/CodeGen/X86/tailcall-structret.ll:6:10: error: expected string not found in input
; CHECK: jmp init
^
<stdin>:1:2: note: scanning from here
.section __TEXT,__text,regular,pure_instructions
^
<stdin>:13:2: note: possible intended match here
jmp _init ## TAILCALL
^
llvm-svn: 175124
Cameron Zwarich [Thu, 14 Feb 2013 02:51:05 +0000 (02:51 +0000)]
Fix RegisterCoalescer::rematerializeTrivialDef() so that it works on flipped
CoalescerPairs. Also, make it take a CoalescerPair directly like other methods
of RegisterCoalescer.
llvm-svn: 175123
Cameron Zwarich [Thu, 14 Feb 2013 02:51:03 +0000 (02:51 +0000)]
Fix some issues with rematerialization in RegisterCoalescer when the destination
of the copy is a subregister def. The current code assumes that it can do a full
def of the destination register, but it is not checking that the def operand is
read-undef. It also doesn't clear the subregister index of the destination in
the new instruction to reflect the full subregister def.
These issues were found running 'make check' with my next commit that enables
rematerialization in more cases.
llvm-svn: 175122
Rafael Espindola [Thu, 14 Feb 2013 02:49:18 +0000 (02:49 +0000)]
Don't assume the mangling of static functions.
llvm-svn: 175121
Rafael Espindola [Thu, 14 Feb 2013 01:58:08 +0000 (01:58 +0000)]
Don't asume that a static function in an extern "C" block will not be mangled.
Since functions with internal linkage don't have language linkage, it is valid
to overload them:
extern "C" {
static int foo();
static int foo(int);
}
So we mangle them.
llvm-svn: 175120
Rafael Espindola [Thu, 14 Feb 2013 01:47:04 +0000 (01:47 +0000)]
merge hasCLanguageLinkage and isExternC. Keep the shorter name.
I added hasCLanguageLinkage while fixing some language linkage bugs some
time ago so that I wouldn't have to check all users of isExternC. It turned
out to be a much longer detour than expected, but this patch finally
merges the two again. The isExternC function now implements just the
standard notion of having C language linkage.
llvm-svn: 175119
Edwin Vane [Thu, 14 Feb 2013 01:34:59 +0000 (01:34 +0000)]
Initial docs directory
Adding a new docs directory, with files auto-generated by sphinx-quickstart,
for user documentation for the various tools in the clang-tools-extra
repository.
Author: Jack Yang <jack.yang@intel.com>
Reviewers: gribozavr, silvas
llvm-svn: 175118
Rafael Espindola [Thu, 14 Feb 2013 01:18:37 +0000 (01:18 +0000)]
Add a getLanguageLinkage method to VarDecls and FunctionDecls. Use it to fix
some cases where functions with no language linkage were being treated as having
C language linkage. In particular, don't warn in
extern "C" {
static NonPod foo();
}
Since getLanguageLinkage checks the language linkage, the linkage computation
cannot use the language linkage. Break the loop by checking just the context
in the linkage computation.
llvm-svn: 175117
Nick Lewycky [Thu, 14 Feb 2013 00:55:17 +0000 (00:55 +0000)]
When marking derived classes' virtual methods ODR-used in order to trigger
instantiation in order to permit devirtualization later in codegen, skip over
pure functions since those can't be devirtualization targets.
llvm-svn: 175116
Ted Kremenek [Thu, 14 Feb 2013 00:32:25 +0000 (00:32 +0000)]
Teach ccc-analyze to pass on -iquote with no spaces between it an the argument.
llvm-svn: 175115
Daniel Malea [Thu, 14 Feb 2013 00:30:40 +0000 (00:30 +0000)]
Fix broken testcase: ignore LLDB output that warns about more elements than can be displayed.
- test could be re-written at some point, but right now just trying to clean up known failures for the linux buildbot.
llvm-svn: 175114
Richard Smith [Thu, 14 Feb 2013 00:23:04 +0000 (00:23 +0000)]
Actually fix rendering of the example code block!
llvm-svn: 175113
Richard Smith [Thu, 14 Feb 2013 00:22:00 +0000 (00:22 +0000)]
Some grammar fixes to 'Format String Checking', and reorder the text slightly to try to make the final code block actually get rendered.
llvm-svn: 175112
Daniel Malea [Thu, 14 Feb 2013 00:20:44 +0000 (00:20 +0000)]
Marking two tests expected-to-fail on Linux
- PR-15260: lldb does not display correct value of 1-bit field
- PR-16261: lldb does not display size of (file/class) static array
llvm-svn: 175111
Richard Smith [Thu, 14 Feb 2013 00:13:34 +0000 (00:13 +0000)]
Document Clang's support for [[gnu::...]] attributes.
llvm-svn: 175110
Argyrios Kyrtzidis [Thu, 14 Feb 2013 00:12:44 +0000 (00:12 +0000)]
[docs] The stat cache has been removed from the PCH, update the docs to remove mentions of it.
llvm-svn: 175109
Greg Clayton [Wed, 13 Feb 2013 23:57:48 +0000 (23:57 +0000)]
Centralized the expression prefixes that are used for both expressions and utility functions.
llvm-svn: 175108
Weiming Zhao [Wed, 13 Feb 2013 23:24:40 +0000 (23:24 +0000)]
temporarily revert the patch due to some conflicts
llvm-svn: 175107
Greg Clayton [Wed, 13 Feb 2013 23:22:26 +0000 (23:22 +0000)]
A lot more cleanup on the AppleObjCRuntimeV2 class.
llvm-svn: 175106
Greg Clayton [Wed, 13 Feb 2013 23:09:45 +0000 (23:09 +0000)]
Cleaned up and removed unused code.
llvm-svn: 175105
Bill Wendling [Wed, 13 Feb 2013 23:00:51 +0000 (23:00 +0000)]
Retain the name of the new internal global that's been shrunk.
It's possible (e.g. after an LTO build) that an internal global may be used for
debugging purposes. If that's the case appending a '.b' to it makes it hard to
find that variable. Steal the name from the old GV before deleting it so that
they can find that variable again.
llvm-svn: 175104
Matt Kopec [Wed, 13 Feb 2013 22:59:58 +0000 (22:59 +0000)]
Skip another two other tests asserting on Linux in RecordLayoutBuilder::updateExternalFieldOffset().
llvm-svn: 175103
Anshuman Dasgupta [Wed, 13 Feb 2013 22:56:34 +0000 (22:56 +0000)]
Hexagon: add support for predicate-GPR copies.
llvm-svn: 175102
Greg Clayton [Wed, 13 Feb 2013 22:56:14 +0000 (22:56 +0000)]
<rdar://problem/
13210494>
Parse objective C information as efficiently as possible and without taking dangerous runtime locks.
Reworked the way objective C information is parsed by:
1 - don't read all class names up front, this is about 500K of data with names
2 - add a 32 bit hash map that maps a hash of a name to the Class pointer (isa)
3 - Improved name lookups by using the new hash map
4 - split up reading the objc runtime info into dynamic and shared cache since the shared cache only needs to be read once.
5 - When reading all isa values, also get the 32 bit hash instead of the name
6 - Read names lazily now that we don't need all names up front
7 - Allow the hash maps to not be there and still have this function correctly
There is dead code in here with all of the various methods I tried. I want to check this in first to not lose any of it in case we need to revert to any of the extra code. I will promptly cleanup and commit again.
llvm-svn: 175101
Fariborz Jahanian [Wed, 13 Feb 2013 22:50:36 +0000 (22:50 +0000)]
objective-C: Make order of ivars which are synthesized
in the course of property synthesis deterministic (ordered
by their type size), instead of having hashtable order
(as it is currently). // rdar://
13192366
llvm-svn: 175100
Alex Rosenberg [Wed, 13 Feb 2013 22:45:52 +0000 (22:45 +0000)]
llvm-gcc -pedantic warns about C++ comments in C90 mode even if they're
inside an #if 0 block.
llvm-svn: 175098
Ryan Govostes [Wed, 13 Feb 2013 22:37:27 +0000 (22:37 +0000)]
Fix iterator definitions for ImmutableSet and ImmutableMap.
llvm-svn: 175097
Tom Stellard [Wed, 13 Feb 2013 22:05:20 +0000 (22:05 +0000)]
R600: Add support for 128-bit parameters
NOTE: This is a candidate for the Mesa stable branch.
llvm-svn: 175096
Daniel Malea [Wed, 13 Feb 2013 22:05:18 +0000 (22:05 +0000)]
Skip test that causes lldb to assert due to inferior getting a SIGSEGV with si_code == SI_KERNEL (PR-15258)
- crash does not always reproduce
llvm-svn: 175095
Daniel Malea [Wed, 13 Feb 2013 22:00:44 +0000 (22:00 +0000)]
Allow expression evaluation to work when multiple threads exist in the inferior (on Linux)
- handle m_resume_state == eStateStopped || eStateSuspended in DoResume rather than asserting
llvm-svn: 175094
Eli Bendersky [Wed, 13 Feb 2013 22:00:37 +0000 (22:00 +0000)]
s/grep/FileCheck/ in some tests
llvm-svn: 175093
Nick Lewycky [Wed, 13 Feb 2013 21:59:15 +0000 (21:59 +0000)]
Don't build tail calls to functions with three inreg arguments on x86-32 PIC.
Fixes PR15250!
llvm-svn: 175092
Sean Callanan [Wed, 13 Feb 2013 21:53:01 +0000 (21:53 +0000)]
Made NULL, nil, and Nil use the appropriate
builtins.
<rdar://problem/
13204027>
llvm-svn: 175091
Tanya Lattner [Wed, 13 Feb 2013 21:51:18 +0000 (21:51 +0000)]
Remove outdated performance.html webpage.
llvm-svn: 175090
Eli Bendersky [Wed, 13 Feb 2013 21:46:38 +0000 (21:46 +0000)]
s/grep/FileCheck/ in some tests
llvm-svn: 175089
Weiming Zhao [Wed, 13 Feb 2013 21:43:02 +0000 (21:43 +0000)]
Bug fix 13622: Add paired register support for inline asm with 64-bit data on ARM
llvm-svn: 175088
Chad Rosier [Wed, 13 Feb 2013 21:41:58 +0000 (21:41 +0000)]
[ms-inline asm] Fix up test case for non-Darwin platforms.
llvm-svn: 175087
Jyotsna Verma [Wed, 13 Feb 2013 21:38:46 +0000 (21:38 +0000)]
Hexagon: Use absolute addressing mode loads/stores for global+offset
instead of redefining separate instructions for them.
llvm-svn: 175086
Ryan Govostes [Wed, 13 Feb 2013 21:38:22 +0000 (21:38 +0000)]
Add iterator_traits to ImmutableMap and ImmutableSet.
llvm-svn: 175085
Chad Rosier [Wed, 13 Feb 2013 21:34:54 +0000 (21:34 +0000)]
[ms-inline asm] Add test case for r175083.
llvm-svn: 175084
Chad Rosier [Wed, 13 Feb 2013 21:33:44 +0000 (21:33 +0000)]
[ms-inline-asm] Add support for memory references that have non-immediate
displacements.
rdar://
12974533
llvm-svn: 175083
Chad Rosier [Wed, 13 Feb 2013 21:27:17 +0000 (21:27 +0000)]
[ms-inline asm] Add a comment about the determinism of the rewrite sort.
llvm-svn: 175082
Dmitri Gribenko [Wed, 13 Feb 2013 21:19:39 +0000 (21:19 +0000)]
configure: remove workaround for gcc's -Wno-maybe-uninitialized
Since r174770 gcc version check is not needed because CXX_FLAG_CHECK
implements the workaround itself.
llvm-svn: 175080
Richard Smith [Wed, 13 Feb 2013 21:18:23 +0000 (21:18 +0000)]
ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11,
base-to-derived casts have undefined behavior if the object is not actually an
instance of the derived type. Runtime library part.
llvm-svn: 175079
Richard Smith [Wed, 13 Feb 2013 21:18:23 +0000 (21:18 +0000)]
ubsan: Add checking for invalid downcasts. Per [expr.static.cast]p2 and p11,
base-to-derived casts have undefined behavior if the object is not actually an
instance of the derived type.
llvm-svn: 175078
Sean Silva [Wed, 13 Feb 2013 21:17:20 +0000 (21:17 +0000)]
[docs] PR15254: Add "AST" to the lexicon.
llvm-svn: 175077
Benjamin Kramer [Wed, 13 Feb 2013 21:12:29 +0000 (21:12 +0000)]
LoopVectorize: Simplify code for clarity.
No functionality change.
llvm-svn: 175076
Bill Wendling [Wed, 13 Feb 2013 21:10:15 +0000 (21:10 +0000)]
Add a blurb about the attributes changes to the release notes.
llvm-svn: 175075
Daniel Jasper [Wed, 13 Feb 2013 20:33:44 +0000 (20:33 +0000)]
Allow breaking after the return type in function declarations.
This has so far been disabled for Google style, but should be done
before breaking at nested name specifiers or in template parameters.
Before (in Google style):
template <typename T>
aaaaaaaa::aaaaa::aaaaaa<T,
aaaaaaaaaaaaaaaaaaaaaaaaa>
aaaaaaaaaaaaaaaaaaaaaaaa<
T>::aaaaaaa() {}
After:
template <typename T>
aaaaaaaa::aaaaa::aaaaaa<T,
aaaaaaaaaaaaaaaaaaaaaaaaa>
aaaaaaaaaaaaaaaaaaaaaaaa<T>::aaaaaaa() {}
llvm-svn: 175074
Reed Kotler [Wed, 13 Feb 2013 20:28:27 +0000 (20:28 +0000)]
For Mips 16, add the optimization where the 16 bit form of addiu sp can be used
if the offset fits in 11 bits. This makes use of the fact that the abi
requires sp to be 8 byte aligned so the actual offset can fit in 8
bits. It will be shifted left and sign extended before being actually used.
The assembler or direct object emitter will shift right the 11 bit
signed field by 3 bits. We don't need to deal with that here.
llvm-svn: 175073
Manman Ren [Wed, 13 Feb 2013 20:23:48 +0000 (20:23 +0000)]
Clean up LDV, no functionality change.
Remove dead functions: renameRegister
Move private member variables from LDV to Impl
Remove ssp/uwtable from testing case
llvm-svn: 175072
Andrew Kaylor [Wed, 13 Feb 2013 19:57:06 +0000 (19:57 +0000)]
Improved handling for DW_AT_const_value
llvm-svn: 175071
Bill Wendling [Wed, 13 Feb 2013 19:44:17 +0000 (19:44 +0000)]
Use 'RC_XBS' instead of 'RC_BUILDIT' to catch all times when it's built in the Apple way.
llvm-svn: 175070
Bill Wendling [Wed, 13 Feb 2013 19:44:08 +0000 (19:44 +0000)]
Use 'RC_XBS' instead of 'RC_BUILDIT' to catch all times when it's built in the Apple way.
llvm-svn: 175069
Daniel Jasper [Wed, 13 Feb 2013 19:25:54 +0000 (19:25 +0000)]
Fix comment alignment close to the column limit.
Due to an error in one of the expressions, we used to not align comments
although it would have been possible.
llvm-svn: 175068
Andrew Trick [Wed, 13 Feb 2013 19:22:27 +0000 (19:22 +0000)]
MIsched: HazardRecognizers are created for each DAG. Free them.
llvm-svn: 175067
David Peixotto [Wed, 13 Feb 2013 19:21:47 +0000 (19:21 +0000)]
PR14992 - Tablegen incorrectly converts ARM tLDMIA_UPD pseudo to tLDMIA
Fixed bug in tablegen conversion when source pseudo instruction has
a different number of arguments than the destination instruction.
llvm-svn: 175066
Daniel Malea [Wed, 13 Feb 2013 18:56:41 +0000 (18:56 +0000)]
Skip tests that assert on Linux in RecordLayoutBuilder::updateExternalFieldOffset()
- Filed bugzilla PR-15256
llvm-svn: 175065
Daniel Malea [Wed, 13 Feb 2013 18:55:17 +0000 (18:55 +0000)]
Disable TestConditionalBreak.py due to Linux crash
- test was passing before r174793...
llvm-svn: 175064
Chad Rosier [Wed, 13 Feb 2013 18:38:58 +0000 (18:38 +0000)]
[ms-inline-asm] Use an array_pod_sort, rather than a std:sort.
llvm-svn: 175063
Argyrios Kyrtzidis [Wed, 13 Feb 2013 18:33:28 +0000 (18:33 +0000)]
[libclang] Fix annotation of a range where the begin or end location
is inside a macro argument.
Previously we would give up and not annotate anything in the range.
rdar://
11891550
llvm-svn: 175062
Greg Clayton [Wed, 13 Feb 2013 18:15:56 +0000 (18:15 +0000)]
Quiet "the missing case in switch statement" warnings from newer clang builds.
llvm-svn: 175061
Pekka Jaaskelainen [Wed, 13 Feb 2013 18:08:57 +0000 (18:08 +0000)]
Metadata for annotating loops as parallel. The first consumer for this
metadata is the loop vectorizer.
See the documentation update for more info.
llvm-svn: 175060
Alexander Potapenko [Wed, 13 Feb 2013 17:52:55 +0000 (17:52 +0000)]
[ASan] When re-executing the process on OS X, make sure we update the existing DYLD_INSERT_LIBRARIES correctly.
Previously ASan used to hang in an exec loop, because it failed to overwrite the env var value
(see https://code.google.com/p/address-sanitizer/issues/detail?id=159).
llvm-svn: 175059
Krzysztof Parzyszek [Wed, 13 Feb 2013 17:40:07 +0000 (17:40 +0000)]
Add registration for PPC-specific passes to allow the IR to be dumped
via -print-after-all.
llvm-svn: 175058
Benjamin Kramer [Wed, 13 Feb 2013 13:40:35 +0000 (13:40 +0000)]
X86: Disable generation of rep;movsl when %esi is used as a base pointer.
This happens when there is both stack realignment and a dynamic alloca in the
function. If we overwrite %esi (rep;movsl uses fixed registers) we'll lose the
base pointer and the next register spill will write into oblivion.
Fixes PR15249 and unbreaks firefox on i386/freebsd. Mozilla uses dynamic allocas
and freebsd a 4 byte stack alignment.
llvm-svn: 175057
Dmitry Vyukov [Wed, 13 Feb 2013 13:05:36 +0000 (13:05 +0000)]
tsan: do not imitate memory write on malloc() (Go)
better memory range access functions (put only 1 event to trace) (Go)
llvm-svn: 175056
Tim Northover [Wed, 13 Feb 2013 12:46:32 +0000 (12:46 +0000)]
Mention AArch64 in release notes for 3.3
llvm-svn: 175055
Timur Iskhodzhanov [Wed, 13 Feb 2013 12:24:19 +0000 (12:24 +0000)]
... and now fix the +Asserts build
llvm-svn: 175054
Timur Iskhodzhanov [Wed, 13 Feb 2013 12:14:25 +0000 (12:14 +0000)]
Fix the microsoft-abi-structors test expectations to match both Release and Release+Asserts builds
llvm-svn: 175053
Manuel Klimek [Wed, 13 Feb 2013 10:54:19 +0000 (10:54 +0000)]
Pull search state out as class members.
Fix some comments.
llvm-svn: 175052
Manuel Klimek [Wed, 13 Feb 2013 10:46:36 +0000 (10:46 +0000)]
An attempt to make the search algorithm easier to understand.
- clear ownership: the SpecificBumpPtrAllocator owns all StateNodes
- this allows us to simplify the memoization data structure into a
std::set (FIXME: figure out whether we want to use a hash based
data structure).
- introduces StateNode as recursive data structure, instead of using
Edge and the Seen-map combined to drill through the graph
- using a count to stabilize the penalty instead of relying on the
container
- pulled out a method to forward-apply states in the end
This leads to a ~40% runtime decrease on Nico's benchmark.
Main FiXME is that the parameter lists of some function get too long.
I'd vote for either pulling the Queue etc into the Formatter proper,
or creating an inner class just for the search algorithm.
llvm-svn: 175051