Dmitri Gribenko [Sat, 15 Dec 2012 20:41:17 +0000 (20:41 +0000)]
Documentation: convert InternalsManual.html to reST
Patch by Anastasi Voitova with with small fixes by me.
llvm-svn: 170275
Kostya Serebryany [Sat, 15 Dec 2012 18:36:23 +0000 (18:36 +0000)]
[sanitizer] fix a bug that has crept into the sanitizer allocator and caused SEGV on allocations between 1Mb and 2Mb, improve the test
llvm-svn: 170274
Benjamin Kramer [Sat, 15 Dec 2012 16:47:44 +0000 (16:47 +0000)]
X86: Add a couple of target-specific dag combines that turn VSELECTS into psubus if possible.
We match the pattern "x >= y ? x-y : 0" into "subus x, y" and two special cases
if y is a constant. DAGCombiner canonicalizes those so we first have to undo the
canonicalization for those cases. The pattern occurs in gzip when the loop
vectorizer is enabled. Part of PR14613.
llvm-svn: 170273
Dmitri Gribenko [Sat, 15 Dec 2012 14:25:25 +0000 (14:25 +0000)]
Documentation: LanguageExtensions.rst: convert link to :doc: style and fix up
heading underline.
llvm-svn: 170272
Chandler Carruth [Sat, 15 Dec 2012 09:31:54 +0000 (09:31 +0000)]
Add a corollary test for PR14572. We got this code path correct already.
llvm-svn: 170271
Chandler Carruth [Sat, 15 Dec 2012 09:26:06 +0000 (09:26 +0000)]
Relax an overly aggressive assert to fix PR14572.
The alloca width is based on the alloc size, not the type size.
llvm-svn: 170270
Chandler Carruth [Sat, 15 Dec 2012 09:01:13 +0000 (09:01 +0000)]
Make '-mtune=x86_64' assume fast unaligned memory accesses.
Not all chips targeted by x86_64 have this feature, but a dramatically
increasing number do. Specifying a chip-specific tuning parameter will
continue to turn the feature on or off as appropriate for that
particular chip, but the generic flag should try to achieve the best
performance on the most widely available hardware. Today, the number of
chips with fast UA access dwarfs those without in the x86-64 space.
Note that this also brings LLVM's code generation for this '-march' flag
more in line with that of modern GCCs. Reviewed by Dan Gohman.
llvm-svn: 170269
Chandler Carruth [Sat, 15 Dec 2012 08:56:20 +0000 (08:56 +0000)]
Actually update the grammar of this sentence to reflect the removal of CellSPU.
llvm-svn: 170268
NAKAMURA Takumi [Sat, 15 Dec 2012 06:11:13 +0000 (06:11 +0000)]
Revert r170246, "Enable the loop vectorizer by default."
llvm-svn: 170267
Jim Ingham [Sat, 15 Dec 2012 02:40:54 +0000 (02:40 +0000)]
Remove the “len” defaulted parameter from CommandReturnObject::AppendMessage, AppendWarning and AppendError. Nobody was using them, and it meant if you accidentally used the AppendWarning when you meant AppendWarningWithFormat with an integer in the format string, it would compile and then return your string plus some unknown amount of junk.
llvm-svn: 170266
Greg Clayton [Sat, 15 Dec 2012 02:08:17 +0000 (02:08 +0000)]
<rdar://problem/
11990131>
Memory read's "repeat" behavior forgets "-t" option. It also formatted the type as hex bytes + ASCII. Now we revert to the default format when displaying types unless the user sets the format option manually.
llvm-svn: 170265
Greg Clayton [Sat, 15 Dec 2012 01:44:51 +0000 (01:44 +0000)]
<rdar://problem/
12156204>
x/a print wouldn't always reset the word size to the size of a pointer if a previous memory read using x/<gdb-format> had been used that set it to another width.
llvm-svn: 170264
Will Dietz [Sat, 15 Dec 2012 01:39:14 +0000 (01:39 +0000)]
[ubsan] Emit branch weight metadata to hint towards common case.
Results in better block placement that helps close the
performance gap when making ubsan checks recoverable.
llvm-svn: 170263
Greg Clayton [Sat, 15 Dec 2012 01:19:07 +0000 (01:19 +0000)]
<rdar://problem/
12582041>
_regexp_attach doesn't handle the case where no arguments are provided. It now also handles the case you were you pass options.
llvm-svn: 170262
Jordan Rose [Sat, 15 Dec 2012 00:37:01 +0000 (00:37 +0000)]
Docs: redirect "static analysis extensions" section to the analyzer site.
The notes on the objc_method_family and ns_returns_retained-type attributes
have been moved to the Objective-C section, since both are used by ARC.
The notes on analyzer_noreturn are now only on the analyzer site.
The inadequacy of these docs was noticed months ago by Jonathan Sauer;
I'm only just now getting around to cleaning them up.
llvm-svn: 170261
Jordan Rose [Sat, 15 Dec 2012 00:36:53 +0000 (00:36 +0000)]
Remove old description of analyzer internals from public docs.
The file still exists in docs/analyzer/, but it won't be linked to from
clang.llvm.org or processed as part of the default Sphinx doc-build.
RegionStore has changed a lot from what Ted and Zhongxing describe here!
llvm-svn: 170260
Reed Kotler [Sat, 15 Dec 2012 00:20:05 +0000 (00:20 +0000)]
This code implements most of mips16 hardfloat as it is done by gcc.
In this case, essentially it is soft float with different library routines.
The next step will be to make this fully interoperational with mips32 floating
point and that requires creating stubs for functions with signatures that
contain floating point types.
I have a more sophisticated design for mips16 hardfloat which I hope to
implement at a later time that directly does floating point without the need
for function calls.
The mips16 encoding has no floating point instructions so one needs to
switch to mips32 mode to execute floating point instructions.
llvm-svn: 170259
Eric Christopher [Sat, 15 Dec 2012 00:04:07 +0000 (00:04 +0000)]
To simplify some code move the unit emission into the holders.
Make emitDIE public accordingly. No functional change.
llvm-svn: 170258
Eric Christopher [Sat, 15 Dec 2012 00:04:04 +0000 (00:04 +0000)]
Use begin and end label names from the section for info.
llvm-svn: 170257
Sean Callanan [Fri, 14 Dec 2012 23:43:03 +0000 (23:43 +0000)]
Fixed two conditionals that I accidentally
reversed in r170152.
<rdar://problem/
12886584>
llvm-svn: 170256
Kevin Enderby [Fri, 14 Dec 2012 23:04:25 +0000 (23:04 +0000)]
Make sure the alternate PC+imm syntax of LDR instruction with a small
immediate generates the narrow version. Needed when doing round-trip
assemble/disassemble testing using the alternate syntax that specifies
'pc' directly.
llvm-svn: 170255
Greg Clayton [Fri, 14 Dec 2012 22:36:35 +0000 (22:36 +0000)]
Switch "disassemble" with no arguments or options to disassemble the current frame instead of around the current PC.
llvm-svn: 170254
Greg Clayton [Fri, 14 Dec 2012 22:35:50 +0000 (22:35 +0000)]
Remove unused variable.
llvm-svn: 170253
Greg Clayton [Fri, 14 Dec 2012 22:17:58 +0000 (22:17 +0000)]
Remove references to files that Sean recently removed from the Xcode project.
llvm-svn: 170250
Michael Ilseman [Fri, 14 Dec 2012 22:08:26 +0000 (22:08 +0000)]
Add back FoldOpIntoPhi optimizations with fix. Included test cases to help catch these errors and to test the presence of the optimization itself
llvm-svn: 170248
Nadav Rotem [Fri, 14 Dec 2012 21:30:23 +0000 (21:30 +0000)]
Enable the loop vectorizer by default.
llvm-svn: 170246
Nadav Rotem [Fri, 14 Dec 2012 21:20:37 +0000 (21:20 +0000)]
TypeLegalizer: Do not generate target specific nodes with illegal types, because we cant type-legalize them.
llvm-svn: 170245
Andrew Kaylor [Fri, 14 Dec 2012 21:11:46 +0000 (21:11 +0000)]
Skipping the DeadStripTestCase.test_with_dwarf test on Linux because the Linux ld lacks support for the -dead_strip option.
llvm-svn: 170244
Duncan Sands [Fri, 14 Dec 2012 21:10:59 +0000 (21:10 +0000)]
Release notes for dragonegg 3.2.
llvm-svn: 170243
Daniel Malea [Fri, 14 Dec 2012 21:07:07 +0000 (21:07 +0000)]
Avoid possible overflow when reading inferior memory (and logging is enabled)
Patch by Matt Kopec!
llvm-svn: 170242
Andrew Kaylor [Fri, 14 Dec 2012 21:03:37 +0000 (21:03 +0000)]
Enabling ItaniumABILanguageRuntime and SymbolFileDWARFDebugMap plugins on non-Apple platforms.
llvm-svn: 170241
Nadav Rotem [Fri, 14 Dec 2012 20:43:49 +0000 (20:43 +0000)]
Fix a crash in ValueTracking on vectors of pointers.
llvm-svn: 170240
Filipe Cabecinhas [Fri, 14 Dec 2012 20:38:58 +0000 (20:38 +0000)]
Fixed a typo.
llvm-svn: 170239
Anton Yartsev [Fri, 14 Dec 2012 20:28:48 +0000 (20:28 +0000)]
fixed line endings
llvm-svn: 170238
Bill Schmidt [Fri, 14 Dec 2012 20:28:38 +0000 (20:28 +0000)]
This patch removes some nondeterminism from direct object file output
for TLS dynamic models on 64-bit PowerPC ELF. The default sort routine
for relocations only sorts on the r_offset field; but with TLS, there
can be two relocations with the same r_offset. For PowerPC, this patch
sorts secondarily on descending r_type, which matches the behavior
expected by the linker.
llvm-svn: 170237
Daniel Malea [Fri, 14 Dec 2012 20:02:21 +0000 (20:02 +0000)]
Add GCC instructions to build section of website
llvm-svn: 170236
Dmitry Vyukov [Fri, 14 Dec 2012 20:01:58 +0000 (20:01 +0000)]
tsan: synchronize connect->accept calls
llvm-svn: 170235
Anton Yartsev [Fri, 14 Dec 2012 19:48:34 +0000 (19:48 +0000)]
added post-statement callback to CXXNewExpr and pre-statement callback to CXXDeleteExpr
llvm-svn: 170234
Argyrios Kyrtzidis [Fri, 14 Dec 2012 19:44:11 +0000 (19:44 +0000)]
Don't hit an assertion failure when calculating the __PRETTY_FUNCTION__
of a member function with parenthesized declarator.
Like this test case:
class Foo {
const char *(baz)() {
return __PRETTY_FUNCTION__;
}
};
llvm-svn: 170233
Anna Zaks [Fri, 14 Dec 2012 19:08:20 +0000 (19:08 +0000)]
[analyzer] Propagate the checker's state from checkBranchCondition
Fixes a bug, where we were dropping the state modifications from the
checkBranchCondition checker callback.
llvm-svn: 170232
Anna Zaks [Fri, 14 Dec 2012 19:08:17 +0000 (19:08 +0000)]
[analyzer] Refactor: Store visited Decls instead of CallGraphNodes.
llvm-svn: 170231
Edwin Vane [Fri, 14 Dec 2012 18:59:24 +0000 (18:59 +0000)]
Use renamed CommonOptionsParser member functions
GetCompilations() and GetSourcePathLists() were renamed to match LLVM style.
Reviewed By: Alexander Kornienko
llvm-svn: 170230
Edwin Vane [Fri, 14 Dec 2012 18:58:25 +0000 (18:58 +0000)]
Style and Doc fix for CommonOptionsParser
- Renaming GetCompilations() and GetSourcePathList() to follow LLVM
style.
- Updating docs to reflect name change.
- Also updating help text to not mention clang-check since this class
can be used by any tool.
Reviewed By: Alexander Kornienko
llvm-svn: 170229
Argyrios Kyrtzidis [Fri, 14 Dec 2012 18:53:47 +0000 (18:53 +0000)]
[preprocessor] For errors at a function macro invocation, also include
a note about where the macro is defined.
llvm-svn: 170228
Pedro Artigas [Fri, 14 Dec 2012 18:52:11 +0000 (18:52 +0000)]
Add more reset methods to make all objects that the backend may use for outputting code have a reset, some are not used but were declared for completeness
llvm-svn: 170227
Shuxin Yang [Fri, 14 Dec 2012 18:46:06 +0000 (18:46 +0000)]
rdar://
12753946
Implement rule : "x * (select cond 1.0, 0.0) -> select cond x, 0.0"
llvm-svn: 170226
NAKAMURA Takumi [Fri, 14 Dec 2012 18:30:20 +0000 (18:30 +0000)]
[CMake] Move libxml2 stuff from clang to llvm/cmake.
llvm-svn: 170225
Andrew Kaylor [Fri, 14 Dec 2012 18:24:34 +0000 (18:24 +0000)]
Change crash handling to use eStateStopped rather than eStateCrashed.
llvm-svn: 170224
Nico Weber [Fri, 14 Dec 2012 18:22:38 +0000 (18:22 +0000)]
Don't require a space between the two ">" in "vector<id<protocol>>" in objc++11.
C++11 allowed writing "vector<vector<int>>" without a space between the two ">".
This change allows this for protocols in template lists too in -std=c++11 mode,
and improves the diagnostic in c++98 mode.
llvm-svn: 170223
Bill Schmidt [Fri, 14 Dec 2012 17:02:38 +0000 (17:02 +0000)]
This patch improves the 64-bit PowerPC InitialExec TLS support by providing
for a wider range of GOT entries that can hold thread-relative offsets.
This matches the behavior of GCC, which was not documented in the PPC64 TLS
ABI. The ABI will be updated with the new code sequence.
Former sequence:
ld 9,x@got@tprel(2)
add 9,9,x@tls
New sequence:
addis 9,2,x@got@tprel@ha
ld 9,x@got@tprel@l(9)
add 9,9,x@tls
Note that a linker optimization exists to transform the new sequence into
the shorter sequence when appropriate, by replacing the addis with a nop
and modifying the base register and relocation type of the ld.
llvm-svn: 170209
Alexey Samsonov [Fri, 14 Dec 2012 15:37:35 +0000 (15:37 +0000)]
[Sanitizer] remove extra typename
llvm-svn: 170208
Dmitry Vyukov [Fri, 14 Dec 2012 14:42:40 +0000 (14:42 +0000)]
tsan: add test for errno spoiling in signal handler
llvm-svn: 170207
Kostya Serebryany [Fri, 14 Dec 2012 14:20:29 +0000 (14:20 +0000)]
[sanitizer] add AllocatorLeakTest
llvm-svn: 170206
Dmitry Vyukov [Fri, 14 Dec 2012 13:56:27 +0000 (13:56 +0000)]
tsan: add test for malloc/free in signal handler
llvm-svn: 170205
Evgeniy Stepanov [Fri, 14 Dec 2012 13:48:55 +0000 (13:48 +0000)]
Add MemorySanitizer.cpp to the lint script.
llvm-svn: 170204
Evgeniy Stepanov [Fri, 14 Dec 2012 13:48:31 +0000 (13:48 +0000)]
Fix lint warnings in MemorySanitizer.cpp.
llvm-svn: 170203
Dmitry Vyukov [Fri, 14 Dec 2012 13:46:03 +0000 (13:46 +0000)]
tsan: fix lint warning
llvm-svn: 170202
Chandler Carruth [Fri, 14 Dec 2012 13:44:05 +0000 (13:44 +0000)]
The CellSPU backend is more than experimental, it's dead now...
llvm-svn: 170201
Chandler Carruth [Fri, 14 Dec 2012 13:43:59 +0000 (13:43 +0000)]
Remove a section that was in 3.1's notes.
Just give a shout if this was actually still important....
llvm-svn: 170200
Evgeniy Stepanov [Fri, 14 Dec 2012 13:43:11 +0000 (13:43 +0000)]
[msan] Origin stores and loads do not need explicit alignment.
Origin address is always 4 byte aligned, and the access type is always i32.
llvm-svn: 170199
Chandler Carruth [Fri, 14 Dec 2012 13:41:02 +0000 (13:41 +0000)]
Remove several entries from the 3.2 release notes that were in the 3.1
release notes already. =]
llvm-svn: 170198
Chandler Carruth [Fri, 14 Dec 2012 13:37:18 +0000 (13:37 +0000)]
Fix the order of these sections of the release notes.
llvm-svn: 170197
Chandler Carruth [Fri, 14 Dec 2012 13:37:17 +0000 (13:37 +0000)]
Add a rough draft of some content about the new SROA. I'll try to proof
read this and clean it up tomorrow, but hopefully it's a good
placeholder.
llvm-svn: 170196
Chandler Carruth [Fri, 14 Dec 2012 13:37:12 +0000 (13:37 +0000)]
Delete a long-stale "if its ready" comment. All indications is that this
will look a bit different when we have time to get it ready to turn on,
and we won't likely need this reminder.
llvm-svn: 170195
Chandler Carruth [Fri, 14 Dec 2012 13:22:57 +0000 (13:22 +0000)]
Provide some actual highlight bullets for Clang.
If anyone has better highlights (I'm obviously biased by the things that
I'm excited about) jump in and add them!
llvm-svn: 170194
Kostya Serebryany [Fri, 14 Dec 2012 13:16:19 +0000 (13:16 +0000)]
[asan] more asan_allocator2 code: record the allocation stack trace, implement calloc/etc
llvm-svn: 170193
Evgeniy Stepanov [Fri, 14 Dec 2012 12:54:18 +0000 (12:54 +0000)]
[msan] Refactor default shadow propagation and origin tracking.
This change moves the code for default shadow propagaition (handleShadowOr)
and origin tracking (setOriginForNaryOp) into a new builder-like class. Also
gets rid of handleShadowOrBinary.
llvm-svn: 170192
Dmitry Vyukov [Fri, 14 Dec 2012 12:24:11 +0000 (12:24 +0000)]
tsan: fix build
llvm-svn: 170191
Kostya Serebryany [Fri, 14 Dec 2012 12:15:09 +0000 (12:15 +0000)]
[asan] more asan_allocator2 code: actually un/poison shadow on malloc/free
llvm-svn: 170190
Alexey Samsonov [Fri, 14 Dec 2012 11:52:02 +0000 (11:52 +0000)]
[Sanitizer] use INTERCEPT_FUNCTION(foo) instead of CHECK(INTERCEPT_FUNCTION(foo)) - this check may fail on Linux if user program defines its own version of library function foo
llvm-svn: 170189
NAKAMURA Takumi [Fri, 14 Dec 2012 10:17:26 +0000 (10:17 +0000)]
Targets.cpp: [cygwin] Add the predefined macro "_X86_", according to newer version of cygwin/w32api.
llvm-svn: 170188
Dmitry Vyukov [Fri, 14 Dec 2012 10:17:22 +0000 (10:17 +0000)]
tsan: disable allocator tests in debug build
llvm-svn: 170187
Dmitry Vyukov [Fri, 14 Dec 2012 10:12:14 +0000 (10:12 +0000)]
tsan: comment out debug output in test
llvm-svn: 170186
Dmitry Vyukov [Fri, 14 Dec 2012 10:10:57 +0000 (10:10 +0000)]
tsan: add sanity test for Go runtime into presubmit test
llvm-svn: 170185
Dmitry Vyukov [Fri, 14 Dec 2012 09:57:42 +0000 (09:57 +0000)]
tsan: add socketpair() interceptor and test
llvm-svn: 170184
Patrik Hagglund [Fri, 14 Dec 2012 09:05:13 +0000 (09:05 +0000)]
Change TargetLowering::getLoadExtAction to take an MVT, instead of
EVT.
llvm-svn: 170183
Argyrios Kyrtzidis [Fri, 14 Dec 2012 06:54:03 +0000 (06:54 +0000)]
Have Sema::ActOnStartOfFunctionDef return the declaration that was passed it.
This fixes the missing warning here:
struct S {
template <typename T>
void meth() {
char arr[3];
arr[4] = 0; // warning: array index 4 is past the end of the array
}
};
template <typename T>
void func() {
char arr[3];
arr[4] = 0; // no warning
}
llvm-svn: 170180
Argyrios Kyrtzidis [Fri, 14 Dec 2012 06:53:58 +0000 (06:53 +0000)]
Remove code from Sema::ActOnStartOfFunctionTemplateDef that duplicates what
Sema::ActOnStartOfFunctionDef is already doing.
llvm-svn: 170179
Nico Weber [Fri, 14 Dec 2012 02:41:18 +0000 (02:41 +0000)]
fix another spello
llvm-svn: 170177
Nico Weber [Fri, 14 Dec 2012 02:40:09 +0000 (02:40 +0000)]
fix spello
llvm-svn: 170176
NAKAMURA Takumi [Fri, 14 Dec 2012 02:35:04 +0000 (02:35 +0000)]
clang/test/CodeGenCXX/lambda-expressions.cpp: Relax expression for -Asserts.
"entry:" is not met in -Asserts build.
llvm-svn: 170175
Greg Clayton [Fri, 14 Dec 2012 02:15:00 +0000 (02:15 +0000)]
Cleaned up the UUID mismatch just printing itself whenever it wants to by allowing an optional feedback stream to be passed along when getting the symbol vendor.
llvm-svn: 170174
Greg Clayton [Fri, 14 Dec 2012 02:13:44 +0000 (02:13 +0000)]
<rdar://problem/
12878674>
Remove debug asserts.
llvm-svn: 170173
Nadav Rotem [Fri, 14 Dec 2012 01:57:00 +0000 (01:57 +0000)]
revert r170166 - disable the loop vectorizer.
llvm-svn: 170172
Sean Callanan [Fri, 14 Dec 2012 01:29:59 +0000 (01:29 +0000)]
Rmoved the old LLVM disassembler based on libedis.
llvm-svn: 170171
Ted Kremenek [Fri, 14 Dec 2012 01:23:13 +0000 (01:23 +0000)]
Refactor dump methods to make RegionBindingsRef printable in the debugger.
llvm-svn: 170170
Sean Callanan [Fri, 14 Dec 2012 00:56:57 +0000 (00:56 +0000)]
Made the struct test case actually use an expression
instead of falling through to the "frame variable"
code.
llvm-svn: 170169
Sean Callanan [Fri, 14 Dec 2012 00:54:13 +0000 (00:54 +0000)]
Some incorrect debug information caused LLDB
to report a structure with an array of size 1
at the end without accounting for that array
when reporting the struct's total size to Clang.
LLDB now coerces such an array to size 0.
<rdar://problem/
12822204>
llvm-svn: 170168
Enrico Granata [Fri, 14 Dec 2012 00:52:54 +0000 (00:52 +0000)]
Trigger the display of error and output in sourced commands from the result object's status instead of the presence of text in the error stream
This should be more consistent with the notion of command success/failure and avoids spewing warnings that the user might not care about
There will need to be an option to specify the level of verbosity desired (never show anything, only show failures, errors and warning, everything)
llvm-svn: 170167
Nadav Rotem [Fri, 14 Dec 2012 00:30:34 +0000 (00:30 +0000)]
Enable the loop vectorizer.
llvm-svn: 170166
Greg Clayton [Fri, 14 Dec 2012 00:26:21 +0000 (00:26 +0000)]
Fix a parser_type to get created with the right AST, and also make variables made from symbols to not be "void * const", but just "void *".
llvm-svn: 170165
Nadav Rotem [Fri, 14 Dec 2012 00:26:15 +0000 (00:26 +0000)]
Make the test less sensitive to the inline threshold.
llvm-svn: 170164
Enrico Granata [Fri, 14 Dec 2012 00:07:09 +0000 (00:07 +0000)]
Fixing the -f option so that one specify multiple filters, e.g.
./dotest.py -C clang --arch x86_64 --arch i386 -v -t -f ObjCDataFormatterTestCase.test_appkit_with_dsym_and_run_command -f ObjCDataFormatterTestCase.test_appkit_with_dwarf_and_run_command -f TestObjCBuiltinTypes.test_with_dsym_and_python_api -f TestObjCBuiltinTypes.test_with_dwarf_and_python_api -f ObjCDataFormatterTestCase.test_appkit_with_dsym_and_run_command -f ObjCDataFormatterTestCase.test_appkit_with_dwarf_and_run_command -f TestObjCBuiltinTypes.test_with_dsym_and_python_api -f -TestObjCBuiltinTypes.test_with_dwarf_and_python_api
The previous implementation would only remember the last filter passed, and consequently break redo.py
llvm-svn: 170163
Nadav Rotem [Fri, 14 Dec 2012 00:02:07 +0000 (00:02 +0000)]
Disable the loop vectorizer.
llvm-svn: 170162
Enrico Granata [Thu, 13 Dec 2012 23:50:33 +0000 (23:50 +0000)]
<rdar://problem/
11689939>
Supporting a compact display syntax for ObjC pointers where 0x00.....0 is replaced by a much more legible "nil"
e.g. this would show:
(NSArray *) $2 = nil
instead of:
(NSArray *) $2 = 0x0000000000000000 <nil>
llvm-svn: 170161
Eli Friedman [Thu, 13 Dec 2012 23:37:17 +0000 (23:37 +0000)]
Make sure the __invoke function for lambdas returns properly. Per bug report on IRC>
llvm-svn: 170160
Jakob Stoklund Olesen [Thu, 13 Dec 2012 23:23:46 +0000 (23:23 +0000)]
Use the new MI bundling API in MachineInstrBundle itself.
The new API is higher level than just manipulating the bundle flags
directly, and the setIsInsideBundle() function will disappear soon.
llvm-svn: 170159
Shuxin Yang [Thu, 13 Dec 2012 23:16:19 +0000 (23:16 +0000)]
Remove two popcount patterns which we are already able to recognize.
llvm-svn: 170158
Nadav Rotem [Thu, 13 Dec 2012 23:11:54 +0000 (23:11 +0000)]
Enable the Loop Vectorizer by default for O2 and O3. Disable if-conversion by default. I plan to revert this patch later today.
llvm-svn: 170157
David Blaikie [Thu, 13 Dec 2012 22:43:07 +0000 (22:43 +0000)]
Debug Info: add support to mark member variables as artificial
This is the LLVM portion of r170154.
llvm-svn: 170156