Benjamin Kramer [Thu, 7 Mar 2013 20:33:29 +0000 (20:33 +0000)]
ArrayRefize some code. No functionality change.
llvm-svn: 176648
Jyotsna Verma [Thu, 7 Mar 2013 20:28:34 +0000 (20:28 +0000)]
Hexagon: Handle i8, i16 and i1 Var Args.
llvm-svn: 176647
Jakub Staszak [Thu, 7 Mar 2013 20:22:39 +0000 (20:22 +0000)]
Simplify code. No functionality change.
llvm-svn: 176646
Jakub Staszak [Thu, 7 Mar 2013 20:21:27 +0000 (20:21 +0000)]
Change Index type from unsigned long to unsigned. This should fix PR14980.
llvm-svn: 176645
Daniel Malea [Thu, 7 Mar 2013 20:13:11 +0000 (20:13 +0000)]
Fix enum_types test on Linux
- Required flag "-std=c++11" needs to be set before including Makefile.rules
llvm-svn: 176644
Jakub Staszak [Thu, 7 Mar 2013 20:04:17 +0000 (20:04 +0000)]
Remove trailing spaces.
llvm-svn: 176643
Jakub Staszak [Thu, 7 Mar 2013 20:01:47 +0000 (20:01 +0000)]
Change NULL to 0.
llvm-svn: 176642
Jakub Staszak [Thu, 7 Mar 2013 20:01:19 +0000 (20:01 +0000)]
ArrayRef ca accept one element. Simplify code a little bit, also it matches now
coding in the other places of the file.
llvm-svn: 176641
Howard Hinnant [Thu, 7 Mar 2013 19:38:08 +0000 (19:38 +0000)]
Albert Wong: definition for regex_traits<_CharT>::__regex_word.
llvm-svn: 176640
Howard Hinnant [Thu, 7 Mar 2013 19:25:03 +0000 (19:25 +0000)]
Change _LIBCPP_TYPE_VIS to use __type_visibility__(default) instead of __visibility__(default) when available. This change makes just the type_info visible so that types like vectors and strings can be used as exception objects across dylib boundaries even when hidden visibility is specified globally (at the command line), and yet this allows clients to hide the member functions of things like vector and string (with global visibility commands).
llvm-svn: 176639
Andrew Trick [Thu, 7 Mar 2013 19:21:08 +0000 (19:21 +0000)]
pre-RA-sched debug-only fix
llvm-svn: 176638
Jyotsna Verma [Thu, 7 Mar 2013 19:10:28 +0000 (19:10 +0000)]
Hexagon: Add support to lower block address.
llvm-svn: 176637
Andrew Trick [Thu, 7 Mar 2013 19:07:57 +0000 (19:07 +0000)]
pre-RA-sched assertion fix. This bug was exposed by r176037.
rdar:
13370002 [pre-RA-sched] assertion: released too many times
I tracked this down to an earlier hack that is no longer applicable
and interfered with normal scheduler logic. With the changes in
r176037, it was causing an instruction to be scheduled multiple times.
I have an external test case that I tried hard to reduce and
failed. I can't even reproduce with llc.
llvm-svn: 176636
Benjamin Kramer [Thu, 7 Mar 2013 18:51:02 +0000 (18:51 +0000)]
Move testcase, this is testing extraction not inserting.
llvm-svn: 176635
Benjamin Kramer [Thu, 7 Mar 2013 18:48:40 +0000 (18:48 +0000)]
X86: Fold EXTRACT_SUBVECTORs of a BUILD_VECTOR into a smaller BUILD_VECTOR.
That can usually be lowered efficiently and is common in sandybridge code.
It would be nice to do this in DAGCombiner but we can't insert arbitrary
BUILD_VECTORs this late.
Fixes PR15462.
llvm-svn: 176634
Daniel Malea [Thu, 7 Mar 2013 18:39:26 +0000 (18:39 +0000)]
Fix issue that causes linux buildbots to ignore errors
- change string "ERROR" to "FAIL" to match clang lit test results
Also, make LLDB tests work on machines that do not have svn and/or git installed
llvm-svn: 176633
Pekka Jaaskelainen [Thu, 7 Mar 2013 16:46:43 +0000 (16:46 +0000)]
Fixed a crash when cloning a function into a function with
different size argument list and without attributes in the
arguments.
llvm-svn: 176632
Edwin Vane [Thu, 7 Mar 2013 16:22:05 +0000 (16:22 +0000)]
Have LoopConvert use 'auto &&' where necessary
For iterators where the dereference operator returns by value, LoopConvert
should use 'auto &&' in the range-based for loop expression.
If the dereference operator returns an rvalue reference, this is deemed too
strange and the for loop is not converted.
Moved test case from iterator_failing.cpp to iterator.cpp and added extra
tests.
Fixes PR15437.
Reviewer: gribozavr
llvm-svn: 176631
Edwin Vane [Thu, 7 Mar 2013 15:44:40 +0000 (15:44 +0000)]
Adding lvalue and rvalue reference type matchers
Updated docs and tests.
Reviewers: klimek, gribozavr
llvm-svn: 176630
Guy Benyei [Thu, 7 Mar 2013 13:06:10 +0000 (13:06 +0000)]
Add 'e' to the SPIR data layout - SPIR is little endian for now.
Add 32/64 bit specific target defines for SPIR.
llvm-svn: 176629
Matthew Curtis [Thu, 7 Mar 2013 12:32:26 +0000 (12:32 +0000)]
Minor refactor of how we get compilation phases.
There is now a single function to get the list of phases for a given
output Type.
No functionality change intended.
llvm-svn: 176628
Chandler Carruth [Thu, 7 Mar 2013 10:09:47 +0000 (10:09 +0000)]
Switch from autogenerating tests to using the preprocessor.
NOTE: You may need to run 'make clean' or 'ninja -t clean' etc!!! This
is due to really nasty bug/interactions between
CMake/configure/make/Ninja/LIT...
This commit tries to back out the support for generating test cases as
part of the build system due to the issues I brought up in post-commit
review:
1) It adds a *lot* of complexity and fragility to the build system. See
the number of commits required to try to get all the bots happy.
2) It isn't really necessary -- we can already run scripts to generate
things with the RUN lines of a test.
3) It makes the tests somewhat harder to debug as they cross between
more domains.
4) In almost all cases it isn't really needed or it can be done directly
using the preprocessor.
I should have been more proactive reviewing this, and I'm really sorry
about the churn here. =/ To help keep track of what commits are going
where, this backs out most of the non-test-changes from these revisions:
r176397
r176373
r176293
r176184
r175744
r175624
r175545
r175544
There were several trivial or cleanup changes to the lit files or other
files. Some of these looked ok, but I didn't try to tease them apart...
Edwin, if you know what to look for, please carry on with the cleanups
there, and sorry for hosing stuff here but I'm not much of a Python
person, and so I was erring on the side of cautiously backing out the
change.
I've tried to preserve the test changes everywhere I could, but review
is appreciated here in case I missed some.
I then re-wrote the tests to use the preprocessor rather than python to
expand to the various bits of code. The nicest part of this is that now
all the files are just C++ code. They edit and behave like C++ code,
etc. RUN lines with different -D flags are used to run the same test
over multiple different configurations, and includes bracketed in
special defines are used to flesh out a collection of standard interface
stubs to test interactions between pieces. These probably aren't perfect
yet, but I think its an improvement (at least in terms of build system
complexity) and will hopefully be a useful demonstration of the
technique I prefer for these types of tests.
llvm-svn: 176627
Christian Konig [Thu, 7 Mar 2013 09:04:14 +0000 (09:04 +0000)]
R600/SI: rework input interpolation v2
v2: update CMakeLists.txt as well
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176626
Christian Konig [Thu, 7 Mar 2013 09:04:04 +0000 (09:04 +0000)]
R600/SI: remove SI_vs_load_buffer_index
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176625
Christian Konig [Thu, 7 Mar 2013 09:03:59 +0000 (09:03 +0000)]
R600/SI: remove SGPR address space v2
v2: fix R600 regressions
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176624
Christian Konig [Thu, 7 Mar 2013 09:03:52 +0000 (09:03 +0000)]
R600/SI: add proper formal parameter handling for SI
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176623
Christian Konig [Thu, 7 Mar 2013 09:03:46 +0000 (09:03 +0000)]
R600/SI: remove shader type intrinsic
Just encode the type as target specific attribute.
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176622
Christian Konig [Thu, 7 Mar 2013 09:03:38 +0000 (09:03 +0000)]
R600/SI: switch types of SGPRs to v*i8
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176621
Christian Konig [Thu, 7 Mar 2013 09:03:30 +0000 (09:03 +0000)]
R600/SI: fix unused variable warning
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 176620
Nick Lewycky [Thu, 7 Mar 2013 08:42:27 +0000 (08:42 +0000)]
Fix build with clang, this was supposed to be part of r176617.
llvm-svn: 176619
Nick Lewycky [Thu, 7 Mar 2013 08:28:54 +0000 (08:28 +0000)]
Take the GCDA version string as an input to llvm_gcda_start_file.
llvm-svn: 176618
Nick Lewycky [Thu, 7 Mar 2013 08:28:53 +0000 (08:28 +0000)]
Add flags for additional control over coverage generation. Pick the version
string to be emitted, and two properties about the files themselves.
Use $PWD to absolut-ify the path to the coverage file. Yes, this is what GCC
does. Reverts my own r175706.
llvm-svn: 176617
Nick Lewycky [Thu, 7 Mar 2013 08:28:49 +0000 (08:28 +0000)]
Switch from a version 4.2/4.4 switch to a four-byte version string to be put
into the actual gcov file.
Instead of using the bottom 4 bytes as the function identifier, use a counter.
This makes the identifier numbers stable across multiple runs.
llvm-svn: 176616
Nadav Rotem [Thu, 7 Mar 2013 06:34:49 +0000 (06:34 +0000)]
No need to go through int64 and APInt when generating a new constant.
llvm-svn: 176615
Jim Grosbach [Thu, 7 Mar 2013 05:47:54 +0000 (05:47 +0000)]
SDAG: Handle scalarizing an extend of a <1 x iN> vector.
Just scalarize the element and rebuild a vector of the result type
from that.
rdar://
13281568
llvm-svn: 176614
Greg Clayton [Thu, 7 Mar 2013 03:25:11 +0000 (03:25 +0000)]
Added new properties to lldb.SBModule classes:
"compile_units" returns an array of all compile units in a module as a list() of lldb.SBCompileUnit objects.
"compile_unit" returns a compile unit accessor object that allows indexed access, search by full or partial path, or by regex:
(lldb) script
comp_unit = lldb.target.module['TextEdit'].compile_unit['Document.m']
comp_unit = lldb.target.module['TextEdit'].compile_unit['/path/to/Document.m']
comp_unit = lldb.target.module['TextEdit'].compile_unit[0]
comp_unit = lldb.target.module['TextEdit'].compile_unit[1]
for comp_unit in lldb.target.module['TextEdit'].compile_unit[re.compile("\.m$")]
print comp_unit
This helps do quick searches and scripting while debugging.
llvm-svn: 176613
Anna Zaks [Thu, 7 Mar 2013 03:02:36 +0000 (03:02 +0000)]
[analyzer] Warn on passing a reference to null pointer as an argument in a call
Warn about null pointer dereference earlier when a reference to a null pointer is
passed in a call. The idea is that even though the standard might allow this, reporting
the issue earlier is better for diagnostics (the error is reported closer to the place where
the pointer was set to NULL). This also simplifies analyzer’s diagnostic logic, which has
to track “where the null came from”. As a consequence, some of our null pointer
warning suppression mechanisms started triggering more often.
TODO: Change the name of the file and class to reflect the new check.
llvm-svn: 176612
Greg Clayton [Thu, 7 Mar 2013 02:58:47 +0000 (02:58 +0000)]
Stop the "module" property from throwing an exception when the module name was not found in the target module list.
llvm-svn: 176611
Greg Clayton [Thu, 7 Mar 2013 02:57:54 +0000 (02:57 +0000)]
Added a new module that can dump all line tables for all compile units in any modules that are specified as arguments to the "dump_module_line_tables" command.
I used this to verify that the debug map line tables were the same as previous LLDB releases prior to my change in the DWARF in .o file linking.
llvm-svn: 176610
Rafael Espindola [Thu, 7 Mar 2013 02:43:24 +0000 (02:43 +0000)]
Don't duplicate function names. Thanks to Jordan Rose for pointing it out.
llvm-svn: 176609
Nick Lewycky [Thu, 7 Mar 2013 02:03:08 +0000 (02:03 +0000)]
Remove GCDAProfiling.c. This copy is old, the copy in compiler-rt is newer and
is the one that should be used.
llvm-svn: 176608
Rafael Espindola [Thu, 7 Mar 2013 02:00:27 +0000 (02:00 +0000)]
Add a hasExternalLinkage helper. No functionality change.
llvm-svn: 176607
Rafael Espindola [Thu, 7 Mar 2013 01:42:44 +0000 (01:42 +0000)]
Add a hasExternalStorageAsWritten helper. No functionality change.
It is possible that some of the current uses of
"getStorageClassAsWritten() == SC_Extern" should use this but I don't know
enough about SC_PrivateExtern to change and test them.
llvm-svn: 176606
Manman Ren [Thu, 7 Mar 2013 01:42:00 +0000 (01:42 +0000)]
Debug Info: store the files and directories for each compile unit.
We now emit a line table for each compile unit. To reduce the prologue size
of each line table, the files and directories used by each compile unit are
stored in std::map<unsigned, std::vector< > > instead of std::vector< >.
The prologue for a lto'ed image can be as big as 93K. Duplicating 93K for each
compile unit causes a huge increase of debug info. With this patch, each
prologue will only emit the files required by the compile unit.
rdar://problem/
13342023
llvm-svn: 176605
Nadav Rotem [Thu, 7 Mar 2013 01:38:04 +0000 (01:38 +0000)]
ArrayRef has a OneElt constructor. Beautify the code.
llvm-svn: 176604
Rafael Espindola [Thu, 7 Mar 2013 01:34:57 +0000 (01:34 +0000)]
Fix typo.
llvm-svn: 176603
Nadav Rotem [Thu, 7 Mar 2013 01:30:40 +0000 (01:30 +0000)]
Switch from std::vector to ArrayRef. Speedup FoldBitCast by 5x.
llvm-svn: 176602
Jordan Rose [Thu, 7 Mar 2013 01:23:25 +0000 (01:23 +0000)]
[analyzer] Check for returning null references in ReturnUndefChecker.
Officially in the C++ standard, a null reference cannot exist. However,
it's still very easy to create one:
int &getNullRef() {
int *p = 0;
return *p;
}
We already check that binds to reference regions don't create null references.
This patch checks that we don't create null references by returning, either.
<rdar://problem/
13364378>
llvm-svn: 176601
Jordan Rose [Thu, 7 Mar 2013 01:23:14 +0000 (01:23 +0000)]
[analyzer] Clean up a few doc comments for ProgramState and CallEvent.
No functionality change.
llvm-svn: 176600
Andrew Trick [Thu, 7 Mar 2013 01:03:35 +0000 (01:03 +0000)]
SimplifyCFG fix for volatile load/store.
Fixes rdar:
13349374.
Volatile loads and stores need to be preserved even if the language
standard says they are undefined. "volatile" in this context means "get
out of the way compiler, let my platform handle it".
Additionally, this is the only way I know of with llvm to write to the
first page (when hardware allows) without dropping to assembly.
llvm-svn: 176599
Michael Liao [Thu, 7 Mar 2013 01:01:29 +0000 (01:01 +0000)]
Fix two remaining issue after fixing PR15355 when CMOV is not available
- Phi nodes should be replaced/updated after lowering CMOV into branch
because 'mainMBB' updating operand in Phi node is changed.
- Add EFLAGS in livein before lowering the 2nd CMOV. It's necessary as
we will reuse the EFLAGS generated before the 1st lowered CMOV, which
won't clobber EFLAGS. However, we need explicitly specify that.
- '-attr=-cmov' test case are added.
llvm-svn: 176598
Daniel Malea [Thu, 7 Mar 2013 00:52:25 +0000 (00:52 +0000)]
Clean up LLDB CMake build output
- remove "-debug" flag from swig scripts
- use "echo -n" instead of "echo" in dummy target to avoid printing a useless newline
llvm-svn: 176597
Daniel Malea [Thu, 7 Mar 2013 00:50:54 +0000 (00:50 +0000)]
Fix running of LLDB tests with cmake+ninja+ccache
- Avoid passing 'ccache' as the test compiler. Instead, use first arg after ccache.
llvm-svn: 176596
Daniel Malea [Thu, 7 Mar 2013 00:48:53 +0000 (00:48 +0000)]
Fix Linux i386 build
- missing definitions were causing different definitions of type 'off_t', resulting in linker errors
- fix is to define _LARGEFILE_SOURCE and _FILE_OFFSET_BITS=64
llvm-svn: 176595
Han Ming Ong [Wed, 6 Mar 2013 23:59:34 +0000 (23:59 +0000)]
<rdar://problem/
13119170>
Reap the child process (debugserver) when it is done.
llvm-svn: 176594
Howard Hinnant [Wed, 6 Mar 2013 23:30:19 +0000 (23:30 +0000)]
No functionality change at this time. I've split _LIBCPP_VISIBLE up into two flags: _LIBCPP_TYPE_VIS and _LIBCPP_FUNC_VIS. This is in preparation for taking advantage of clang's new __type_visibility__ attribute.
llvm-svn: 176593
Greg Clayton [Wed, 6 Mar 2013 23:23:27 +0000 (23:23 +0000)]
<rdar://problem/
13362109>
LLDB was not parsing line tables correctly for DWARF in .o files after recent debug map changes. This has now been fixed.
llvm-svn: 176592
Jason Molenda [Wed, 6 Mar 2013 23:19:17 +0000 (23:19 +0000)]
Remove some tabs and extraneous space chars from ObjectFileMachO.cpp.
Noticed these while working on the last commit.
llvm-svn: 176590
Jason Molenda [Wed, 6 Mar 2013 23:17:36 +0000 (23:17 +0000)]
Retrieve the dyld shared cache mapping offset from the shared cache instead of hardcoding the value.
Read the version number of the dyld shared cache.
<rdar://problem/
13311882>
llvm-svn: 176589
Adrian Prantl [Wed, 6 Mar 2013 22:03:30 +0000 (22:03 +0000)]
Ensure that DIType is regenerated after we visit an implementation
that adds ivars to an interface.
Fixes rdar://
13175234
This is an update to r176116 that performs a smart caching of interfaces.
llvm-svn: 176584
Shankar Easwaran [Wed, 6 Mar 2013 21:59:27 +0000 (21:59 +0000)]
[ELF] Order weak symbols properly
llvm-svn: 176583
Andrew Kaylor [Wed, 6 Mar 2013 21:37:41 +0000 (21:37 +0000)]
Skipping TestConstVariables on Linux.
The test was marked as expectedFailureLinux, but now it passes with gcc and some versions of clang. Newer versions of clang introduce a failure due to bad DWARF information.
llvm-svn: 176581
Akira Hatanaka [Wed, 6 Mar 2013 21:32:03 +0000 (21:32 +0000)]
[mips] Custom-legalize BR_JT.
In N64-static, GOT address is needed to compute the branch address.
llvm-svn: 176580
Greg Clayton [Wed, 6 Mar 2013 21:17:06 +0000 (21:17 +0000)]
Added tests to test for "settings set" with trailing whitespaces after the values.
llvm-svn: 176579
Andrew Kaylor [Wed, 6 Mar 2013 21:13:09 +0000 (21:13 +0000)]
Adding support for DW_OP_stack_value in DWARFExpression::Evaluate
llvm-svn: 176578
Anna Zaks [Wed, 6 Mar 2013 20:26:02 +0000 (20:26 +0000)]
[analyzer] Pass the correct Expr to the bug reporter visitors when dealing with CompoundLiteralExpr
This allows us to trigger the IDC visitor in the added test case.
llvm-svn: 176577
Anna Zaks [Wed, 6 Mar 2013 20:25:59 +0000 (20:25 +0000)]
[analyzer] IDC: Add config option; perform the idc check on first “null node” rather than last “non-null”.
The second modification does not lead to any visible result, but, theoretically, is what we should
have been looking at to begin with since we are checking if the node was assumed to be null in
an inlined function.
llvm-svn: 176576
Sean Callanan [Wed, 6 Mar 2013 19:57:25 +0000 (19:57 +0000)]
Added a little bit of logging to ClangFunction to
make it more obvious what's going on.
llvm-svn: 176575
Andrew Kaylor [Wed, 6 Mar 2013 19:35:33 +0000 (19:35 +0000)]
Add support for non-register scalar values in DoMaterializeOneVariable.
llvm-svn: 176574
Howard Hinnant [Wed, 6 Mar 2013 19:27:56 +0000 (19:27 +0000)]
Have basic_istream::read call sgetn intead of sbumpc individual characters. This addresses llvm.org/bugs/show_bug.cgi?id=15427.
llvm-svn: 176573
Andrew Trick [Wed, 6 Mar 2013 19:04:56 +0000 (19:04 +0000)]
Generalize my previous fix for -print-options.
Always print options that differ from their implicit default. At least
for simple option types.
llvm-svn: 176572
Michael Liao [Wed, 6 Mar 2013 18:24:34 +0000 (18:24 +0000)]
Remove tailing whitespaces
llvm-svn: 176570
Andrew Trick [Wed, 6 Mar 2013 18:22:22 +0000 (18:22 +0000)]
Give -loop-vectorize an explicit default.
This way, clang -mllvm -print-options shows that the driver is overriding it.
llvm-svn: 176569
Howard Hinnant [Wed, 6 Mar 2013 18:16:12 +0000 (18:16 +0000)]
Correct silly type-o. Thanks Richard.
llvm-svn: 176568
Argyrios Kyrtzidis [Wed, 6 Mar 2013 18:12:50 +0000 (18:12 +0000)]
[PCH] When pre-validating the headers from the PCH, only validate non-system headers.
Stat'ing all the headers from the PCH to make sure they are up-to-date takes significant time.
In a particular source file (whose PCH file included Cocoa.h) from total -fsyntax-only time
12% was just stat calls. Change pre-validation to only check non-system headers.
There are some notable disadvantages:
-If a system header, that is not include-guarded, changes after the PCH was created, we will not
find it in the header info table and we will #import it, effectively #importing it twice, thus
we will emit some error due to a multiple definition and after that the "header was modified" error will likely
be emitted, for example something like:
NSDictionary.h:12:1: error: duplicate interface definition for class 'NSDictionary'
@interface NSDictionary : NSObject <NSCopying, NSMutableCopying, NSSecureCoding, NSFastEnumeration>
^
NSDictionary.h:12:12: note: previous definition is here
@interface NSDictionary : NSObject <NSCopying, NSMutableCopying, NSSecureCoding, NSFastEnumeration>
^
fatal error: file 'NSDictionary.h' has been modified since the precompiled header was built
Though we get the "header was modified" error, this is a bit confusing.
-Theoretically it is possible that such a system header will cause no errors but it will just cause an
unfortunate semantic change, though I find this rather unlikely.
The advantages:
-Reduces compilation time when using a huge PCH like the Cocoa ones
-System headers change very infrequent and when they do, users/build systems should be able to know that
re-building from scratch is needed.
Addresses rdar://
13056262
llvm-svn: 176567
Argyrios Kyrtzidis [Wed, 6 Mar 2013 18:12:47 +0000 (18:12 +0000)]
[PCH] For HeaderFileInfoTrait, hash the key using size & time of the file.
Previously the hash would be the filename portion of the path, which could be
different for a filename with different case or a symbolic link with a different
name completely.
This did not actually create any issue so far because by validating all headers
in the PCH we created uniqued FileEntries based on inodes, so an #include of
a symbolic link (refering to a file from the PCH) would end up with a FileEntry
with filename same as the one recorded in the PCH.
llvm-svn: 176566
Argyrios Kyrtzidis [Wed, 6 Mar 2013 18:12:44 +0000 (18:12 +0000)]
[PCH] We don't need to instantiate HeaderFileInfoTrait in ASTReader more than once.
We can just re-use the one from HeaderFileInfoLookupTable.
llvm-svn: 176565
Argyrios Kyrtzidis [Wed, 6 Mar 2013 18:12:41 +0000 (18:12 +0000)]
[PCH] Remove a couple of fields from serialization::reader::HeaderFileInfoTrait that are not used for anything.
llvm-svn: 176564
Fariborz Jahanian [Wed, 6 Mar 2013 17:55:30 +0000 (17:55 +0000)]
Spell Objective-C correctly.
llvm-svn: 176563
Shuxin Yang [Wed, 6 Mar 2013 17:48:48 +0000 (17:48 +0000)]
Memory Dependence Analysis (not mem-dep test) take advantage of "invariant.load" metadata.
The "invariant.load" metadata indicates the memory unit being accessed is immutable.
A load annotated with this metadata can be moved across any store.
As I am not sure if it is legal to move such loads across barrier/fence, this
change dose not allow such transformation.
rdar://
11311484
Thank Arnold for code review.
llvm-svn: 176562
Fariborz Jahanian [Wed, 6 Mar 2013 17:36:51 +0000 (17:36 +0000)]
say objective-C in the warning and streamline
several diagnostics into one. // rdar://
13094352
llvm-svn: 176560
Howard Hinnant [Wed, 6 Mar 2013 17:30:26 +0000 (17:30 +0000)]
The bitset(unsigned long long) constructor was broken by the constexpr additions only on 32 bit platforms. Fixed. This addresses llvm.org/bugs/show_bug.cgi?id=15444.
llvm-svn: 176559
Matt Kopec [Wed, 6 Mar 2013 17:20:48 +0000 (17:20 +0000)]
Improve/Cleanup ptrace wrapper and remove dependency on user.h
Patch by Ashok Thirumurthi.
llvm-svn: 176558
Dmitri Gribenko [Wed, 6 Mar 2013 17:14:05 +0000 (17:14 +0000)]
Driver: add support for new style multiarch in Ubuntu 13.04
Patch by Sanne Wouda.
llvm-svn: 176557
Edwin Vane [Wed, 6 Mar 2013 17:02:57 +0000 (17:02 +0000)]
New ASTMatchers and enhancement to hasOverloadedOperatorName
Added two new narrowing matchers:
* hasMethod: aplies a matcher to a CXXRecordDecl's methods until a match is made
or there are no more methods.
* hasCanonicalType: applies a matcher to a QualType's canonicalType.
Enhanced hasOverloadedOperatorName to work on CXXMethodDecl as well as
CXXOperatorCallExpr.
Updated tests and docs.
Reviewers: klimek, gribozavr
llvm-svn: 176556
Reid Kleckner [Wed, 6 Mar 2013 16:11:58 +0000 (16:11 +0000)]
[msan] Block reports from interceptors during _Unwind_Backtrace
Summary:
I'm hitting a big recursive report from:
uninit on strlen ->
__msan::PrintWarningWithOrigin() ->
__msan::GetStackTrace() ->
__sanitizer::StackTrace::SlowUnwindStack() ->
_Unwind_Backtrace() ->
... libgcc calls ->
uninit on strlen() ->
... repeats
Reviewers: eugenis
Differential Revision: http://llvm-reviews.chandlerc.com/D497
llvm-svn: 176555
Reid Kleckner [Wed, 6 Mar 2013 15:41:35 +0000 (15:41 +0000)]
[msan] pthread_join() returns int, not void*
pthread_join() returns an error code, not the result of the
pthread_create() function like I thought.
llvm-svn: 176554
Stefanus Du Toit [Wed, 6 Mar 2013 15:32:20 +0000 (15:32 +0000)]
Test case for PR-15414 - nullptr_t functions
The use-null-ptr transform will transform calls to functions that return a
nullptr_t. Even if the function were to only return a null pointer and do
nothing else, this replacement would still be undesired as the behavior and
signature of the function could change in the future.
This adds an XFAILed test case to demonstrate the issue.
Reviewed by: Edwin Vane, Tareq Siraj
llvm-svn: 176553
Reid Kleckner [Wed, 6 Mar 2013 14:54:08 +0000 (14:54 +0000)]
[sanitizers] Fix check failure on dealloc from new thread
Summary:
Adds a test for this case, which was reduced from a chromium build of
WebKit's DumpRenderTree.
Reviewers: eugenis
CC: glider
Differential Revision: http://llvm-reviews.chandlerc.com/D495
llvm-svn: 176552
Edwin Vane [Wed, 6 Mar 2013 14:51:34 +0000 (14:51 +0000)]
Fixed Use-Nullptr when replacing return(0)
Before fix, the paren expression was being replaced resulting in returnnullptr.
ParenExpr and implicit casts now ignored so we get return(nullptr) instead.
Added new test cases.
Fixes PR15398
Author: Ariel Bernal <ariel.j.bernal@intel.com>
llvm-svn: 176551
Alexey Samsonov [Wed, 6 Mar 2013 10:54:18 +0000 (10:54 +0000)]
[Sanitize] Don't emit function attribute sanitize_address/thread/memory if the function is blacklisted.
llvm-svn: 176550
Greg Clayton [Wed, 6 Mar 2013 06:23:54 +0000 (06:23 +0000)]
Fixed enum printing for negative enums. There previously was no testing to validate that enum values were being displayed correctly.
Also added C++11 enum test cases to cover enums as int8_t, int16_t int32_t, int64_t, uint8_t, uint16_t, uint32_t, and uint64_t both for DWARF and dSYM cases. The DWARF being emitted by clang is missing the enum integer type, but the code is now ready to accept and deal with the integral type if it is supplied.
llvm-svn: 176548
Jim Grosbach [Wed, 6 Mar 2013 05:44:53 +0000 (05:44 +0000)]
InstCombine: Don't shrink allocas when combining with a bitcast.
When considering folding a bitcast of an alloca into the alloca itself,
make sure we don't shrink the amount of memory being allocated, or
things rapidly go sideways.
rdar://
13324424
llvm-svn: 176547
Greg Clayton [Wed, 6 Mar 2013 02:34:51 +0000 (02:34 +0000)]
Fixed test suite errors due to new clang -v output.
llvm-svn: 176546
Greg Clayton [Wed, 6 Mar 2013 02:19:38 +0000 (02:19 +0000)]
Now that "settings set" will strip leading and trailing spaces, we need a way to be able to specify string values that contain spaces. So now settings setting <property> <value>" can have a <value> that is quoted:
settings set prompt "(lldb2) "
llvm-svn: 176545
Weiming Zhao [Wed, 6 Mar 2013 02:08:54 +0000 (02:08 +0000)]
revert r176531 due to clan-native-arm fails
llvm-svn: 176544
Akira Hatanaka [Wed, 6 Mar 2013 01:58:03 +0000 (01:58 +0000)]
[mips] Add a line which checks function name. Rename file.
llvm-svn: 176543
Richard Smith [Wed, 6 Mar 2013 01:37:38 +0000 (01:37 +0000)]
PR15390: If a function returns a pointer to a function, that function type
can't have default arguments even though it's a parameter-declaration-clause in
a function declaration.
llvm-svn: 176542
Greg Clayton [Wed, 6 Mar 2013 00:59:41 +0000 (00:59 +0000)]
Disable JITed code in any processes that use DynamicLoaderStatic.
llvm-svn: 176541
Richard Smith [Wed, 6 Mar 2013 00:46:00 +0000 (00:46 +0000)]
After issuing a diagnostic for undefining or redefining a builtin macro,
continue parsing the directive rather than silently discarding it.
Allowing undef or redef of __TIME__ and __DATE__ is important to folks
who want stable, reproducible builds.
llvm-svn: 176540