Commit Graph

119076 Commits

Author SHA1 Message Date
Richard Smith 13f6718b42 PR11594: Don't blindly build a UnaryOperator UO_Minus on an expression which
might not be an rvalue when checking array accesses. Instead, pass through a
flag indicating the array index is negated.

llvm-svn: 146753
2011-12-16 19:31:14 +00:00
Kostya Serebryany 6bb2f1d5b5 Recently the GCD tests started failing because of the invalid size of
FakeStack on the worker threads.
This patch moves the AsanThread initialization into a separate
procedure that's called when AsanThread objects are called for worker
threads.
Patch by glider@google.com

llvm-svn: 146752
2011-12-16 19:13:35 +00:00
Jakob Stoklund Olesen 5af144809e Don't adjust for alignment padding in OffsetIsInRange.
This adjustment is already included in the block offsets computed by
BasicBlockInfo, and adjusting again here can cause the pass to loop.

When CreateNewWater splits a basic block, OffsetIsInRange would reject
the new CPE on the next pass because of the too conservative alignment
adjustment. This caused the block to be split again, and so on.

llvm-svn: 146751
2011-12-16 19:10:00 +00:00
Benjamin Kramer 9ca2e7293b Hexagon: Fix a nasty order-of-initialization bug.
Reenable the tests.

llvm-svn: 146750
2011-12-16 19:08:59 +00:00
Richard Smith f6f003af6a C++11 constexpr: Add note stacks containing backtraces if constant evaluation
fails within a call to a constexpr function. Add -fconstexpr-backtrace-limit
argument to driver and frontend, to control the maximum number of notes so
produced (default 10). Fix APValue printing to be able to pretty-print all
APValue types, and move the testing for this functionality from a unittest to
a -verify test now that it's visible in clang's output.

llvm-svn: 146749
2011-12-16 19:06:07 +00:00
Anna Zaks e48ee50324 [analyzer] Better stdin support.
llvm-svn: 146748
2011-12-16 18:28:50 +00:00
Anna Zaks cf2f8aab62 Fixup to the relaxed diagnostic verification option r146633.
We should exit when matching the '+' even if nothing was found to
prevent searching for all UNIT_MAX items.

llvm-svn: 146747
2011-12-16 18:28:45 +00:00
Greg Clayton b5c39fe9cc Handle all of the "thumb" target triple architecture variants that llvm
handles.

llvm-svn: 146746
2011-12-16 18:15:52 +00:00
Devang Patel fbfd303dc4 Update tests to match r146744.
llvm-svn: 146745
2011-12-16 17:56:38 +00:00
Devang Patel 78847f0bbe In DICompositeType, referenced to derived type is either metadata or null.
llvm-svn: 146744
2011-12-16 17:51:31 +00:00
Devang Patel c16e1ed4bb Clarify and fix subprogram description.
llvm-svn: 146743
2011-12-16 17:50:04 +00:00
Douglas Gregor 7671e53b66 If there is a definition of an ObjCInterfaceDecl, make it the Decl returned from the corresponding ObjCInterfaceType
llvm-svn: 146740
2011-12-16 16:34:57 +00:00
Jakob Stoklund Olesen 2a05f691ab Note ARM constant island alignment in the release notes.
The command line option should be removed, but not until the feature has
gotten a lot of testing. The ARMConstantIslandPass tends to have subtle
bugs that only show up after a while.

llvm-svn: 146739
2011-12-16 16:07:41 +00:00
David Blaikie 1d578782fa Support decltype in pseudo destructors and dependent destructor calls.
Reviewed by Eli Friedman.

llvm-svn: 146738
2011-12-16 16:03:09 +00:00
Fariborz Jahanian 21c2484454 objc: after issuing the warning on direct use of __attribute__((NSObject))
on properties, prevent consequential error  diagnostics. // rdar://10591336

llvm-svn: 146737
2011-12-16 15:54:29 +00:00
Howard Hinnant d6f44b6601 Allow unique_ptr<T const []> to be constructed with a T* (in addition to a const T*)
llvm-svn: 146736
2011-12-16 15:37:23 +00:00
Manuel Klimek 2c899a181c Adds a JSON parser and a benchmark (json-bench) to catch performance regressions.
llvm-svn: 146735
2011-12-16 13:09:10 +00:00
Chandler Carruth fd1f79871a Put the '*' in the right place in the unit test. Forgot to fix up this
bit of style, sorry.

llvm-svn: 146733
2011-12-16 09:37:55 +00:00
Chandler Carruth 2fd51946ee Make GCC happy by using makeAraryRef instead of the implicit conversion.
I have no idea why GCC can't cope with the implicit conversion and Clang
can, or whose bug it is. Grr.

llvm-svn: 146732
2011-12-16 09:36:16 +00:00
Chandler Carruth 1d47523a8e Add a generic collection of class templates to ADT for building
variadic-like functions in C++98. See the comments in the header file
for a more detailed description of how these work. We plan to use these
extensively in the AST matching library. This code and idea were
originally authored by Zhanyong Wan. I've condensed it using macros
to reduce repeatition and adjusted it to fit better with LLVM's ADT.

Thanks to both David Blaikie and Doug Gregor for the review!

llvm-svn: 146729
2011-12-16 08:58:59 +00:00
Chris Lattner 3fdf98c60f By popular demand, link up types by name if they are isomorphic and one is an
autorenamed version of the other.   This makes the IR easier to read, because
we don't end up with random renamed versions of the types after LTO'ing a large app.

llvm-svn: 146728
2011-12-16 08:36:07 +00:00
Tobias Grosser f5c8ae5f4b Adapt to move of isSafeToSpeculativelyExecute into another header.
llvm-svn: 146727
2011-12-16 08:27:42 +00:00
Craig Topper a4d411cb1b Don't try to match 'unpackl/h v, v' for 32xi8 and 16xi16 when only AVX1 is supported. Fix 'unpackh v, v' for 256-bit types to understand 128-bit lanes.
llvm-svn: 146726
2011-12-16 08:06:31 +00:00
NAKAMURA Takumi 1b745d0086 Tweak CMake build on Cygwin.
llvm-svn: 146725
2011-12-16 06:21:08 +00:00
NAKAMURA Takumi 93d990bd61 Target/Hexagon: Fix CMake build.
llvm-svn: 146724
2011-12-16 06:21:02 +00:00
Jason Molenda 9d828ac0aa When we're unwinding out of frame 0 and we end up with a bogus frame
1 -- an address pointing off into non-executable memory -- don't
abort the unwind.  We'll use the ABI's default UnwindPlan to try
to get out of frame 1 and on many platforms with a standard frame
chain stack layout we can get back on track and get a valid frame
2.  This preserves the lldb behavior to-date; the change last week
to require the memory region to be executable broke it.

I'd like to mark this frame specially when displayed to the user;
I tried to override the places where the frame's pc value is returned
to change it to a sentinel value (e.g. LLDB_INVALID_ADDRESS) but
couldn't get that to work cleanly so I backed that part out for
now.  When this happens we'll often miss one of the user's actual
frames, the one that's of most interest to the user, so I'd like
to make this visually distinctive.

Note that a frame in non-executable memory region is only allowed
for frame 1.  After that we should be solid on the unwind and any
pc address in non-executable memory indicates a failure and we
should stop unwinding.

llvm-svn: 146723
2011-12-16 04:30:31 +00:00
Douglas Gregor ab1ec82e04 Fix chaining of ObjCInterfaceDecl redeclarations
llvm-svn: 146722
2011-12-16 03:12:41 +00:00
Andrew Trick ca3417e932 Avoid a confusing assert for silly options: -unroll-runtime -unroll-count=1.
No need for an explicit test case for an unsupported combination of options.

llvm-svn: 146721
2011-12-16 02:03:48 +00:00
Johnny Chen 7cc3d31e15 Simplify the setup leading to the testing of ReadMemory(), ReadCStringFromMemory(), and ReadUnsignedFromMemory().
Instead of getting the location of the variable and converting the hex string to an int, just use
val.AddressOf().GetValueAsUnsigned() to compute the address of the memory region to read from.

llvm-svn: 146719
2011-12-16 01:56:27 +00:00
Kostya Serebryany 561dade58d [asan] add a test for instrumenting globals
llvm-svn: 146718
2011-12-16 01:28:19 +00:00
Jim Ingham 30b74fa886 Remove unnecessary #include.
llvm-svn: 146717
2011-12-16 00:46:12 +00:00
Johnny Chen e7e8af807a Add a test sequence of SBProcess.ReadCStringFromMemory() with (char *)my_char_ptr as the address to read from.
char *my_char_ptr = (char *)"Does it work?";

llvm-svn: 146716
2011-12-16 00:25:30 +00:00
Sean Callanan caa69f0570 Sema::RequireCompleteType currently attempts to
instantiate a class from its template pattern
before it consults the ExternalASTSource.  LLDB
in particular will sometimes provide patterns
that need to be completed first.

To make this possible, I have moved the
completion before the code that does the
instantiation, allowing the ExternalASTSource
to provide the required information.

llvm-svn: 146715
2011-12-16 00:20:31 +00:00
Jim Grosbach 4a29971f02 ARM NEON aliases for vmovq.f*
llvm-svn: 146714
2011-12-16 00:12:22 +00:00
Jakob Stoklund Olesen 6a5f0a1906 Extract a method. No functional change.
llvm-svn: 146713
2011-12-16 00:12:05 +00:00
Jim Ingham 918533bcfe Fix a bug where when debugging with .o files, we end up with two symbols for each real OBJC_CLASS_$_whatever, one of which is correctly classified as an ObjCClass symbol, and the other is just a data symbol. This was messing up the ObjC dynamic type detection.
<rdar://problem/10589527>

llvm-svn: 146712
2011-12-16 00:05:58 +00:00
Daniel Dunbar 30a8976127 llvm-config: Fix --targets-built, I changed this to use the registry but wasn't
properly initializing the target infos. I decided it wasn't worth linking them
in for this, so just switched back to using the Makefile variable for now. We
can reconsider later if we ever get pluggable targets.

llvm-svn: 146711
2011-12-16 00:04:43 +00:00
Jim Grosbach 66886253a7 Thumb2 ADR assembly parsing w/o the .w suffix.
llvm-svn: 146710
2011-12-15 23:52:17 +00:00
Eli Friedman 64944090ff Make sure we correctly note the existence of an i8 immediate for vblendvps and friends, so we compute fixups correctly. PR11586.
llvm-svn: 146709
2011-12-15 23:46:18 +00:00
Daniel Dunbar c7037f3f81 llvm-config: Update help text for removal of "backend" pseudo component.
llvm-svn: 146708
2011-12-15 23:43:17 +00:00
Argyrios Kyrtzidis af41b3f0f2 In SourceManager::~SourceManager do a sanity check to make sure we
don't try to destruct a null ContentCache.

rdar://10567159

llvm-svn: 146707
2011-12-15 23:37:55 +00:00
Daniel Dunbar 06bb798803 build/unittests: Fix llvm-config names for gtest libraries, and bring Makefile
library names in line with those used by CMake.
 - Patch by Johannes Obermayr, with tweaks by me.

llvm-svn: 146706
2011-12-15 23:35:08 +00:00
Douglas Gregor c8b0c9d8fd When generating debug info for an ObjCInterfaceDecl, try to dig out the definition.
llvm-svn: 146705
2011-12-15 23:32:29 +00:00
Johnny Chen 6e55cd5e29 Add test scenario for newly added SBProcess APIs: ReadCStringFromMemory() and ReadUnsignedFromMemory().
llvm-svn: 146704
2011-12-15 23:30:05 +00:00
Kostya Serebryany 6d37656c1a [asan] revert r146529: we do need to build asan-rt as a shared library. Will need to find a different way to modify preinit_array
llvm-svn: 146703
2011-12-15 23:08:00 +00:00
Nick Lewycky c9e935c7e2 Move parts of lib/Target that use CodeGen into lib/CodeGen.
llvm-svn: 146702
2011-12-15 22:58:58 +00:00
Kostya Serebryany 7ce9686921 [asan] added a test to check that a global-buffer-overflow mentioned the module name; revert incorrect change in the test runner
llvm-svn: 146701
2011-12-15 22:57:32 +00:00
Eli Friedman c9bf1b1bff Make check a bit more strict so we don't call ARM_AM::getFP32Imm with a value that isn't a 32-bit value. (This is just to be safe; I don't think this actually causes any issues in practice.)
llvm-svn: 146700
2011-12-15 22:56:53 +00:00
Jim Grosbach a47294e24d ARM NEON VCLE is an alias for VCGE w/ the source operands reversed.
llvm-svn: 146699
2011-12-15 22:56:33 +00:00
Kostya Serebryany 7a9eb49a47 [asan] add the name of the module to the description of a global variable. This improves the readability of global-buffer-overflow reports.
llvm-svn: 146698
2011-12-15 22:55:55 +00:00