Sean Callanan [Tue, 19 Mar 2013 01:45:02 +0000 (01:45 +0000)]
Fixed handling of function pointers in the IR
interpreter. They now have correct values, even
when the process is not running.
llvm-svn: 177372
Nick Lewycky [Tue, 19 Mar 2013 01:37:55 +0000 (01:37 +0000)]
Emit the linkage name instead of the function name, when available. This means
that we'll prefer to emit the mangled C++ name (pending a clang change).
llvm-svn: 177371
Douglas Gregor [Tue, 19 Mar 2013 00:41:04 +0000 (00:41 +0000)]
Remove stray semicolon
llvm-svn: 177370
Douglas Gregor [Tue, 19 Mar 2013 00:38:50 +0000 (00:38 +0000)]
Minor optimization to r177367 to treat a module with missing dependencies as out-of-date rather than missing.
llvm-svn: 177369
Rafael Espindola [Tue, 19 Mar 2013 00:36:57 +0000 (00:36 +0000)]
Finish refactoring the tool selection logic.
The general pattern now is that Foobar::constructTool only creates tools
defined in the tools::foobar namespace and then delegates to the parent.
The remaining duplicated code is now in the tools themselves.
llvm-svn: 177368
Douglas Gregor [Tue, 19 Mar 2013 00:28:20 +0000 (00:28 +0000)]
<rdar://problem/
13363214> Eliminate race condition between module rebuild and the global module index.
The global module index was querying the file manager for each of the
module files it knows about at load time, to prune out any out-of-date
information. The file manager would then cache the results of the
stat() falls used to find that module file.
Later, the same translation unit could end up trying to import one of the
module files that had previously been ignored by the module cache, but
after some other Clang instance rebuilt the module file to bring it
up-to-date. The stale stat() results in the file manager would
trigger a second rebuild of the already-up-to-date module, causing
failures down the line.
The global module index now lazily resolves its module file references
to actual AST reader module files only after the module file has been
loaded, eliminating the stat-caching race. Moreover, the AST reader
can communicate to its caller that a module file is missing (rather
than simply being out-of-date), allowing us to simplify the
module-loading logic and allowing the compiler to recover if a
dependent module file ends up getting deleted.
llvm-svn: 177367
Enrico Granata [Tue, 19 Mar 2013 00:27:22 +0000 (00:27 +0000)]
This checkin removes the last Cocoa formatters that were implemented in Python and reimplements them in C++. The Python Cocoa formatters are not shipped as part of LLDB anymore, but still exist in the source repository for user reference. Python formatters still exist for STL classes and users can still define their own Python formatters
llvm-svn: 177366
Greg Clayton [Tue, 19 Mar 2013 00:20:55 +0000 (00:20 +0000)]
<rdar://problem/
13443931>
Fixed a crasher in the SourceManager where it wasn't checking the m_target member variable for NULL.
In doing this fix, I hardened this class to have weak pointers to the debugger and target in case they do go away. I also changed SBSourceManager to hold onto weak pointers to the debugger and target so they don't keep objects alive by holding a strong reference to them.
llvm-svn: 177365
Sean Callanan [Tue, 19 Mar 2013 00:10:07 +0000 (00:10 +0000)]
Refactored the expression parser so that the IR
and the JITted code are managed by a standalone
class that handles memory management itself.
I have removed RecordingMemoryManager and
ProcessDataAllocator, which filled similar roles
and had confusing ownership, with a common class
called IRExecutionUnit. The IRExecutionUnit
manages all allocations ever made for an expression
and frees them when it goes away. It also contains
the code generator and can vend the Module for an
expression to other clases.
The end goal here is to make the output of the
expression parser re-usable; that is, to avoid
re-parsing when re-parsing isn't necessary.
I've also cleaned up some code and used weak pointers
in more places. Please let me know if you see any
leaks; I checked myself as well but I might have
missed a case.
llvm-svn: 177364
Richard Smith [Tue, 19 Mar 2013 00:04:20 +0000 (00:04 +0000)]
Update ubsan test for clang change r177362.
llvm-svn: 177363
Richard Smith [Tue, 19 Mar 2013 00:01:12 +0000 (00:01 +0000)]
PR15383: When -fsanitize=float-cast-overflow checks a float-to-int conversion,
it wasn't taking into account that the float should be truncated *before* the
range check happens. Thus (unsigned)-0.99 and (unsigned char)255.9 have defined
behavior and should not be trapped.
llvm-svn: 177362
Hal Finkel [Mon, 18 Mar 2013 23:58:28 +0000 (23:58 +0000)]
Fix a sign-extension bug in PPCCTRLoops
Don't sign extend the immediate value from the OR instruction in
an LIS/OR pair.
llvm-svn: 177361
Rafael Espindola [Mon, 18 Mar 2013 23:56:07 +0000 (23:56 +0000)]
Centralize the logic for using the integrated assembler.
llvm-svn: 177360
Argyrios Kyrtzidis [Mon, 18 Mar 2013 23:54:50 +0000 (23:54 +0000)]
[libclang] Modify clang_getCursorType to be able to handle a function template decl.
llvm-svn: 177359
Fariborz Jahanian [Mon, 18 Mar 2013 23:45:52 +0000 (23:45 +0000)]
documentation comment parsing. Added couple of
top-level HeaderDoc tags @functiongroup and
@methodgroup to doc. tags recognized.
// rdar://
12379114
llvm-svn: 177358
Jakub Staszak [Mon, 18 Mar 2013 23:45:45 +0000 (23:45 +0000)]
Move #include of BitVector from .h to .cpp file.
Also remove unneeded #include and forward declaration.
llvm-svn: 177357
Jakub Staszak [Mon, 18 Mar 2013 23:40:46 +0000 (23:40 +0000)]
Add some constantness.
llvm-svn: 177356
Marshall Clow [Mon, 18 Mar 2013 23:39:36 +0000 (23:39 +0000)]
Fix bug in test; found by AddressSanitizer
llvm-svn: 177355
Richard Smith [Mon, 18 Mar 2013 23:37:25 +0000 (23:37 +0000)]
Diagnose uses of 'alignof' on functions in -pedantic mode.
llvm-svn: 177354
Jordan Rose [Mon, 18 Mar 2013 23:34:37 +0000 (23:34 +0000)]
[analyzer] Do part of the work to find shortest bug paths up front.
Splitting the graph trimming and the path-finding (r177216) already
recovered quite a bit of performance lost to increased suppression.
We can still do better by also performing the reverse BFS up front
(needed for shortest-path-finding) and only walking the shortest path
for each report. This does mean we have to walk back up the path and
invalidate all the BFS numbers if the report turns out to be invalid,
but it's probably still faster than redoing the full BFS every time.
More performance work for <rdar://problem/
13433687>
llvm-svn: 177353
Jordan Rose [Mon, 18 Mar 2013 23:34:32 +0000 (23:34 +0000)]
[analyzer] Replace uses of assume() with isNull() in BR visitors.
Also, replace a std::string with a SmallString.
No functionality change.
llvm-svn: 177352
Jakub Staszak [Mon, 18 Mar 2013 23:33:44 +0000 (23:33 +0000)]
Remove unneeded #includes.
llvm-svn: 177351
Jakub Staszak [Mon, 18 Mar 2013 23:33:14 +0000 (23:33 +0000)]
Make methods const.
llvm-svn: 177350
Chad Rosier [Mon, 18 Mar 2013 23:32:33 +0000 (23:32 +0000)]
Test case for r177347.
llvm-svn: 177349
Jakub Staszak [Mon, 18 Mar 2013 23:31:30 +0000 (23:31 +0000)]
Make method private. Keep coding standard.
llvm-svn: 177348
Chad Rosier [Mon, 18 Mar 2013 23:31:24 +0000 (23:31 +0000)]
[ms-inline asm] Avoid emitting a redundant sizing directive, if we've already
parsed one. Test case coming shortly.
rdar://
13446980
llvm-svn: 177347
Greg Clayton [Mon, 18 Mar 2013 23:25:00 +0000 (23:25 +0000)]
Fixed the README to match the current code.
llvm-svn: 177346
David Blaikie [Mon, 18 Mar 2013 23:22:28 +0000 (23:22 +0000)]
PR15539: Record "evaluating if/elif condition" flag in the right place
The previous implementation missed the case where the elif condition was
evaluated from the context of an #ifdef that was false causing PR15539.
llvm-svn: 177345
Bill Wendling [Mon, 18 Mar 2013 23:11:54 +0000 (23:11 +0000)]
Update testcase for r177340.
llvm-svn: 177344
Jim Ingham [Mon, 18 Mar 2013 23:08:30 +0000 (23:08 +0000)]
Add a comment about how to add a target for a test case.
llvm-svn: 177343
Jakub Staszak [Mon, 18 Mar 2013 23:08:01 +0000 (23:08 +0000)]
Change NULL to 0.
llvm-svn: 177342
Jim Ingham [Mon, 18 Mar 2013 23:05:00 +0000 (23:05 +0000)]
Move the performance test cases into their own project.
llvm-svn: 177341
Bill Wendling [Mon, 18 Mar 2013 23:04:39 +0000 (23:04 +0000)]
Register the flush function for each compile unit.
For each compile unit, we want to register a function that will flush that
compile unit. Otherwise, __gcov_flush() would only flush the counters within the
current compile unit, and not any outside of it.
PR15191 & <rdar://problem/
13167507>
llvm-svn: 177340
Jakub Staszak [Mon, 18 Mar 2013 23:04:30 +0000 (23:04 +0000)]
Remove trailing spaces.
llvm-svn: 177339
Hal Finkel [Mon, 18 Mar 2013 23:00:58 +0000 (23:00 +0000)]
Fix PPC unaligned 64-bit loads and stores
PPC64 supports unaligned loads and stores of 64-bit values, but
in order to use the r+i forms, the offset must be a multiple of 4.
Unfortunately, this cannot always be determined by examining the
immediate itself because it might be available only via a TOC entry.
In order to get around this issue, we additionally predicate the
selection of the r+i form on the alignment of the load or store
(forcing it to be at least 4 in order to select the r+i form).
llvm-svn: 177338
Bill Wendling [Mon, 18 Mar 2013 22:59:47 +0000 (22:59 +0000)]
Add some GCOV functions that register all of the __llvm_gcov_flush() functions.
The __llvm_gcov_flush() functions only work for the local compile unit. However,
when __gcov_flush() is called, the user expects all of the counters to be
flushed, not just the ones in the current compile unit.
This adds some library functions that register the flush functions. It also
defined __gcov_flush() so that loops through that list and calls the functions.
PR15191 & <rdar://problem/
13167507>
llvm-svn: 177337
Argyrios Kyrtzidis [Mon, 18 Mar 2013 22:55:24 +0000 (22:55 +0000)]
[frontend] Initialize the diagnostic client before loading an ast file.
Issue reported by Tom Honermann!
http://llvm.org/bugs/show_bug.cgi?id=15377
llvm-svn: 177336
Richard Smith [Mon, 18 Mar 2013 22:52:47 +0000 (22:52 +0000)]
Add missing diagnostic for a nested-name-specifier on a free-standing type definition. Bump some related diagnostics from warning to extension in C++, since they're errors there. Add some missing checks for function specifiers on non-function declarations.
llvm-svn: 177335
Arnold Schwaighofer [Mon, 18 Mar 2013 22:47:09 +0000 (22:47 +0000)]
ARM cost model: Make some vector integer to float casts cheaper
The default logic marks them as too expensive.
For example, before this patch we estimated:
cost of 16 for instruction: %r = uitofp <4 x i16> %v0 to <4 x float>
While this translates to:
vmovl.u16 q8, d16
vcvt.f32.u32 q8, q8
All other costs are left to the values assigned by the fallback logic. Theses
costs are mostly reasonable in the sense that they get progressively more
expensive as the instruction sequences emitted get longer.
radar://
13445992
llvm-svn: 177334
Arnold Schwaighofer [Mon, 18 Mar 2013 22:47:06 +0000 (22:47 +0000)]
ARM cost model: Correct cost for some cheap float to integer conversions
Fix cost of some "cheap" cast instructions. Before this patch we used to
estimate for example:
cost of 16 for instruction: %r = fptoui <4 x float> %v0 to <4 x i16>
While we would emit:
vcvt.s32.f32 q8, q8
vmovn.i32 d16, q8
vuzp.8 d16, d17
All other costs are left to the values assigned by the fallback logic. Theses
costs are mostly reasonable in the sense that they get progressively more
expensive as the instruction sequences emitted get longer.
radar://
13434072
llvm-svn: 177333
Greg Clayton [Mon, 18 Mar 2013 22:34:00 +0000 (22:34 +0000)]
Code cleanup:
- don't use preprocessor macros
- use switch statements
- don't put anything in the lldb namespace, use "lldb_perf" namespace.
- Pass the action struct into each TestStep() for each step fill in
- Modify the ActionWanted class to have accessors to make the continue, next, finish, kill instead of using preproc macros
llvm-svn: 177332
Quentin Colombet [Mon, 18 Mar 2013 22:30:07 +0000 (22:30 +0000)]
Extend global merge pass to optionally consider global constant variables.
Also add some checks to not merge globals used within landing pad instructions or marked as "used".
llvm-svn: 177331
Argyrios Kyrtzidis [Mon, 18 Mar 2013 22:23:49 +0000 (22:23 +0000)]
[PCH] Fix assertion hit related to enum decls inside templated funtions.
Report and suggested fix by Tom Honermann!
http://llvm.org/bugs/show_bug.cgi?id=13020
llvm-svn: 177330
Reed Kotler [Mon, 18 Mar 2013 22:18:00 +0000 (22:18 +0000)]
This code works around what appears to be a bug in another part of clang.
I have filed http://llvm.org/bugs/show_bug.cgi?id=15538 against clang.
This code is safer anyway because "cast" assumes you really know that
it's okay to make the cast. In this case isa should not be false and
dyn_cast should not return null as far as I understand. But everything
else is valid so I did not want to revert my previous patch for attributes
mips16/nomips16 or use an llvm_unreachable here which would make a number
of our tests fail for mips.
llvm-svn: 177329
Bill Schmidt [Mon, 18 Mar 2013 22:12:04 +0000 (22:12 +0000)]
Change test cases to handle unaligned references.
Hal Finkel recently added code to allow unaligned memory references
for PowerPC. Two tests were temporarily modified with
-disable-ppc-unaligned to keep them from failing. This patch adjusts
the expected code generation for the unaligned references.
llvm-svn: 177328
David Blaikie [Mon, 18 Mar 2013 22:08:16 +0000 (22:08 +0000)]
Remove unnecessary leading comment characters in lit-only file
llvm-svn: 177327
Jakob Stoklund Olesen [Mon, 18 Mar 2013 22:01:35 +0000 (22:01 +0000)]
Add SchedRW annotations to most of X86InstrSSE.td.
We hitch a ride with the existing OpndItins class that was used to add
instruction itinerary classes in the many multiclasses in this file.
Use the link provided by the X86FoldableSchedWrite.Folded to find the
right SchedWrite for folded loads.
llvm-svn: 177326
Richard Smith [Mon, 18 Mar 2013 21:57:52 +0000 (21:57 +0000)]
Add a linebreak. Linebreaks are cool.
llvm-svn: 177325
Richard Smith [Mon, 18 Mar 2013 21:44:56 +0000 (21:44 +0000)]
C++11 status: mark 'extended integral types' as N/A, since we don't support
any, and add a note that we cannot treat __int128 as an extended integral type.
llvm-svn: 177324
Jakob Stoklund Olesen [Mon, 18 Mar 2013 21:32:39 +0000 (21:32 +0000)]
Annotate X86 arithmetic instructions with SchedRW lists.
This new-style scheduling information is going to replace the
instruction iteneraries.
This also serves as a test case for Andy's fix in r177317.
llvm-svn: 177323
Manman Ren [Mon, 18 Mar 2013 21:23:25 +0000 (21:23 +0000)]
Check whether a pointer is non-null (isKnownNonNull) in isKnownNonZero.
This handles the case where we have an inbounds GEP with alloca as the pointer.
This fixes the regression in PR12750 and rdar://
13286434.
Note that we can also fix this by handling some GEP cases in isKnownNonNull.
llvm-svn: 177321
Richard Smith [Mon, 18 Mar 2013 21:12:30 +0000 (21:12 +0000)]
Bring inheriting constructor implementation up-to-date with current defect
reports, and implement implicit definition of inheriting constructors.
Remaining missing features: inheriting constructor templates, implicit
exception specifications for inheriting constructors, inheriting constructors
from dependent bases.
llvm-svn: 177320
Rafael Espindola [Mon, 18 Mar 2013 20:48:54 +0000 (20:48 +0000)]
Centralize the recording of which tools have been constructed.
llvm-svn: 177319
Anna Zaks [Mon, 18 Mar 2013 20:46:56 +0000 (20:46 +0000)]
[analyzer] Warn when a ‘nil’ object is added to NSArray or NSMutableArray.
llvm-svn: 177318
Andrew Trick [Mon, 18 Mar 2013 20:42:25 +0000 (20:42 +0000)]
TableGen fix for the new machine model.
Properly handle cases where a group of instructions have different
SchedRW lists with the same itinerary class.
This was supposed to work, but I left in an early break.
llvm-svn: 177317
David Blaikie [Mon, 18 Mar 2013 20:31:44 +0000 (20:31 +0000)]
Include '.test' suffix in target specific lit configs that need it
Apparently my final cleanup to use a relevant suffix for these tests before
committing r176831 caused them to stop running since lit wasn't configured to
run tests with that suffix in those directories (why don't we just have a
global suffix list?). So, add the suffix to the relevant directories & fix the
test that has bitrotted over the last week due to my debug info schema changes.
llvm-svn: 177315
Eric Christopher [Mon, 18 Mar 2013 20:21:47 +0000 (20:21 +0000)]
Make the fields in the diagram match the descriptive text above them.
llvm-svn: 177314
Sebastian Pop [Mon, 18 Mar 2013 20:21:13 +0000 (20:21 +0000)]
Correct function to decide if a SCEV can be ignored
When doing SCEV based code generation, we ignore instructions calculating values
that are fully defined by a SCEV expression. The values that are calculated by
this instructions are recalculated on demand.
This commit improves the check to verify if certain instructions can be ignored
and recalculated on demand.
llvm-svn: 177313
Dmitry Vyukov [Mon, 18 Mar 2013 19:47:36 +0000 (19:47 +0000)]
tsan: smaller memory block headers (32b->16b)
llvm-svn: 177312
Fariborz Jahanian [Mon, 18 Mar 2013 19:41:18 +0000 (19:41 +0000)]
Objective-C modern translator. Don't put line info
into the pre-preprocessed file to be passed to
modern translator when compiling in no debug mode.
// rdar://
13138170
llvm-svn: 177311
Marshall Clow [Mon, 18 Mar 2013 19:34:07 +0000 (19:34 +0000)]
Removed raw references to __sun__, __FreeBSD__, __GLIBC__ and __linux__; now just check to see if they are defined.
llvm-svn: 177310
Tobias Grosser [Mon, 18 Mar 2013 19:19:57 +0000 (19:19 +0000)]
ScopHelper: clang-format
llvm-svn: 177309
Tobias Grosser [Mon, 18 Mar 2013 19:17:07 +0000 (19:17 +0000)]
ScopHelper: Remove some dead code
llvm-svn: 177307
Sebastian Pop [Mon, 18 Mar 2013 19:09:49 +0000 (19:09 +0000)]
use the canonical IV only when it exists
llvm-svn: 177306
Rafael Espindola [Mon, 18 Mar 2013 18:50:01 +0000 (18:50 +0000)]
Use early returns when checking if we already constructed a tool and when
delegating to Generic_GCC::SelectTool (it already updates the tool map).
llvm-svn: 177305
Marshall Clow [Mon, 18 Mar 2013 18:20:48 +0000 (18:20 +0000)]
Removed raw references to _MSC_VER; now just check to see if it is defined.
llvm-svn: 177304
Rafael Espindola [Mon, 18 Mar 2013 18:19:46 +0000 (18:19 +0000)]
Remove unused argument.
llvm-svn: 177303
Rafael Espindola [Mon, 18 Mar 2013 18:10:27 +0000 (18:10 +0000)]
Pass an ArgList to every toolchain constructor. Remove the useIntegratedAs
argument.
llvm-svn: 177301
Rafael Espindola [Mon, 18 Mar 2013 17:58:22 +0000 (17:58 +0000)]
Inline ShouldUseIntegratedAssembler and move the documentation to
useIntegratedAs.
llvm-svn: 177300
Rafael Espindola [Mon, 18 Mar 2013 17:52:57 +0000 (17:52 +0000)]
Refactor a bit of duplicated code to useIntegratedAs.
llvm-svn: 177299
Bill Wendling [Mon, 18 Mar 2013 17:47:33 +0000 (17:47 +0000)]
Update
llvm-svn: 177298
Marshall Clow [Mon, 18 Mar 2013 17:45:34 +0000 (17:45 +0000)]
Removed raw references to __APPLE__; now just check to see if it is defined.
llvm-svn: 177297
Hal Finkel [Mon, 18 Mar 2013 17:40:46 +0000 (17:40 +0000)]
Fix 80-col. violations in PPCCTRLoops
llvm-svn: 177296
Hal Finkel [Mon, 18 Mar 2013 17:40:44 +0000 (17:40 +0000)]
Fix large count and negative constant count handling in PPCCTRLoops
This commit fixes an assert that would occur on loops with large constant counts
(like looping for ((uint32_t) -1) iterations on PPC64). The existing code did
not handle counts that it computed to be negative (asserting instead), but
these can be created with valid inputs.
This bug was discovered by bugpoint while I was attempting to isolate a
completely different problem.
Also, in writing test cases for the negative-count problem, I discovered that
the ori/lsi handling was broken (there was a typo which caused the logic that
was supposed to detect these pairs and extract the iteration count to always
fail). This has now also been corrected (and is covered by one of the new test
cases).
llvm-svn: 177295
Hal Finkel [Mon, 18 Mar 2013 17:40:27 +0000 (17:40 +0000)]
Cleanup initial-value constants in PPCCTRLoops
Because the initial-value constants had not been added to the list
of instructions considered for DCE the resulting code had redundant
constant-materialization instructions.
llvm-svn: 177294
Rafael Espindola [Mon, 18 Mar 2013 17:25:58 +0000 (17:25 +0000)]
Remove unused argument.
llvm-svn: 177293
Dmitry Vyukov [Mon, 18 Mar 2013 17:21:15 +0000 (17:21 +0000)]
tsan: add support for idle threads
llvm-svn: 177292
Marshall Clow [Mon, 18 Mar 2013 17:04:29 +0000 (17:04 +0000)]
Removed raw references to _WIN32; now just check to see if it is defined.
llvm-svn: 177291
Hans Wennborg [Mon, 18 Mar 2013 17:03:05 +0000 (17:03 +0000)]
Fix integer comparison in DIEInteger::BestForm.
The always-true "(int)Int == (signed)Int" comparison was found
while experimenting with a potential new Clang warning.
llvm-svn: 177290
Dmitry Vyukov [Mon, 18 Mar 2013 16:56:48 +0000 (16:56 +0000)]
tsan: touch less shadow memory during operations on big memory ranges
greatly reduces memory consumption
llvm-svn: 177289
Dmitry Vyukov [Mon, 18 Mar 2013 15:49:07 +0000 (15:49 +0000)]
tsan: mark shadow for thread stack as "don't need" when thread exits
llvm-svn: 177288
Rafael Espindola [Mon, 18 Mar 2013 15:33:26 +0000 (15:33 +0000)]
Remove unused argument.
llvm-svn: 177287
Dmitry Vyukov [Mon, 18 Mar 2013 13:55:33 +0000 (13:55 +0000)]
tsan: better memory profiler
llvm-svn: 177286
Matthew Curtis [Mon, 18 Mar 2013 13:08:24 +0000 (13:08 +0000)]
ReleaseNotes: Tweak hexagonv2/hexagonv3 removal note.
llvm-svn: 177284
Daniel Jasper [Mon, 18 Mar 2013 12:50:26 +0000 (12:50 +0000)]
Fix clang-format segfault.
When annotating "lines" starting with ":", clang-format would segfault.
This could actually happen in valid code, e.g.
#define A :
llvm-svn: 177283
Benjamin Kramer [Mon, 18 Mar 2013 12:41:25 +0000 (12:41 +0000)]
ObjCDictionaryElements are pod-like.
llvm-svn: 177282
Benjamin Kramer [Mon, 18 Mar 2013 12:14:30 +0000 (12:14 +0000)]
Remove default copy ctor/assignment, makes AttributeSet trivially copyable.
And enables SmallVector's pod optimizations.
llvm-svn: 177281
David Tweed [Mon, 18 Mar 2013 12:07:24 +0000 (12:07 +0000)]
Initially forgotten-to-svn-add test case for r177279.
llvm-svn: 177280
David Tweed [Mon, 18 Mar 2013 11:54:44 +0000 (11:54 +0000)]
The optimization a + (-0.0f) -> a was being misapplied to a + (+0.0f) in the vector case (because
we weren't differntiating floating-point zeroinitializers from other zero-initializers)
which was causing problems for code relying upon a + (+0.0f) to, eg, flush denormals to
0. Make the scalar and vector cases have the same behaviour.
llvm-svn: 177279
Dmitry Vyukov [Mon, 18 Mar 2013 11:51:51 +0000 (11:51 +0000)]
tsan: madvise(DONTNEED) shadow memory on munmap()
llvm-svn: 177278
Christian Konig [Mon, 18 Mar 2013 11:34:16 +0000 (11:34 +0000)]
R600/SI: implement indirect adressing for SI
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177277
Christian Konig [Mon, 18 Mar 2013 11:34:10 +0000 (11:34 +0000)]
R600/SI: add float vector types
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177276
Christian Konig [Mon, 18 Mar 2013 11:34:05 +0000 (11:34 +0000)]
R600/SI: add shl pattern
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177275
Christian Konig [Mon, 18 Mar 2013 11:34:00 +0000 (11:34 +0000)]
R600/SI: add BUFFER_LOAD_DWORD pattern
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177274
Christian Konig [Mon, 18 Mar 2013 11:33:55 +0000 (11:33 +0000)]
R600/SI: implement SI.load.const intrinsic
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177273
Christian Konig [Mon, 18 Mar 2013 11:33:50 +0000 (11:33 +0000)]
R600/SI: enable all S_LOAD and S_BUFFER_LOAD opcodes
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177272
Christian Konig [Mon, 18 Mar 2013 11:33:45 +0000 (11:33 +0000)]
R600/SI: fix inserting waits for all defines
Unfortunately the previous fix for inserting waits for unordered
defines wasn't sufficient, cause it's possible that even ordered
defines are only partially used (or not used at all).
Signed-off-by: Christian König <christian.koenig@amd.com>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
llvm-svn: 177271
Kostya Serebryany [Mon, 18 Mar 2013 10:52:25 +0000 (10:52 +0000)]
[asan] make the __asan_get_heap_size more robust
llvm-svn: 177270
Dmitry Vyukov [Mon, 18 Mar 2013 10:32:21 +0000 (10:32 +0000)]
tsan: move implementation out of h file
llvm-svn: 177269
Dmitry Vyukov [Mon, 18 Mar 2013 10:10:15 +0000 (10:10 +0000)]
tsan: fix clang -Wall build
Clang does not like classes with virtual functions but w/o virtual dtor.
Go does not like libstdc++ (operator delete).
llvm-svn: 177267