When lowering build_vector to an insertps, we would still lower it, even
if the source vectors weren't v4x32. This would break on avx if the source
was a v8x32. We now check the type of the source vectors.
llvm-svn: 208487
Add ARM support for dllexport and dllimport attributes. This is a relatively
conservative change. The alternative is to entirely drop the architecture
requirement. The dllimport and dllexport attributes are not restricted to any
architecture, simply to platforms that support this attribute (currently
Windows).
llvm-svn: 208486
The loop body used to contain a switch statement; it looks like r96685 replaced
that with an if/else if/else but accidentally left one of the three break
statements from the switch behind, skipping the clang_disposeString() call
for replacements (and the rest of the loop too, which apparently doesn't make
a differences for the test cases we have).
r96685: http://lists.cs.uiuc.edu/pipermail/cfe-commits/Week-of-Mon-20100215/027754.html
This too might possibly the last leak in clang (PR19521).
llvm-svn: 208483
Fixes PR19689, and possibly the last leak in clang (PR19521). PR19689 contains
a few suggestions on how to make the default argument token code a bit more
solid, these changes can possibly be landed in follow-ups.
llvm-svn: 208481
We were swapping the true & false results while testing for FMAX/FMIN,
but not putting them back to the original state if the later checks
failed.
Should fix PR19700.
llvm-svn: 208469
Also tidy up, simplify, and extend the test coverage to demonstrate the
limitations. This test should now fail if the bugs are fixed (&
hopefully whoever ends up in this situation sees the FIXMEs and realizes
that the test needs to be updated to positively test their change that
has fixed some or all of these issues).
I do wonder whether I could demonstrate breakage without a macro here,
but any way I slice it I can't think of a way to get two calls to the
same function on the same line/column in non-macro C++ - implicit
conversions happen at the same location as an explicit function, but
you'd never get an implicit conversion on the result of an explicit call
to the same implicit conversion operator (since the value is already
converted to the desired result)...
llvm-svn: 208468
To allow the compilation with gcc, we can't use the
__has_feature(cxx_noexcept) to detect the default destructor
exception specification, which is noexcept(true) by default
in C++11. We should use __cplusplus >= 201103L instead.
llvm-svn: 208465
The implementation might be better to have a method is64Bit() in the class
SymbolicFile instead of having the static routine isSymbolList64Bit() in
llvm-nm.cpp . But this is very much in the sprit of isObject() and
getNMTypeChar() in llvm-nm.cpp that has a series of if else statements
based on the specific class of the SymbolicFile. I can update this if
folks would like.
Also the tests were updated to be explicit about checking the address for
64-bits or 32-bits from object files.
llvm-svn: 208463
There is no total ordering if the CFG is disconnected. We don't care if we
catch all CSE opportunities in dead code either so just exclude ignore them in
the assert.
PR19646
llvm-svn: 208461
Check that the profile runtime works as expected. This tests the
functions that are meant to be available to advanced users.
In particular, check that the `atexit()` hook can be disabled by
defining a custom `__llvm_profile_runtime` variable, that the libc
dependencies are optional, and that the various functions for writing
out files work for basic cases.
llvm-svn: 208460
This reverts commit r200561.
This calling convention was an attempt to match the MSVC C++ ABI for
methods that return structures by value. This solution didn't scale,
because it would have required splitting every CC available on Windows
into two: one for methods and one for free functions.
Now that we can put sret on the second arg (r208453), and Clang does
that (r208458), revert this hack.
llvm-svn: 208459
Summary:
MSVC always passes 'sret' after 'this', unlike GCC. This required
changing a number of places in Clang that assumed the sret parameter was
always first in LLVM IR.
This fixes win64 MSVC ABI compatibility for methods returning structs.
Reviewers: rsmith, majnemer
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3618
llvm-svn: 208458
we do not use the information from SCEVAddRecExpr to compute the shape of the array,
so a better place for this function is in ScalarEvolution.
llvm-svn: 208456
r144269 changed clang_disposeDiagnostic() to be a no-op, but didn't update
code completion diagnostics. Let CXCodeCompleteResults store all diagnostics
returned by clang_codeCompleteGetDiagnostic() and then free them up in
clang_disposeCodeCompleteResults().
Code completion diagnostics referred to data stored in CXCodeCompleteResults
before already, so it wasn't possible to refer to the results of
clang_codeCompleteGetDiagnostic() after clang_disposeCodeCompleteResults()
before this change already -- hence this should be a safe, backwards-compatible
change.
Leak found by LSan, fixes PR19690.
llvm-svn: 208454
MSVC always places the implicit sret parameter after the implicit this
parameter of instance methods. We used to handle this for
x86_thiscallcc by allocating the sret parameter on the stack and leaving
the this pointer in ecx, but that doesn't handle alternative calling
conventions like cdecl, stdcall, fastcall, or the win64 convention.
Instead, change the verifier to allow sret on the second parameter.
This also requires changing the Mips and X86 backends to return the
argument with the sret parameter, instead of assuming that the sret
parameter comes first.
The Sparc backend also returns sret parameters in a register, but I
wasn't able to update it to handle secondary sret parameters. It
currently calls report_fatal_error if you feed it an sret in the second
parameter.
Reviewers: rafael.espindola, majnemer
Differential Revision: http://reviews.llvm.org/D3617
llvm-svn: 208453
We did not actively try to resolve dllexported symbols specified
by /export or by a module definition file. So if exported symbols
would be resolved for other reasons, like other symbols refer to
them, that was fine, but if (unreferenced) exported symbols were
in an archive file, and no one refers to that file in the archive,
they remained unresolved.
That would obviously cause the issue that dllexported symbols are
not in a resultant DLL.
In this patch, we create an undefined symbol for each dllexported
symbol, to let the core linker to resolve it.
llvm-svn: 208452
This change lets MSan rely on libcxx's own build system instead of manually
compiling its sources and setting up all the necessary compile flags. It would
also simplify compiling libcxx with another sanitizers (in particular, TSan).
The tricky part is to make sure libcxx is reconfigured/rebuilt when Clang or
MSan runtime library is changed. "clobber" step used in this patch works well
for me, but it's possible it would break for other configurations - will
watch the buildbots.
llvm-svn: 208451
The base class is the culprit/risk here - a sealed/final derived class
with virtual functions and a non-virtual dtor can't accidentally be
polymorphically destroyed (if the base class's dtor is protected - which
also suppresses this warning).
llvm-svn: 208449
One error we were not deleting the alias or putting it in the Module. The
end result is that there was an use left of the aliasee when the module was
deleted.
llvm-svn: 208447
Previously the handling of exported symbol was wrong if it's
specified in a module definition file in the form of
<externalname>=<internalname>. Export the correct symbol.
llvm-svn: 208446
MacroArgs are owned by TokenLexer, and when a TokenLexer is destroyed, it'll
call its MacroArgs's destroy() method. destroy() only appends the MacroArg to
Preprocessor's MacroArgCache list, and Preprocessor's destructor then calls
deallocate() on all MacroArgs in that list. This method then ends up freeing
the MacroArgs's memory.
In a code completion context, Parser::cutOffParsing() gets called when a code
completion token is hit, which changes the type of the current token to
tok::eof. eof tokens aren't always ConsumeToken()ed, so
Preprocessor::HandleEndOfFile() isn't always called, and that function is
responsible for popping the macro stack.
Due to this, Preprocessor::CurTokenLexer can be non-NULL when
~Preprocessor runs. It's a unique_ptr, so it ended up being destructed after
~Preprocessor completed, and its MacroArgs thus got added to the freelist after
the code freeing things on the freelist had already completed. The fix is to
explicitly call reset() before the freelist processing happens. (See the bug
for more notes.)
llvm-svn: 208438
The thread safety analysis isn't very useful in ObjC (you can't annotate
ObjC classes or methods) but we can still analyze the actual code and
show violations in usage of C/C++ functions.
Fixes PR19541, which does not use thread safety attributes but crashes
with -Weverything.
llvm-svn: 208436
This patch adds support to ARM for custom lowering of the
llvm.{u|s}add.with.overflow.i32 intrinsics for i32/i64. This is particularly useful
for handling idiomatic saturating math functions as generated by
InstCombineCompare.
Test cases included.
rdar://14853450
llvm-svn: 208435
Since ExtractValue is not included in ComputeSpeculationCost CFGs containing
ExtractValueInsts cannot be simplified. In particular this interacts with
InstCombineCompare's tendency to insert add.with.overflow intrinsics for
certain idiomatic math operations, preventing optimization.
This patch adds ExtractValue to the ComputeSpeculationCost. Test case included
rdar://14853450
llvm-svn: 208434
Filipe Cabecinhas