We were previously checking at every destructor declaration, but that was a bit
excessive. Since the deleting destructor is emitted with the vtable, do the
check when the vtable is marked used.
Differential Revision: http://llvm-reviews.chandlerc.com/D2851
llvm-svn: 202046
Fix an unintentional stdin read in the darwin-asan-nofortify.c test and replace
it with an explicit test for multiple -E inputs passed to the driver.
Noticed while working on the in-process driver patch.
llvm-svn: 202007
Most 64-bit targets define int64_t as long int, and AArch64 should
make same definition to follow LP64 model. In GNU tool chain, int64_t
is defined as long int for 64-bit target. So to get consistent with GNU,
it's better Changing int64_t from 'long long int' to 'long int',
otherwise clang will get different name mangling suffix compared with g++.
llvm-svn: 202004
We still don't use the PGO to set branch weights for these loops, but at
least this keeps the compiler from crashing. <rdar://problem/16137778>
llvm-svn: 202002
- Only include offsets with local (in function scope) symbols, where we don't encode scoping
- Only include the filename with non-system symbols. Presumably the system headers will not provide conflicting definitions.
rdar://15976823
llvm-svn: 201990
The DiagnosticBuilder's lifetime in parser typo recovery was overlapping with
the subsequent consume which can itself emit PP diagnostics.
Patch by Olivier Goffart!
llvm-svn: 201965
The integrated assembler is a feature. This makes the new flags the default
option, and the previous versions aliases. Ideally, at some point the aliases
would be entirely removed.
llvm-svn: 201963
In r199283 I switched the name of this variable to CCC_OVERRIDE_OPTIONS, but
I kept some code to continue recognizing the old name temporarily. As far as
I know, the only use of this was for some internal testing at Apple, and we've
now switched to use the new name. If anyone else is still using this and needs
more time to switch names, I guess we'll find out! <rdar://problem/15821425>
llvm-svn: 201962
Forward the -no-integrated-as option to -cc1 rather than simply invoking the
appropriate tool. This is useful since this option has been overloaded to
permit disabling of parsing inline assembly at the MC layer.
This re-applies the previous version of the patch with a renaming of the driver
option to the public name rather than the internal name (-target vs -triple).
The actual failure is fixed separately of an overly aggressive negative pattern
match in the MIPS driver tests. It also fixes the incorrect test for targets
that have the integrated assembler disabled by default.
llvm-svn: 201960
The tests attempt to validate the invocation of the assembler program with the
integrated assembler disabled. However, the match pattern for the negative
tests are lax and will match both the driver invocation as well as the assembler
invocation. Make the tests more strict by ensuring that we only match the
assembler invocation.
llvm-svn: 201959
Forward the -no-integrated-as option to -cc1 rather than simply invoking the
appropriate tool. This is useful since this option has been overloaded to
permit disabling of parsing inline assembly at the MC layer.
llvm-svn: 201952
The language forbids defining enums in prototypes, so this check is normally
redundant, but if an enum is defined during template instantiation it should
not be added to the prototype scope.
While at it, clean up the code that deals with tag definitions in prototype
scope and expand the visibility warning to cover the case where an anonymous
enum is defined.
Differential Revision: http://llvm-reviews.chandlerc.com/D2742
llvm-svn: 201927
A recent change caused multi-dimensional arrays not to be handled
correctly, this patch fixes that. Also, it adds a lit test for
multi-dimensional arrays.
llvm-svn: 201917
It was previously thought that Sema::InstantiateClass could not fail
from within this point in instantiate.
However, it can happen if the class is invalid some way (i.e. invalid
base specifier).
This fixes PR18907.
Differential Revision: http://llvm-reviews.chandlerc.com/D2850
llvm-svn: 201913
In particular, if we see an @property within the @interface of a class
conforming to a protocol with this attribute, we treat that as
if the implementation were available, per the rules of default
property synthesis.
llvm-svn: 201911
CGRecordLayoutBuilder was aging, complex, multi-pass, and shows signs of
existing before ASTRecordLayoutBuilder. It redundantly performed many
layout operations that are now performed by ASTRecordLayoutBuilder and
asserted that the results were the same. With the addition of support
for the MS-ABI, such as placement of vbptrs, vtordisps, different
bitfield layout and a variety of other features, CGRecordLayoutBuilder
was growing unwieldy in its redundancy.
This patch re-architects CGRecordLayoutBuilder to not perform any
redundant layout but rather, as directly as possible, lower an
ASTRecordLayout to an llvm::type. The new architecture is significantly
smaller and simpler than the CGRecordLayoutBuilder and contains fewer
ABI-specific code paths. It's also one pass.
The architecture of the new system is described in the comments. For the
most part, the new system simply takes all of the fields and bases from
an ASTRecordLayout, sorts them, inserts padding and dumps a record.
Bitfields, unions and primary virtual bases make this process a bit more
complicated. See the inline comments.
In addition, this patch updates a few lit tests due to the fact that the
new system computes more accurate llvm types than CGRecordLayoutBuilder.
Each change is commented individually in the review.
Differential Revision: http://llvm-reviews.chandlerc.com/D2795
llvm-svn: 201907
Added two new options for -mfpu when targetting ARM:
* fpv4-sp-d16
* fp4-sp-d16
The first is the same spelling as gcc.
The lack of a leading `v' is correct, this is consistent with ARM's
documentation and gcc's spelling of the option.
llvm-svn: 201846
Virtual methods expect 'this' to point to the vfptr containing the
virtual method, and this extends to virtual member pointer thunks. The
relevant vfptr is always at offset zero on entry to the thunk, and no
this adjustment is needed.
Previously we would not include the vfptr adjustment in the member
pointer, and we'd look at the vfptr offset when loading from the vftable
in the thunk.
Fixes PR18917.
llvm-svn: 201835
Slight change to the way zero-sized sub-objects are tracked in the
presence of virtual bases.
In addition we correctly distinguish between dsize and nvsize.
addresses http://llvm.org/bugs/show_bug.cgi?id=18826
Unit tests are included.
llvm-svn: 201832
There is no bound architecture for the dsymutil action in the driver. Trying
to check various properties of the target will cause an assertion failure
because the target doesn't get initialized without a bound architecture.
<rdar://problem/16111555>
llvm-svn: 201830
Somehow both Daniel and I missed the fact that while loops are only identical
if they have identical bodies.
Patch by Daniel Fahlgren!
llvm-svn: 201829
The MS ABI requires that we determine the vbptr offset if have a
virtual inheritance model. Instead, raise an error pointing to the
diagnostic when this happens.
This fixes PR18583.
Differential Revision: http://llvm-reviews.chandlerc.com/D2842
llvm-svn: 201824
gcc never expands macros in pragmas and MSVC always expands macros
before processing pragmas. Clang usually allows macro expansion, except
in a handful of pragmas, most of which are handled by the lexer.
Also remove PPCallbacks for pragmas that are currently handled in the
parser. Without a Parser, such as with clang -E, these callbacks would
never be called.
Fixes PR18576.
llvm-svn: 201821
This breaks backwards compatibility with existing code. Previously, this
was defined as
#define _mm_prefetch(a, sel) (__builtin_prefetch((void *)(a), 0, (sel)))
Which basically accepts any pointer. Changing this to char* simply
breaks a lot of existing code. I have tried changing char* to
"const void*", which seems to be the right thing as per Intel
specification this should work on basically any pointer. However,
apparently this breaks windows compatibility (because of a conflicting
declaration in windows.h).
So, we probably need to #ifdef this based on whether clang is compiling
for windows. According to Chandler, this might be done by introducing an
additional symbol to a fake type in BuiltinsX86.def and then condition
the type expansion on the platform.
llvm-svn: 201775
This patch adds several built-ins that are required for ms
compatibility. _mm_prefetch must be a built-in because it takes a
compile-time constant argument and our prior approach of using a #define
to the current built-in doesn't work in the presence of re-declaration
of _mm_prefetch. The others can be obtained by including the windows
system headers. If a user includes the windows system headers but not
intrin.h they still need to work and therefore must be built-in because
we don't get a chance to implement them in intrin.h in this case.
llvm-svn: 201734
This definition is not chosen idly. There is an unfortunate reality with
max_align_t -- the specific nature of its definition leaks into the ABI
almost immediately. Because it is part of C11 and C++11 it becomes
essential for it to match with other systems on that ABI. There is an
effort to discourage any further use of this construct as a consequence
-- using max_align_t introduces an immediate ABI problem. We can never
update it to have larger alignment even as the microarchitecture changes
to necessitate higher alignment. =/
The particular definition here exactly matches the ABI of GCC's chosen
::max_align_t definition, for better or worse. This was written with the
help of Richard Smith who was decoding the exact ABI implications of the
selected definition in GCC. Notably, in-register arguments are impacted
by the particular definition chosen. =/
No one is under the illusion that this is a "good" or "useful"
definition of max_align_t, and we are working with the standards
committee to specify a more useful interface to address this need.
llvm-svn: 201729
Hans Wennborg