Summary:
An _Atomic of an empty struct is pretty silly. In general we just widen empty
structs to hold a byte's worth of storage, and we represent size and alignment
as 0 internally and let LLVM figure out what to do. For _Atomic it's a bit
different: the memory model mandates concrete effects occur when atomic
operations occur, so in most cases actual instructions need to get emitted. It's
really not worth trying to optimize empty struct atomics by figuring out e.g.
that a fence would do, even though sane compilers should do optimize atomics.
Further, wg21.link/p0528 will fix C++20 atomics with padding bits so that
cmpxchg on them works, which means that we'll likely need to do the zero-init
song and dance for empty atomic structs anyways (and I think we shouldn't
special-case this behavior to C++20 because prior standards are just broken).
This patch therefore makes a minor change to r176658 "Promote atomic type sizes
up to a power of two": if the width of the atomic's value type is 0, just use 1
byte for width and leave alignment as-is (since it should never be zero, and
over-aligned zero-width structs are weird but fine).
This fixes an assertion:
(NumBits >= MIN_INT_BITS && "bitwidth too small"), function get, file ../lib/IR/Type.cpp, line 241.
It seems like this has run into other assertions before (namely the unreachable
Kind check in ImpCastExprToType), but I haven't reproduced that issue with
tip-of-tree.
<rdar://problem/39678063>
Reviewers: arphaman, rjmccall
Subscribers: aheejin, cfe-commits
Differential Revision: https://reviews.llvm.org/D46613
llvm-svn: 331845
This is similar to the LLVM change https://reviews.llvm.org/D46290.
We've been running doxygen with the autobrief option for a couple of
years now. This makes the \brief markers into our comments
redundant. Since they are a visual distraction and we don't want to
encourage more \brief markers in new code either, this patch removes
them all.
Patch produced by
for i in $(git grep -l '\@brief'); do perl -pi -e 's/\@brief //g' $i & done
for i in $(git grep -l '\\brief'); do perl -pi -e 's/\\brief //g' $i & done
Differential Revision: https://reviews.llvm.org/D46320
llvm-svn: 331834
Summary:
This patch tackles long hanging fruit for the builtin operator<=> expressions. It is currently needs some cleanup before landing, but I want to get some initial feedback.
The main changes are:
* Lookup, build, and store the required standard library types and expressions in `ASTContext`. By storing them in ASTContext we don't need to store (and duplicate) the required expressions in the BinaryOperator AST nodes.
* Implement [expr.spaceship] checking, including diagnosing narrowing conversions.
* Implement `ExprConstant` for builtin spaceship operators.
* Implement builitin operator<=> support in `CodeGenAgg`. Initially I emitted the required comparisons using `ScalarExprEmitter::VisitBinaryOperator`, but this caused the operand expressions to be emitted once for every required cmp.
* Implement [builtin.over] with modifications to support the intent of P0946R0. See the note on `BuiltinOperatorOverloadBuilder::addThreeWayArithmeticOverloads` for more information about the workaround.
Reviewers: rsmith, aaron.ballman, majnemer, rnk, compnerd, rjmccall
Reviewed By: rjmccall
Subscribers: rjmccall, rsmith, aaron.ballman, junbuml, mgorny, cfe-commits
Differential Revision: https://reviews.llvm.org/D45476
llvm-svn: 331677
FunctionProtoType.
We previously re-evaluated the expression each time we wanted to know whether
the type is noexcept or not. We now evaluate the expression exactly once.
This is not quite "no functional change": it fixes a crasher bug during AST
deserialization where we would try to evaluate the noexcept specification in a
situation where we have not deserialized sufficient portions of the AST to
permit such evaluation.
llvm-svn: 331428
This is not yet part of any C++ working draft, and so is controlled by the flag
-fchar8_t rather than a -std= flag. (The GCC implementation is controlled by a
flag with the same name.)
This implementation is experimental, and will be removed or revised
substantially to match the proposal as it makes its way through the C++
committee.
llvm-svn: 331244
As reported here: https://bugs.llvm.org/show_bug.cgi?id=37033
Any usage of a builtin function that uses a va_list by reference
will cause an assertion when redeclaring it.
After discussion in the review, it was concluded that the correct
way of accomplishing this fix is to make attempts to redeclare certain
builtins an error. Unfortunately, doing this limitation for all builtins
is likely a breaking change, so this commit simply limits it to
types with custom type checking and those that take a reference.
Two tests needed to be updated to make this work.
Differential Revision: https://reviews.llvm.org/D45383
llvm-svn: 330160
Summary:
This change consolidates the always/never lists that may be provided to
clang to externally control which functions should be XRay instrumented
by imbuing attributes. The files follow the same format as defined in
https://clang.llvm.org/docs/SanitizerSpecialCaseList.html for the
sanitizer blacklist.
We also deprecate the existing `-fxray-instrument-always=` and
`-fxray-instrument-never=` flags, in favour of `-fxray-attr-list=`.
This fixes http://llvm.org/PR34721.
Reviewers: echristo, vlad.tsyrklevich, eugenis
Reviewed By: vlad.tsyrklevich
Subscribers: llvm-commits, cfe-commits
Differential Revision: https://reviews.llvm.org/D45357
llvm-svn: 329543
Found via codespell -q 3 -I ../clang-whitelist.txt
Where whitelist consists of:
archtype
cas
classs
checkk
compres
definit
frome
iff
inteval
ith
lod
methode
nd
optin
ot
pres
statics
te
thru
Patch by luzpaz! (This is a subset of D44188 that applies cleanly with a few
files that have dubious fixes reverted.)
Differential revision: https://reviews.llvm.org/D44188
llvm-svn: 329399
ObjC and ObjC++ pass non-trivial structs in a way that is incompatible
with each other. For example:
typedef struct {
id f0;
__weak id f1;
} S;
// this code is compiled in c++.
extern "C" {
void foo(S s);
}
void caller() {
// the caller passes the parameter indirectly and destructs it.
foo(S());
}
// this function is compiled in c.
// 'a' is passed directly and is destructed in the callee.
void foo(S a) {
}
This patch fixes the incompatibility by passing and returning structs
with __strong or weak fields using the C ABI in C++ mode. __strong and
__weak fields in a struct do not cause the struct to be destructed in
the caller and __strong fields do not cause the struct to be passed
indirectly.
Also, this patch fixes the microsoft ABI bug mentioned here:
https://reviews.llvm.org/D41039?id=128767#inline-364710
rdar://problem/38887866
Differential Revision: https://reviews.llvm.org/D44908
llvm-svn: 328731
r327219 added wrappers to std::sort which randomly shuffle the container before
sorting. This will help in uncovering non-determinism caused due to undefined
sorting order of objects having the same key.
To make use of that infrastructure we need to invoke llvm::sort instead of
std::sort.
llvm-svn: 328636
If the link clause is used on the declare target directive, the object
should be linked on target or target data directives, not during the
codegen. Patch adds support for this clause.
llvm-svn: 328544
The patch adds nocf_check target independent attribute for disabling checks that were enabled by cf-protection flag.
The attribute can be appertained to functions and function pointers.
Attribute name follows GCC's similar attribute name.
Differential Revision: https://reviews.llvm.org/D41880
llvm-svn: 327768
Added initial codegen for device side of declarations inside `omp
declare target` construct + codegen for implicit `declare target`
functions, which are used in the target regions.
llvm-svn: 327636
When parsing comments, for example, for -Wdocumentation, slightly different
behaviour occurs when -fparse-all-comments is specified. However, these
differences are subtle:
1. All comments are saved during parsing, regardless of whether they are doc
comments or not.
2. "Maybe-doc" comments, like <, !, etc, are saved as such, instead of marking
them as ordinary comments. The maybe-doc type of comment is never saved
otherwise. (Warning on these is the impetus of -Wdocumentation.)
3. All comments are treated as doc comments in ASTContext, even if they are ordinary.
This change moves the logic for checking CommentOptions.ParseAllComments closer
to where it has an effect. The overall logic is unchanged, but checks of the
ParseAllComments flag are now done where the effect will be clearer.
Subscribers: cfe-commits
llvm-svn: 326512
Original change:
[NFC] Move CommentOpts checks to the call sites that depend on it.
When parsing comments, for example, for -Wdocumentation, slightly different
behaviour occurs when -fparse-all-comments is specified. However, these
differences are subtle:
1. All comments are saved during parsing, regardless of whether they are doc comments or not.
2. "Maybe-doc" comments, like //<, //!, etc, are saved as such, instead of marking them as ordinary comments. The maybe-doc type of comment is never saved otherwise. (Warning on these is the impetus of -Wdocumentation.)
3. All comments are treated as doc comments in ASTContext, even if they are ordinary.
This change moves the logic for checking CommentOptions.ParseAllComments closer
to where it has an effect. The overall logic is unchanged, but checks of the
ParseAllComments flag are now done where the effect will be clearer.
llvm-svn: 326508
When parsing comments, for example, for -Wdocumentation, slightly different
behaviour occurs when -fparse-all-comments is specified. However, these
differences are subtle:
1. All comments are saved during parsing, regardless of whether they are doc
comments or not.
2. "Maybe-doc" comments, like //<, //!, etc, are saved as such, instead of
marking them as ordinary comments. The maybe-doc type of comment is never
saved otherwise. (Warning on these is the impetus of -Wdocumentation.)
3. All comments are treated as doc comments in ASTContext, even if they are
ordinary.
This change moves the logic for checking CommentOptions.ParseAllComments closer
to where it has an effect. The overall logic is unchanged, but checks of the
ParseAllComments flag are now done where the effect will be clearer.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D43663
llvm-svn: 326501
So I wrote a clang-tidy check to lint out redundant `isa`, `cast`, and
`dyn_cast`s for fun. This is a portion of what it found for clang; I
plan to do similar cleanups in LLVM and other subprojects when I find
time.
Because of the volume of changes, I explicitly avoided making any change
that wasn't highly local and obviously correct to me (e.g. we still have
a number of foo(cast<Bar>(baz)) that I didn't touch, since overloading
is a thing and the cast<Bar> did actually change the type -- just up the
class hierarchy).
I also tried to leave the types we were cast<>ing to somewhere nearby,
in cases where it wasn't locally obvious what we were dealing with
before.
llvm-svn: 326416
ARC mode.
Declaring __strong pointer fields in structs was not allowed in
Objective-C ARC until now because that would make the struct non-trivial
to default-initialize, copy/move, and destroy, which is not something C
was designed to do. This patch lifts that restriction.
Special functions for non-trivial C structs are synthesized that are
needed to default-initialize, copy/move, and destroy the structs and
manage the ownership of the objects the __strong pointer fields point
to. Non-trivial structs passed to functions are destructed in the callee
function.
rdar://problem/33599681
Differential Revision: https://reviews.llvm.org/D41228
llvm-svn: 326307
This broke Clang bootstrap on Windows, PR36453.
> This handles them exactly the same way that we handle const integral
> static data members with inline definitions, which is what MSVC does.
>
> As a follow-up, now that we have a way to mark variables inline in the
> AST, we should consider marking them implicitly inline there instead of
> only treating them as inline in CodeGen. Unfortunately, this breaks a
> lot of dllimport test cases, so that is future work for now.
>
> Fixes PR36125.
llvm-svn: 325576
This handles them exactly the same way that we handle const integral
static data members with inline definitions, which is what MSVC does.
As a follow-up, now that we have a way to mark variables inline in the
AST, we should consider marking them implicitly inline there instead of
only treating them as inline in CodeGen. Unfortunately, this breaks a
lot of dllimport test cases, so that is future work for now.
Fixes PR36125.
llvm-svn: 325375
The 'trivial_abi' attribute can be applied to a C++ class, struct, or
union. It makes special functions of the annotated class (the destructor
and copy/move constructors) to be trivial for the purpose of calls and,
as a result, enables the annotated class or containing classes to be
passed or returned using the C ABI for the underlying type.
When a type that is considered trivial for the purpose of calls despite
having a non-trivial destructor (which happens only when the class type
or one of its subobjects is a 'trivial_abi' class) is passed to a
function, the callee is responsible for destroying the object.
For more background, see the discussions that took place on the mailing
list:
http://lists.llvm.org/pipermail/cfe-dev/2017-November/055955.htmlhttp://lists.llvm.org/pipermail/cfe-commits/Week-of-Mon-20180101/thread.html#214043
rdar://problem/35204524
Differential Revision: https://reviews.llvm.org/D41039
llvm-svn: 324269
Summary:
Clang incorrectly reports empty unions as having a unique object representation. However, this is not correct since `sizeof(EmptyUnion) == 1` AKA it has 8 bits of padding. Therefore it should be treated the same as an empty struct and report `false`.
@erichkeane also suggested this fix should be merged into the 6.0 release branch, so the initial release of `__has_unique_object_representations` is as bug-free as possible.
Reviewers: erichkeane, rsmith, aaron.ballman, majnemer
Reviewed By: erichkeane
Subscribers: cfe-commits, erichkeane
Differential Revision: https://reviews.llvm.org/D42863
llvm-svn: 324134
GCC's attribute 'target', in addition to being an optimization hint,
also allows function multiversioning. We currently have the former
implemented, this is the latter's implementation.
This works by enabling functions with the same name/signature to coexist,
so that they can all be emitted. Multiversion state is stored in the
FunctionDecl itself, and SemaDecl manages the definitions.
Note that it ends up having to permit redefinition of functions so
that they can all be emitted. Additionally, all versions of the function
must be emitted, so this also manages that.
Note that this includes some additional rules that GCC does not, since
defining something as a MultiVersion function after a usage has been made illegal.
The only 'history rewriting' that happens is if a function is emitted before
it has been converted to a multiversion'ed function, at which point its name
needs to be changed.
Function templates and virtual functions are NOT yet supported (not supported
in GCC either).
Additionally, constructors/destructors are disallowed, but the former is
planned.
llvm-svn: 322028
These just overloads for _Float128. They're supported by GCC 7 and used
by glibc. APFloat support is already there so just add the overloads.
__builtin_copysignf128
__builtin_fabsf128
__builtin_huge_valf128
__builtin_inff128
__builtin_nanf128
__builtin_nansf128
This is the same support that GCC has, according to the documentation,
but limited to _Float128.
llvm-svn: 321948
As discussed in the mail thread <https://groups.google.com/a/isocpp.org/forum/
#!topic/std-discussion/T64_dW3WKUk> "Calling noexcept function throug non-
noexcept pointer is undefined behavior?", such a call should not be UB.
However, Clang currently warns about it.
This change removes exception specifications from the function types recorded
for -fsanitize=function, both in the functions themselves and at the call sites.
That means that calling a non-noexcept function through a noexcept pointer will
also not be flagged as UB. In the review of this change, that was deemed
acceptable, at least for now. (See the "TODO" in compiler-rt
test/ubsan/TestCases/TypeCheck/Function/function.cpp.)
To remove exception specifications from types, the existing internal
ASTContext::getFunctionTypeWithExceptionSpec was made public, and some places
otherwise unrelated to this change have been adapted to call it, too.
This is the cfe part of a patch covering both cfe and compiler-rt.
Differential Revision: https://reviews.llvm.org/D40720
llvm-svn: 321859
Commit 7ac28eb0a5 / r310911 ("[OpenCL] Allow targets to select address
space per type", 2017-08-15) made Basic depend on AST, introducing a
circular dependency. Break this dependency by adding the
OpenCLTypeKind enum in Basic and map from AST types to this enum in
ASTContext.
Differential Revision: https://reviews.llvm.org/D40838
llvm-svn: 319883
As rsmith pointed out, the original implementation of this intrinsic
missed a number of important situations. This patch fixe a bunch of
shortcomings and implementation details to make it work correctly.
Differential Revision: https://reviews.llvm.org/D39347
llvm-svn: 319446
These functions were defined as static members of TemplateSpecializationType.
Now they are moved to namespace level. Previously there were different
implementations for lists containing TemplateArgument and TemplateArgumentLoc,
now these implementations share the same code.
This change is a result of refactoring patch D40508. NFC.
llvm-svn: 319178
explicitly instantiated, still emit it with each use.
We don't emit a definition of the member with an explicit instantiation
definition (and indeed it appears that we're not allowed to, since an explicit
instantiation definition does not constitute an odr-use and only odr-use
permits definition for defaulted special members). So we still need to emit a
weak definition with each use.
This also makes defaulted-in-class declarations behave more like
implicitly-declared special members, which matches their design intent.
And it matches the way this problem was solved in GCC.
llvm-svn: 318474
Added support for regcall as default calling convention. Also added code to
exclude main when applying default calling conventions.
Patch-By: eandrews
Differential Revision: https://reviews.llvm.org/D39210
llvm-svn: 317268
Fastcall doesn't support variadic function calls, so
setting the default calling convention to Fastcall would
result in incorrect code being emitted for these conditions.
This patch adds a 'variadic' test to the default calling conv
test, as well as fixes the behavior of fastcall.
llvm-svn: 316528
Summary:
Convert clang::LangAS to a strongly typed enum
Currently both clang AST address spaces and target specific address spaces
are represented as unsigned which can lead to subtle errors if the wrong
type is passed. It is especially confusing in the CodeGen files as it is
not possible to see what kind of address space should be passed to a
function without looking at the implementation.
I originally made this change for our LLVM fork for the CHERI architecture
where we make extensive use of address spaces to differentiate between
capabilities and pointers. When merging the upstream changes I usually
run into some test failures or runtime crashes because the wrong kind of
address space is passed to a function. By converting the LangAS enum to a
C++11 we can catch these errors at compile time. Additionally, it is now
obvious from the function signature which kind of address space it expects.
I found the following errors while writing this patch:
- ItaniumRecordLayoutBuilder::LayoutField was passing a clang AST address
space to TargetInfo::getPointer{Width,Align}()
- TypePrinter::printAttributedAfter() prints the numeric value of the
clang AST address space instead of the target address space.
However, this code is not used so I kept the current behaviour
- initializeForBlockHeader() in CGBlocks.cpp was passing
LangAS::opencl_generic to TargetInfo::getPointer{Width,Align}()
- CodeGenFunction::EmitBlockLiteral() was passing a AST address space to
TargetInfo::getPointerWidth()
- CGOpenMPRuntimeNVPTX::translateParameter() passed a target address space
to Qualifiers::addAddressSpace()
- CGOpenMPRuntimeNVPTX::getParameterAddress() was using
llvm::Type::getPointerTo() with a AST address space
- clang_getAddressSpace() returns either a LangAS or a target address
space. As this is exposed to C I have kept the current behaviour and
added a comment stating that it is probably not correct.
Other than this the patch should not cause any functional changes.
Reviewers: yaxunl, pcc, bader
Reviewed By: yaxunl, bader
Subscribers: jlebar, jholewinski, nhaehnle, Anastasia, cfe-commits
Differential Revision: https://reviews.llvm.org/D38816
llvm-svn: 315871
Currently Clang uses default address space (0) to represent private address space for OpenCL
in AST. There are two issues with this:
Multiple address spaces including private address space cannot be diagnosed.
There is no mangling for default address space. For example, if private int* is emitted as
i32 addrspace(5)* in IR. It is supposed to be mangled as PUAS5i but it is mangled as
Pi instead.
This patch attempts to represent OpenCL private address space explicitly in AST. It adds
a new enum LangAS::opencl_private and adds it to the variable types which are implicitly
private:
automatic variables without address space qualifier
function parameter
pointee type without address space qualifier (OpenCL 1.2 and below)
Differential Revision: https://reviews.llvm.org/D35082
llvm-svn: 315668
This patch relates to: https://reviews.llvm.org/D33666 This adds support
for template parameters to be passed to the address_space attribute.
The main goal is to add further flexibility to the attribute and allow
for it to be used easily with templates.
The main additions are a new type (DependentAddressSpaceType) alongside
its TypeLoc and its mangling. As well as the logic required to support
dependent address spaces which mainly resides in TreeTransform.h and
SemaType.cpp.
llvm-svn: 314649
C11 standard refers to the unsigned counterpart of the type ptrdiff_t
in the paragraph 7.21.6.1p7 where it defines the format specifier %tu.
In Clang (in PrintfFormatString.cpp, lines 508-510) there is a FIXME for this case,
in particular, Clang didn't diagnose %tu issues at all, i.e.
it didn't emit any warnings on the code printf("%tu", 3.14).
In this diff we add a method getUnsignedPointerDiffType for getting the corresponding type
similarly to how it's already done in the other analogous cases (size_t, ssize_t, ptrdiff_t etc)
and fix -Wformat diagnostics for %tu plus the emitted fix-it as well.
Test plan: make check-all
Differential revision: https://reviews.llvm.org/D38270
llvm-svn: 314470
The attribute informs the compiler that the annotated pointer parameter
of a function cannot escape and enables IRGen to attach attribute
'nocapture' to parameters that are annotated with the attribute. That is
the only optimization that currently takes advantage of 'noescape', but
there are other optimizations that will be added later that improves
IRGen for ObjC blocks.
This recommits r313722, which was reverted in r313725 because clang
couldn't build compiler-rt. It failed to build because there were
function declarations that were missing 'noescape'. That has been fixed
in r313929.
rdar://problem/19886775
Differential Revision: https://reviews.llvm.org/D32210
llvm-svn: 313945
This reverts commit r313722.
It looks like compiler-rt/lib/tsan/rtl/tsan_libdispatch_mac.cc cannot be
compiled because some of the functions declared in the file do not match
the ones in the SDK headers (which are annotated with 'noescape').
llvm-svn: 313725
The attribute informs the compiler that the annotated pointer parameter
of a function cannot escape and enables IRGen to attach attribute
'nocapture' to parameters that are annotated with the attribute. That is
the only optimization that currently takes advantage of 'noescape', but
there are other optimizations that will be added later that improves
IRGen for ObjC blocks.
rdar://problem/19886775
Differential Revision: https://reviews.llvm.org/D32210
llvm-svn: 313722
The attribute informs the compiler that the annotated pointer parameter
of a function cannot escape and enables IRGen to attach attribute
'nocapture' to parameters that are annotated with the attribute. That is
the only optimization that currently takes advantage of 'noescape', but
there are other optimizations that will be added later that improves
IRGen for ObjC blocks.
rdar://problem/19886775
Differential Revision: https://reviews.llvm.org/D32520
llvm-svn: 313720
This is a recommit of r312781; in some build configurations
variable names are omitted, so changed the new regression
test accordingly.
llvm-svn: 312794
This adds _Float16 as a source language type, which is a 16-bit floating point
type defined in C11 extension ISO/IEC TS 18661-3.
In follow up patches documentation and more tests will be added.
Differential Revision: https://reviews.llvm.org/D33719
llvm-svn: 312781
Generalize getOpenCLImageAddrSpace into getOpenCLTypeAddrSpace, such
that targets can select the address space per type.
No functional changes intended.
Initial patch by Simon Perretta.
Differential Revision: https://reviews.llvm.org/D33989
llvm-svn: 310911
OpenCL 2.0 atomic builtin functions have a scope argument which is ideally
represented as synchronization scope argument in LLVM atomic instructions.
Clang supports translating Clang atomic builtin functions to LLVM atomic
instructions. However it currently does not support synchronization scope
of LLVM atomic instructions. Without this, users have to use LLVM assembly
code to implement OpenCL atomic builtin functions.
This patch adds OpenCL 2.0 atomic builtin functions as Clang builtin
functions, which supports generating LLVM atomic instructions with
synchronization scope operand.
Currently only constant memory scope argument is supported. Support of
non-constant memory scope argument will be added later.
Differential Revision: https://reviews.llvm.org/D28691
llvm-svn: 310082
C11 standard refers to the signed counterpart of the type size_t in
the paragraph 7.21.6.1 where it defines d, i, o, u, x, or x conversion specifiers
(in printf format string).
In Clang there is a FIXME (in lib/Analysis/PrintfFormatString.cpp) for this case
(which is not handled correctly at the moment).
This diff adds getSignedSizeType method to TargetInfo and exposes it
in ASTContext similarly to how it is done for getSizeType.
lib/Analysis/PrintfFormatString.cpp will be changed in a separate commit.
Differential revision: https://reviews.llvm.org/D35378
Test plan: make check-all
llvm-svn: 308037
According to the documentation, when encoding a bit-field, GNU runtime
needs its starting position in addition to its type and size.
https://gcc.gnu.org/onlinedocs/gcc/Type-encoding.html
Prior to r297702, the starting position information was not being
encoded, which is incorrect, and after r297702, an assertion started to
fail because an ObjCIvarDecl was being passed to a function expecting a
FieldDecl.
This commit moves LookupFieldBitOffset to ASTContext and uses the
function to encode the starting position of bit-fields.
llvm-svn: 306364
declarations that are owned but unconditionally visible.
This allows us to set declarations as visible even if they have a local owning
module, without losing information. In turn, that means that our Objective-C
support can keep on incorrectly assuming the "hidden" bit on the declaration is
the whole story with regard to name visibility. This will also be useful once
we support the C++ Modules TS export semantics.
Objective-C name visibility is still incorrect in any case where the "hidden"
bit is not the complete story: for instance, in Objective-C++ the set of
visible categories will be wrong during template instantiation, and with local
submodule visibility enabled it will be wrong when building modules. Fixing that
will require a major overhaul of how visibility is handled for Objective-C (and
particularly for categories).
llvm-svn: 306075
This allows for -fms-extensions to work the same on LP64. For example,
_BitScanReverse is expected to be 32-bit, matching Windows/LLP64, even
though long is 64-bit on x86_64 Darwin or Linux (LP64).
Implement this by adding a new character code 'N', which is 'int' if
the target is LP64 and the same 'L' otherwise
Differential Revision: https://reviews.llvm.org/D34377
rdar://problem/32599746
llvm-svn: 305875
We were doing FindNodeOrInsertPos on SubstTemplateTypeParmPackTypes, so
we should presumably be inserting into SubstTemplateTypeParmPackTypes.
Looks like the FoldingSet API can be tweaked a bit so that we can catch
things like this at compile-time. I'll look into that shortly.
I'm unsure of how to test this; suggestions welcome.
Thanks to Vladimir Voskresensky for bringing this up!
llvm-svn: 305207
Pipes are now the size of pointers rather than the size
of the type that they contain.
Patch by Simon Perretta!
Differential Revision: https://reviews.llvm.org/D33597
llvm-svn: 304708
Alloca always returns a pointer in alloca address space, which may
be different from the type defined by the language. For example,
in C++ the auto variables are in the default address space. Therefore
cast alloca to the expected address space when necessary.
Differential Revision: https://reviews.llvm.org/D32248
llvm-svn: 303370
Some decls are created not where they are written, but in other module
files/users (implicit special members and function template implicit
specializations). To correctly identify them, use a bit next to the definition
to track the modular codegen property.
Discussed whether the module file bit could be omitted in favor of
reconstituting from the modular codegen decls list - best guess today is that
the efficiency improvement of not having to deserialize the whole list whenever
any function is queried by a module user is worth it for the small size
increase of this redundant (list + bit-on-def) representation.
Reviewers: rsmith
Differential Revision: https://reviews.llvm.org/D29901
llvm-svn: 299982
For OpenCL, the private address space qualifier is 0 in AST. Before this change, 0 address space qualifier
is always mapped to target address space 0. As now target private address space is specified by
alloca address space in data layout, address space qualifier 0 needs to be mapped to alloca addr space specified by the data layout.
This change has no impact on targets whose alloca addr space is 0.
With contributions from Matt Arsenault, Tony Tye and Wen-Heng (Jack) Chung
Differential Revision: https://reviews.llvm.org/D31404
llvm-svn: 299965
Summary:
The -fxray-always-instrument= and -fxray-never-instrument= flags take
filenames that are used to imbue the XRay instrumentation attributes
using a whitelist mechanism (similar to the sanitizer special cases
list). We use the same syntax and semantics as the sanitizer blacklists
files in the implementation.
As implemented, we respect the attributes that are already defined in
the source file (i.e. those that have the
[[clang::xray_{always,never}_instrument]] attributes) before applying
the always/never instrument lists.
Reviewers: rsmith, chandlerc
Subscribers: jfb, mgorny, cfe-commits
Differential Revision: https://reviews.llvm.org/D30388
llvm-svn: 299041
attributes.
These patches don't work because we can't currently access the parameter
information in a reliable way when building attributes. I thought this
would be relatively straightforward to fix, but it seems not to be the
case. Fixing this will requrie a substantial re-plumbing of machinery to
allow attributes to be handled in this location, and several other fixes
to the attribute machinery should probably be made at the same time. All
of this will make the patch .... substantially more complicated.
Reverting for now as there are active miscompiles caused by the current
version.
llvm-svn: 298695
declarations and calls instead of just definitions, and then teach it to
*not* attach such attributes even if the source code contains them.
This follows the design direction discussed on cfe-dev here:
http://lists.llvm.org/pipermail/cfe-dev/2017-January/052066.html
The idea is that for C standard library builtins, even if the library
vendor chooses to annotate their routines with __attribute__((nonnull)),
we will ignore those attributes which pertain to pointer arguments that
have an associated size. This allows the widespread (and seemingly
reasonable) pattern of calling these routines with a null pointer and
a zero size. I have only done this for the library builtins currently
recognized by Clang, but we can now trivially add to this set. This will
be controllable with -fno-builtin if anyone should care to do so.
Note that this does *not* change the AST. As a consequence, warnings,
static analysis, and source code rewriting are not impacted.
This isn't even a regression on any platform as neither Clang nor LLVM
have ever put 'nonnull' onto these arguments for declarations. All this
patch does is enable it on other declarations while preventing us from
ever accidentally enabling it on these libc functions due to a library
vendor.
It will also allow any other libraries using this annotation to gain
optimizations based on the annotation even when only a declaration is
visible.
llvm-svn: 298491
For variables in generic address spaces, for example:
```
unsigned char V[6442450944];
...
```
the address space is not yet known when we get into
*getConstantArrayType*, it is 0. AMDGCN target's
address space 0 has 32 bits pointers, so when we
call *getPointerWidth* with 0, the array size is
trimmed to 32 bits, which is not right.
Differential Revision: https://reviews.llvm.org/D30845
llvm-svn: 298420
1. Reimplemented conditional operator so that it checks
compatibility of unqualified pointees of the 2nd and
the 3rd operands (C99, OpenCL v2.0 6.5.15).
Define QualTypes compatibility for OpenCL as following:
- corresponding types are compatible (C99 6.7.3)
- CVR-qualifiers are equal (C99 6.7.3)
- address spaces are equal (implementation defined)
2. Added generic address space to Itanium mangling.
Review: D30037
Patch by Dmitry Borisenkov!
llvm-svn: 297468
This patch honors the unaligned type qualifier (currently available through he
keyword __unaligned and -fms-extensions) in CodeGen. In the current form the
patch affects declarations and expressions. It does not affect fields of
classes.
Differential Revision: https://reviews.llvm.org/D30166
llvm-svn: 297276
Fix an assertion that is hit when a redeclaration with differing types only
differs in the unaligned type-qualifier.
Differential Revision: https://reviews.llvm.org/D29986
llvm-svn: 296099
Removed ndrange_t as Clang builtin type and added
as a struct type in the OpenCL header.
Use type name to do the Sema checking in enqueue_kernel
and modify IR generation accordingly.
Review: D28058
Patch by Dmitry Borisenkov!
llvm-svn: 295311
such guides below explicit ones, and ensure that references to the class's
template parameters are not treated as forwarding references.
We make a few tweaks to the wording in the current standard:
1) The constructor parameter list is copied faithfully to the deduction guide,
without losing default arguments or a varargs ellipsis (which the standard
wording loses by omission).
2) If the class template declares no constructors, we add a T() -> T<...> guide
(which will only ever work if T has default arguments for all non-pack
template parameters).
3) If the class template declares nothing that looks like a copy or move
constructor, we add a T(T<...>) -> T<...> guide.
#2 and #3 follow from the "pretend we had a class type with these constructors"
philosophy for deduction guides.
llvm-svn: 295007
First pass at generating weak definitions of inline functions from module files
(& skipping (-O0) or emitting available_externally (optimizations)
definitions where those modules are used).
External functions defined in modules are emitted into the modular
object file as well (this may turn an existing ODR violation (if that
module were imported into multiple translations) into valid/linkable
code).
Internal symbols (static functions, for example) are not correctly
supported yet. The symbol will be produced, internal, in the modular
object - unreferenceable from the users.
Reviewers: rsmith
Differential Revision: https://reviews.llvm.org/D28845
llvm-svn: 293456
As Mehdi put it, entities should either be
available_externally+weak_odr, or linkonce_odr+linkonce_odr. While some
functions are emitted a_e/weak, their local variables were emitted
a_e/linkonce_odr.
While it might be nice to emit them a_e/weak, the Itanium ABI (& best
guess at MSVC's behavior as well) requires the local to be
linkonce/linkonce.
Reviewers: rsmith, mehdi_amini
Differential Revision: https://reviews.llvm.org/D29233
llvm-svn: 293344
This change adds a new type node, DeducedTemplateSpecializationType, to
represent a type template name that has been used as a type. This is modeled
around AutoType, and shares a common base class for representing a deduced
placeholder type.
We allow deduced class template types in a few more places than the standard
does: in conditions and for-range-declarators, and in new-type-ids. This is
consistent with GCC and with discussion on the core reflector. This patch
does not yet support deduced class template types being named in typename
specifiers.
llvm-svn: 293207
Under this defect resolution, the injected-class-name of a class or class
template cannot be used except in very limited circumstances (when declaring a
constructor, in a nested-name-specifier, in a base-specifier, or in an
elaborated-type-specifier). This is apparently done to make parsing easier, but
it's a pain for us since we don't know whether a template-id using the
injected-class-name is valid at the point when we annotate it (we don't yet
know whether the template-id will become part of an elaborated-type-specifier).
As a tentative resolution to a perceived language defect, mem-initializer-ids
are added to the list of exceptions here (they generally follow the same rules
as base-specifiers).
When the reference to the injected-class-name uses the 'typename' or 'template'
keywords, we permit it to be used to name a type or template as an extension;
other compilers also accept some cases in this area. There are also a couple of
corner cases with dependent template names that we do not yet diagnose, but
which will also get this treatment.
llvm-svn: 292518
FunctionDecls.
This commit silences an incorrect warning that is issued when a function
pointer is cast to another function pointer type. The warning gets
issued because alignments of the source and destination do not match in
Sema::CheckCastAlign, which happens because ASTContext::getTypeInfoImpl
and ASTContext::getDeclAlign return different values for functions (the
former returns 4 while the latter returns 1).
This should fix PR31558.
rdar://problem/29533528
Differential Revision: https://reviews.llvm.org/D27478
llvm-svn: 291253
Summary:
We compile user opencl kernel code with spir triple. But built-ins are written in OpenCL and we compile it with triple x86_64 to be able to use x86 intrinsics. And we need address spaces to match in both cases. So, we change fake address space map in OpenCL for matching with spir.
On CPU address spaces are not really important but we'd like to preserve address space information in order to perform optimizations relying on this info like enhanced alias analysis.
Reviewers: pekka.jaaskelainen, Anastasia
Subscribers: pekka.jaaskelainen, yaxunl, bader, cfe-commits
Differential Revision: https://reviews.llvm.org/D28048
llvm-svn: 290436
* In C++, never create a FunctionNoProtoType for a builtin (fixes C++1z
crasher from r289754).
* Fix type of __sync_synchronize to be a no-parameter function rather than a
varargs function. This matches GCC.
* Fix type of vfprintf to match its actual type. We gave it a wrong type due
to PR4290 (apparently autoconf generates invalid code and expects compilers
to choke it down or it miscompiles the program; the relevant error in clang
was downgraded to a warning in r122744 to fix other occurrences of this
autoconf brokenness, so we don't need this workaround any more).
* Turn off vararg argument checking for __noop, since it's not *really* a
varargs function. Alternatively we could add custom type checking for it
and synthesize parameter types matching the actual arguments in each call,
but that seemed like overkill.
llvm-svn: 290146
* a dependent non-type using-declaration within a function template can be
valid, as it can refer to an enumerator, so don't reject it in the template
definition
* we can partially substitute into a dependent using-declaration if it appears
within a (local class in a) generic lambda within a function template, which
means an UnresolvedUsing*Decl doesn't necessarily instantiate to a UsingDecl.
llvm-svn: 290071
In amdgcn target, null pointers in global, constant, and generic address space take value 0 but null pointers in private and local address space take value -1. Currently LLVM assumes all null pointers take value 0, which results in incorrectly translated IR. To workaround this issue, instead of emit null pointers in local and private address space, a null pointer in generic address space is emitted and casted to local and private address space.
Tentative definition of global variables with non-zero initializer will have weak linkage instead of common linkage since common linkage requires zero initializer and does not have explicit section to hold the non-zero value.
Virtual member functions getNullPointer and performAddrSpaceCast are added to TargetCodeGenInfo which by default returns ConstantPointerNull and emitting addrspacecast instruction. A virtual member function getNullPointerValue is added to TargetInfo which by default returns 0. Each target can override these virtual functions to get target specific null pointer and the null pointer value for specific address space, and perform specific translations for addrspacecast.
Wrapper functions getNullPointer is added to CodegenModule and getTargetNullPointerValue is added to ASTContext to facilitate getting the target specific null pointers and their values.
This change has no effect on other targets except amdgcn target. Other targets can provide support of non-zero null pointer in a similar way.
This change only provides support for non-zero null pointer for C and OpenCL. Supporting for other languages will be added later incrementally.
Differential Revision: https://reviews.llvm.org/D26196
llvm-svn: 289252
This adds the access qualifier to the Pipe Type, rather than using a class
hierarchy.
It also fixes mergeTypes for Pipes, by disallowing merges. Only identical
pipe types can be merged. The test case in invalid-pipes-cl2.0.cl is added
to check that.
llvm-svn: 288332
Summary:
We don't need a side table in ASTContext to hold CXXDefaultArgExprs. The
important part of building the CXXDefaultArgExprs was to ODR use the
default argument expressions, not to make AST nodes. Refactor the code
to only check the default argument, and remove the side table in
ASTContext which wasn't being serialized.
Fixes PR31121
Reviewers: thakis, rsmith, majnemer
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D27007
llvm-svn: 287774
Since array parameters decay to pointers, '_Nullable' and friends
should be available for use there as well. This is especially
important for parameters that are typedefs of arrays. The unsugared
syntax for this follows the syntax for 'static'-sized arrays in C:
void test(int values[_Nullable]);
This syntax was previously accepted but the '_Nullable' (and any other
attributes) were silently discarded. However, applying '_Nullable' to
a typedef was previously rejected and is now accepted; therefore, it
may be necessary to test for the presence of this feature:
#if __has_feature(nullability_on_arrays)
One important change here is that DecayedTypes don't always
immediately contain PointerTypes anymore; they may contain an
AttributedType instead. This only affected one place in-tree, so I
would guess it's not likely to cause problems elsewhere.
This commit does not change -Wnullability-completeness just yet. I
want to think about whether it's worth doing something special to
avoid breaking existing clients that compile with -Werror. It also
doesn't change '#pragma clang assume_nonnull' behavior, which
currently treats the following two declarations as equivalent:
#pragma clang assume_nonnull begin
void test(void *pointers[]);
#pragma clang assume_nonnull end
void test(void * _Nonnull pointers[]);
This is not the desired behavior, but changing it would break
backwards-compatibility. Most likely the best answer is going to be
adding a new warning.
Part of rdar://problem/25846421
llvm-svn: 286519
Extend lifetime of ExceptionTypeStorage, as it is referenced by
CanonicalEPI and used outside the block (ExceptionSpec.Exceptions is an
ArrayRef)
Patch by Sam McCall!
Differential Revision: https://reviews.llvm.org/D25983
llvm-svn: 285192
This has the following ABI impact:
1) Functions whose parameter or return types are non-throwing function pointer
types have different manglings in c++1z mode from prior modes. This is
necessary because c++1z permits overloading on the noexceptness of function
pointer parameter types. A warning is issued for cases that will change
manglings in c++1z mode.
2) Functions whose parameter or return types contain instantiation-dependent
exception specifications change manglings in all modes. This is necessary
to support overloading on / SFINAE in these exception specifications, which
a careful reading of the standard indicates has essentially always been
permitted.
Note that, in order to be affected by these changes, the code in question must
specify an exception specification on a function pointer/reference type that is
written syntactically within the declaration of another function. Such
declarations are very rare, and I have so far been unable to find any code
that would be affected by this. (Note that such things will probably become
more common in C++17, since it's a lot easier to get a noexcept function type
as a function parameter / return type there.)
This change does not affect the set of symbols produced by a build of clang,
libc++, or libc++abi.
llvm-svn: 285150
dependent noexcept specifications result in the same canonical function type.
We still use non-canonical hashing when deduplicating type sugar so that
diagnostics will point to the right place.
llvm-svn: 284457
This reverts commit r284176. It still marks some modules as invisible
that should be visible. Will follow up with the author with a test case.
llvm-svn: 284382
compiles without -fmodules-local-submodule-visibility. Original commit message:
[modules] When merging one definition into another, propagate the list of
re-exporting modules from the discarded definition to the retained definition.
llvm-svn: 284176
Summary: It doesn't need to be refcounted anymore, either.
Reviewers: timshen
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D25420
llvm-svn: 283768
For ObjC type parameter, we used to have TypedefType that is canonicalized to
id or the bound type. We can't represent "T <protocol>" and thus will lose
the type information in the following example:
@interface MyMutableDictionary<KeyType, ObjectType> : NSObject
- (void)setObject:(ObjectType)obj forKeyedSubscript:(KeyType <NSCopying>)key;
@end
MyMutableDictionary<NSString *, NSString *> *stringsByString;
NSNumber *n1, *n2;
stringsByString[n1] = n2;
--> no warning on type mismatch of the key.
To fix the problem, we introduce a new type ObjCTypeParamType that supports
a list of protocol qualifiers.
We create ObjCTypeParamType for ObjCTypeParamDecl when we create
ObjCTypeParamDecl. We also substitute ObjCTypeParamType instead of TypedefType
on an ObjCTypeParamDecl.
rdar://24619481
rdar://25060179
Differential Revision: http://reviews.llvm.org/D23080
llvm-svn: 281358
We also need to add ObjCTypeParamTypeLoc. ObjCTypeParamType supports the
representation of "T <protocol>" where T is a type parameter. Before this,
we use TypedefType to represent the type parameter for ObjC.
ObjCTypeParamType has "ObjCTypeParamDecl *OTPDecl" and it extends from
ObjCProtocolQualifiers. It is a non-canonical type and is canonicalized
to the underlying type with the protocol qualifiers.
rdar://24619481
rdar://25060179
Differential Revision: http://reviews.llvm.org/D23079
llvm-svn: 281355
To construct the canonical type of ObjCTypeParamType, we need to apply
qualifiers on ObjCObjectPointerType. The updated applyObjCProtocolQualifiers
handles this case by merging the protocol lists, constructing a new
ObjCObjectType, then a new ObjCObjectPointerType.
rdar://24619481
rdar://25060179
Differential Revision: http://reviews.llvm.org/D24059
llvm-svn: 281353
Check that ExpandStructures is true before visiting the list of ivars.
rdar://problem/27135221
Differential revision: https://reviews.llvm.org/D22929
llvm-svn: 278956
The size of image type is reported incorrectly as size of a pointer to address space 0, which causes error when casting image type to pointers by __builtin_astype.
The fix is to get image address space from TargetInfo then report the size accordingly.
Differential Revision: https://reviews.llvm.org/D22927
llvm-svn: 277647
Currently Clang use int32 to represent sampler_t, which have been a source of issue for some backends, because in some backends sampler_t cannot be represented by int32. They have to depend on kernel argument metadata and use IPA to find the sampler arguments and global variables and transform them to target specific sampler type.
This patch uses opaque pointer type opencl.sampler_t* for sampler_t. For each use of file-scope sampler variable, it generates a function call of __translate_sampler_initializer. For each initialization of function-scope sampler variable, it generates a function call of __translate_sampler_initializer.
Each builtin library can implement its own __translate_sampler_initializer(). Since the real sampler type tends to be architecture dependent, allowing it to be initialized by a library function simplifies backend design. A typical implementation of __translate_sampler_initializer could be a table lookup of real sampler literal values. Since its argument is always a literal, the returned pointer is known at compile time and easily optimized to finally become some literal values directly put into image read instructions.
This patch is partially based on Alexey Sotkin's work in Khronos Clang (3d4eec6162).
Differential Revision: https://reviews.llvm.org/D21567
llvm-svn: 277024
we first touch any part of that module. Instead, defer them until the first
time that module is (transitively) imported. The initializer step for a module
then recursively initializes modules that its own headers imported.
For example, this avoids running the <iostream> global initializer in programs
that don't actually use iostreams, but do use other parts of the standard
library.
llvm-svn: 276159
Summary:
Space for storing the //constraint-expression// of the
//requires-clause// associated with a `TemplateParameterList` is
arranged by taking a bit out of the `NumParams` field for the purpose
of determining whether there is a //requires-clause// or not, and by
adding to the trailing objects tied to the `TemplateParameterList`. An
accessor is provided.
An appropriate argument is supplied to `TemplateParameterList::Create`
at the various call sites.
Serialization changes will addressed as the Concepts implementation
becomes more solid.
Drive-by fix:
This change also replaces the custom
`FixedSizeTemplateParameterListStorage` implementation with one that
follows the interface provided by `llvm::TrailingObjects`.
Reviewers: aaron.ballman, faisalv, rsmith
Subscribers: cfe-commits, nwilson
Differential Revision: https://reviews.llvm.org/D19322
llvm-svn: 276069
member is redundantly redeclared outside the class definition in code built in
c++17 mode, ensure we emit a non-discardable definition of the data member for
c++11 and c++14 compilations to use.
llvm-svn: 274416
This patch adds a __nth_element builtin that allows fetching the n-th type of a
parameter pack with very little compile-time overhead. The patch was inspired by
r252036 and r252115 by David Majnemer, which add a similar __make_integer_seq
builtin for efficiently creating a std::integer_sequence.
Reviewed as D15421. http://reviews.llvm.org/D15421
llvm-svn: 274316
This patch corresponds to reviews:
http://reviews.llvm.org/D15120http://reviews.llvm.org/D19125
It adds support for the __float128 keyword, literals and target feature to
enable it. Based on the latter of the two aforementioned reviews, this feature
is enabled on Linux on i386/X86 as well as SystemZ.
This is also the second attempt in commiting this feature. The first attempt
did not enable it on required platforms which caused failures when compiling
type_traits with -std=gnu++11.
If you see failures with compiling this header on your platform after this
commit, it is likely that your platform needs to have this feature enabled.
llvm-svn: 268898
__global__ functions are a special case in CUDA.
Even when the symbol would normally not be externally
visible according to C++ rules, they still must be visible
in CUDA GPU object so host-side stub can launch them.
Differential Revision: http://reviews.llvm.org/D19748
llvm-svn: 268299
When comparing unqualified types, canonical types should be used, otherwise equivalent types may be treated as different type.
Differential Revision: http://reviews.llvm.org/D19662
llvm-svn: 267906
Since this patch provided support for the __float128 type but disabled it
on all platforms by default, some platforms can't compile type_traits with
-std=gnu++11 since there is a specialization with __float128.
This reverts the patch until D19125 is approved (i.e. we know which platforms
need this support enabled).
llvm-svn: 266460
This patch corresponds to review:
http://reviews.llvm.org/D15120
It adds support for the __float128 keyword, literals and a target feature to
enable it. This support is disabled by default on all targets and any target
that has support for this type is free to add it.
Based on feedback that I've received from target maintainers, this appears to
be the right thing for most targets. I have not heard from the maintainers of
X86 which I believe supports this type. I will subsequently investigate the
impact of enabling this on X86.
llvm-svn: 266186
Putting OpenCLImageTypes.def to clangAST library violates layering requirement: "It's not OK for a Basic/ header to include an AST/ header".
This fixes the modules build.
Differential revision: http://reviews.llvm.org/D18954
Reviewers: Richard Smith, Vassil Vassilev.
llvm-svn: 266180
I. Current implementation of images is not conformant to spec in the following points:
1. It makes no distinction with respect to access qualifiers and therefore allows to use images with different access type interchangeably. The following code would compile just fine:
void write_image(write_only image2d_t img);
kernel void foo(read_only image2d_t img) { write_image(img); } // Accepted code
which is disallowed according to s6.13.14.
2. It discards access qualifier on generated code, which leads to generated code for the above example:
call void @write_image(%opencl.image2d_t* %img);
In OpenCL2.0 however we can have different calls into write_image with read_only and wite_only images.
Also generally following compiler steps have no easy way to take different path depending on the image access: linking to the right implementation of image types, performing IR opts and backend codegen differently.
3. Image types are language keywords and can't be redeclared s6.1.9, which can happen currently as they are just typedef names.
4. Default access qualifier read_only is to be added if not provided explicitly.
II. This patch corrects the above points as follows:
1. All images are encapsulated into a separate .def file that is inserted in different points where image handling is required. This avoid a lot of code repetition as all images are handled the same way in the code with no distinction of their exact type.
2. The Cartesian product of image types and image access qualifiers is added to the builtin types. This simplifies a lot handling of access type mismatch as no operations are allowed by default on distinct Builtin types. Also spec intended access qualifier as special type qualifier that are combined with an image type to form a distinct type (see statement above - images can't be created w/o access qualifiers).
3. Improves testing of images in Clang.
Author: Anastasia Stulova
Reviewers: bader, mgrang.
Subscribers: pxli168, pekka.jaaskelainen, yaxunl.
Differential Revision: http://reviews.llvm.org/D17821
llvm-svn: 265783
Add parsing, sema analysis for 'declare target' construct for OpenMP 4.0
(4.5 support will be added in separate patch).
The declare target directive specifies that variables, functions (C, C++
and Fortran), and subroutines (Fortran) are mapped to a device. The declare
target directive is a declarative directive. In Clang declare target is
implemented as implicit attribute for the declaration.
The syntax of the declare target directive is as follows:
#pragma omp declare target
declarations-definition-seq
#pragma omp end declare target
Based on patch from Michael Wong http://reviews.llvm.org/D15321
llvm-svn: 265530
Summary:
When the code is compiled for arm32 and the builtin `__va_list` declaration is created by `CreateAAPCSABIBuiltinVaListDecl`, the declaration is not saved in the `ASTContext` which may lead to a compilation error or crash.
Minimal reproducer I was able to find:
**header.h**
```
#include <stdarg.h>
typedef va_list va_list_1;
```
**test.cpp**
```
typedef __builtin_va_list va_list_2;
void foo(const char* format, ...) { va_list args; va_start( args, format ); }
```
Steps to reproduce:
```
clang -x c++-header --target=armv7l-linux-eabihf header.h
clang -c -include header.h --target=armv7l-linux-eabihf test.cpp
```
Compilation error:
```
error: non-const lvalue reference to type '__builtin_va_list'
cannot bind to a value of unrelated type 'va_list' (aka '__builtin_va_list')
```
Compiling the same code as a C source leads to a crash:
```
clang --target=armv7l-linux-eabihf header.h
clang -c -x c -include header.h --target=armv7l-linux-eabihf test.cpp
```
Reviewers: logan, rsmith
Subscribers: cfe-commits, asl, aemerson, rengolin
Differential Revision: http://reviews.llvm.org/D18557
llvm-svn: 264930
This is like r262493, but for pragma detect_mismatch instead of pragma comment.
The two pragmas have similar behavior, so use the same approach for both.
llvm-svn: 262506
`#pragma comment` was handled by Sema calling a function on ASTConsumer, and
CodeGen then implementing this function and writing things to its output.
Instead, introduce a PragmaCommentDecl AST node and hang one off the
TranslationUnitDecl for every `#pragma comment` line, and then use the regular
serialization machinery. (Since PragmaCommentDecl has codegen relevance, it's
eagerly deserialized.)
http://reviews.llvm.org/D17799
llvm-svn: 262493
to allow arbitrary data to be associated with a parameter.
Also, fix a bug where we apparently haven't been serializing
this information for the last N years.
llvm-svn: 262278
Per review feedback the name was wrong and it can be used outside
Objective-C.
Unfortunately, making the internal struct visible broke some ASTMatchers
tests that assumed that the first record decl would be from user code,
rather than a builtin type. I'm worried that this will also affect
users' code. So this patch adds a typedef to wrap the internal struct
and only makes the typedef visible to namelookup. This is sufficient to
allow the ASTReader to merge the decls we need without making the struct
itself visible.
rdar://problem/24425801
llvm-svn: 259734
Original message:
Make CF constant string decl visible to name lookup to fix module errors
The return type of the __builtin___*StringMakeConstantString functions
is a pointer to a struct, so we need that struct to be visible to name
lookup so that we will correctly merge multiple declarations of that
type if they come from different modules.
Incidentally, to make this visible to name lookup we need to rename the
type to __NSConstantString, since the real NSConstantString is an
Objective-C interface type. This shouldn't affect anyone outside the
compiler since users of the constant string builtins cast the result
immediately to CFStringRef.
Since this struct type is otherwise implicitly created by the AST
context and cannot access namelookup, we make this a predefined type
and initialize it in Sema.
Note: this issue of builtins that refer to types not visible to name
lookup technically also affects other builtins (e.g. objc_msgSendSuper),
but in all other cases the builtin is a library builtin and the issue
goes away if you include the library that defines the types it uses,
unlike for these constant string builtins.
rdar://problem/24425801
llvm-svn: 259721
The return type of the __builtin___*StringMakeConstantString functions
is a pointer to a struct, so we need that struct to be visible to name
lookup so that we will correctly merge multiple declarations of that
type if they come from different modules.
Incidentally, to make this visible to name lookup we need to rename the
type to __NSConstantString, since the real NSConstantString is an
Objective-C interface type. This shouldn't affect anyone outside the
compiler since users of the constant string builtins cast the result
immediately to CFStringRef.
Since this struct type is otherwise implicitly created by the AST
context and cannot access namelookup, we make this a predefined type
and initialize it in Sema.
Note: this issue of builtins that refer to types not visible to name
lookup technically also affects other builtins (e.g. objc_msgSendSuper),
but in all other cases the builtin is a library builtin and the issue
goes away if you include the library that defines the types it uses,
unlike for these constant string builtins.
rdar://problem/24425801
llvm-svn: 259624
Summary:
Support for OpenCL 2.0 pipe type.
This is a bug-fix version for bader's patch reviews.llvm.org/D14441
Reviewers: pekka.jaaskelainen, Anastasia
Subscribers: bader, Anastasia, cfe-commits
Differential Revision: http://reviews.llvm.org/D15603
llvm-svn: 257254
for types which are used as pointees in PointerUnions, PointerIntPairs,
and DenseMap pointer keys.
This is part of a series of patches to allow LLVM to check for complete
pointee types when computing its pointer traits. This is absolutely
necessary to get correct (or reproducible) results for things like how
many low bits are guaranteed to be zero.
I think this is the last patch for getting Clang clean here!!!
llvm-svn: 256615
keys, and PointerIntPairs where the pointee types are incomplete
out-of-line to where we have the complete type.
This is the standard pattern used throughout the AST library to address
the inherently mutually cross referenced nature of the AST.
This is part of a series of patches to allow LLVM to check for complete
pointee types when computing its pointer traits. This is absolutely
necessary to get correct (or reproducible) results for things like how
many low bits are guaranteed to be zero.
llvm-svn: 256612
SmallVector<.., 16> (16!!!!) objects to a simple SmallVector of pairs.
This no longer de-duplicates the common function pointers used during
deallocation, but this doesn't really seem worth the complexity and
overhead of managing the map-of-vectors. Notably, there is no reason to
assume that functions have the 4-byte alignment that DenseMap relies on,
and indeed this prevents checking the alignment of the DenseMap keys
because we can't even meaningfully query the alignment of functions
using our existing alignment tools.
Generally, function pointers don't seem like a great idea for keys in
a DenseMap. =]
I chatted with Richard Smith about this a bit as well and have written
down a FIXME because this *does* waste some memory and in general seems
a very clumsy memory management strategy. He would like to see a more
fundamental fix eventually here that tries to build a better pattern.
llvm-svn: 256610
When RAV traverses a Stmt or Expr node, if the corresponding Traverse*
functions have not been overridden, it will now use data recursion to walk
those nodes. We arrange this to be an unobservable optimization to RAV
subclasses, and to gracefully degrade as parts of the visitation are overridden
with functions that might observe the visitation.
For instance, if an RAV subclass overrides TraverseUnaryNot, we will ensure
that there are real recursive stack frames for those traversals, but we'll
use data recursion for all other traversals.
This removes the need for DataRecursiveASTVisitor, and for the
'shouldUseDataRecursionFor' extension point, both of which are removed by this
change.
llvm-svn: 253948
https://gcc.gnu.org/onlinedocs/gcc/Typeof.html
Differences from the GCC extension:
* __auto_type is also permitted in C++ (but only in places where
it could appear in C), allowing its use in headers that might
be shared across C and C++, or used from C++98
* __auto_type can be combined with a declarator, as with C++ auto
(for instance, "__auto_type *p")
* multiple variables can be declared in a single __auto_type
declaration, with the C++ semantics (the deduced type must be
the same in each case)
This patch also adds a missing restriction on applying typeof to
a bit-field, which GCC has historically rejected in C (due to
lack of clarity as to whether the operand should be promoted).
The same restriction also applies to __auto_type in C (in both
GCC and Clang).
This also fixes PR25449.
Patch by Nicholas Allegra!
llvm-svn: 252690
This new builtin template allows for incredibly fast instantiations of
templates like std::integer_sequence.
Performance numbers follow:
My work station has 64 GB of ram + 20 Xeon Cores at 2.8 GHz.
__make_integer_seq<std::integer_sequence, int, 90000> takes 0.25
seconds.
std::make_integer_sequence<int, 90000> takes unbound time, it is still
running. Clang is consuming gigabytes of memory.
Differential Revision: http://reviews.llvm.org/D13786
llvm-svn: 252036
This sets the mostly expected Darwin default ABI options for these two
platforms. Active changes from these defaults for watchOS are in a later patch.
llvm-svn: 251708
This relands r250831 after some fixes to shrink the ParentMap overall
with one addtional tweak: nodes with pointer identity (e.g. Decl* and
friends) can be store more efficiently so I put them in a separate map.
All other nodes (so far only TypeLoc and NNSLoc) go in a different map
keyed on DynTypedNode. This further uglifies the code but significantly
reduces memory overhead.
Overall this change still make ParentMap significantly larger but it's
nowhere as bad as before. I see about 25 MB over baseline (pre-r251008)
on X86ISelLowering.cpp. If this becomes an issue we could consider
splitting the maps further as DynTypedNode is still larger (32 bytes)
than a single TypeLoc (16 bytes) but I didn't want to introduce even
more complexity now.
Differential Revision: http://reviews.llvm.org/D14011
llvm-svn: 251101
Previously, __weak was silently accepted and ignored in MRC mode.
That makes this a potentially source-breaking change that we have to
roll out cautiously. Accordingly, for the time being, actual support
for __weak references in MRC is experimental, and the compiler will
reject attempts to actually form such references. The intent is to
eventually enable the feature by default in all non-GC modes.
(It is, of course, incompatible with ObjC GC's interpretation of
__weak.)
If you like, you can enable this feature with
-Xclang -fobjc-weak
but like any -Xclang option, this option may be removed at any point,
e.g. if/when it is eventually enabled by default.
This patch also enables the use of the ARC __unsafe_unretained qualifier
in MRC. Unlike __weak, this is being enabled immediately. Since
variables are essentially __unsafe_unretained by default in MRC,
the only practical uses are (1) communication and (2) changing the
default behavior of by-value block capture.
As an implementation matter, this means that the ObjC ownership
qualifiers may appear in any ObjC language mode, and so this patch
removes a number of checks for getLangOpts().ObjCAutoRefCount
that were guarding the processing of these qualifiers. I don't
expect this to be a significant drain on performance; it may even
be faster to just check for these qualifiers directly on a type
(since it's probably in a register anyway) than to do N dependent
loads to grab the LangOptions.
rdar://9674298
llvm-svn: 251041
These are by far the most common types to be parents in the AST so it makes
sense to optimize for them. Put them directly into the value of the map.
This currently saves 32 bytes per parent in the map and a pointer
indirection at the cost of some additional complexity in the code.
Sadly this means we cannot return an ArrayRef from getParents anymore, add
a proxy class that can own a single DynTypedNode and otherwise behaves
exactly the same as ArrayRef.
For example on a random large file (X86ISelLowering.cpp) this reduces the
size of the parent map by 24 MB.
Differential Revision: http://reviews.llvm.org/D13976
llvm-svn: 251008
Putting DynTypedNode in the ParentMap bloats its memory foot print.
Before the void* key had 8 bytes, now we're at 40 bytes per key which
can mean multiple gigabytes increase for large ASTs and this count
doesn't even include all the added TypeLoc nodes. Revert until I come
up with a better data structure.
This reverts commit r250831.
llvm-svn: 250889
Firstly this changes the type of parent map to be keyed on DynTypedNode to
simplify the following changes. This comes with a DenseMapInfo for
DynTypedNode, which is a bit incomplete still and will probably only work
for parentmap right now.
Then the RecursiveASTVisitor in ASTContext is updated and finally
ASTMatchers hasParent and hasAncestor learn about the new functionality.
Now ParentMap is only missing TemplateArgumentLocs and CXXCtorInitializers.
Differential Revision: http://reviews.llvm.org/D13897
llvm-svn: 250831
Summary: It breaks the build for the ASTMatchers
Subscribers: klimek, cfe-commits
Differential Revision: http://reviews.llvm.org/D13893
llvm-svn: 250827
Out-of-line definitions of static data members which have an inline
initializer must get GVA_DiscardableODR linkage instead of
GVA_StrongExternal linkage.
MSVC 2013's behavior is different with respect to this and would cause
link errors if one TU provided a definition while another did not.
MSVC 2015 fixed this bug, making this OK. Note that the 2015 behavior
is always compatible with 2013: it never produces a strong definition.
This essentially reverts r237787.
llvm-svn: 250757
No ABI for C++ currently makes it possible to implement the standard
100% perfectly. We wrongly hid some of our compatible behavior behind
-fms-compatibility instead of tying it to the compiler ABI.
llvm-svn: 249656
Objective-C ARC lifetime qualifiers are dropped when canonicalizing
function types. Perform the same adjustment before comparing the
deduced result types of lambdas. Fixes rdar://problem/22344904.
llvm-svn: 249065
Adjust __global__ functions with DiscardableODR linkage to use
StrongODR linkage instead, so they are visible externally.
Differential Revision: http://reviews.llvm.org/D13067
llvm-svn: 248400
* adds -aux-triple option to specify target triple
* propagates aux target info to AST context and Preprocessor
* pulls in target specific preprocessor macros.
* pulls in target-specific builtins from aux target.
* sets appropriate host or device attribute on builtins.
Differential Revision: http://reviews.llvm.org/D12917
llvm-svn: 248299
Summary:
This change adds support for `__builtin_ms_va_list`, a GCC extension for
variadic `ms_abi` functions. The existing `__builtin_va_list` support is
inadequate for this because `va_list` is defined differently in the Win64
ABI vs. the System V/AMD64 ABI.
Depends on D1622.
Reviewers: rsmith, rnk, rjmccall
CC: cfe-commits
Differential Revision: http://reviews.llvm.org/D1623
llvm-svn: 247941
This implements basic support for compiling (though not yet assembling
or linking) for a WebAssembly target. Note that ABI details are not yet
finalized, and may change.
Differential Revision: http://reviews.llvm.org/D12002
llvm-svn: 246814
A class without a name for linkage purposes gets a name along the lines
of <unnamed-type-foo> where foo is either the name of a declarator which
defined it (like a variable or field) or a
typedef-name (like a typedef or alias-declaration).
We handled the declarator case correctly but it would fall down during
template instantiation if the declarator didn't share the tag's type.
We failed to handle the typedef-name case at all.
Instead, keep track of the association between the two and keep it up to
date in the face of template instantiation.
llvm-svn: 246469
Adds parsing/sema analysis/serialization/deserialization for array sections in OpenMP constructs (introduced in OpenMP 4.0).
Currently it is allowed to use array sections only in OpenMP clauses that accepts list of expressions.
Differential Revision: http://reviews.llvm.org/D10732
llvm-svn: 245937
We risk iterator invalidation issues if we use a DenseMap to hold the
backing storage for an APValue. Instead, BumpPtrAllocate them and
use APValue * as our DenseMap value.
Also, don't assume that MaterializedGlobalTemporaryMap won't regrow
between when we initially perform a lookup and later on when we actually
try to insert into it.
This fixes PR24289.
Differential Revision: http://reviews.llvm.org/D11629
llvm-svn: 244989
chain and fix the cases where it fires.
* Handle the __va_list_tag as a predefined decl. Previously we failed to merge
sometimes it because it's not visible to name lookup. (In passing, remove
redundant __va_list_tag typedefs that we were creating for some ABIs. These
didn't affect the mangling or representation of the type.)
* For Decls derived from Redeclarable that are not in fact redeclarable
(implicit params, function params, ObjC type parameters), remove them from
the list of expected redeclarable decls.
llvm-svn: 243259
We treated DependentSizedArrayTypes with the same element type but
differing size expressions as equivalently canonical. This would lead
to bizarre behavior during template instantiation.
This fixes PR24212.
llvm-svn: 243093
Normally, explicit specializations are treated like strong external
definitions. However, MSVC treats explicit specializations of static
data members as weak. MSVC 2013's <regex> implementation has such an
explicit specialization which leads to clang emitting a strong
definition in each translation unit which includes it. Tweak clang's
linkage calculation to give such entities GVA_StrongODR linkage instead.
This fixes PR24165.
llvm-svn: 242592
Introduce co- and contra-variance for Objective-C type parameters,
which allows us to express that (for example) an NSArray is covariant
in its type parameter. This means that NSArray<NSMutableString *> * is
a subtype of NSArray<NSString *> *, which is expected of the immutable
Foundation collections.
Type parameters can be annotated with __covariant or __contravariant
to make them co- or contra-variant, respectively. This feature can be
detected by __has_feature(objc_generics_variance). Implements
rdar://problem/20217490.
llvm-svn: 241549
The __kindof type qualifier can be applied to Objective-C object
(pointer) types to indicate id-like behavior, which includes implicit
"downcasting" of __kindof types to subclasses and id-like message-send
behavior. __kindof types provide better type bounds for substitutions
into unspecified generic types, which preserves more type information.
llvm-svn: 241548
The Objective-C common-type computation had a few problems that
required a significant rework, including:
- Quadradic behavior when finding the common base type; now it's
linear.
- Keeping around type arguments when computing the common type
between a specialized and an unspecialized type
- Introducing redundant protocol qualifiers.
Part of rdar://problem/6294649. Also fixes rdar://problem/19572837 by
addressing a longstanding bug in
ASTContext::CollectInheritedProtocols().
llvm-svn: 241544
When messaging a method that was defined in an Objective-C class (or
category or extension thereof) that has type parameters, substitute
the type arguments for those type parameters. Similarly, substitute
into property accesses, instance variables, and other references.
This includes general infrastructure for substituting the type
arguments associated with an ObjCObject(Pointer)Type into a type
referenced within a particular context, handling all of the
substitutions required to deal with (e.g.) inheritance involving
parameterized classes. In cases where no type arguments are available
(e.g., because we're messaging via some unspecialized type, id, etc.),
we substitute in the type bounds for the type parameters instead.
Example:
@interface NSSet<T : id<NSCopying>> : NSObject <NSCopying>
- (T)firstObject;
@end
void f(NSSet<NSString *> *stringSet, NSSet *anySet) {
[stringSet firstObject]; // produces NSString*
[anySet firstObject]; // produces id<NSCopying> (the bound)
}
When substituting for the type parameters given an unspecialized
context (i.e., no specific type arguments were given), substituting
the type bounds unconditionally produces type signatures that are too
strong compared to the pre-generics signatures. Instead, use the
following rule:
- In covariant positions, such as method return types, replace type
parameters with “id” or “Class” (the latter only when the type
parameter bound is “Class” or qualified class, e.g,
“Class<NSCopying>”)
- In other positions (e.g., parameter types), replace type
parameters with their type bounds.
- When a specialized Objective-C object or object pointer type
contains a type parameter in its type arguments (e.g.,
NSArray<T>*, but not NSArray<NSString *> *), replace the entire
object/object pointer type with its unspecialized version (e.g.,
NSArray *).
llvm-svn: 241543
Objective-C type arguments can be provided in angle brackets following
an Objective-C interface type. Syntactically, this is the same
position as one would provide protocol qualifiers (e.g.,
id<NSCopying>), so parse both together and let Sema sort out the
ambiguous cases. This applies both when parsing types and when parsing
the superclass of an Objective-C class, which can now be a specialized
type (e.g., NSMutableArray<T> inherits from NSArray<T>).
Check Objective-C type arguments against the type parameters of the
corresponding class. Verify the length of the type argument list and
that each type argument satisfies the corresponding bound.
Specializations of parameterized Objective-C classes are represented
in the type system as distinct types. Both specialized types (e.g.,
NSArray<NSString *> *) and unspecialized types (NSArray *) are
represented, separately.
llvm-svn: 241542
Produce type parameter declarations for Objective-C type parameters,
and attach lists of type parameters to Objective-C classes,
categories, forward declarations, and extensions as
appropriate. Perform semantic analysis of type bounds for type
parameters, both in isolation and across classes/categories/extensions
to ensure consistency.
Also handle (de-)serialization of Objective-C type parameter lists,
along with sundry other things one must do to add a new declaration to
Clang.
Note that Objective-C type parameters are typedef name declarations,
like typedefs and C++11 type aliases, in support of type erasure.
Part of rdar://problem/6294649.
llvm-svn: 241541
Virtual inheritance member pointers are always relative to the vbindex,
even when the member pointer doesn't point into a virtual base. This is
corrected by adjusting the non-virtual offset backwards from the vbptr
back to the top of the most derived class. While we performed this
adjustment when manifesting member pointers as constants or when
performing conversions, we didn't perform the adjustment when mangling
them.
llvm-svn: 240453
While the rest of the Objective-C metadata seems to honor
objc_runtime_name, the encoding strings produced by, e.g., @encode and
property meta, were not. Fixes rdar://problem/21408305.
llvm-svn: 239852
There are 3 cases of defining static const member:
initialized inside the class, not defined outside the class.
initialized inside the class, defined outside the class.
not initialized inside the class, defined outside the class.
Revision r213304 was supposed to fix the linkage problem of case (1), but mistakenly it made case (2) behave the same.
As a result, out-of-line definition of static data member is not handled correctly.
Proposed patch distinguishes between cases (1) and (2) and allows to properly emit static const members under –fms-compatibility option.
This fixes http://llvm.org/PR21164.
Differential Revision: http://reviews.llvm.org/D9850
llvm-svn: 237787
With this change, enabling -fmodules-local-submodule-visibility results in name
visibility rules being applied to submodules of the current module in addition
to imported modules (that is, names no longer "leak" between submodules of the
same top-level module). This also makes it much safer to textually include a
non-modular library into a module: each submodule that textually includes that
library will get its own "copy" of that library, and so the library becomes
visible no matter which including submodule you import.
llvm-svn: 237473
The GCC construct __attribute__((aligned)) is defined to set alignment
to "the default alignment for the target architecture" according to
the GCC documentation:
The default alignment is sufficient for all scalar types, but may not be
enough for all vector types on a target that supports vector operations.
The default alignment is fixed for a particular target ABI.
clang currently hard-coded an alignment of 16 bytes for that construct,
which is correct on some platforms (including X86), but wrong on others
(including SystemZ). Since this value is ABI-relevant, it is important
to get correct for compatibility purposes.
This patch adds a new TargetInfo member "DefaultAlignForAttributeAligned"
that targets can set to the appropriate default __attribute__((aligned))
value.
Note that I'm deliberately *not* using the existing "SuitableAlign"
value, which is used to set the pre-defined macro __BIGGEST_ALIGNMENT__,
since those two values may not be the same on all platforms. In fact,
on X86, __attribute__((aligned)) always uses 16-byte alignment, while
__BIGGEST_ALIGNMENT__ may be larger if AVX-2 or AVX-512 are supported.
(This is actually not yet correctly implemented in clang either.)
The patch provides a value for DefaultAlignForAttributeAligned only for
SystemZ, and leaves the default for all other targets at 16, which means
no visible change in behavior on all other targets. (The value is still
wrong for some other targets, but I'd prefer to leave it to the target
maintainers for those platforms to fix.)
llvm-svn: 235397
SystemZ prefers to align all global variables to two bytes, which is
implemented by setting the TargetInfo member MinGlobalAlign.
However, for compatibility with existing compilers this should *not*
change the ABI alignment value as retrieved via __alignof__, which
it currently does.
This patch fixes the issue by having ASTContext::getDeclAlign ignore
the MinGlobalAlign setting in the ForAlignof case.
Since SystemZ is the only platform setting MinGlobalAlign, this should
cause no change for any other target.
llvm-svn: 235395
It was documented as 8 and operator new[] defaults to 8, but the normal
operator new was never updated and happily wasted bytes on every other
allocation.
We still have to allocate all Types with 16 byte alignment, update the
allocation calls for Types that were missing explicit alignment.
llvm-svn: 233922
for a DeclContext, and fix propagation of exception specifications along
redeclaration chains.
This reverts r232905, r232907, and r232907, which reverted r232793, r232853,
and r232853.
One additional change is present here to resolve issues with LLDB: distinguish
between whether lexical decls missing from the lookup table are local or are
provided by the external AST source, and still look in the external source if
that's where they came from.
llvm-svn: 232928
When we need to build the lookup table for a DeclContext, we used to pull in
all lexical declarations for the context; instead, just build a lookup table
for the local lexical declarations. We previously didn't guarantee that the
imported declarations would be in the returned map, but in some cases we'd
happen to put them all in there regardless. Now we're even lazier about this.
This unnecessary work was papering over some other bugs:
- LookupVisibleDecls would use the DC for name lookups in the TU in C, and
this was not guaranteed to find all imported names (generally, the DC for
the TU in C is not a reliable place to perform lookups). We now use an
identifier-based lookup mechanism for this.
- We didn't actually load in the list of eagerly-deserialized declarations
when importing a module (so external definitions in a module wouldn't be
emitted by users of those modules unless they happened to be deserialized
by the user of the module).
llvm-svn: 232793
We assumed that the most recent declaration of an inline function would
also be inline. However, a more recent declaration can come from a
friend declaration in a class template that is instantiated at the
definition of the function.
llvm-svn: 232786
std::make_exception_ptr calls std::__GetExceptionInfo in order to figure
out how to properly copy the exception object.
Differential Revision: http://reviews.llvm.org/D8280
llvm-svn: 232188
This adds support for copy-constructor closures. These are generated
when the C++ runtime has to call a copy-constructor with a particular
calling convention or with default arguments substituted in to the call.
Because the runtime has no mechanism to call the function with a
different calling convention or know-how to evaluate the default
arguments at run-time, we create a thunk which will do all the
appropriate work and package it in a way the runtime can use.
Differential Revision: http://reviews.llvm.org/D8225
llvm-svn: 231952
of extern "C" declarations. This is simpler and vastly more efficient for
modules builds (we no longer need to load *all* extern "C" declarations to
determine if we have a redeclaration).
No functionality change intended.
llvm-svn: 231538
Find all unambiguous public classes of the exception object's class type
and reference all of their copy constructors. Yes, this is not
conforming but it is necessary in order to implement their ABI. This is
because the copy constructor is actually referenced by the metadata
describing which catch handlers are eligible to handle the exception
object.
N.B. This doesn't yet handle the copy constructor closure case yet,
that work is ongoing.
Differential Revision: http://reviews.llvm.org/D8101
llvm-svn: 231499
Summary:
Allow user to provide multiple blacklists by passing several
-fsanitize-blacklist= options. These options now don't override
default blacklist from Clang resource directory, which is always
applied (which fixes PR22431).
-fno-sanitize-blacklist option now disables all blacklists that
were specified earlier in the command line (including the default
one).
This change depends on http://reviews.llvm.org/D7367.
Test Plan: regression test suite
Reviewers: timurrrr
Subscribers: cfe-commits, kcc, pcc
Differential Revision: http://reviews.llvm.org/D7368
llvm-svn: 228156
We didn't consider any alignment attributes on an EnumDecl when
calculating alignment.
While we are here, ignore alignment specifications on typedef types if
one is used as the underlying type. Otherwise, weird things happen:
enum Y : int;
Y y;
typedef int __attribute__((aligned(64))) u;
enum Y : u {};
What is the alignment of 'Y'? It would be more consistent with the
overall design of enums with fixed underlying types to consider the
underlying type's UnqualifiedDesugaredType.
This fixes PR22279.
llvm-svn: 226653
The test was fixed after a discussion with the revision author: the check
pattern was made more flexible as the "%call" part is not what we actually want
to check strictly there.
The original patch description:
===
Introduce SPIR calling conventions.
This implements Section 3.7 from the SPIR 1.2 spec:
SPIR kernels should use "spir_kernel" calling convention.
Non-kernel functions use "spir_func" calling convention. All
other calling conventions are disallowed.
The patch works only for OpenCL source. Any other uses will need
to ensure that kernels are assigned the spir_kernel calling
convention correctly.
===
llvm-svn: 226561
This implements Section 3.7 from the SPIR 1.2 spec:
SPIR kernels should use "spir_kernel" calling convention.
Non-kernel functions use "spir_func" calling convention. All
other calling conventions are disallowed.
The patch works only for OpenCL source. Any other uses will need
to ensure that kernels are assigned the spir_kernel calling
convention correctly.
llvm-svn: 226548
Types composed with certain implicit record types would have their RTTI
marked as hidden because the implicit record type didn't have any
visibility.
This manifests itself as triggering false positives from tools like
clang's -fsantize=function feature. The RTTI for a function type's
return type wouldn't match if the return type was an implicit record
type.
Patch by Stephan Bergmann!
llvm-svn: 226148
Also add a few asserts for this. The existing code assumes this in a bunch
of places already (see e.g. the assert at the top of ParseTypedefDecl(), and
there are many unchecked calls on the result of GetTypeForDeclarator()), and
from looking through the code this should always be true from what I can tell.
This allows removing ASTContext::getNullTypeSourceInfo() too as that's now
unused.
No behavior change intended.
llvm-svn: 225125
To support it in the frontend, the following has been added:
- generic address space type attribute;
- documentation for the OpenCL address space attributes;
- parsing of __generic(generic) keyword;
- test code for the parser and diagnostics.
llvm-svn: 222831
For all threadprivate variables which have constructor/destructor emit call to void __kmpc_threadprivate_register(ident_t * <Current Location>, void *<Original Global Addr>, kmpc_ctor <Constructor>, kmpc_cctor NULL, kmpc_dtor <Destructor>);
In expressions all references to such variables are replaced by calls to void *__kmpc_threadprivate_cached(ident_t *<Current Location>, kmp_int32 <Current Thread Id>, void *<Original Global Addr>, size_t <Size of Data>, void ***<Pointer to autogenerated cache – array of private copies of threadprivate variable>);
Test test/OpenMP/threadprivate_codegen.cpp checks that codegen is correct. Also it checks that codegen is correct after serialization/deserialization and one of passes verifies debug info.
Differential Revision: http://reviews.llvm.org/D4002
llvm-svn: 221663
Get rid of ugly SanitizerOptions class thrust into LangOptions:
* Make SanitizeAddressFieldPadding a regular language option,
and rely on default behavior to initialize/reset it.
* Make SanitizerBlacklistFile a regular member LangOptions.
* Introduce the helper class "SanitizerSet" to represent the
set of enabled sanitizers and make it a member of LangOptions.
It is exactly the entity we want to cache and modify in CodeGenFunction,
for instance. We'd also be able to reuse SanitizerSet in
CodeGenOptions for storing the set of recoverable sanitizers,
and in the Driver to represent the set of sanitizers
turned on/off by the commandline flags.
No functionality change.
llvm-svn: 221653
Plumb through the full QualType of the TemplateArgument::Declaration, as
it's insufficient to only know whether the type is a reference or
pointer (that was necessary for mangling, but insufficient for debug
info). This shouldn't increase the size of TemplateArgument as
TemplateArgument::Integer is still longer by another 32 bits.
Several bits of code were testing that the reference-ness of the
parameters matched, but this seemed to be insufficient (various other
features of the type could've mismatched and wouldn't've been caught)
and unnecessary, at least insofar as removing those tests didn't cause
anything to fail.
(Richard - perchaps you can hypothesize why any of these checks might
need to test reference-ness of the parameters (& explain why
reference-ness is part of the mangling - I would've figured that for the
reference-ness to be different, a prior template argument would have to
be different). I'd be happy to add them in/beef them up and add test
cases if there's a reason for them)
llvm-svn: 219900
Soon we'll need to have access to blacklist before the CodeGen
phase (see http://reviews.llvm.org/D5687), so parse and construct
the blacklist earlier.
llvm-svn: 219857
We build a NestedNameSpecifier that records the CXXRecordDecl in which
__super appeared. Name lookup is performed in all base classes of the
recorded CXXRecordDecl. Use of __super is allowed only inside class and
member function scope.
llvm-svn: 218484
we were failing to find that bit-field when performing integer promotions. This
brings us closer to following the standard, and closer to GCC.
In C, this change is technically a regression: we get bit-field promotions
completely wrong in C, promoting cases that are categorically not bit-field
designators. This change makes us do so slightly more consistently, though.
llvm-svn: 218428
FunctionProtoType::ExtProtoInfo. Most of the users of these fields don't care
about the other ExtProtoInfo bits and just want to talk about the exception
specification.
llvm-svn: 214450
Iterator invalidation issues already force us to do one lookup and one
insert.
Don't use the particular bit-pattern of the 'Align' field to determine
whether or not we have already inserted into the TypeInfo DenseMap;
instead ask for an iterator to the TypeInfo entry.
llvm-svn: 214293
The MS ABI has a notion of 'required alignment' for fields; this
alignment supercedes pragma pack directives.
MSVC takes into account alignment attributes on typedefs when
determining whether or not a field has a certain required alignment.
Do the same in clang by tracking whether or not we saw such an attribute
when calculating the type's bitwidth and alignment.
This fixes PR20418.
Reviewers: rnk
Differential Revision: http://reviews.llvm.org/D4714
llvm-svn: 214274
This makes us emit dllexported in-class initialized static data members (which
are treated as definitions in MSVC), even when they're not referenced.
It also makes their special linkage reflected in the GVA linkage instead of
getting massaged in CodeGen.
Differential Revision: http://reviews.llvm.org/D4563
llvm-svn: 213304
Summary:
Do not store duplicate parents when memoization data is available.
This does not solve the duplication problem, but ameliorates it.
Reviewers: klimek
Subscribers: klimek, cfe-commits
Differential Revision: http://reviews.llvm.org/D4124
llvm-svn: 210902
The backing store of thread local variables is internal for OS X and all
accesses must go through the thread wrapper.
However, individual TUs may have inlined through the thread wrapper.
To fix this, give the thread wrapper functions WeakAnyLinkage. This
prevents them from getting inlined into call-sites.
This fixes PR19989.
llvm-svn: 210632
expression of array-of-unknown-bound type, don't try to complete the array
bound, and return the alignment of the element type rather than 1.
llvm-svn: 210608
MSVC delays parsing of default arguments until instantiation. If the
default argument is never used, it is never parsed. We don't model
this.
Instead, if lookup of a type name fails in a template argument context,
we form a DependentNameType, which will be looked up at instantiation
time.
This fixes errors about 'CControlWinTraits' in atlwin.h.
Reviewers: rsmith
Differential Revision: http://reviews.llvm.org/D3995
llvm-svn: 210382
On test files I ran this on, memory consumption overall went down from
2.5G to 2G, without performance regressions.
I also investigated making DynTypedNode by itself smaller (by pulling
out pointers for everything that doesn't fit in 8 bytes). This led to
another 200-300MB saved, but also introduced a significant regression in
performance due to the memory management overhead.
llvm-svn: 209297
ensure that querying the first declaration for its most recent declaration
checks for redeclarations from the imported module.
This works as follows:
* The 'most recent' pointer on a canonical declaration grows a pointer to the
external AST source and a generation number (space- and time-optimized for
the case where there is no external source).
* Each time the 'most recent' pointer is queried, if it has an external source,
we check whether it's up to date, and update it if not.
* The ancillary data stored on the canonical declaration is allocated lazily
to avoid filling it in for declarations that end up being non-canonical.
We'll still perform a redundant (ASTContext) allocation if someone asks for
the most recent declaration from a decl before setPreviousDecl is called,
but such cases are probably all bugs, and are now easy to find.
Some finessing is still in order here -- in particular, we use a very general
mechanism for handling the DefinitionData pointer on CXXRecordData, and a more
targeted approach would be more compact.
Also, the MayHaveOutOfDateDef mechanism should now be expunged, since it was
addressing only a corner of the full problem space here. That's not covered
by this patch.
Early performance benchmarks show that this makes no measurable difference to
Clang performance without modules enabled (and fixes a major correctness issue
with modules enabled). I'll revert if a full performance comparison shows any
problems.
llvm-svn: 209046
GetGVALinkageForFunction handles TSK_ExplicitInstantiationDeclaration
twice, remove the redundant code trying to handle it again.
While we are here, update the reference we make to the standard. It
seems like another paragraph was added causing this text to get
renumbered.
llvm-svn: 208850
Having various possible states of initialization following construction doesn't
add value here.
Also remove the unused size_reserve parameter.
llvm-svn: 207897
Summary:
A reference temporary should inherit the linkage of the variable it
initializes. Otherwise, we may hit cases where a reference temporary
wouldn't have the same value in all translation units.
Reviewers: rsmith
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D3515
llvm-svn: 207451
When instantiating an array that has an alignment attribute on it, we
were looking through the array type and only considering the element
type for the resulting alignment. We need to make sure we take the
array's requirements into account too.
llvm-svn: 206317
Summary:
MSVC always emits inline functions marked with the extern storage class
specifier. The result is something similar to the opposite of
__attribute__((gnu_inline)).
This extension is also available in C.
This fixes PR19264.
Reviewers: rnk, rsmith
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D3207
llvm-svn: 205485
This adds Clang support for the ARM64 backend. There are definitely
still some rough edges, so please bring up any issues you see with
this patch.
As with the LLVM commit though, we think it'll be more useful for
merging with AArch64 from within the tree.
llvm-svn: 205100
correctly order comments in SourceManager::isBeforeInTranslationUnit() order
Unfortunately, this is not as simple as it was implemented previously, and
actually requires doing a merge sort.
llvm-svn: 204936
Summary:
The MSVC ABI appears to mangle the lexical scope into the names of
statics. Specifically, a counter is incremented whenever a scope is
entered where things can be declared in such a way that an ambiguity can
arise. For example, a class scope inside of a class scope doesn't do
anything interesting because the nested class cannot collide with
another nested class.
There are problems with this scheme:
- It is unreliable. The counter is only incremented when a previously
never encountered scope is entered. There are cases where this will
cause ambiguity amongst declarations that have the same name where one
was introduced in a deep scope while the other was introduced right
after in the previous lexical scope.
- It is wasteful. Statements like: {{{{{{{ static int foo = a; }}}}}}}
will make the mangling of "foo" larger than it need be because the
scope counter has been incremented many times.
Because of these problems, and practical implementation concerns. We
choose not to implement this scheme if the local static or local type
isn't visible. The mangling of these declarations will look very
similar but the numbering will make far more sense, this scheme is
lifted from the Itanium ABI implementation.
Reviewers: rsmith, doug.gregor, rnk, eli.friedman, cdavis5x
Reviewed By: rnk
CC: cfe-commits
Differential Revision: http://llvm-reviews.chandlerc.com/D2953
llvm-svn: 202951
With r197755 we started reading the contents of buffer file entries, but the
buffers may point to ASTReader blobs that have been disposed.
Fix this by having the CompilerInstance object keep a reference to the ASTReader
as well as having the ASTContext keep reference to the ExternalASTSource.
This was very difficult to construct a test case for.
rdar://16149782
llvm-svn: 202346
When calculating the preferred alignment of a type, consider if a alignment
attribute came from a typedef declaration. If one did, do not naturally align
the type.
Patch by Stephan Tolksdorf, with a little tweaking and an additional testcase by me.
llvm-svn: 202088
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
Member pointers are mangled as they would be represented at runtime.
They can be a single integer literal, single decl, or a tuple with some
more numbers tossed in. With Clang today, most of those numbers will be
zero because we reject pointers to members of virtual bases.
This change required moving VTableContextBase ownership from
CodeGenVTables to ASTContext, because mangling now depends on vtable
layout.
I also hoisted the inheritance model helpers up to be inline static
methods of MSInheritanceAttr. This makes the AST code that deals with
member pointers much more readable.
MSVC doesn't appear to have stable manglings of null member pointers:
- Null data memptrs in function templates have a mangling collision with
the first field of a non-polymorphic single inheritance class.
- The mangling of null data memptrs changes if you add casts.
- Large null function memptrs in class templates crash MSVC.
Clang uses the class template mangling for null data memptrs and the
function template mangling for null function memptrs to deal with this.
Reviewers: majnemer
Differential Revision: http://llvm-reviews.chandlerc.com/D2695
llvm-svn: 200857
A return type is the declared or deduced part of the function type specified in
the declaration.
A result type is the (potentially adjusted) type of the value of an expression
that calls the function.
Rule of thumb:
* Declarations have return types and parameters.
* Expressions have result types and arguments.
llvm-svn: 200082
not using backing ivar warning, ignore when
property is not being synthesized (user declared its
implementation @dynamic). // rdar://1583425
llvm-svn: 199820
Fix a perennial source of confusion in the clang type system: Declarations and
function prototypes have parameters to which arguments are supplied, so calling
these 'arguments' was a stretch even in C mode, let alone C++ where default
arguments, templates and overloading make the distinction important to get
right.
Readability win across the board, especially in the casting, ADL and
overloading implementations which make a lot more sense at a glance now.
Will keep an eye on the builders and update dependent projects shortly.
No functional change.
llvm-svn: 199686
There's been long-standing confusion over the role of these two options. This
commit makes the necessary changes to differentiate them clearly, following up
from r198936.
MicrosoftExt (aka. fms-extensions):
Enable largely unobjectionable Microsoft language extensions to ease
portability. This mode, also supported by gcc, is used for building software
like FreeBSD and Linux kernel extensions that share code with Windows drivers.
MSVCCompat (aka. -fms-compatibility, formerly MicrosoftMode):
Turn on a special mode supporting 'heinous' extensions for drop-in
compatibility with the Microsoft Visual C++ product. Standards-compilant C and
C++ code isn't guaranteed to work in this mode. Implies MicrosoftExt.
Note that full -fms-compatibility mode is currently enabled by default on the
Windows target, which may need tuning to serve as a reasonable default.
See cfe-commits for the full discourse, thread 'r198497 - Move MS predefined
type_info out of InitializePredefinedMacros'
No change in behaviour.
llvm-svn: 199209
__builtin_va_list and friends have been showing up where they shouldn't for way
to long, making unwanted appearences in -ast-print, tooling and source level
visitors and even the hello world tutorial on the clang website.
This commit factors down the implicit typedef and record creation facilities to
ensure they're marked implicit.
Also fixes a unit test that was testing incorrect behaviour, and removes old
hacks in the DeclPrinter that tried to skip implicit declarations manually.
llvm-svn: 197336
Summary:
In general, this type node can be used to represent any type adjustment
that occurs implicitly without losing type sugar. The immediate use of
this is to adjust the calling conventions of member function pointer
types without breaking template instantiation.
Fixes PR17996.
Reviewers: rsmith
Differential Revision: http://llvm-reviews.chandlerc.com/D2332
llvm-svn: 196451
The xcore llvm backend does not handle 8 byte alignment viz:
"%BadAlignment = alloca i64, align 8"
So getPreferredTypeAlign() must never overalign.
llvm-svn: 194462
Differential Revision: http://llvm-reviews.chandlerc.com/D2090
Clang was "improperly" over-aligning arrays with sizes are not a multiple of
their alignment.
This behavior was removed in microsoft 32 bit mode.
In addition, after examination of ASTContext::getTypeInfoImpl, a redundant code block in
MicrosoftRecordLayoutBuilder::getAdjustedFieldInfo was deleted.
llvm-svn: 193898
address spaces which is both (1) a "semantic" concept and
(2) possibly a hardware level restriction. It is desirable to
be able to discard/merge the LLVM-level address spaces on arguments for which
there is no difference to the current backend while keeping
track of the semantic address spaces in a funciton prototype. To do this
enable addition of the address space into the name-mangling process. Add
some tests to document this behaviour against inadvertent changes.
Patch by Michele Scandale!
llvm-svn: 190684
Static locals requiring initialization are not thread safe on Windows.
Unfortunately, it's possible to create static locals that are actually
externally visible with inline functions and templates. As a result, we
have to implement an initialization guard scheme that is compatible with
TUs built by MSVC, which makes thread safety prohibitively difficult.
MSVC's scheme is that every function that requires a guard gets an i32
bitfield. Each static local is assigned a bit that indicates if it has
been initialized, up to 32 bits, at which point a new bitfield is
created. MSVC rejects inline functions with more than 32 static locals,
and the externally visible mangling (?_B) only allows for one guard
variable per function.
On Eli's recommendation, I used MangleNumberingContext to track which
bit each static corresponds to.
Implements PR16888.
Reviewers: rjmccall, eli.friedman
Differential Revision: http://llvm-reviews.chandlerc.com/D1416
llvm-svn: 190427
getRealTypeByWidth and getIntTypeByWidth
for ASTContext names are almost same(invokes new methods from TargetInfo):
getIntTypeForBitwidth and getRealTypeForBitwidth.
As first commit for PR16752 fix: 'mode' attribute for unusual targets doesn't work properly
Description:
Troubles could be happened due to some assumptions in handleModeAttr function (see SemaDeclAttr.cpp).
For example, it assumes that 32 bit integer is 'int', while it could be 16 bit only.
Instead of asking target: 'which type do you want to use for int32_t ?' it just hardcodes general opinion. That doesn't looks pretty correct.
Please consider the next solution:
1. In Basic/TargetInfo add getIntTypeByWidth and getRealTypeByWidth virtual methods. By default current behaviour could be implemented here.
2. Fix handleModeAttr according to new methods in TargetInfo.
This approach is implemented in the patch attached to this post.
Fixes:
1st Commit (Current): Add new methods for TargetInfo:
getRealTypeByWidth and getIntTypeByWidth
for ASTContext names are almost same(invokes new methods from TargetInfo):
getIntTypeForBitwidth and getRealTypeForBitwidth
2nd Commit (Next): Fix SemaDeclAttr, handleModeAttr function.
llvm-svn: 190044
Summary:
Makes functions with implicit calling convention compatible with
function types with a matching explicit calling convention. This fixes
things like calls to qsort(), which has an explicit __cdecl attribute on
the comparator in Windows headers.
Clang will now infer the calling convention from the declarator. There
are two cases when the CC must be adjusted during redeclaration:
1. When defining a non-inline static method.
2. When redeclaring a function with an implicit or mismatched
convention.
Fixes PR13457, and allows clang to compile CommandLine.cpp for the
Microsoft C++ ABI.
Excellent test cases provided by Alexander Zinenko!
Reviewers: rsmith
Differential Revision: http://llvm-reviews.chandlerc.com/D1231
llvm-svn: 189412
Specifically, the following features are not included in this commit:
- any sort of capturing within generic lambdas
- nested lambdas
- conversion operator for captureless lambdas
- ensuring all visitors are generic lambda aware
As an example of what compiles:
template <class F1, class F2>
struct overload : F1, F2 {
using F1::operator();
using F2::operator();
overload(F1 f1, F2 f2) : F1(f1), F2(f2) { }
};
auto Recursive = [](auto Self, auto h, auto ... rest) {
return 1 + Self(Self, rest...);
};
auto Base = [](auto Self, auto h) {
return 1;
};
overload<decltype(Base), decltype(Recursive)> O(Base, Recursive);
int num_params = O(O, 5, 3, "abc", 3.14, 'a');
Please see attached tests for more examples.
Some implementation notes:
- Add a new Declarator context => LambdaExprParameterContext to
clang::Declarator to allow the use of 'auto' in declaring generic
lambda parameters
- Augment AutoType's constructor (similar to how variadic
template-type-parameters ala TemplateTypeParmDecl are implemented) to
accept an IsParameterPack to encode a generic lambda parameter pack.
- Add various helpers to CXXRecordDecl to facilitate identifying
and querying a closure class
- LambdaScopeInfo (which maintains the current lambda's Sema state)
was augmented to house the current depth of the template being
parsed (id est the Parser calls Sema::RecordParsingTemplateParameterDepth)
so that Sema::ActOnLambdaAutoParameter may use it to create the
appropriate list of corresponding TemplateTypeParmDecl for each
auto parameter identified within the generic lambda (also stored
within the current LambdaScopeInfo). Additionally,
a TemplateParameterList data-member was added to hold the invented
TemplateParameterList AST node which will be much more useful
once we teach TreeTransform how to transform generic lambdas.
- SemaLambda.h was added to hold some common lambda utility
functions (this file is likely to grow ...)
- Teach Sema::ActOnStartOfFunctionDef to check whether it
is being called to instantiate a generic lambda's call
operator, and if so, push an appropriately prepared
LambdaScopeInfo object on the stack.
- Teach Sema::ActOnStartOfLambdaDefinition to set the
return type of a lambda without a trailing return type
to 'auto' in C++1y mode, and teach the return type
deduction machinery in SemaStmt.cpp to process either
C++11 and C++14 lambda's correctly depending on the flag.
- various tests were added - but much more will be needed.
A greatful thanks to all reviewers including Eli Friedman,
James Dennett and the ever illuminating Richard Smith. And
yet I am certain that I have allowed unidentified bugs to creep in;
bugs, that I will do my best to slay, once identified!
Thanks!
llvm-svn: 188977
Add support for half (a.k.a. __fp16) in builtin descriptions.
The second argument to BUILTIN() now accepts 'h' to represent half.
Patch by Daniel Sanders
llvm-svn: 188464
We were exposing the extra alignment given to large arrays. The new behavior
matches gcc, which is a good thing since this is a gcc extension.
Thanks to Joerg Sonnenberger for noticing it.
While at it, centralize the method description in the .h file.
llvm-svn: 187999
JF Bastien