This re-lands r328845 with fixes for crbug.com/827810.
The initial motiviation was to hoist MethodVFTableLocation to global
scope so it could be forward declared.
In this patch, I noticed that MicrosoftVTableContext uses some risky
patterns. It has methods that return references to data stored in
DenseMaps. I've made some of them return by value for trivial structs
and I've moved some things into separate allocations.
llvm-svn: 329007
A recent addition to Coroutines TS (https://wg21.link/p0913) adds a pre-defined
coroutine noop_coroutine that does nothing. To implement this feature, we implemented
an llvm.coro.noop intrinsic that returns a coroutine handle to a coroutine that
does nothing when resumed or destroyed.
This patch adds a builtin __builtin_coro_noop() that maps to llvm.coro.noop intrinsic.
Related llvm change: https://reviews.llvm.org/D45114
llvm-svn: 328993
Summary:
https://reviews.llvm.org/rL325291 implemented Coroutines TS N4723
section [dcl.fct.def.coroutine]/7, but it performed lookup of allocator
functions within both the global and class scope, whereas the specified
behavior is to perform lookup for custom allocators within just the
class scope.
To fix, add parameters to the `Sema::FindAllocationFunctions` function
such that it can be used to lookup allocators in global scope,
class scope, or both (instead of just being able to look up in just global
scope or in both global and class scope). Then, use those parameters
from within the coroutine Sema.
This incorrect behavior had the unfortunate side-effect of causing the
bug https://bugs.llvm.org/show_bug.cgi?id=36578 (or at least the reports
of that bug in C++ programs). That bug would occur for any C++ user with
a coroutine frame that took a single pointer argument, since it would
then find the global placement form `operator new`, described in the
C++ standard 18.6.1.3.1. This patch prevents Clang from generating code
that triggers the LLVM assert described in that bug report.
Test Plan: `check-clang`
Reviewers: GorNishanov, eric_niebler, lewissbaker
Reviewed By: GorNishanov
Subscribers: EricWF, cfe-commits
Differential Revision: https://reviews.llvm.org/D44552
llvm-svn: 328949
Achieves almost a 200% speedup on the example where the performance of
visitors was problematic.
Performance on sqlite3 is unaffected.
rdar://38818362
Differential Revision: https://reviews.llvm.org/D45113
llvm-svn: 328911
C++ structured bindings for non-tuple-types are defined in a peculiar
way, where the resulting declaration is not a VarDecl, but a
BindingDecl.
That means a lot of existing machinery stops working.
rdar://36912381
Differential Revision: https://reviews.llvm.org/D44956
llvm-svn: 328910
Sometimes template instantiation causes CXXBindTemporaryExpr to be missing in
its usual spot. In CFG, temporary destructors work by relying on
CXXBindTemporaryExprs, so they won't work in this case.
Avoid the crash and notify the clients that we've encountered an unsupported AST
by failing to provide the ill-formed construction context for the temporary.
Differential Revision: https://reviews.llvm.org/D44955
llvm-svn: 328895
This allows forward declaring it so that we can add it to
MicrosoftMangleContext::mangleVirtualMemPtrThunk without including
VTableBuilder.h. That saves a hashtable lookup when emitting virtual
member pointer functions.
It also shortens a really long type name. This struct has "VFtable" in
the name, so it seems pretty unlikely that someone will assume it is
generally useful for non-MS C++ ABI stuff.
llvm-svn: 328845
Deprecation replacement can be any text but if it looks like a name of
ObjC method and has the same number of arguments as original method,
replace all slot names so after applying a fix-it you have valid code.
rdar://problem/36660853
Reviewers: aaron.ballman, erik.pilkington, rsmith
Reviewed By: erik.pilkington
Subscribers: cfe-commits, jkorous-apple
Differential Revision: https://reviews.llvm.org/D44589
llvm-svn: 328807
This patch sets target specific calling convention for CUDA kernels in IR.
Patch by Greg Rodgers.
Revised and lit test added by Yaxun Liu.
Differential Revision: https://reviews.llvm.org/D44747
llvm-svn: 328795
The conversion of operatios to bitcode helps to eliminate an additional
store in certain cases. We used to lower these load intrinsics in DAG to
DAG conversion by which time, the "Dead Store Elimination" pass is
already run. There is an associated LLVM patch.
Patch by Sumanth Gundapaneni.
llvm-svn: 328776
The AST for the fragment
```
@interface I
@end
template <typename>
void decode(I *p) {
for (I *k in p) {}
}
void decode(I *p) {
decode<int>(p);
}
```
differs heavily when templatized and non-templatized:
```
|-FunctionTemplateDecl 0x7fbfe0863940 <line:4:1, line:7:1> line:5:6 decode
| |-TemplateTypeParmDecl 0x7fbfe0863690 <line:4:11> col:11 typename depth 0 index 0
| |-FunctionDecl 0x7fbfe08638a0 <line:5:1, line:7:1> line:5:6 decode 'void (I *__strong)'
| | |-ParmVarDecl 0x7fbfe08637a0 <col:13, col:16> col:16 referenced p 'I *__strong'
| | `-CompoundStmt 0x7fbfe0863b88 <col:19, line:7:1>
| | `-ObjCForCollectionStmt 0x7fbfe0863b50 <line:6:3, col:20>
| | |-DeclStmt 0x7fbfe0863a50 <col:8, col:13>
| | | `-VarDecl 0x7fbfe08639f0 <col:8, col:11> col:11 k 'I *const __strong'
| | |-ImplicitCastExpr 0x7fbfe0863a90 <col:16> 'I *' <LValueToRValue>
| | | `-DeclRefExpr 0x7fbfe0863a68 <col:16> 'I *__strong' lvalue ParmVar 0x7fbfe08637a0 'p' 'I *__strong'
| | `-CompoundStmt 0x7fbfe0863b78 <col:19, col:20>
| `-FunctionDecl 0x7fbfe0863f80 <line:5:1, line:7:1> line:5:6 used decode 'void (I *__strong)'
| |-TemplateArgument type 'int'
| |-ParmVarDecl 0x7fbfe0863ef8 <col:13, col:16> col:16 used p 'I *__strong'
| `-CompoundStmt 0x7fbfe0890cf0 <col:19, line:7:1>
| `-ObjCForCollectionStmt 0x7fbfe0890cc8 <line:6:3, col:20>
| |-DeclStmt 0x7fbfe0890c70 <col:8, col:13>
| | `-VarDecl 0x7fbfe0890c00 <col:8, col:11> col:11 k 'I *__strong' callinit
| | `-ImplicitValueInitExpr 0x7fbfe0890c60 <<invalid sloc>> 'I *__strong'
| |-ImplicitCastExpr 0x7fbfe0890cb0 <col:16> 'I *' <LValueToRValue>
| | `-DeclRefExpr 0x7fbfe0890c88 <col:16> 'I *__strong' lvalue ParmVar 0x7fbfe0863ef8 'p' 'I *__strong'
| `-CompoundStmt 0x7fbfe0863b78 <col:19, col:20>
```
Note how in the instantiated version ImplicitValueInitExpr unexpectedly appears.
While objects are auto-initialized under ARC, it does not make sense to
have an initializer for a for-loop variable, and it makes even less
sense to have such a different AST for instantiated and non-instantiated
version.
Digging deeper, I have found that there are two separate Sema* files for
dealing with templates and for dealing with non-templatized code.
In a non-templatized version, an initialization was performed only for
variables which are not loop variables for an Objective-C loop and not
variables for a C++ for-in loop:
```
if (FRI && (Tok.is(tok::colon) || isTokIdentifier_in())) {
bool IsForRangeLoop = false;
if (TryConsumeToken(tok::colon, FRI->ColonLoc)) {
IsForRangeLoop = true;
if (Tok.is(tok::l_brace))
FRI->RangeExpr = ParseBraceInitializer();
else
FRI->RangeExpr = ParseExpression();
}
Decl *ThisDecl = Actions.ActOnDeclarator(getCurScope(), D);
if (IsForRangeLoop)
Actions.ActOnCXXForRangeDecl(ThisDecl);
Actions.FinalizeDeclaration(ThisDecl);
D.complete(ThisDecl);
return Actions.FinalizeDeclaratorGroup(getCurScope(), DS, ThisDecl);
}
SmallVector<Decl *, 8> DeclsInGroup;
Decl *FirstDecl = ParseDeclarationAfterDeclaratorAndAttributes(
D, ParsedTemplateInfo(), FRI);
```
However the code in SemaTemplateInstantiateDecl was inconsistent,
guarding only against C++ for-in loops.
rdar://38391075
Differential Revision: https://reviews.llvm.org/D44989
llvm-svn: 328749
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
These instructions have been around for a long time, but we
haven't supported intrinsics for them. The "new" vesrions use
the CSx register for the start of the buffer instead of the K
field in the Mx register.
There is a related llvm patch.
Patch by Brendon Cahoon.
llvm-svn: 328725
When the declare target variables are emitted for the device,
constructors|destructors for these variables must emitted and registered
by the runtime in the offloading sections.
llvm-svn: 328705
The diagnostic system for Clang can already handle many AST nodes. Instead
of converting them to strings first, just hand the AST node directly to
the diagnostic system and let it handle the output. Minor changes in some
diagnostic output.
llvm-svn: 328688
structs.
r326307 and r327870 made changes that allowed using non-trivial C
structs with fields qualified with __strong or __weak. This commit makes
the following C++ triviality type traits available to non-trivial C
structs:
__has_trivial_assign
__has_trivial_move_assign
__has_trivial_copy
__has_trivial_move_constructor
__has_trivial_constructor
__has_trivial_destructor
rdar://problem/33599681
Differential Revision: https://reviews.llvm.org/D44913
llvm-svn: 328680
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
Adding a matcher for BinaryOperator and cxxOperatorCallExpr to be able to
decide whether it is any kind of assignment operator or not. This would be
useful since allows us to easily detect assignments via matchers for static
analysis (Tidy, SA) purposes.
Differential Revision: https://reviews.llvm.org/D44893
llvm-svn: 328618
Summary:
This fixes PR33561 and PR34185.
Don't store pending template instantiations for late-parsed templates in
the normal PendingInstantiations queue. Instead, use a separate list
that will only be parsed and instantiated at end of TU when late
template parsing actually works and doesn't infinite loop.
Reviewers: rsmith, thakis, hans
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D44846
llvm-svn: 328567
This make -ivfsoverlay behave more like other fatal errors (e.g. missing
-include file) by skipping the missing file instead of bailing out of
the whole compilation. This makes it possible for libclang to still
provide some functionallity as well as to correctly produce the fatal
error diagnostic (previously we lost the diagnostic in libclang since
there was no TU to tie it to).
rdar://33385423
llvm-svn: 328337
Changes the analyzer to believe that methods annotated with _Nonnull
from system frameworks indeed return non null objects.
Local methods with such annotation are still distrusted.
rdar://24291919
Differential Revision: https://reviews.llvm.org/D44341
llvm-svn: 328282
Putting back the code in commit r327189 that was reverted in r322737. The code is being committed in three stages and this one is the last stage: 1) r327455 fp16 feature flags, 2) r327836 pass half type or i16 based on FullFP16, and 3) the code here which the front-end fp16 vector intrinsic for ARM.
Differential revision https://reviews.llvm.org/D43650
llvm-svn: 328277
CXXCtorInitializer-based constructors are also affected by the C++17 mandatory
copy elision, like variable constructors and return value constructors.
Extend r328248 to support those.
Differential Revision: https://reviews.llvm.org/D44763
llvm-svn: 328255
In C++17 copy elision is mandatory for variable and return value constructors
(as long as it doesn't involve type conversion) which results in AST that does
not contain elidable constructors in their usual places. In order to provide
construction contexts in this scenario we need to cover more AST patterns.
This patch makes the CFG prepared for these scenarios by:
- Fork VariableConstructionContext and ReturnedValueConstructionContext into
two different sub-classes (each) one of which indicates the C++17 case and
contains a reference to an extra CXXBindTemporaryExpr.
- Allow CFGCXXRecordTypedCall element to accept VariableConstructionContext and
ReturnedValueConstructionContext as its context.
Differential Revision: https://reviews.llvm.org/D44597
llvm-svn: 328248
If a memory region (or an SVal that represents a pointer to that memory region)
is a (direct or indirect, not necessarily proper) sub-region of a SymbolicRegion
then it is said to have a symbolic base.
For now SVal::symbol_iterator explores the symbol within a symbolic region
only when the SVal represents a pointer to the symbolic region itself,
not to any of its sub-regions.
This behavior is not indended by any user of symbol_iterator; all users who
cared about such behavior were expecting the iterator to descend into the
symbolic base of an arbitrary region, find the parent symbol of the symbolic
base region, and iterate over that symbol. Lack of such behavior resulted in
bugs demonstarted by the test cases.
Hence the decision to change the API to behave more intuitively.
Differential Revision: https://reviews.llvm.org/D44347
llvm-svn: 328247
This fixes host-side LTO during CUDA compilation. Before, LTO
pipeline construction was clashing with CUDA pipeline construction.
At the moment there's no point doing LTO on device side as each
device-side TU is a complete program. We will need to figure out
compilation pipeline construction for the device-side LTO when we
have working support for multi-TU device-side CUDA compilation.
Differential Revision: https://reviews.llvm.org/D44691
llvm-svn: 328161
Summary:
Libc++'s default allocator uses `__builtin_operator_new` and `__builtin_operator_delete` in order to allow the calls to new/delete to be ellided. However, libc++ now needs to support over-aligned types in the default allocator. In order to support this without disabling the existing optimization Clang needs to support calling the aligned new overloads from the builtins.
See llvm.org/PR22634 for more information about the libc++ bug.
This patch changes `__builtin_operator_new`/`__builtin_operator_delete` to call any usual `operator new`/`operator delete` function. It does this by performing overload resolution with the arguments passed to the builtin to determine which allocation function to call. If the selected function is not a usual allocation function a diagnostic is issued.
One open issue is if the `align_val_t` overloads should be considered "usual" when `LangOpts::AlignedAllocation` is disabled.
In order to allow libc++ to detect this new behavior the value for `__has_builtin(__builtin_operator_new)` has been updated to `201802`.
Reviewers: rsmith, majnemer, aaron.ballman, erik.pilkington, bogner, ahatanak
Reviewed By: rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D43047
llvm-svn: 328134
When skipping building the module for a private framework module,
LangOpts.CurrentModule isn't enough for implict modules builds; for
instance, in case a private module is built while building a public one,
LangOpts.CurrentModule doesn't reflect the -fmodule-name being passed
down, but instead the module name which triggered the build.
Store the actual -fmodule-name in LangOpts.ModuleName and actually
check a name was provided during compiler invocation in order to
skip building the private module.
rdar://problem/38434694
llvm-svn: 328053
Eugene Zelenko