(or of a lambda init-capture, which is sort-of such a variable). The semantics
of such constructs will change when we implement N3922, so we intend to warn on
this in Clang 3.6 then change the semantics in Clang 3.7.
llvm-svn: 228792
We'd give the VarDecl a CXXConstructExpr even though it is annotated
with an alias attribute. This would make us trip over sanity checking
asserts.
This fixes PR22493.
llvm-svn: 228523
After r228258, Clang started emitting C++ EH IR that LLVM wasn't ready
to deal with, even when exceptions were disabled with /EHs-. This time,
make /EHs- turn off -fexceptions while still emitting exceptional
constructs in functions using __try. Since Sema rejects C++ exception
handling constructs before CodeGen, landingpads should only appear in
such functions as the result of a __try.
llvm-svn: 228329
It caused a chromium base unittest that tests throwing and catching SEH
exceptions to fail (http://crbug.com/455488) and I suspect it might also
be the cause of the chromium clang win 64-bit shared release builder timing
out during compiles. So revert to see if that's true.
llvm-svn: 228262
When visiting AssignmentOps, keep evaluating after a failure (when possible) in
order to identify overflow in subexpressions.
Differential Revision: http://reviews.llvm.org/D1238
llvm-svn: 228202
std::lock_guard. If EXCLUSIVE_LOCKS_REQUIRED is placed on the constructor of
a SCOPED_LOCKABLE class, then that constructor is assumed to adopt the lock;
e.g. the lock must be held on construction, and will be released on destruction.
llvm-svn: 228194
Some standard header files from MSVC2012 use 'mutable' on references, though it is directly prohibited by the standard.
Fix for http://llvm.org/PR22444
Differential Revision: http://reviews.llvm.org/D7370
llvm-svn: 228113
There are four major kinds of declarations that cause code generation:
- FunctionDecl (includes CXXMethodDecl etc)
- ObjCMethodDecl
- BlockDecl
- CapturedDecl
This patch tracks __try usage on FunctionDecls and diagnoses __try usage
in other decls. If someone wants to use __try from ObjC, they can use it
from a free function, since the ObjC code will need an ObjC-style EH
personality.
Eventually we will want to look through CapturedDecls and track SEH
usage on the parent FunctionDecl, if present.
llvm-svn: 228058
These checks detect potential deadlocks caused by inconsistent lock
ordering. The checks are implemented under the -Wthread-safety-beta flag.
This patch also replaces calls to getAttrs() with calls to attrs() throughout
ThreadSafety.cpp, which fixes the earlier issue that cause assert failures.
llvm-svn: 228051
These checks detect potential deadlocks caused by inconsistent lock
ordering. The checks are implemented under the -Wthread-safety-beta flag.
llvm-svn: 227997
Thou shall not jump into SEH blocks. Jumping out of SEH __try and __excepts
is A-ok. Jumping out of __finally blocks is B-ok (msvc doesn't error about it,
but warns that it has undefined behavior).
I've checked that clang's behavior with this patch matches msvc's behavior.
We don't have the warning on jumping out of a __finally yet, see the FIXME
in the test. clang also currently crashes on codegen for a jump out of a
__finally block, see PR22414 comment 7.
I also added a few tests for the interaction of indirect jumps and SEH blocks.
MSVC doesn't support indirect jumps, so there's no way to know if clang behave
the same way as msvc here. clang's behavior with this patch does make sense
to me, but maybe it could be argued that it should be more permissive (see
FIXME in the indirect jump tests -- shout if you have an opinion on this).
llvm-svn: 227982
This check does not apply when Borland extensions are enabled, as they
have a checked in test case indicating that mixed usage of SEH and C++
is supported.
llvm-svn: 227876
list-initialization that gets converted to some form other than an
InitListExpr. CXXTemporaryObjectExpr is a special case here, because it
represents a fused CXXFunctionalCastExpr + CXXConstructExpr. That, in
itself, is probably a design error...
llvm-svn: 227377
infinite recursion.
Also guard against said infinite recursion by adding an assert that will
trigger if CorrectDelayedTyposInExpr is called before a previous call to
CorrectDelayedTyposInExpr returns (i.e. if the TreeTransform run by
CorrectDelayedTyposInExpr calls a sequence of methods that
end up calling CorrectDelayedTyposInExpr, as the new test case had done
prior to this commit). Fixes PR22292.
llvm-svn: 227368
Under certain circumstances, the identifier mentioned in the diagnostic
won't match the intended correction even though the replacement
expression and the note pointing to the decl are both correct.
Basically, the TreeTransform assumes the TypoExpr's Consumer points to
the correct TypoCorrection, but the handling of typos that appear to be
ambiguous from the point of view of TransformTypoExpr would cause that
assumption to be violated by altering the Consumer's correction stream.
This fix allows the Consumer's correction stream to be reset to the
right TypoCorrection after successfully resolving the percieved ambiguity.
Included is a fix to suppress correcting the RHS of an assignment to the
LHS of that assignment for non-C++ code, to prevent a regression in
test/SemaObjC/provisional-ivar-lookup.m.
This fixes PR22297.
llvm-svn: 227251
clang currently calls MarkVTableUsed() for classes that get their virtual
methods called or that participate in a dynamic_cast. This is unnecessary,
since CodeGen only emits vtables when it generates constructor, destructor, and
vtt code. (*)
Note that Sema::MarkVTableUsed() doesn't cause the emission of a vtable.
Its main user-visible effect is that it instantiates virtual member functions
of template classes, to make sure that if codegen decides to write a vtable
all the entries in the vtable are defined.
While this shouldn't change the behavior of codegen (other than being faster),
it does make clang more permissive: virtual methods of templates (in particular
destructors) end up being instantiated less often. In particular, classes that
have members that are smart pointers to incomplete types will now get their
implicit virtual destructor instantiated less frequently. For example, this
used to not compile but does now compile:
template <typename T> struct OwnPtr {
~OwnPtr() { static_assert((sizeof(T) > 0), "TypeMustBeComplete"); }
};
class ScriptLoader;
struct Base { virtual ~Base(); };
struct Sub : public Base {
virtual void someFun() const {}
OwnPtr<ScriptLoader> m_loader;
};
void f(Sub *s) { s->someFun(); }
The more permissive behavior matches both gcc (where this is not often
observable, since in practice most things with virtual methods have a key
function, and Sema::DefineUsedVTables() skips vtables for classes with key
functions) and cl (which is my motivation for this change) – this fixes
PR20337. See this issue and the review thread for some discussions about
optimizations.
This is similar to r213109 in spirit. r225761 was a prerequisite for this
change.
Various tests relied on "a->f()" marking a's vtable as used (in the sema
sense), switch these to just construct a on the stack. This forces
instantiation of the implicit constructor, which will mark the vtable as used.
(*) The exception is -fapple-kext mode: In this mode, qualified calls to
virtual functions (`a->Base::f()`) still go through the vtable, and since the
vtable pointer off this doesn't point to Base's vtable, this needs to reference
Base's vtable directly. To keep this working, keep referencing the vtable for
virtual calls in apple kext mode.
llvm-svn: 227073
The lowering looks a lot like normal EH lowering, with the exception
that the exceptions are caught by executing filter expression code
instead of matching typeinfo globals. The filter expressions are
outlined into functions which are used in landingpad clauses where
typeinfo would normally go.
Major aspects that still need work:
- Non-call exceptions in __try bodies won't work yet. The plan is to
outline the __try block in the frontend to keep things simple.
- Filter expressions cannot use local variables until capturing is
implemented.
- __finally blocks will not run after exceptions. Fixing this requires
work in the LLVM SEH preparation pass.
The IR lowering looks like this:
// C code:
bool safe_div(int n, int d, int *r) {
__try {
*r = normal_div(n, d);
} __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) {
return false;
}
return true;
}
; LLVM IR:
define i32 @filter(i8* %e, i8* %fp) {
%ehptrs = bitcast i8* %e to i32**
%ehrec = load i32** %ehptrs
%code = load i32* %ehrec
%matches = icmp eq i32 %code, i32 u0xC0000094
%matches.i32 = zext i1 %matches to i32
ret i32 %matches.i32
}
define i1 zeroext @safe_div(i32 %n, i32 %d, i32* %r) {
%rr = invoke i32 @normal_div(i32 %n, i32 %d)
to label %normal unwind to label %lpad
normal:
store i32 %rr, i32* %r
ret i1 1
lpad:
%ehvals = landingpad {i8*, i32} personality i32 (...)* @__C_specific_handler
catch i8* bitcast (i32 (i8*, i8*)* @filter to i8*)
%ehptr = extractvalue {i8*, i32} %ehvals, i32 0
%sel = extractvalue {i8*, i32} %ehvals, i32 1
%filter_sel = call i32 @llvm.eh.seh.typeid.for(i8* bitcast (i32 (i8*, i8*)* @filter to i8*))
%matches = icmp eq i32 %sel, %filter_sel
br i1 %matches, label %eh.except, label %eh.resume
eh.except:
ret i1 false
eh.resume:
resume
}
Reviewers: rjmccall, rsmith, majnemer
Differential Revision: http://reviews.llvm.org/D5607
llvm-svn: 226760
Currently we emit DeferredDeclsToEmit in reverse order. This patch changes that.
The advantages of the change are that
* The output order is a bit closer to the source order. The change to
test/CodeGenCXX/pod-member-memcpys.cpp is a good example.
* If we decide to deffer more, it will not cause as large changes in the
estcases as it would without this patch.
llvm-svn: 226751
be corrected.
This fixes PR22250, which exposed the bug where if there's more than one
TypoExpr in the arguments, once one failed to be corrected none of the
TypoExprs after it would be handled at all thanks to an early return.
llvm-svn: 226624
The test case is based on the reduction from PR21679 and has to be
freestanding to work correctly, since some of the expected errors (and
some of the problems that were fixed) only occur when the end of the
file is reached.
llvm-svn: 226603
If an unscoped enum is used as a nested name specifier and the language dialect
is not C++ 11, issue an extension warning.
This fixes PR16951.
Differential Revision: http://reviews.llvm.org/D6389
llvm-svn: 226413
Previously if an enumeration was used in a nested name specifier in pre-C++11
language dialect, error message was 'XXX is not a class, namespace, or scoped
enumeration'. This patch removes the word 'scoped' as in C++11 any enumeration
may be used in this context.
llvm-svn: 226410
Clang currently crashes on
class C {
C() = default;
C() = delete;
};
My cunning plan for fixing this was to change the `if (!FnD)` in
Parser::ParseCXXInlineMethodDef() to `if (!FnD || FnD->isInvalidDecl)` – but
alas, the second constructor decl wasn't marked as invalid. This lets
Sema::MergeFunctionDecl() return true on function redeclarations, which leads
to them being marked invalid.
This also improves error messages when functions are redeclared.
llvm-svn: 226365
ambiguity but wasn't.
In the new test case, "click" wasn't being corrected properly because
Sema::ClassifyName would call CorrectTypo for "click" then later
Sema::DiagnoseEmptyLookup would call CorrectTypoDelayed for the same use
of "click" (the former by the parser needing to determine what the
identifier is so it knows how to parse the statement, i.e. is it the
beginning of a declaration or an expression). CorrectTypo would record
that typo correction for "click" failed and CorrectTypoDelayed would see
that and not even try to correct the typo, even though in this case
CorrectTypo failed due to an ambiguity (both "Click" and "clock" having
an edit distance of one from "click") that could be resolved with more
information. The fix is two-fold:
1) Have CorrectTypo not record failed corrections if the reason for
the failure was two or more corrections with the same edit
distance, and
2) Make the CorrectionCandidateCallback used by
Parser::ParseCastExpression reject FunctionDecl candidates when the
next token after the identifier is a ".", "=", or "->" since
functions cannot be assigned to and do not have members that can be
referenced.
The reason for two correction spots is that from r222549 until r224375
landed, the first correction attempt would fail completely but the
second would suggest "clock" while having the note point to the
declaration of "Click".
llvm-svn: 226334
Clang would previously become confused and crash here.
It does not make a lot of sense to export these, so warning seems appropriate.
MSVC will export some member functions for this kind of specializations, whereas
MinGW ignores the dllexport-edness. The latter behaviour seems better.
Differential Revision: http://reviews.llvm.org/D6984
llvm-svn: 226208
conflicting attribute, warn about the conflict and pick a "winning"
attribute to preserve, instead of emitting an error. This matches the
behavior when the conflicting attributes are on different declarations.
Along the way I discovered that conflicts involving __forceinline were
reported as 'always_inline' (alternate spelling, same attribute) so
fixed that up to report the attribute as spelled in the source.
llvm-svn: 225813
-Wself-move is similiar to -Wself-assign. This warning is triggered when
a value is attempted to be moved to itself. See r221008 for a bug that
would have been caught with this warning.
llvm-svn: 225581
We have a diagnostic describing that constexpr changed in C++14 when
compiling in C++11 mode. While doing this, it examines the previous
declaration and assumes that it is a function. However it is possible,
in the context of error recovery, for this to not be the case.
llvm-svn: 225518
We forgot to mark designated initializer expression that contain type
dependent array designators as type dependent. This would lead to
crashes when we try to determine which array element we were trying to
initialize.
This fixes PR22056.
llvm-svn: 225494
we're instantiating, if there's a ParmVarDecl within a FunctionDecl context
that is not a parameter of that function. Add some asserts to catch this kind
of issue more generally, and fix another bug exposed by those asserts where we
were missing a local instantiation scope around substitution of
explicitly-specified template arguments.
llvm-svn: 225490
transform.
Also diagnose typos in the initializer of an invalid C++ declaration.
Both issues were hit using the same line of test code, depending on
whether the code was treated as C or C++.
Fixes PR22092.
llvm-svn: 225389
r185773 added an assert that checked that a CXXUnresolvedConstructExpr either
has a valid rparen, or exactly one argument. This doesn't have to be true for
invalid inputs. Convert the assert to an if, and add a test for this case.
Found by SLi's afl bot.
llvm-svn: 225140
Treat volatile accesses as "maybe" instead of "definite" side effects for the purposes of warning on evaluations in an unevaluated context. No longer diagnose on idiomatic code like:
int * volatile v;
(void)sizeof(*v);
llvm-svn: 225116
GCC permits array l-values in asm output operands even though they
aren't modifiable l-values. We used to permit it but this behavior
regressed in r224916.
llvm-svn: 224918
Functions are l-values in C++ but shouldn't be available as output
parameters in inline assembly. Neither should overloaded function
l-values.
This fixes PR21949.
llvm-svn: 224916
Clang has a hack to accept definitions of structs with tag names which
have the same name as intrinsics. However, this hack didn't guard
against annotation tokens showing up in the token stream.
llvm-svn: 224909
We expected the type of a TagDecl to be a TagType, not an
InjectedClassNameType. Introduced a helper method, Type::getAsTagDecl,
to abstract away the difference; redefine Type::getAsCXXRecordDecl to be
in terms of it.
llvm-svn: 224898
The lit.cfg files only add .cpp to suffixes, so these tests used to never run,
oops. (Also tweak to of these tests in minor ways to make the actually pass.)
llvm-svn: 224718
Previously we thought the instance member was a function, not a field,
and we'd say something silly like:
t.cpp:4:27: error: call to non-static member function without an object argument
static int f() { return n; }
^
Noticed in PR21923.
llvm-svn: 224480
There are a few cases where unqualified lookup can find C++ methods.
Unfortunately, none of them seem to have illegal access paths, so I
can't excercise the diagnostic source range code that I am changing
here.
Fixes PR21851, which was a crash on valid.
llvm-svn: 224471
pessimistic about when to do so.
This also fixes PR21905 as the initialization argument was no longer
viewed as being type dependent due to the TypoExpr being type-cast.
llvm-svn: 224386
Remove Sema::UnqualifiedTyposCorrected, a cache of corrected typos. It would only cache typo corrections that didn't provide ValidateCandidate of which there were few left, and it had a bug when we had the same identifier spelled wrong twice. See the last two tests in typo-correction.cpp for cases this fires.
llvm-svn: 224375
Turning our _Atomic L-value into an R-value removes its _Atomic-ness.
However, we didn't update our 'FromType' which made
ScalarTypeToBooleanCastKind think we were trying to pass it a
non-scalar.
This fixes PR21836.
llvm-svn: 224322
We know that const_cast<char *>((void)Something) is ill-formed, even if
'Something' is dependent because you can't cast from void to a pointer
type.
This fixes PR21845.
llvm-svn: 224299
Previously we would attempt to build a TypeSourceInfo for a null type,
and then we would forget to pop the function scope before returning an
error.
Reviewers: rsmith
Differential Revision: http://reviews.llvm.org/D6665
llvm-svn: 224271
Don't send a value dependent expression into the expression evaluator,
HandleSizeof would crash. Making HandleSizeof handle dependent types
would noisily warn about the operation even if everything turns out OK
after instantiation.
This fixes PR21848.
llvm-svn: 224240
clang lets programmers be pretty cavalier when it comes to void return
statements in functions which have non-void return types. However, we
cannot be so forgiving in constexpr functions: evaluation will go off
the rails very quickly.
Instead, keep the return statement in the AST but mark the function as
invalid. Doing so gives us nice diagnostics while making constexpr
evaluation halt.
This fixes PR21859.
llvm-svn: 224189
The extension has the following syntax:
__builtin_call_with_static_chain(Call, Chain)
where Call must be a function call expression and Chain must be of pointer type
This extension performs a function call Call with a static chain pointer
Chain passed to the callee in a designated register. This is useful for
calling foreign language functions whose ABI uses static chain pointers
(e.g. to implement closures).
Differential Revision: http://reviews.llvm.org/D6332
llvm-svn: 224167
having OptimizeNone remove them again, just don't add them in the
first place if the function already has OptimizeNone.
Note that MinSize can still appear due to attributes on different
declarations; a future patch will address that.
llvm-svn: 224047
Comparing the address of an object with an incomplete type might return
true with a 'distinct' object if the former has a size of zero.
However, such an object should compare unequal with null.
llvm-svn: 224040
getLVForNamespaceScopeDecl believed that it wasn't possible for it to
ever see an IndirectFieldDecl. However, this can occur when determining
whether or not something is a redeclaration of a member of an anonymous
static union.
This fixes PR21858.
llvm-svn: 223975
This attribute serves as a hint to improve warnings about the ranges of
enumerators used as flag types. It currently has no working C++ implementation
due to different semantics for enums in C++. For more explanation, see the docs
and testcases.
Reviewed by Aaron Ballman.
llvm-svn: 222906
GCC and ICC both reject this and the 'Runtime-sized arrays with
automatic storage duration' (N3639) paper forbade this as well.
Previously, we would crash on our way to mangling.
This fixes PR21632.
llvm-svn: 222569
Clang r181627 moved a check for block-scope variables into this code for
handling thread storage class specifiers, but in the process, it broke the
logic for checking if the target supports TLS. Fix this with some simple
restructuring of the code. rdar://problem/18796883
llvm-svn: 222512
for lambda expressions. That can't ever work; we need to transform the
parameters in order to create new ones in the new call operator context.
Fixes a rejects-valid when transforming a context containing a
lambda-expression that uses its function parameters in C++14 mode.
llvm-svn: 222482
std::X::swap exception specifications (allowing parsing of non-conforming code
in libstdc++). The old conditions also matched the functions in MSVC's STL,
which were relying on deferred parsing here.
llvm-svn: 222471
If there is more than one TypoExpr within the expr being transformed and
any but the last TypoExpr seen don't have any viable candidates, the
tree transform will be aborted early and the remaining TypoExprs are
never seen and hence never diagnosed. This adds a simple
RecursiveASTVisitor to find all of the TypoExprs to be diagnosed in the
case where typo correction of the entire expr fails (and the result of
the tree transform is an ExprError).
llvm-svn: 222465
Sema::ActOnIdExpression to use the new functionality.
Among other things, this allows recovery in several cases where it
wasn't possible before (e.g. correcting a mistyped static_cast<>).
llvm-svn: 222464
The bug is that ExprCleanupObjects isn't always empty
in a fresh evaluation context. New evaluation contexts just
track the current depth of the stack.
The assertion will misfire whenever we finish processing
a function body inside an expression that contained an earlier
block literal with non-trivial captures. That's actually
a lot less likely than you'd think, though, because it has
to be a real function declaration, not just another block.
Mixed block/lambda code would work, as would a template
instantiation or a local class definition.
The code works correctly if the assertion is disabled.
rdar://16356628
llvm-svn: 222194
Specifically, when we have this situation:
struct A {
template <typename T> struct B {
int m1 = sizeof(A);
};
B<int> m2;
};
We can't parse m1's initializer eagerly because we need A to be
complete. Therefore we wait until the end of A's class scope to parse
it. However, we can trigger instantiation of B before the end of A,
which will attempt to instantiate the field decls eagerly, and it would
build a bad field decl instantiation that said it had an initializer but
actually lacked one.
Fixed by deferring instantiation of default member initializers until
they are needed during constructor analysis. This addresses a long
standing FIXME in the code.
Fixes PR19195.
Reviewed By: rsmith
Differential Revision: http://reviews.llvm.org/D5690
llvm-svn: 222192
a member named 'swap' and then expect unqualified lookup for the name 'swap' in
its exception specification to find anything else.
Without delay-parsed exception specifications, this was ill-formed (NDR) by
[basic.scope.class]p1, rule 2. With delay-parsed exception specifications, the
call to 'swap' unambiguously finds the function being declared, which then
fails because the arguments don't work for that function.
llvm-svn: 221955
Without this, -Wunused-local-typedef would incorrectly warn on the two typedefs
in this program:
void foo() {
struct A {};
struct B : public A {
typedef A INHERITED;
B() : INHERITED() {}
typedef B SELF;
B(int) : SELF() {}
};
}
llvm-svn: 221765
This is a new form of expression of the form:
(expr op ... op expr)
where one of the exprs is a parameter pack. It expands into
(expr1 op (expr2onwards op ... op expr))
(and likewise if the pack is on the right). The non-pack operand can be
omitted; in that case, an empty pack gives a fallback value or an error,
depending on the operator.
llvm-svn: 221573
Trying to import or export such classes doesn't make sense, and Clang
would assert trying to export vtables for them.
This is consistent with how we treat functions with internal linkage,
but it is stricter than MSVC so we may have to back down if it breaks
real code.
llvm-svn: 221160
It turns out that MinGW never dllimports of exports inline functions.
This means that code compiled with Clang would fail to link with
MinGW-compiled libraries since we might try to import functions that
are not imported.
To fix this, make Clang never dllimport inline functions when targeting
MinGW.
llvm-svn: 221154
but MSABI was never defined in the test. It seems we are erroring
on code that we should be accepting when compiling for MSVC compatibility.
This should make the test less confusing until PR21406 is fixed.
llvm-svn: 220825
This includes adding the new TypoExpr-based lazy typo correction to
LookupMemberExprInRecord as an alternative to the existing eager typo
correction.
llvm-svn: 220698
Wire it through everywhere we have support for fastcall, essentially.
This allows us to parse the MSVC "14" CTP headers, but we will
miscompile them because LLVM doesn't support __vectorcall yet.
Reviewed By: Aaron Ballman
Differential Revision: http://reviews.llvm.org/D5808
llvm-svn: 220573
If a templated class is not instantiated, then the AST for it could be missing
some things that would throw the field checker off. Wait until specialization
before emitting these warnings.
llvm-svn: 220363
complete object to a pointer to the start of another complete object does
not evaluate to the constant 'false'. All other comparisons between the
addresses of subobjects of distinct complete objects still do.
llvm-svn: 220343
A second instance of attributed types escaped the previous change, identified
thanks to Richard Smith! When deducing the void case, we would also assume that
the type would not be attributed. Furthermore, properly handle multiple
attributes being applied to a single TypeLoc.
Properly handle this case and future-proof a bit by ignoring parenthesis
further. The test cases do use the additional parenthesis to ensure that this
case remains properly handled.
Addresses post-commit review comments from Richard Smith to SVN r219851.
llvm-svn: 219974
Specifically, avoid typo-correcting the variable name into a type before
typo-correcting the actual type name in the declaration. Doing so
results in a very unpleasant cascade of errors, with the typo correction
of the actual type name being buried in the middle.
llvm-svn: 219732
and !=) to support mixed complex and real operand types.
This requires removing an assert from SemaChecking, and adding support
both to the constant evaluator and the code generator to synthesize the
imaginary part when needed. This seemed somewhat cleaner than having
just the comparison operators force real-to-complex conversions.
I've added test cases for these operations. I'm really terrified that
there were *no* tests in-tree which exercised this.
This turned up when trying to build R after my change to the complex
type lowering.
llvm-svn: 219570
operators where one type is a C complex type, and to emit both the
efficient and correct implementation for complex arithmetic according to
C11 Annex G using this extra information.
For both multiply and divide the old code was writing a long-hand
reduced version of the math without any of the special handling of inf
and NaN recommended by the standard here. Instead of putting more
complexity here, this change does what GCC does which is to emit
a libcall for the fully general case.
However, the old code also failed to do the proper minimization of the
set of operations when there was a mixed complex and real operation. In
those cases, C provides a spec for much more minimal operations that are
valid. Clang now emits the exact suggested operations. This change isn't
*just* about performance though, without minimizing these operations, we
again lose the correct handling of infinities and NaNs. It is critical
that this happen in the frontend based on assymetric type operands to
complex math operations.
The performance implications of this change aren't trivial either. I've
run a set of benchmarks in Eigen, an open source mathematics library
that makes heavy use of complex. While a few have slowed down due to the
libcall being introduce, most sped up and some by a huge amount: up to
100% and 140%.
In order to make all of this work, also match the algorithm in the
constant evaluator to the one in the runtime library. Currently it is
a broken port of the simplifications from C's Annex G to the long-hand
formulation of the algorithm.
Splitting this patch up is very hard because none of this works without
the AST change to preserve non-complex operands. Sorry for the enormous
change.
Follow-up changes will include support for sinking the libcalls onto
cold paths in common cases and fastmath improvements to allow more
aggressive backend folding.
Differential Revision: http://reviews.llvm.org/D5698
llvm-svn: 219557
Summary: This fixes PR21235.
Test Plan: Includes an automated test.
Reviewers: hansw
Subscribers: cfe-commits
Differential Revision: http://reviews.llvm.org/D5718
llvm-svn: 219551
Assertion failed: "Computed __func__ length differs from type!"
Reworked PredefinedExpr representation with internal StringLiteral field for function declaration.
Differential Revision: http://reviews.llvm.org/D5365
llvm-svn: 219393
Windows TLS relies on indexing through a tls_index in order to get at
the DLL's thread local variables. However, this index is not exported
along with the variable: it is assumed that all accesses to thread local
variables are inside the same module which created the variable in the
first place.
While there are several implementation techniques we could adopt to fix
this (notably, the Itanium ABI gets this for free), it is not worth the
heroics.
Instead, let's just ban this combination. We could revisit this in the
future if we need to.
This fixes PR21111.
llvm-svn: 219049
Richard noted in the review of r217349 that extra handling of
__builtin_assume_aligned inside of the expression evaluator was needed. He was
right, and this should address the concerns raised, namely:
1. The offset argument to __builtin_assume_aligned can have side effects, and
we need to make sure that all arguments are properly evaluated.
2. If the alignment assumption does not hold, that introduces undefined
behavior, and undefined behavior cannot appear inside a constexpr.
and hopefully the diagnostics produced are detailed enough to explain what is
going on.
llvm-svn: 218992
This CL has caused bootstrap failures on Linux and OSX buildbots running with -Werror.
Example report from http://lab.llvm.org:8011/builders/sanitizer-x86_64-linux/builds/13183/steps/bootstrap%20clang/logs/stdio:
================================================================
[ 91%] Building CXX object tools/clang/tools/diagtool/CMakeFiles/diagtool.dir/ShowEnabledWarnings.cpp.o
In file included from /home/dtoolsbot/build/sanitizer-x86_64-linux/build/llvm/lib/Target/R600/AMDGPUISelDAGToDAG.cpp:20:
In file included from /home/dtoolsbot/build/sanitizer-x86_64-linux/build/llvm/lib/Target/R600/SIISelLowering.h:19:
/home/dtoolsbot/build/sanitizer-x86_64-linux/build/llvm/lib/Target/R600/SIInstrInfo.h:71:8: error: 'getLdStBaseRegImmOfs' overrides a member function but is not marked 'override' [-Werror,-Winconsistent-missing-override]
bool getLdStBaseRegImmOfs(MachineInstr *LdSt,
^
/home/dtoolsbot/build/sanitizer-x86_64-linux/build/llvm/include/llvm/Target/TargetInstrInfo.h:815:16: note: overridden virtual function is here
virtual bool getLdStBaseRegImmOfs(MachineInstr *LdSt,
^
================================================================
llvm-svn: 218969
for an overriding method if class has at least one
'override' specified on one of its methods.
Reviewed by Doug Gregor. rdar://18295240
(I have already checked in all llvm files with missing 'override'
methods and Bob Wilson has fixed a TableGen of FastISel so
no warnings are expected from build of llvm after this patch.
I have already verified this).
llvm-svn: 218925
This adds support for the align_value attribute. This attribute is supported by
Intel's compiler (versions 14.0+), and several of my HPC users have requested
support in Clang. It specifies an alignment assumption on the values to which a
pointer points, and is used by numerical libraries to encourage efficient
generation of vector code.
Of course, we already have an aligned attribute that can specify enhanced
alignment for a type, so why is this additional attribute important? The
problem is that if you want to specify that an input array of T is, say,
64-byte aligned, you could try this:
typedef double aligned_double attribute((aligned(64)));
void foo(aligned_double *P) {
double x = P[0]; // This is fine.
double y = P[1]; // What alignment did those doubles have again?
}
the access here to P[1] causes problems. P was specified as a pointer to type
aligned_double, and any object of type aligned_double must be 64-byte aligned.
But if P[0] is 64-byte aligned, then P[1] cannot be, and this access causes
undefined behavior. Getting round this problem requires a lot of awkward
casting and hand-unrolling of loops, all of which is bad.
With the align_value attribute, we can accomplish what we'd like in a well
defined way:
typedef double *aligned_double_ptr attribute((align_value(64)));
void foo(aligned_double_ptr P) {
double x = P[0]; // This is fine.
double y = P[1]; // This is fine too.
}
This attribute does not create a new type (and so it not part of the type
system), and so will only "propagate" through templates, auto, etc. by
optimizer deduction after inlining. This seems consistent with Intel's
implementation (thanks to Alexey for confirming the various Intel-compiler
behaviors).
As a final note, I would have chosen to call this aligned_value, not
align_value, for better naming consistency with the aligned attribute, but I
think it would be more useful to users to adopt Intel's name.
llvm-svn: 218910
Clang warns (treated as error by default, but still ignored in system headers)
when passing non-POD arguments to variadic functions, and generates a trap
instruction to crash the program if that code is ever run.
Unfortunately, MSVC happily generates code for such calls without a warning,
and there is code in system headers that use it.
This makes Clang not insert the trap instruction when in -fms-compatibility
mode, while still generating the warning/error message.
Differential Revision: http://reviews.llvm.org/D5492
llvm-svn: 218640
In addition to __builtin_assume_aligned, GCC also supports an assume_aligned
attribute which specifies the alignment (and optional offset) of a function's
return value. Here we implement support for the assume_aligned attribute by making
use of the @llvm.assume intrinsic.
llvm-svn: 218500
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
A record which contains a flexible array member is itself a flexible
array member. A struct which contains such a record should also
consider itself to be a flexible array member.
llvm-svn: 218378
lists. Since the fields are inititalized one at a time, using a field with
lower index to initialize a higher indexed field should not be warned on.
llvm-svn: 218339
that function, and apart from being slow, this is unnecessary: ADL can trigger
instantiations that are not permitted here. The standard isn't *completely*
clear here, but this seems like the intent, and in any case this approach is
permitted by [temp.inst]p7.
llvm-svn: 218330
warns when a guarded variable is passed by reference as a function argument.
This is released as a separate warning flag, because it could potentially
break existing code that uses thread safety analysis.
llvm-svn: 218087
The reasoning is that this construct is accepted by all compilers and valid in
C++11, so it doesn't seem like a useful warning to have enabled by default.
Building with -pedantic, -Wbind-to-temporary-copy, or -Wc++98-compat still
shows the warning.
The motivation is that I built re2, and this was the only warning that was
emitted during the build. Both changing re2 to fix the warning and detecting
clang and suppressing the warning in re2's build seem inferior than just giving
the compiler a good default for this warning.
Also move the cxx98compat version of this warning to CXX98CompatPedantic, and
update tests accordingly.
llvm-svn: 218008
We would end up marking the vtable of the derived class as used for no
reason. Because the call itself is qualified, it is never virtual, and
the vtable of the derived class isn't helpful. We would end up rejecting
code that MSVC accepts for no benefit.
See http://crbug.com/413478
llvm-svn: 217910
constexpr function. Part of this fix is a tentative fix for an as-yet-unfiled
core issue (we're missing a prohibition against reading mutable members from
unions via a trivial constructor/assignment, since that doesn't perform an
lvalue-to-rvalue conversion on the members).
llvm-svn: 217852
In Parser::ParseCXXClassMemberDeclaration(), it was possible to change
isAccessDecl = NextToken().is(tok::kw_operator);
to
isAccessDecl = false;
and no tests would fail. Now there's coverage for this.
llvm-svn: 217519
This makes use of the recently-added @llvm.assume intrinsic to implement a
__builtin_assume(bool) intrinsic (to provide additional information to the
optimizer). This hooks up __assume in MS-compatibility mode to mirror
__builtin_assume (the semantics have been intentionally kept compatible), and
implements GCC's __builtin_assume_aligned as assume((p - o) & mask == 0). LLVM
now contains special logic to deal with assumptions of this form.
llvm-svn: 217349
The warning warns on TypedefNameDecls -- typedefs and C++11 using aliases --
that are !isReferenced(). Since the isReferenced() bit on TypedefNameDecls
wasn't used for anything before this warning it wasn't always set correctly,
so this patch also adds a few missing MarkAnyDeclReferenced() calls in
various places for TypedefNameDecls.
This is made a bit complicated due to local typedefs possibly being used only
after their local scope has closed. Consider:
template <class T>
void template_fun(T t) {
typename T::Foo s3foo; // YYY
(void)s3foo;
}
void template_fun_user() {
struct Local {
typedef int Foo; // XXX
} p;
template_fun(p);
}
Here the typedef in XXX is only used at end-of-translation unit, when YYY in
template_fun() gets instantiated. To handle this, typedefs that are unused when
their scope exits are added to a set of potentially unused typedefs, and that
set gets checked at end-of-TU. Typedefs that are still unused at that point then
get warned on. There's also serialization code for this set, so that the
warning works with precompiled headers and modules. For modules, the warning
is emitted when the module is built, for precompiled headers each time the
header gets used.
Finally, consider a function using C++14 auto return types to return a local
type defined in a header:
auto f() {
struct S { typedef int a; };
return S();
}
Here, the typedef escapes its local scope and could be used by only some
translation units including the header. To not warn on this, add a
RecursiveASTVisitor that marks all delcs on local types returned from auto
functions as referenced. (Except if it's a function with internal linkage, or
the decls are private and the local type has no friends -- in these cases, it
_is_ safe to warn.)
Several of the included testcases (most of the interesting ones) were provided
by Richard Smith.
(gcc's spelling -Wunused-local-typedefs is supported as an alias for this
warning.)
llvm-svn: 217298
"protected scope" is very unhelpful here and actively confuses users. Instead,
simply state the nature of the problem in the diagnostic: we cannot jump from
here to there. The notes explain nicely why not.
llvm-svn: 217293
Originally, self reference checking made a double pass over some expressions
to handle reference type checking. Now, allow HandleValue to also check
reference types, and fallback to Visit for unhandled expressions.
llvm-svn: 217203
before retrying the initialization to produce diagnostics. Otherwise, we may
fail to produce any diagnostics, and silently produce invalid AST in a -Asserts
build. Also add a note to this codepath to make it more clear why we were
trying to create a temporary.
llvm-svn: 217197
Scoped lockable objects (mutex guards) are implemented as if it is a
lock itself that is acquired upon construction and unlocked upon
destruction. As it if course needs to be used to actually lock down
something else (a mutex), it keeps track of this knowledge through its
underlying mutex field in its FactEntry.
The problem with this approach is that this only allows us to lock down
a single mutex, so extend the code to use a vector of underlying
mutexes. This, however, makes the code a bit more complex than
necessary, so subclass FactEntry into LockableFactEntry and
ScopedLockableFactEntry and move all the logic that differs between
regular locks and scoped lockables into member functions.
llvm-svn: 217016
don't mark the field as initialized until the next initializer instead of
instantly. Since this checker is AST based, statements are processed in tree
order instead of following code flow. This can result in different warnings
from just reordering the code. Also changed to use one checker per constructor
instead of creating a new checker per field.
class T {
int x, y;
// Already warns
T(bool b) : x(!b ? (1 + y) : (y = 5)) {}
// New warning added here, previously (1 + y) comes after (y = 5) in the AST
// preventing the warning.
T(bool b) : x(b ? (y = 5) : (1 + y)) {}
};
llvm-svn: 216641
Fix r216438 to catch more complicated self-initialized in std::move. For
instance, "Foo f = std::move(cond ? OtherFoo : (UNUSED_VALUE, f));"
Make sure that BinaryConditionalOperator, ConditionalOperator, BinaryOperator
with comma operator, and OpaqueValueExpr perform the correct usage forwarding
across the three uninitialized value checkers.
llvm-svn: 216627
This shouldn't really be allowed, but it comes up in real code (see PR). As
long as the decl hasn't been used there's no technical difficulty in supporting
it, so downgrade the error to a warning.
Differential Revision: http://reviews.llvm.org/D5087
llvm-svn: 216619
the no-arguments case. Don't expand this to an __attribute__((nonnull(A, B,
C))) attribute, since that does the wrong thing for function templates and
varargs functions.
In passing, fix a grammar error in the diagnostic, a crash if
__attribute__((nonnull(N))) is applied to a varargs function,
a bug where the same null argument could be diagnosed multiple
times if there were multiple nonnull attributes referring to it,
and a bug where nonnull attributes would not be accumulated correctly
across redeclarations.
llvm-svn: 216520
We would previously assert (a decl cannot have two DLL attributes) on this code:
template <typename T> struct __declspec(dllimport) S { T f() { return T(); } };
template struct __declspec(dllexport) S<int>;
The problem was that when instantiating, we would take the attribute from the
template even if the instantiation itself already had an attribute.
Also, don't inherit DLL attributes from the template to its members before
instantiation, as the attribute may change.
I couldn't figure out what MinGW does here, so I'm leaving that open. At least
we're not asserting anymore.
llvm-svn: 216340
The situation it is warning about (see PR20725) is not very likely
to be a real problem, and it is unclear what action the user should take
if the warning does fire.
llvm-svn: 216283
Normally we mark all members of exported classes referenced to get them emitted.
However, MSVC doesn't do this for class templates that are implicitly specialized or
just have an explicit instantiation declaration. For such specializations, the members
are emitted when referenced.
The exception is the case when the dllexport attribute is propagated from a base class
to a base class template that doesn't have an explicit attribute: in this case all
methods of the base class template do get instantiated.
llvm-svn: 216145
Changes diagnostic options, language standard options, diagnostic identifiers, diagnostic wording to use c++14 instead of c++1y. It also modifies related test cases to use the updated diagnostic wording.
llvm-svn: 215982
-Wthread-safety umbrella flag, pending updates to documentation. The flag
works, but is likely to be confusing to existing users of -Wthread-safety.
llvm-svn: 215679
Richard Smith