This patch add __builtin_matrix_transpose to Clang, as described in
clang/docs/MatrixTypes.rst.
Reviewers: rjmccall, jfb, rsmith, Bigcheese
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D72778
Reliably mark the loop variable declaration in a range for as having an
invalid initializer if anything goes wrong building the initializer. We
previously based this determination on whether an error was emitted,
which is not a reliable signal due to error suppression (during error
recovery etc).
Also, properly mark the variable as having initializer errors rather
than simply marking it invalid. This is necessary to mark any structured
bindings as invalid too.
This generalizes the previous fix in
936ec89e91.
Summary:
When getting a warning that we release a capability that isn't held it's
sometimes not clear why. So just like we do for double locking, we add a
note on the previous release operation, which marks the point since when
the capability isn't held any longer.
We can find this previous release operation by looking up the
corresponding negative capability.
Reviewers: aaron.ballman, delesley
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D81352
Summary:
The standard std::unique_lock can be constructed to manage a lock without
initially acquiring it by passing std::defer_lock as second parameter.
It can be acquired later by calling lock().
To support this, we use the locks_excluded attribute. This might seem
like an odd choice at first, but its consistent with the other
annotations we support on scoped capability constructors. By excluding
the lock we state that it is currently not in use and the function
doesn't change that, which is exactly what the constructor does.
Along the way we slightly simplify handling of scoped capabilities.
Reviewers: aaron.ballman, delesley
Reviewed By: aaron.ballman
Differential Revision: https://reviews.llvm.org/D81332
Summary:
We lost errorBit for StmtExpr if a recoveryExpr is the result
expr of a CompoundStmt, which will lead to crashes.
```
// `-StmtExpr
// `-CompoundStmt
// `-RecoveryExp
({ invalid(); });
```
Reviewers: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D81154
This patch implements the * binary operator for values of
MatrixType. It adds support for matrix * matrix, scalar * matrix and
matrix * scalar.
For the matrix, matrix case, the number of columns of the first operand
must match the number of rows of the second. For the scalar,matrix variants,
the element type of the matrix must match the scalar type.
Reviewers: rjmccall, anemet, Bigcheese, rsmith, martong
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D76794
trivial.
We previously took a shortcut by assuming that if a subobject had a
trivial copy assignment operator (with a few side-conditions), we would
always invoke it, and could avoid going through overload resolution.
That turns out to not be correct in the presenve of ref-qualifiers (and
also won't be the case for copy-assignments with requires-clauses
either). Use the same logic for lazy declaration of copy-assignments
that we use for all other special member functions.
Previously committed as c57f8a3a20. This
now also includes an extension of LLDB's workaround for handling special
members without the help of Sema to cover copy assignments.
trivial.
We previously took a shortcut by assuming that if a subobject had a
trivial copy assignment operator (with a few side-conditions), we would
always invoke it, and could avoid going through overload resolution.
That turns out to not be correct in the presenve of ref-qualifiers (and
also won't be the case for copy-assignments with requires-clauses
either). Use the same logic for lazy declaration of copy-assignments
that we use for all other special member functions.
In C++17 the postfix-expression of a call expression is sequenced before
each expression in the expression-list and any default argument.
Differential Revision: https://reviews.llvm.org/D58579
Reviewed By: rsmith
parameters with default arguments.
Directly follow the wording by relaxing the AST invariant that all
parameters after one with a default arguemnt also have default
arguments, and removing the diagnostic on missing default arguments
on a pack-expanded parameter following a parameter with a default
argument.
Testing also revealed that we need to special-case explicit
specializations of templates with a pack following a parameter with a
default argument, as such explicit specializations are otherwise
impossible to write. The standard wording doesn't address this case; a
issue has been filed.
This exposed a bug where we would briefly consider a parameter to have
no default argument while we parse a delay-parsed default argument for
that parameter, which is also fixed.
Partially incorporates a patch by Raul Tambre.
Summary:
Add a new warning -Wuninitialized-const-reference as a subgroup of -Wuninitialized to address a bug filed here: https://bugs.llvm.org/show_bug.cgi?id=45624
This warning is controlled by -Wuninitialized and can be disabled by -Wno-uninitialized-const-reference.
The warning is diagnosed when passing uninitialized variables as const reference parameters to a function.
Differential Revision: https://reviews.llvm.org/D79895
This patch implements matrix index expressions
(matrix[RowIdx][ColumnIdx]).
It does so by introducing a new MatrixSubscriptExpr(Base, RowIdx, ColumnIdx).
MatrixSubscriptExprs are built in 2 steps in ActOnMatrixSubscriptExpr. First,
if the base of a subscript is of matrix type, we create a incomplete
MatrixSubscriptExpr(base, idx, nullptr). Second, if the base is an incomplete
MatrixSubscriptExpr, we create a complete
MatrixSubscriptExpr(base->getBase(), base->getRowIdx(), idx)
Similar to vector elements, it is not possible to take the address of
a MatrixSubscriptExpr.
For CodeGen, a new MatrixElt type is added to LValue, which is very
similar to VectorElt. The only difference is that we may need to cast
the type of the base from an array to a vector type when accessing it.
Reviewers: rjmccall, anemet, Bigcheese, rsmith, martong
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D76791
This patch implements the + and - binary operators for values of
MatrixType. It adds support for matrix +/- matrix, scalar +/- matrix and
matrix +/- scalar.
For the matrix, matrix case, the types must initially be structurally
equivalent. For the scalar,matrix variants, the element type of the
matrix must match the scalar type.
Reviewers: rjmccall, anemet, Bigcheese, rsmith, martong
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D76793
We didn't properly build default argument expressions previously -- we
failed to build the wrapper CXXDefaultArgExpr node, which meant that
std::source_location misbehaved, and we didn't perform default argument
instantiation when necessary, which meant that dependent default
arguments in function templates didn't work at all.
arr is a volatile non-local array.
This fixes a recent regression exposed by removing lvalue-to-rvalue
conversion of discarded volatile arrays. In passing, regularize the
rules we use to determine whether '(void)expr;' warns when expr is a
volatile glvalue.
We currently diagnose static data members directly contained in unnamed classes,
but we should also diagnose when they're in a class that is nested (directly or
indirectly) in an unnamed class. Do this by iterating up the list of parent
DeclContexts and checking if any is an unnamed class.
Similarly also check for function or method DeclContexts (which includes things
like blocks and openmp captured statements) as then the class is considered to
be a local class, which means static data members aren't allowed.
Differential Revision: https://reviews.llvm.org/D80295
alignment information on VarDecls in more cases
This commit improves upon https://reviews.llvm.org/D21099. The code that
computes the source alignment now understands array subscript
expressions, binary operators, derived-to-base casts, and several more
expressions.
rdar://problem/59242343
Differential Revision: https://reviews.llvm.org/D78767
My test needs a requires target clause to support inline assembly. This
patch splits out the asm tests into a separate test so we don't skip the
rest of the conditions.
The backends don't seem to properly handle the _ExtInt type in inline
assembly with crashes occurring in many. While the ones I tested seem to
work for powers of 2 < 64 (and some any multiple of 64 greater than
that), it seemed like a better idea to just use of this type in inline
assembly prohibited.
the expression that is passed to it if it has a function type or array
type
lvalue-to-rvalue conversion should only be applied to non-function,
non-array types, but clang was applying the conversion to discarded
value expressions of array types.
rdar://problem/61203170
Differential Revision: https://reviews.llvm.org/D78134
Summary: This allows for suppressing warnings about the conversion function never being called if it overrides a virtual function in a base class.
Differential Revision: https://reviews.llvm.org/D78444
We should check non-dependent element types before creating a
DependentSizedMatrixType. Otherwise we do not generate an error message
for dependent-sized matrix types with invalid non-dependent element
types, if the template is never instantiated. See the make5 struct in
the tests.
It also moves the SEMA template tests to
clang/test/SemaTemplate/matrix-type.cpp and introduces a few more test
cases.
This patch adds a matrix type to Clang as described in the draft
specification in clang/docs/MatrixSupport.rst. It introduces a new option
-fenable-matrix, which can be used to enable the matrix support.
The patch adds new MatrixType and DependentSizedMatrixType types along
with the plumbing required. Loads of and stores to pointers to matrix
values are lowered to memory operations on 1-D IR arrays. After loading,
the loaded values are cast to a vector. This ensures matrix values use
the alignment of the element type, instead of LLVM's large vector
alignment.
The operators and builtins described in the draft spec will will be added in
follow-up patches.
Reviewers: martong, rsmith, Bigcheese, anemet, dexonsmith, rjmccall, aaron.ballman
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D72281
Objective-C++11 and under MS extensions.
This matches the MSVC behavior, and means that Objective-C behaves as a
set of extensions to the base language, rather than replacing the base
language rule with a different one.
The 'class' or 'struct' keyword is only permitted as part of either an
enum definition or a standalone opaque-enum-declaration, not as part of
an elaborated type specifier. We previously failed to diagnose this, and
generally didn't properly implement the restrictions on elaborated type
specifiers for enumeration types.
In passing, also fixed incorrect parsing for enum-bases, which we
previously parsed as a type-name, but are actually a type-specifier-seq.
This matters for cases like 'enum E : int *p;', which is valid as a
Microsoft extension.
Plus some minor parse diagnostic improvements.
Bumped the recently-added ExtWarn for 'enum E : int x;' to be
DefaultError; this is not an intentional extension, so producing an
error by default seems appropriate, but the warning flag to disable it
may still be useful for code written against old Clang. The same
treatment is given here to the diagnostic for 'enum class E x;', which
we similarly have incorrectly accepted for many years. These diagnostics
continue to be suppressed under -fms-extensions and when compiling
Objective-C code. We will need to decide separately whether Objective-C
should follow the C++ rules or the (older) MSVC rules.
Previously we implemented non-standard disambiguation rules to
distinguish an enum-base from a bit-field but otherwise treated a :
after an elaborated-enum-specifier as introducing an enum-base. That
misparses various examples (anywhere an elaborated-type-specifier can
appear followed by a colon, such as within a ternary operator or
_Generic).
We now implement the C++11 rules, with the old cases accepted as
extensions where that seemed reasonable. These amount to:
* an enum-base must always be accompanied by an enum definition (except
in a standalone declaration of the form 'enum E : T;')
* in a member-declaration, 'enum E :' always introduces an enum-base,
never a bit-field
* in a type-specifier (or similar context), 'enum E :' is not
permitted; the colon means whatever else it would mean in that
context.
Fixed underlying types for enums are also permitted in Objective-C and
under MS extensions, plus as a language extension in all other modes.
The behavior in ObjC and MS extensions modes is unchanged (but the
bit-field disambiguation is a bit better); remaining language modes
follow the C++11 rules.
Fixes PR45726, PR39979, PR19810, PR44941, and most of PR24297, plus C++
core issues 1514 and 1966.
Implicit methods for structs can confuse the warning, so exclude checking
the Decl's that are implicit. Implicit Decl's for lambdas still need to
be checked, so skipping all implicit Decl's won't work.
Differential Revision: https://reviews.llvm.org/D79548
The built-in SVE types are supposed to be treated as opaque types.
This means that for initialisation purposes they should be treated
as a single unit, much like a scalar type.
However, as Eli pointed out, actually using "scalar" in the diagnostics
is likely to cause confusion, given the types are logically vectors.
The patch therefore uses custom diagnostics or generalises existing
ones. Some of the messages use the word "indivisible" to try to make
it clear(er) that these types can't be initialised elementwise.
I don't think it's possible to trigger warn_braces_around_(scalar_)init
for sizeless types as things stand, since the types can't be used as
members or elements of more complex types. But it seemed better to be
consistent with ext_many_braces_around_(scalar_)init, so the patch
changes it anyway.
Differential Revision: https://reviews.llvm.org/D76689
Summary:
The crash is triggered on accessing a null InitExpr.
For group declaration, e.g. `auto c = a, &d = {a};`, what's happening:
1. each VarDecl is built separately during the parsing stage.
2. perform the semantic analysis (Sema::BuildDeclaratorGroup) to check
whether the type of the two VarDecl is the same, if not mark it as invalid.
in step 1, VarDecl c and d are built, both of them are valid (after D77395),
but d is without the InitExpr attached (under -fno-recovery-ast), crash
happens in step 2 when accessing the source range of d's InitExpr.
Reviewers: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D79473
There are some lookup oddities with these as reported in PR45780, and
GCC doesn't support these behaviors at all. To be more consistent with
GCC and prevent the crashes caused by our lookup issues, nip the problem
in the bud and prohibit enums here.
I added a limit to make sure that _ExtInt isn't exposed on systems that
haven't considered it in their ABI. The ext-int.cpp Sema test didn't
have a triple, so on non x86/x86_64 it would fail with this new error.
This patch adds said triple to make sure this passes.
Fix a few bugs where we would fail to properly determine header to
module correspondence when determining whether to suggest a #include or
import, and suggest a #include more often in language modes where there
is no import syntax. Generally, if the target is in a header with
include guards or #pragma once, we should suggest either #including or
importing that header, and not importing a module that happens to
textually include it.
In passing, improve the notes we attach to the corresponding
diagnostics: calling an entity that we couldn't see "previous" is
confusing.
Looks like this was just a copy & paste mistake from
getDependentSizedExtVectorType. rdar://60092165
Differential revision: https://reviews.llvm.org/D79012
Summary:
We extend the behavior for local functions and methods of local classes
to lambdas in variable initializers. The initializer is not a separate
scope, but we treat it as such.
We also remove the (faulty) instantiation of default arguments in
TreeTransform::TransformLambdaExpr, because it doesn't do proper
initialization, and if it did, we would do it twice (and thus also emit
eventual errors twice).
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D76038
Summary:
This patch contains 2 separate changes:
1) the initializer of a variable should play no part in decl "invalid" bit;
2) preserve the invalid initializer via recovery exprs;
With 1), we will regress the diagnostics (one big regression is that we loose
the "selected 'begin' function with iterator type" diagnostic in for-range stmt;
but with 2) together, we don't have regressions (the new diagnostics seems to be
improved).
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78116
This reverts commit 61ba1481e2.
I'm reverting this because it breaks the lldb build with
incomplete switch coverage warnings. I would fix it forward,
but am not familiar enough with lldb to determine the correct
fix.
lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp:3958:11: error: enumeration values 'DependentExtInt' and 'ExtInt' not handled in switch [-Werror,-Wswitch]
switch (qual_type->getTypeClass()) {
^
lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp:4633:11: error: enumeration values 'DependentExtInt' and 'ExtInt' not handled in switch [-Werror,-Wswitch]
switch (qual_type->getTypeClass()) {
^
lldb/source/Plugins/TypeSystem/Clang/TypeSystemClang.cpp:4889:11: error: enumeration values 'DependentExtInt' and 'ExtInt' not handled in switch [-Werror,-Wswitch]
switch (qual_type->getTypeClass()) {
Introduction/Motivation:
LLVM-IR supports integers of non-power-of-2 bitwidth, in the iN syntax.
Integers of non-power-of-two aren't particularly interesting or useful
on most hardware, so much so that no language in Clang has been
motivated to expose it before.
However, in the case of FPGA hardware normal integer types where the
full bitwidth isn't used, is extremely wasteful and has severe
performance/space concerns. Because of this, Intel has introduced this
functionality in the High Level Synthesis compiler[0]
under the name "Arbitrary Precision Integer" (ap_int for short). This
has been extremely useful and effective for our users, permitting them
to optimize their storage and operation space on an architecture where
both can be extremely expensive.
We are proposing upstreaming a more palatable version of this to the
community, in the form of this proposal and accompanying patch. We are
proposing the syntax _ExtInt(N). We intend to propose this to the WG14
committee[1], and the underscore-capital seems like the active direction
for a WG14 paper's acceptance. An alternative that Richard Smith
suggested on the initial review was __int(N), however we believe that
is much less acceptable by WG14. We considered _Int, however _Int is
used as an identifier in libstdc++ and there is no good way to fall
back to an identifier (since _Int(5) is indistinguishable from an
unnamed initializer of a template type named _Int).
[0]https://www.intel.com/content/www/us/en/software/programmable/quartus-prime/hls-compiler.html)
[1]http://www.open-std.org/jtc1/sc22/wg14/www/docs/n2472.pdf
Differential Revision: https://reviews.llvm.org/D73967
Summary:
Previously, we treated CXXUuidofExpr as quite a special case: it was the
only kind of expression that could be a canonical template argument, it
could be a constant lvalue base object, and so on. In addition, we
represented the UUID value as a string, whose source form we did not
preserve faithfully, and that we partially parsed in multiple different
places.
With this patch, we create an MSGuidDecl object to represent the
implicit object of type 'struct _GUID' created by a UuidAttr. Each
UuidAttr holds a pointer to its 'struct _GUID' and its original
(as-written) UUID string. A non-value-dependent CXXUuidofExpr behaves
like a DeclRefExpr denoting that MSGuidDecl object. We cache an APValue
representation of the GUID on the MSGuidDecl and use it from constant
evaluation where needed.
This allows removing a lot of the special-case logic to handle these
expressions. Unfortunately, many parts of Clang assume there are only
a couple of interesting kinds of ValueDecl, so the total amount of
special-case logic is not really reduced very much.
This fixes a few bugs and issues:
* PR38490: we now support reading from GUID objects returned from
__uuidof during constant evaluation.
* Our Itanium mangling for a non-instantiation-dependent template
argument involving __uuidof no longer depends on which CXXUuidofExpr
template argument we happened to see first.
* We now predeclare ::_GUID, and permit use of __uuidof without
any header inclusion, better matching MSVC's behavior. We do not
predefine ::__s_GUID, though; that seems like a step too far.
* Our IR representation for GUID constants now uses the correct IR type
wherever possible. We will still fall back to using the
{i32, i16, i16, [8 x i8]}
layout if a definition of struct _GUID is not available. This is not
ideal: in principle the two layouts could have different padding.
Reviewers: rnk, jdoerfert
Subscribers: arphaman, cfe-commits, aeubanks
Tags: #clang
Differential Revision: https://reviews.llvm.org/D78171
Summary:
This patch would cause clang emit more diagnostics, but it is much better than https://reviews.llvm.org/D76831
```cpp
struct A {
A(int);
~A() = delete;
};
void k() {
A a;
}
```
before the patch:
/tmp/t3.cpp:24:5: error: no matching constructor for initialization of 'A'
A a;
^
/tmp/t3.cpp:20:3: note: candidate constructor not viable: requires 1 argument, but 0 were provided
A(int);
^
/tmp/t3.cpp:19:8: note: candidate constructor (the implicit copy constructor) not viable: requires 1 argument, but 0 were provided
struct A {
After the patch:
/tmp/t3.cpp:24:5: error: no matching constructor for initialization of 'A'
A a;
^
/tmp/t3.cpp:20:3: note: candidate constructor not viable: requires 1 argument, but 0 were provided
A(int);
^
/tmp/t3.cpp:19:8: note: candidate constructor (the implicit copy constructor) not viable: requires 1 argument, but 0 were provided
struct A {
^
/tmp/t3.cpp:24:5: error: attempt to use a deleted function
A a;
^
/tmp/t3.cpp:21:3: note: '~A' has been explicitly marked deleted here
~A() = delete;
Reviewers: sammccall
Reviewed By: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D77395
In the MS C++ ABI, the complete destructor variant for a class with
virtual bases is emitted whereever it is needed, instead of directly
alongside the base destructor variant. The complete destructor calls the
base destructor of the current class and the base destructors of each
virtual base. In order for this to work reliably, translation units that
use the destructor of a class also need to mark destructors of virtual
bases of that class used.
Fixes PR38521
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D77081
constructor with default arguments.
We used to try to rebuild the call as a call to the faked-up inherited
constructor, which is only a placeholder and lacks (for example) default
arguments. Instead, build the call by reference to the original
constructor.
In passing, add a note to say where a call that recursively uses a
default argument from within itself occurs. This is usually pretty
obvious, but still at least somewhat useful, and would have saved
significant debugging time for this particular bug.
as invalid.
We create those when forming trivial type source information with no
associated location, which, unfortunately, we do create in some cases
(when a TreeTransform with no base location is used to transform a
QualType).
This would previously lead to rejects-valid bugs when we misinterpreted
these constructs as having no nested-name-specifier.
memchr consistent and comprehensible, and document them.
We previously allowed evaluation of memcmp on arrays of integers of any
size, so long as the call evaluated to 0, and allowed evaluation of
memchr on any array of integral type of size 1 (including enums). The
purpose of constant-evaluating these builtins is only to support
constexpr std::char_traits, so we now consistently allow them on arrays
of (possibly signed or unsigned) char only.
Summary:
This matches llvm::VectorType.
It moves the size from the type bitfield into VectorType, increasing size by 8
bytes (including padding of 4). This is OK as we don't expect to create terribly
many of these types.
c.f. D77313 which enables large power-of-two sizes without growing VectorType.
Reviewers: efriedma, hokein
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D77335
adb290d974 added a new case to
err_atomic_specifier_bad_type. The diagnostic has two %select's
controlled by the same argument, but only the first was updated to have
the new case. Add the extra case for the second %select and add a
test case that exercises the last case.
-Wthread-safety was failing to detect certain AST patterns it should
detect. Make the pattern detection a bit more comprehensive.
Due to an unrelated bug involving template instantiation, this showed up
as a regression in 10.0 vs. 9.0 in the original bug report. The included
testcase fails on older versions of clang, though.
Fixes https://bugs.llvm.org/show_bug.cgi?id=45323 .
Differential Revision: https://reviews.llvm.org/D76943
The __builtin_stdarg_start is the legacy spelling of __builtin_va_start.
It should behave exactly the same, but for the last 9 years it would
behave subtly different for diagnostics. Follow the change from
29ad95b232 to require custom type checking.
scope.
There are a few contexts in which we assume a name is a template name;
if such a context is one where we should perform an unqualified lookup,
and lookup finds nothing, we would form a dependent template name even
if the name is not dependent. This happens in particular for the lookup
of a pseudo-destructor.
In passing, rename ActOnDependentTemplateName to just ActOnTemplateName
given that we apply it for non-dependent template names too.
Instead of bailing out of parsing when we encounter an invalid
template-name or template arguments in a template-id, produce an
annotation token describing the invalid construct.
This avoids duplicate errors and generally allows us to recover better.
In principle we should be able to extend this to store some kinds of
invalid template-id in the AST for tooling use, but that isn't handled
as part of this change.
Built-in SVE types are trivial, since they're trivially copyable
and support default construction.
Differential Revision: https://reviews.llvm.org/D76692
SVE types are trivially copyable: they can be copied simply
by reproducing the byte representation of the source object.
Differential Revision: https://reviews.llvm.org/D76691
This reverts commit 0788acbccb.
This reverts commit c2d7a1f79cedfc9fcb518596aa839da4de0adb69: Revert "[clangd] Add test for FindTarget+RecoveryExpr (which already works). NFC"
It causes a crash on invalid code:
class X {
decltype(unresolved()) foo;
};
constexpr int s = sizeof(X);
Summary: Previously, we dropped the AST node for nonexistent member exprs.
Reviewers: sammccall
Subscribers: cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76764
As reported in PR45298 and PR45299, vector_size type checking would
crash when done in a situation where the scalar is dependent, such as
a member of the current instantiation.
This is because the scalar checking ensures that you can implicitly
convert a value to a vector-type as long as it doesn't require
truncation. It does this by using the constant evaluator to get the
value as a float. Unfortunately, if the scalar is dependent (such as a
member of the current instantiation), we would hit the assert in the
evaluator.
This patch suppresses the truncation- of-value check in the first phase
of translation. All values are properly errored upon instantiation. This
has one minor regression, in that previously in a non-asserts build,
template<typename T>
struct S {
float4 f(float4 f) {
return k + f;
}
static constexpr k = 1.1; // causes a truncation on conversion.
};
would error immediately. Because 'k' is value dependent (as a
member-of-the-current-instantiation), this would still be evaluatable
(despite normally asserting). Due to this patch, this diagnostic is
delayed until instantiation time.
Summary:
Changes:
- handle immediate invocations for constructors.
- add tests
after this patch i believe the implementation of consteval is nearly standard compliant, but IR-gen still needs to be taught not to emit consteval declarations.
Reviewers: rsmith
Reviewed By: rsmith
Subscribers: wchilders
Differential Revision: https://reviews.llvm.org/D74007
Sizeless types can't be used with "new", so it doesn't make sense
to use them with "delete" either. The SVE ACLE therefore doesn't
allow that.
This is slightly stronger than for normal incomplete types, since:
struct S;
void f(S *s) { delete s; }
is (by necessity) just a default-on warning rather than an error.
Differential Revision: https://reviews.llvm.org/D76219
new-expressions for a type T require sizeof(T) to be computable,
so the SVE ACLE does not allow them for sizeless types. At the moment:
auto f() { return new __SVInt8_t; }
creates a call to operator new with a zero size:
%call = call noalias nonnull i8* @_Znwm(i64 0)
This patch reports an appropriate error instead.
Differential Revision: https://reviews.llvm.org/D76218
In the current SVE ACLE spec, the usual rules for throwing and
catching incomplete types also apply to sizeless types. However,
throwing pointers to sizeless types should not pose any real difficulty,
so as an extension, the clang implementation allows that.
This patch enforces these rules for catch statements.
Differential Revision: https://reviews.llvm.org/D76090
Summary:
The same rules for throwing and catching incomplete types also apply
to sizeless types. This patch enforces that for throw statements.
It also make sure that we use "sizeless type" rather "incomplete type"
in the associated message. (Both are correct, but "sizeless type" is
more specific and hopefully more user-friendly.)
The SVE ACLE simply extends the rule for incomplete types to
sizeless types. However, throwing pointers to sizeless types
should not pose any real difficulty, so as an extension,
the clang implementation allows that.
Reviewers: sdesmalen, efriedma, rovka, rjmccall
Subscribers: tschuett, rkruppe, psnobl, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76088
In the current SVE ACLE spec, the usual rules for throwing and
catching incomplete types also apply to sizeless types. However,
throwing pointers to sizeless types should not pose any real difficulty,
so as an extension, the clang implementation allows that.
This patch enforces these rules for explicit exception specs.
Differential Revision: https://reviews.llvm.org/D76087
This patch completes a trio of changes related to arrays of
sizeless types. It rejects various forms of arithmetic on
pointers to sizeless types, in the same way as for other
incomplete types.
Differential Revision: https://reviews.llvm.org/D76086
clang currently accepts:
__SVInt8_t &foo1(__SVInt8_t *x) { return *x; }
__SVInt8_t &foo2(__SVInt8_t *x) { return x[1]; }
The first function is valid ACLE code and generates correct LLVM IR
(and assembly code). But the second function is invalid for the
same reason that arrays of sizeless types are. Trying to code-generate
the function leads to:
llvm/include/llvm/Support/TypeSize.h:126: uint64_t llvm::TypeSize::getFixedSize() const: Assertion `!IsScalable && "Request for a fixed size on a s
calable object"' failed.
Another problem is that:
template<typename T>
constexpr __SIZE_TYPE__ f(T *x) { return &x[1] - x; }
typedef int arr1[f((int *)0) - 1];
typedef int arr2[f((__SVInt8_t *)0) - 1];
produces:
a.cpp:2:48: warning: subtraction of pointers to type '__SVInt8_t' of zero size has undefined behavior [-Wpointer-arith]
constexpr __SIZE_TYPE__ f(T *x) { return &x[1] - x; }
~~~~~ ^ ~
a.cpp:4:18: note: in instantiation of function template specialization 'f<__SVInt8_t>' requested here
typedef int arr2[f((__SVInt8_t *)0) - 1];
This patch reports an appropriate diagnostic instead.
Differential Revision: https://reviews.llvm.org/D76084
TryAnnotateTypeConstraint could annotate a template-id which doesn't end up being a type-constraint,
in which case control flow would incorrectly flow into ParseImplicitInt.
Reenter the loop in this case.
Enable relevant tests for C++20. This required disabling typo-correction during TryAnnotateTypeConstraint
and changing a test case which is broken due to a separate bug (will be reported and handled separately).
transparent context.
(The same crash would happen if a class template with a friend was
declared in an 'export' block, but there are more issues with that
case.)
The SVE ACLE doesn't allow arrays of sizeless types. At the moment
clang accepts the TU:
__SVInt8_t x[2];
but trying to code-generate it triggers the LLVM assertion:
llvm/lib/IR/Type.cpp:588: static llvm::ArrayType* llvm::ArrayType::get(llvm::Type*, uint64_t): Assertion `isValidElementType(ElementType) && "Invalid type for array element!"' failed.
This patch reports an appropriate error instead.
The rules are slightly more restrictive than for general incomplete types.
For example:
struct s;
typedef struct s arr[2];
is valid as far as it goes, whereas arrays of sizeless types are
invalid in all contexts. BuildArrayType therefore needs a specific
check for isSizelessType in addition to the usual handling of
incomplete types.
Differential Revision: https://reviews.llvm.org/D76082
Since fields can't have sizeless type, it also doesn't make sense
to capture sizeless types by value in lambda expressions. This patch
makes sure that we diagnose that and that we use "sizeless type" rather
"incomplete type" in the associated message. (Both are correct, but
"sizeless type" is more specific and hopefully more user-friendly.)
Differential Revision: https://reviews.llvm.org/D75738
The SVE ACLE doesn't allow fields to have sizeless type. At the moment
clang accepts things like:
struct s { __SVInt8_t x; } y;
but trying to code-generate it leads to LLVM asserts like:
llvm/include/llvm/Support/TypeSize.h:126: uint64_t llvm::TypeSize::getFixedSize() const: Assertion `!IsScalable && "Request for a fixed size on a scalable object"' failed.
This patch adds an associated clang diagnostic.
Differential Revision: https://reviews.llvm.org/D75737
function and an overridden function until we know whether the overriding
function is deleted.
We previously did these checks when we first built the declaration,
which was too soon in some cases. We now defer all these checks to the
end of the class.
Also add missing check that a consteval function cannot override a
non-consteval function and vice versa.
clang accepts a TU containing just:
__SVInt8_t x;
However, sizeless types are not allowed to have static or thread-local
storage duration and trying to code-generate the TU triggers an LLVM
fatal error:
Globals cannot contain scalable vectors
<vscale x 16 x i8>* @x
fatal error: error in backend: Broken module found, compilation aborted!
This patch adds an associated clang diagnostic.
Differential Revision: https://reviews.llvm.org/D75736
It would be difficult to guarantee atomicity for sizeless types,
so the SVE ACLE makes atomic sizeless types invalid. As it happens,
we already rejected them before the patch, but for the wrong reason:
error: _Atomic cannot be applied to type 'svint8_t' (aka '__SVInt8_t')
which is not trivially copyable
The SVE types should be treated as trivially copyable; a later
patch fixes that.
Differential Revision: https://reviews.llvm.org/D75734
A previous patch rejected alignof for sizeless types. This patch
extends that to cover the "aligned" attribute and _Alignas. Since
sizeless types are not meant to be used for long-term data, cannot
be used in aggregates, and cannot have static storage duration,
there shouldn't be any need to fiddle with their alignment.
Like with alignof, this is a conservative position that can be
relaxed in future if it turns out to be too restrictive.
Differential Revision: https://reviews.llvm.org/D75573
clang current accepts:
void foo1(__SVInt8_t *x, __SVInt8_t *y) { *x = *y; }
void foo2(__SVInt8_t *x, __SVInt8_t *y) {
memcpy(y, x, sizeof(__SVInt8_t));
}
The first function is valid ACLE code and generates correct LLVM IR.
However, the second function is invalid ACLE code and generates a
zero-length memcpy. The point of this patch is to reject the use
of sizeof in the second case instead.
There's no similar wrong-code bug for alignof. However, the SVE ACLE
conservatively treats alignof in the same way as sizeof, just as the
C++ standard does for incomplete types. The idea is that layout of
sizeless types is an implementation property and isn't defined at
the language level.
Implementation-wise, the patch adds a new CompleteTypeKind enum
that controls whether RequireCompleteType & friends accept sizeless
built-in types. For now the default is to maintain the status quo
and accept sizeless types. However, the end of the series will flip
the default and remove the Default enum value.
The patch also adds new ...CompleteSized... wrappers that callers can
use if they explicitly want to reject sizeless types. The callers then
use diagnostics that have an extra 0/1 parameter to indicats whether
the type is sizeless or not.
The idea is to have three cases:
1. calls that explicitly reject sizeless types, with a tweaked diagnostic
for the sizeless case
2. calls that explicitly allow sizeless types
3. normal/old-style calls that don't make an explicit choice either way
Once the default is flipped, the 3. calls will conservatively reject
sizeless types, using the same diagnostic as for other incomplete types.
Differential Revision: https://reviews.llvm.org/D75572
This patch adds C and C++ tests for various uses of SVE types.
The tests cover valid uses that are already (correctly) accepted and
invalid uses that are already (correctly) rejected. Later patches
will expand the tests as they fix other cases.[*]
Some of the tests for invalid uses aren't obviously related to
scalable vectors. Part of the reason for having them is to make
sure that the quality of the error message doesn't regress once/if
the types are treated as incomplete types.
[*] These later patches all fix invalid uses that are being incorrectly
accepted. I don't know of any cases in which valid uses are being
incorrectly rejected. In other words, this series is all about
diagnosing invalid code rather than enabling something new.
Differential Revision: https://reviews.llvm.org/D75571
Bruno Ricci