This reverts commit b55d723ed6.
Reapply Modify FPFeatures to use delta not absolute settings
To solve https://bugs.llvm.org/show_bug.cgi?id=46166 where the
floating point settings in PCH files aren't compatible, rewrite
FPFeatures to use a delta in the settings rather than absolute settings.
With this patch, these floating point options can be benign.
Reviewers: rjmccall
Differential Revision: https://reviews.llvm.org/D81869
Summary:
Check that the co_await promise.final_suspend() does not potentially throw again after we have resolved dependent types.
This takes care of the cases where promises types are templated.
Added test cases for this scenario and confirmed that the checks happen now.
Also run libcxx tests locally to make sure all tests pass.
Reviewers: Benabik, lewissbaker, junparser, modocache
Reviewed By: modocache
Subscribers: modocache, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D82332
C++ unqualified name lookup searches template parameter scopes
immediately after finishing searching the entity the parameters belong
to. (Eg, for a class template, you search the template parameter scope
after looking in that class template and its base classes and before
looking in the scope containing the class template.) This is complicated
by the fact that scope lookup within a template parameter scope looks in
a different sequence of places prior to reaching the end of the
declarator-id in the template declaration.
We used to approximate the proper lookup rule with a hack in the scope /
decl context walk inside name lookup. Now we instead compute the lookup
parent for each template parameter scope.
In order to get this right, we now make sure to enter a distinct Scope
for each template parameter scope, and make sure to re-enter the
enclosing class scopes properly when handling delay-parsed regions
within a class.
templated class.
When a defaulted operator<=> results in the injection of a defaulted
operator==, that operator== can be named by unqualified name within the
same class, even if the class is templated. To make this work, perform
the transform from defaulted operator<=> to defaulted operator== in the
template definition context instead of the template instantiation
context.
This results in our substituting into a declaration from a context where
we don't have a full list of template arguments (or indeed any), for
which we are now more careful to not spuriously instantiate declarations
that are not dependent on the arguments we're substituting.
specializations and those that are done as part of rewrites.
Do not create Subst* nodes in the latter. We previously had a hybrid of
these two behaviors where we would only create some Subst* nodes but not
others during deduction guide rewrites.
No functional change intended, but the resulting ASTs are more
principled.
Summary:
This patch addresses https://bugs.llvm.org/show_bug.cgi?id=46256
The spec of coroutine requires that the expression co_await promise.final_suspend() shall not be potentially-throwing.
To check this, we recursively look at every call (including Call, MemberCall, OperatorCall and Constructor) in all code
generated by the final suspend, and ensure that the callees are declared with noexcept. We also look at any returned data
type that requires explicit destruction, and check their destructors for noexcept.
This patch does not check declarations with dependent types yet, which will be done in future patches.
Updated all tests to add noexcept to the required functions, and added a dedicated test for this patch.
This patch might start to cause existing codebase fail to compile because most people may not have been strict in tagging
all the related functions noexcept.
Reviewers: lewissbaker, modocache, junparser
Reviewed By: modocache
Subscribers: arphaman, junparser, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D82029
Add -fpch-instantiate-templates which makes template instantiations be
performed already in the PCH instead of it being done in every single
file that uses the PCH (but every single file will still do it as well
in order to handle its own instantiations). I can see 20-30% build
time saved with the few tests I've tried.
The change may reorder compiler output and also generated code, but
should be generally safe and produce functionally identical code.
There are some rare cases that do not compile with it,
such as test/PCH/pch-instantiate-templates-forward-decl.cpp. If
template instantiation bailed out instead of reporting the error,
these instantiations could even be postponed, which would make them
work.
Enable this by default for clang-cl. MSVC creates PCHs by compiling
them using an empty .cpp file, which means templates are instantiated
while building the PCH and so the .h needs to be self-contained,
making test/PCH/pch-instantiate-templates-forward-decl.cpp to fail
with MSVC anyway. So the option being enabled for clang-cl matches this.
Differential Revision: https://reviews.llvm.org/D69585
This patch add __builtin_matrix_column_major_store to Clang,
as described in clang/docs/MatrixTypes.rst. In the initial version,
the stride is not optional yet.
Reviewers: rjmccall, jfb, rsmith, Bigcheese
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D72782
This patch add __builtin_matrix_column_major_load to Clang,
as described in clang/docs/MatrixTypes.rst. In the initial version,
the stride is not optional yet.
Reviewers: rjmccall, rsmith, jfb, Bigcheese
Reviewed By: rjmccall
Differential Revision: https://reviews.llvm.org/D72781
templated class.
When a defaulted operator<=> results in the injection of a defaulted
operator==, that operator== can be named by unqualified name within the
same class, even if the class is templated. To make this work, perform
the transform from defaulted operator<=> to defaulted operator== in the
template definition context instead of the template instantiation
context.
This results in our substituting into a declaration from a context where
we don't have a full list of template arguments (or indeed any), for
which we are now more careful to not spuriously instantiate declarations
that are not dependent on the arguments we're substituting.
We weren't re-entering template scopes in the right order, causing this
to break self-host with -fdelayed-template-parsing.
This reverts commit 237c2a23b6.
C++ unqualified name lookup searches template parameter scopes
immediately after finishing searching the entity the parameters belong
to. (Eg, for a class template, you search the template parameter scope
after looking in that class template and its base classes and before
looking in the scope containing the class template.) This is complicated
by the fact that scope lookup within a template parameter scope looks in
a different sequence of places prior to reaching the end of the
declarator-id in the template declaration.
We used to approximate the proper lookup rule with a hack in the scope /
decl context walk inside name lookup. Now we instead compute the lookup
parent for each template parameter scope. This gets the right answer and
as a bonus is substantially simpler and more uniform.
In order to get this right, we now make sure to enter a distinct Scope
for each template parameter scope. (The fact that we didn't before was
already a bug, but not really observable most of the time, since
template parameters can't shadow each other.)
In 1eddce41, I fixed a non-deterministic result problem by switching a
SmallPtrSet to a SmallSetVector to ensure we iterated it
deterministically. Unfortunately, this seems to show a surprisingly
significant compiletime impact.
This patch does 2 things in an attempt to fix this:
First, it makes the 'small size' optimization 4 instead of 2. As these
are pointers, this only increases the size of Sema by 4
sizeof(pointer)s (2 for the set, 2 for the vector).
Second, instead of using SmallSetVector, which is a SmallVector +
SmallDenseSet, it uses a SetVector of SmallVector + SmallPtrSet. The
hope is that the pointer-specific optimizations of the SmallPtrSet will
minimize the impact on compile-time.
We were using llvm::SmallPtrSet for our ODR-use set which was also used
for instantiating the implicit lambda captures. The order in which the
captures are added depends on this, so the lambda's layout ended up
changing. The test just uses floats, but this was noticed with other
types as well.
This test replaces the short-lived SmallPtrSet (it lasts only for an
expression, which, though is a long time for lambdas, is at least not
forever) with a SmallSetVector.
Also invert the sense of the return value.
As pointed out by the FIXME that this change resolves, isHidden() wasn't
a very accurate name for this function.
I haven't yet changed any of the strings that are output in
ASTDumper.cpp / JSONNodeDumper.cpp / TextNodeDumper.cpp in response to
whether isHidden() is set because
a) I'm not sure whether it's actually desired to change these strings
(would appreciate feedback on this), and
b) In any case, I'd like to get this pure rename out of the way first,
without any changes to tests. Changing the strings that are output in
the various ...Dumper.cpp files will require changes to quite a few
tests, and I'd like to make those in a separate change.
Differential Revision: https://reviews.llvm.org/D81392
Reviewed By: rsmith
...before checking that the default argument is valid with
CheckDefaultArgumentVisitor.
Currently the restrictions on a default argument are checked with the visitor
CheckDefaultArgumentVisitor in ActOnParamDefaultArgument before
performing the conversion to the parameter type in SetParamDefaultArgument.
This was fine before the previous patch but now some valid code post-CWG 2346
is rejected:
void test() {
const int i2 = 0;
extern void h2a(int x = i2); // FIXME: ok, not odr-use
extern void h2b(int x = i2 + 0); // ok, not odr-use
}
This is because the reference to i2 in h2a has not been marked yet with
NOUR_Constant. i2 is marked NOUR_Constant when the conversion to the parameter
type is done, which is done just after.
The solution is to do the conversion to the parameter type before checking
the restrictions on default arguments with CheckDefaultArgumentVisitor.
This has the side-benefit of improving some diagnostics.
Differential Revision: https://reviews.llvm.org/D81616
Reviewed By: rsmith
This reverts commit 263390d4f5.
This can still cause bogus errors:
eigen3/Eigen/src/Core/CoreEvaluators.h:94:38: error: call to implicitly-deleted copy constructor of 'unary_evaluator<Eigen::Inverse<Eigen::Matrix<double, 4, 4, 0, 4, 4>>>'
thrust/system/detail/generic/for_each.h:49:3: error: implicit instantiation of undefined template
'thrust::detail::STATIC_ASSERTION_FAILURE<false>'
`noderef` was failing to trigger warnings in some cases related to c++ style
casting. This patch addresses them.
Differential Revision: https://reviews.llvm.org/D77836
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
DiagnosticErrorTrap is usually inappropriate because it indicates
whether an error message was rendered in a given region (and is
therefore affected by -ferror-limit and by suppression of errors if we
see an invalid declaration).
hasErrorOccurred() is usually inappropriate because it indicates
whethere an "error:" message was displayed, regardless of whether the
message was a warning promoted to an error, and therefore depends on
things like -Werror that are usually irrelevant.
Where applicable, CodeSynthesisContexts are used to attach notes to
the first diagnostic produced in a region of code, isnstead of using an
error trap and then attaching a note to whichever diagnostic happened to
be produced last (or suppressing the note if the final diagnostic is a
disabled warning!).
This is mostly NFC.
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
This patch addresses the review comments on r352930:
- Removes redundant diagnostic checking code
- Removes errnoneous use of diag::err_alias_is_definition, which
turned out to be ineffective anyway since functions can be defined later
in the translation unit and avoid detection.
- Adds a test for various invalid cases for import_name and import_module.
This reapplies D59520, with the addition of adding
`InGroup<IgnoredAttributes>` to the new warnings, to fix the
Misc/warning-flags.c failure.
Differential Revision: https://reviews.llvm.org/D59520
This patch addresses the review comments on r352930:
- Removes redundant diagnostic checking code
- Removes errnoneous use of diag::err_alias_is_definition, which
turned out to be ineffective anyway since functions can be defined later
in the translation unit and avoid detection.
- Adds a test for various invalid cases for import_name and import_module.
Differential Revision: https://reviews.llvm.org/D59520
Summary:
This patch upstreams support for a new storage only bfloat16 C type.
This type is used to implement primitive support for bfloat16 data, in
line with the Bfloat16 extension of the Armv8.6-a architecture, as
detailed here:
https://community.arm.com/developer/ip-products/processors/b/processors-ip-blog/posts/arm-architecture-developments-armv8-6-a
The bfloat type, and its properties are specified in the Arm Architecture
Reference Manual:
https://developer.arm.com/docs/ddi0487/latest/arm-architecture-reference-manual-armv8-for-armv8-a-architecture-profile
In detail this patch:
- introduces an opaque, storage-only C-type __bf16, which introduces a new bfloat IR type.
This is part of a patch series, starting with command-line and Bfloat16
assembly support. The subsequent patches will upstream intrinsics
support for BFloat16, followed by Matrix Multiplication and the
remaining Virtualization features of the armv8.6-a architecture.
The following people contributed to this patch:
- Luke Cheeseman
- Momchil Velikov
- Alexandros Lamprineas
- Luke Geeson
- Simon Tatham
- Ties Stuij
Reviewers: SjoerdMeijer, rjmccall, rsmith, liutianle, RKSimon, craig.topper, jfb, LukeGeeson, fpetrogalli
Reviewed By: SjoerdMeijer
Subscribers: labrinea, majnemer, asmith, dexonsmith, kristof.beyls, arphaman, danielkiss, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76077
recommit e03394c6a6 with fix
When implicit HD function calls a function in device compilation,
if one candidate is an implicit HD function, current resolution rule is:
D wins over HD and H
HD and H are equal
this caused regression when there is an otherwise worse D candidate
This patch changes that to
D, HD and H are all equal
The rationale is that we already know for host compilation there is already
a valid candidate in HD and H candidates that will not cause error. Allowing
HD and H gives us a fall back candidate that will not cause error. If D wins,
that means D has to be a better match otherwise, therefore D should also
be a valid candidate that will not cause error. In this way, we can guarantee
no regression.
Differential Revision: https://reviews.llvm.org/D80450
constexpr variables are compile time constants and implicitly const, therefore
they are safe to emit on both device and host side. Besides, in many cases
they are intended for both device and host, therefore it makes sense
to emit them on both device and host sides if necessary.
In most cases constexpr variables are used as rvalue and the variables
themselves do not need to be emitted. However if their address is taken,
then they need to be emitted.
For C++14, clang is able to handle that since clang emits them with
available_externally linkage together with the initializer.
However for C++17, the constexpr static data member of a class or template class
become inline variables implicitly. Therefore they become definitions with
linkonce_odr or weak_odr linkages. As such, they can not have available_externally
linkage.
This patch fixes that by adding implicit constant attribute to
file scope constexpr variables and constexpr static data members
in device compilation.
Differential Revision: https://reviews.llvm.org/D79237
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.
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
Summary:
Diagnostic is emitted if some declaration of unsupported type
declaration is used inside device code.
Memcpy operations for structs containing member with unsupported type
are allowed. Fixed crash on attempt to emit diagnostic outside of the
functions.
The approach is generalized between SYCL and OpenMP.
CUDA/OMP deferred diagnostic interface is going to be used for SYCL device.
Reviewers: rsmith, rjmccall, ABataev, erichkeane, bader, jdoerfert, aaron.ballman
Reviewed By: jdoerfert
Subscribers: guansong, sstefan1, yaxunl, mgorny, bader, ebevhan, Anastasia, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D74387
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.
When I fixed the targets specific builtins to make sure that aux-targets
are checked, it seems I didn't consider cases where the builtins check
the target info for further info. This patch bubbles the target-info
down to the individual checker functions to ensure that they validate
against the aux-target as well.
For non-aux-target invocations, this is an NFC.
This operator is intended for casting between
pointers to objects in different address spaces
and follows similar logic as const_cast in C++.
Tags: #clang
Differential Revision: https://reviews.llvm.org/D60193
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
constraint expressions.
We create overloaded `&&` and `||` operators to hold the possible
unqualified lookup results (if any) when the operands are dependent. We
could avoid building these in some cases (we will never use the stored
lookup results, and it would be better to not store them or perform the
lookups), but in the general case we will probably still need to handle
overloaded operators even with that optimization.
recommit c77a4078e0 with fix
https://reviews.llvm.org/D77954 caused regressions due to diagnostics in implicit
host device functions.
For now, it seems the most feasible workaround is to treat implicit host device function and explicit host
device function differently. Basically in device compilation for implicit host device functions, keep the
old behavior, i.e. give host device candidates and wrong-sided candidates equal preference. For explicit
host device functions, favor host device candidates against wrong-sided candidates.
The rationale is that explicit host device functions are blessed by the user to be valid host device functions,
that is, they should not cause diagnostics in both host and device compilation. If diagnostics occur, user is
able to fix them. However, there is no guarantee that implicit host device function can be compiled in
device compilation, therefore we need to preserve its overloading resolution in device compilation.
Differential Revision: https://reviews.llvm.org/D79526
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
RecoveryExprs are modeled as dependent type to prevent bogus diagnostics
and crashes in clang.
This patch allows to preseve the type for broken calls when the
RecoveryEprs have a known type, e.g. a broken non-overloaded call, a
overloaded call when the all candidates have the same return type, so
that more features (code completion still work on "take2args(x).^") still
work.
However, adding the type is risky, which may result in more clang code being
affected leading to new crashes and hurt diagnostic, and it requires large
effort to minimize the affect (update all sites in clang to handle errorDepend
case), so we add a new flag (off by default) to allow us to develop/test
them incrementally.
This patch also has some trivial fixes to suppress diagnostics (to prevent regressions).
Tested:
all existing tests are passed (when both "-frecovery-ast", "-frecovery-ast-type" flags are flipped on);
Reviewed By: sammccall
Subscribers: rsmith, arphaman, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D79160
I discovered that when using an aux-target builtin, it was recognized as
a builtin but never checked. This patch checks for an aux-target builtin
and instead validates it against the correct target.
It does this by extracting the checking code for Target-specific
builtins into its own function, then calls with either targetInfo or
AuxTargetInfo.
test cases
Add support for #pragma float_control
Reviewers: rjmccall, erichkeane, sepavloff
Differential Revision: https://reviews.llvm.org/D72841
This reverts commit 85dc033cac, and makes
corrections to the test cases that failed on buildbots.
Expose llvm fence instruction as clang builtin for AMDGPU target
__builtin_amdgcn_fence(unsigned int memoryOrdering, const char *syncScope)
The first argument of this builtin is one of the memory-ordering specifiers
__ATOMIC_ACQUIRE, __ATOMIC_RELEASE, __ATOMIC_ACQ_REL, or __ATOMIC_SEQ_CST
following C++11 memory model semantics. This is mapped to corresponding
LLVM atomic memory ordering for the fence instruction using LLVM atomic C
ABI. The second argument is an AMDGPU-specific synchronization scope
defined as string.
Reviewed By: sameerds
Differential Revision: https://reviews.llvm.org/D75917
This is a code clean up of the PropertyAttributeKind and
ObjCPropertyAttributeKind enums in ObjCPropertyDecl and ObjCDeclSpec that are
exactly identical. This non-functional change consolidates these enums
into one. The changes are to many files across clang (and comments in LLVM) so
that everything refers to the new consolidated enum in DeclObjCCommon.h.
2nd Landing Attempt...
Differential Revision: https://reviews.llvm.org/D77233
This is a code clean up of the PropertyAttributeKind and
ObjCPropertyAttributeKind enums in ObjCPropertyDecl and ObjCDeclSpec that are
exactly identical. This non-functional change consolidates these enums
into one. The changes are to many files across clang (and comments in LLVM) so
that everything refers to the new consolidated enum in DeclObjCCommon.h.
Differential Revision: https://reviews.llvm.org/D77233
This fixes a common mistake (the 3 should be @3): NSNumber *n = 3. This extends
an existing check for NSString. Also, this only errs if the initializer isn't a
null pointer constant, so NSNumber *n = 0; continues to work. rdar://47029572
Differential revision: https://reviews.llvm.org/D78066
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 adds a mechanism to easily add range checks for a builtin's
immediate operands. This patch is tested with the qdech intrinsic, which takes
both an enum for the predicate pattern, as well as an immediate for the
multiplier.
Reviewers: efriedma, SjoerdMeijer, rovka
Reviewed By: efriedma, SjoerdMeijer
Subscribers: mgorny, tschuett, mgrang, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D76678
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
Now compiler defines 5 sets of constants to represent rounding mode.
These are:
1. `llvm::APFloatBase::roundingMode`. It specifies all 5 rounding modes
defined by IEEE-754 and is used in `APFloat` implementation.
2. `clang::LangOptions::FPRoundingModeKind`. It specifies 4 of 5 IEEE-754
rounding modes and a special value for dynamic rounding mode. It is used
in clang frontend.
3. `llvm::fp::RoundingMode`. Defines the same values as
`clang::LangOptions::FPRoundingModeKind` but in different order. It is
used to specify rounding mode in in IR and functions that operate IR.
4. Rounding mode representation used by `FLT_ROUNDS` (C11, 5.2.4.2.2p7).
Besides constants for rounding mode it also uses a special value to
indicate error. It is convenient to use in intrinsic functions, as it
represents platform-independent representation for rounding mode. In this
role it is used in some pending patches.
5. Values like `FE_DOWNWARD` and other, which specify rounding mode in
library calls `fesetround` and `fegetround`. Often they represent bits
of some control register, so they are target-dependent. The same names
(not values) and a special name `FE_DYNAMIC` are used in
`#pragma STDC FENV_ROUND`.
The first 4 sets of constants are target independent and could have the
same numerical representation. It would simplify conversion between the
representations. Also now `clang::LangOptions::FPRoundingModeKind` and
`llvm::fp::RoundingMode` do not contain the value for IEEE-754 rounding
direction `roundTiesToAway`, although it is supported natively on
some targets.
This change defines all the rounding mode type via one `llvm::RoundingMode`,
which also contains rounding mode for IEEE rounding direction `roundTiesToAway`.
Differential Revision: https://reviews.llvm.org/D77379
Move function emitDeferredDiags from Sema to DeferredDiagsEmitter since it
is only used by DeferredDiagsEmitter.
Also skip visited functions to avoid exponential compile time.
Differential Revision: https://reviews.llvm.org/D77028
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.
See rational here: https://reviews.llvm.org/D76173#1922916
Time to compile Attr.h in isolation goes from 2.6s to 1.8s.
Original patch by Johannes, plus some additions from Reid to fix some
clang tooling targets.
Effect on transitive includes is marginal, though:
$ diff -u <(sort thedeps-before.txt) <(sort thedeps-after.txt) \
| grep '^[-+] ' | sort | uniq -c | sort -nr
104 - /usr/local/google/home/rnk/llvm-project/clang/include/clang/AST/OpenMPClause.h
87 - /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/Frontend/OpenMP/OMPContext.h
19 - /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/ADT/SmallSet.h
19 - /usr/local/google/home/rnk/llvm-project/llvm/include/llvm/ADT/SetVector.h
14 - /usr/include/c++/9/set
...
Differential Revision: https://reviews.llvm.org/D76184
Summary:
The basic idea is to walk through the concept definition, looking for
t.foo() where t has the constrained type.
In this patch:
- nested types are recognized and offered after ::
- variable/function members are recognized and offered after the correct
dot/arrow/colon trigger
- member functions are recognized (anything directly called). parameter
types are presumed to be the argument types. parameters are unnamed.
- result types are available when a requirement has a type constraint.
These are printed as constraints, except same_as<T> which prints as T.
Not in this patch:
- support for merging/overloading when two locations describe the same member.
The last one wins, for any given name. This is probably important...
- support for nested template members (T::x<int>)
- support for completing members of (instantiations of) template template parameters
Reviewers: nridge, saar.raz
Subscribers: mgrang, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D73649
Summary:
Added basic representation and parsing/sema handling of array-shaping
operations. Array shaping expression is an expression of form ([s0]..[sn])base,
where s0, ..., sn must be a positive integer, base - a pointer. This
expression is a kind of cast operation that converts pointer expression
into an array-like kind of expression.
Reviewers: rjmccall, rsmith, jdoerfert
Subscribers: guansong, arphaman, cfe-commits, caomhin, kkwli0
Tags: #clang
Differential Revision: https://reviews.llvm.org/D74144
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.
This is the second part loosely extracted from D71179 and cleaned up.
This patch provides semantic analysis support for `omp begin/end declare
variant`, mostly as defined in OpenMP technical report 8 (TR8) [0].
The sema handling makes code generation obsolete as we generate "the
right" calls that can just be handled as usual. This handling also
applies to the existing, albeit problematic, `omp declare variant
support`. As a consequence a lot of unneeded code generation and
complexity is removed.
A major purpose of this patch is to provide proper `math.h`/`cmath`
support for OpenMP target offloading. See PR42061, PR42798, PR42799. The
current code was developed with this feature in mind, see [1].
The logic is as follows:
If we have seen a `#pragma omp begin declare variant match(<SELECTOR>)`
but not the corresponding `end declare variant`, and we find a function
definition we will:
1) Create a function declaration for the definition we were about to generate.
2) Create a function definition but with a mangled name (according to
`<SELECTOR>`).
3) Annotate the declaration with the `OMPDeclareVariantAttr`, the same
one used already for `omp declare variant`, using and the mangled
function definition as specialization for the context defined by
`<SELECTOR>`.
When a call is created we inspect it. If the target has an
`OMPDeclareVariantAttr` attribute we try to specialize the call. To this
end, all variants are checked, the best applicable one is picked and a
new call to the specialization is created. The new call is used instead
of the original one to the base function. To keep the AST printing and
tooling possible we utilize the PseudoObjectExpr. The original call is
the syntactic expression, the specialized call is the semantic
expression.
[0] https://www.openmp.org/wp-content/uploads/openmp-TR8.pdf
[1] https://reviews.llvm.org/D61399#change-496lQkg0mhRN
Reviewers: kiranchandramohan, ABataev, RaviNarayanaswamy, gtbercea, grokos, sdmitriev, JonChesterfield, hfinkel, fghanim, aaron.ballman
Subscribers: bollu, guansong, openmp-commits, cfe-commits
Tags: #clang
Differential Revision: https://reviews.llvm.org/D75779
In order to support non-user-named kernels, SYCL needs some way in the
integration headers to name the kernel object themselves. Initially, the
design considered just RTTI naming of the lambdas, this results in a
quite unstable situation in light of some device/host macros.
Additionally, this ends up needing to use RTTI, which is a burden on the
implementation and typically unsupported.
Instead, we've introduced a builtin, __builtin_unique_stable_name, which
takes a type or expression, and results in a constexpr constant
character array that uniquely represents the type (or type of the
expression) being passed to it.
The implementation accomplishes that simply by using a slightly modified
version of the Itanium Mangling. The one exception is when mangling
lambdas, instead of appending the index of the lambda in the function,
it appends the macro-expansion back-trace of the lambda itself in the
form LINE->COL[~LINE->COL...].
Differential Revision: https://reviews.llvm.org/D76620
Normally clang avoids creating expressions when it encounters semantic
errors, even if the parser knows which expression to produce.
This works well for the compiler. However, this is not ideal for
source-level tools that have to deal with broken code, e.g. clangd is
not able to provide navigation features even for names that compiler
knows how to resolve.
The new RecoveryExpr aims to capture the minimal set of information
useful for the tools that need to deal with incorrect code:
source range of the expression being dropped,
subexpressions of the expression.
We aim to make constructing RecoveryExprs as simple as possible to
ensure writing code to avoid dropping expressions is easy.
Producing RecoveryExprs can result in new code paths being taken in the
frontend. In particular, clang can produce some new diagnostics now and
we aim to suppress bogus ones based on Expr::containsErrors.
We deliberately produce RecoveryExprs only in the parser for now to
minimize the code affected by this patch. Producing RecoveryExprs in
Sema potentially allows to preserve more information (e.g. type of an
expression), but also results in more code being affected. E.g.
SFINAE checks will have to take presence of RecoveryExprs into account.
Initial implementation only works in C++ mode, as it relies on compiler
postponing diagnostics on dependent expressions. C and ObjC often do not
do this, so they require more work to make sure we do not produce too
many bogus diagnostics on the new expressions.
See documentation of RecoveryExpr for more details.
original patch from Ilya
This change is based on https://reviews.llvm.org/D61722
Reviewers: sammccall, rsmith
Reviewed By: sammccall, rsmith
Tags: #clang
Differential Revision: https://reviews.llvm.org/D69330
This makes it possible for plugin attributes to actually do something, and also
removes a lot of boilerplate for simple attributes in SemaDeclAttr.cpp.
Differential Revision: https://reviews.llvm.org/D31342
accept as an extension.
This attempts to accept the same cases a GCC, plus cases where a
comparison is rewritten to an operator== with an integral but non-bool
return type; this is sufficient to avoid most problems with various
major open-source projects (such as ICU) and appears to fix all but one
of the comparison-related C++20 build breaks in LLVM.
This approach is being pursued for standardization.
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).
Melanie Blower