This is a Clang-only change and depends on the existing "musttail"
support already implemented in LLVM.
The [[clang::musttail]] attribute goes on a return statement, not
a function definition. There are several constraints that the user
must follow when using [[clang::musttail]], and these constraints
are verified by Sema.
Tail calls are supported on regular function calls, calls through a
function pointer, member function calls, and even pointer to member.
Future work would be to throw a warning if a users tries to pass
a pointer or reference to a local variable through a musttail call.
Reviewed By: rsmith
Differential Revision: https://reviews.llvm.org/D99517
The previous implementation was insufficient for checking statement
attribute mutual exclusion because attributed statements do not collect
their attributes one-at-a-time in the same way that declarations do. So
the design that was attempting to check for mutual exclusion as each
attribute was processed would not ever catch a mutual exclusion in a
statement. This was missed due to insufficient test coverage, which has
now been added for the [[likely]] and [[unlikely]] attributes.
The new approach is to check all of attributes that are to be applied
to the attributed statement in a group. This required generating
another DiagnoseMutualExclusions() function into AttrParsedAttrImpl.inc.
This changes our approach to processing statement attributes to be more
similar to how we process declaration attributes. Namely,
ActOnAttributedStmt() now calls ProcessStmtAttributes() instead of
vice-versa, and there is now an interface split between building an
attributed statement where you already have a list of semantic
attributes and building an attributed statement with attributes from
the parser.
This should make it easier to support statement attributes that are
dependent on a template. In that case, you would add a
TransformFooAttr() function in TreeTransform.h to perform the semantic
checking (morally similar to how Sema::InstantiateAttrs() already works
for declaration attributes) when transforming the semantic attribute at
instantiation time.
Currently, when one or more attributes are mutually exclusive, the
developer adding the attribute has to manually emit diagnostics. In
practice, this is highly error prone, especially for declaration
attributes, because such checking is not trivial. Redeclarations
require you to write a "merge" function to diagnose mutually exclusive
attributes and most attributes get this wrong.
This patch introduces a table-generated way to specify that a group of
two or more attributes are mutually exclusive:
def : MutualExclusions<[Attr1, Attr2, Attr3]>;
This works for both statement and declaration attributes (but not type
attributes) and the checking is done either from the common attribute
diagnostic checking code or from within mergeDeclAttribute() when
merging redeclarations.
Clang currently automates a fair amount of diagnostic checking for
declaration attributes based on the declarations in Attr.td. It checks
for things like subject appertainment, number of arguments, language
options, etc. This patch uses the same machinery to perform diagnostic
checking on statement attributes.
The attribute definition claimed the attribute was inheritable (which
only applies to declaration attributes) and not a statement attribute.
Further, it treats subject appertainment errors as being parse errors
rather than semantic errors, which leads to us accepting invalid code.
For instance, we currently fail to reject:
void foo() {
int i = 1000;
__attribute__((nomerge, opencl_unroll_hint(8)))
if (i) { foo(); }
}
This addresses the issues by clarifying that opencl_unroll_hint is a
statement attribute and handles its appertainment checks in the
semantic layer instead of the parsing layer. This changes the output of
the diagnostic text to be more consistent with other appertainment
errors.
This patch adds support for two new variants of the vectorize_width
pragma:
1. vectorize_width(X[, fixed|scalable]) where an optional second
parameter is passed to the vectorize_width pragma, which indicates if
the user wishes to use fixed width or scalable vectorization. For
example the user can now write something like:
#pragma clang loop vectorize_width(4, fixed)
or
#pragma clang loop vectorize_width(4, scalable)
In the absence of a second parameter it is assumed the user wants
fixed width vectorization, in order to maintain compatibility with
existing code.
2. vectorize_width(fixed|scalable) where the width is left unspecified,
but the user hints what type of vectorization they prefer, either
fixed width or scalable.
I have implemented this by making use of the LLVM loop hint attribute:
llvm.loop.vectorize.scalable.enable
Tests were added to
clang/test/CodeGenCXX/pragma-loop.cpp
for both the 'fixed' and 'scalable' optional parameter.
See this thread for context: http://lists.llvm.org/pipermail/cfe-dev/2020-November/067262.html
Differential Revision: https://reviews.llvm.org/D89031
This is the initial part of the implementation of the C++20 likelihood
attributes. It handles the attributes in an if statement.
Differential Revision: https://reviews.llvm.org/D85091
Reapply 49e5f603d4
which had been reverted in c94332919b.
Originally reverted because I hadn't updated it in quite a while when I
got around to committing it, so there were a bunch of missing changes to
new code since I'd written the patch.
Reviewers: aaron.ballman
Differential Revision: https://reviews.llvm.org/D76646
Sometimes we also want to avoid merging inline assembly. This patch add
the nomerge function attribute to inline assembly.
Reviewed By: zequanwu
Differential Revision: https://reviews.llvm.org/D84225
There is a version that just tests (also called
isIntegerConstantExpression) & whereas this version is specifically used
when the value is of interest (a few call sites were actually refactored
to calling the test-only version) so let's make the API look more like
it.
Reviewers: aaron.ballman
Differential Revision: https://reviews.llvm.org/D76646
In order to enable future improvements to our attribute diagnostics,
this moves info from ParsedAttr into CommonAttributeInfo, then makes
this type the base of the *Attr and ParsedAttr types. Quite a bit of
refactoring took place, including removing a bunch of redundant Spelling
Index propogation.
Differential Revision: https://reviews.llvm.org/D67368
llvm-svn: 371875
This adds a new vectorize predication loop hint:
#pragma clang loop vectorize_predicate(enable)
that can be used to indicate to the vectoriser that all (load/store)
instructions should be predicated (masked). This allows, for example, folding
of the remainder loop into the main loop.
This patch will be followed up with D64916 and D65197. The former is a
refactoring in the loopvectorizer and the groundwork to make tail loop folding
a more general concept, and in the latter the actual tail loop folding
transformation will be implemented.
Differential Revision: https://reviews.llvm.org/D64744
llvm-svn: 366989
I would like to add some pragma handling here, but couldn't resist a little NFC
and tidy up first.
Differential Revision: https://reviews.llvm.org/D64471
llvm-svn: 365629
to reflect the new license.
We understand that people may be surprised that we're moving the header
entirely to discuss the new license. We checked this carefully with the
Foundation's lawyer and we believe this is the correct approach.
Essentially, all code in the project is now made available by the LLVM
project under our new license, so you will see that the license headers
include that license only. Some of our contributors have contributed
code under our old license, and accordingly, we have retained a copy of
our old license notice in the top-level files in each project and
repository.
llvm-svn: 351636
This patch adds #pragma clang loop pipeline and #pragma clang loop pipeline_initiation_interval for debugging or reducing compile time purposes. It is possible to disable SWP for concrete loops to save compilation time or to find bugs by not doing SWP to certain loops. It is possible to set value of initiation interval to concrete number to save compilation time by not doing extra pipeliner passes or to check created schedule for specific initiation interval.
Patch by Alexey Lapshin.
llvm-svn: 350414
struct LoopHint was only used within Parse and not in any of the Sema or
Codegen files. In the non-Parse files where it was included, it either wasn't
used or LoopHintAttr was used, so its inclusion did nothing.
llvm-svn: 347728
Summary:
Merged the recently added `err_attribute_argument_negative` diagnostic
with existing `err_attribute_requires_positive_integer` diagnostic:
the former allows only strictly positive integer, while the latter
also allows zero.
Reviewers: aaron.ballman
Reviewed By: aaron.ballman
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D51853
llvm-svn: 342367
As a part of attempting to clean up the way attributes are
printed, this patch adds an operator << to the diagnostics/
partialdiagnostics so that ParsedAttr can be sent directly.
This patch also rewrites a large amount* of the times when
ParsedAttr was printed using its IdentifierInfo object instead
of being printed itself.
*"a large amount" == "All I could find".
llvm-svn: 339344
This adds support for the unroll_and_jam pragma, to go with the recently
added unroll and jam pass. The name of the pragma is the same as is used
in the Intel compiler, and most of the code works the same as for unroll.
#pragma clang loop unroll_and_jam has been separated into a different
patch. This part adds #pragma unroll_and_jam with an optional count, and
#pragma no_unroll_and_jam to disable the transform.
Differential Revision: https://reviews.llvm.org/D47267
llvm-svn: 338566
Basically, "AttributeList" loses all list-like mechanisms, ParsedAttributes is
switched to use a TinyPtrVector (and a ParsedAttributesView is created to
have a non-allocating attributes list). DeclaratorChunk gets the later kind,
Declarator/DeclSpec keep ParsedAttributes.
Iterators are added to the ParsedAttribute types so that for-loops work.
llvm-svn: 336945
We noticed when implementing a new pragma that the TableGen-generated function
getAttributeSpellingListIndex() did not work for pragma attributes. It relies
on the values in the enum AttributeList::Syntax and a new value
AS_ContextSensitiveKeyword was added changing the value for AS_Pragma.
Apparently no tests failed since no pragmas currently make use of the
generated function.
To fix this we can move AS_Pragma back to the value that TableGen code expects.
Also to prevent changes in the enum from breaking that routine again I added
calls to getAttributeSpellingListIndex() in the unroll pragma code. That will
cause some lit test failures if the order is changed. I added a comment to
remind of this issue in the future.
This assumes we don’t need/want full TableGen support for
AS_ContextSensitiveKeyword. It currently only appears in getAttrKind and no
other TableGen-generated routines.
Patch by: mikerice
Differential Revision: https://reviews.llvm.org/D36473
llvm-svn: 310483
Summary:
This patch implements parsing of [[clang::suppress(rule, ...)]]
and [[gsl::suppress(rule, ...)]] attributes.
C++ Core Guidelines depend heavily on tool support for
rule enforcement. They also propose a way to suppress
warnings [1] which is by annotating any ancestor in AST
with the C++11 attribute [[gsl::suppress(rule1,...)]].
To have a mechanism to suppress non-C++ Core
Guidelines specific, an additional spelling of [[clang::suppress]]
is defined.
For example, to suppress the warning cppcoreguidelines-slicing,
one could do
```
[[clang::suppress("cppcoreguidelines-slicing")]]
void f() { ... code that does slicing ... }
```
or
```
void g() {
Derived b;
[[clang::suppress("cppcoreguidelines-slicing")]]
Base a{b};
[[clang::suppress("cppcoreguidelines-slicing")]] {
doSomething();
Base a2{b};
}
}
```
This parsing can then be used by clang-tidy, which includes multiple
C++ Core Guidelines rules, to suppress warnings (see
https://reviews.llvm.org/D24888).
For the exact naming of the rule in the attribute, there
are different possibilities, which will be defined in the
corresponding clang-tidy patch.
Currently, clang-tidy supports suppressing of warnings through "//
NOLINT" comments. There are some advantages that the attribute has:
- Suppressing specific warnings instead of all warnings
- Suppressing warnings in a block (namespace, function, compound
statement)
- Code formatting may split a statement into multiple lines,
thus a "// NOLINT" comment may be on the wrong line
I'm looking forward to your comments!
[1] https://github.com/isocpp/CppCoreGuidelines/blob/master/CppCoreGuidelines.md#inforce-enforcement
Reviewers: alexfh, aaron.ballman, rsmith
Subscribers: cfe-commits
Differential Revision: https://reviews.llvm.org/D24886
llvm-svn: 298880
Summary: Although the feature was introduced only in OpenCL C v2.0 spec., it's useful for OpenCL 1.x too and doesn't require HW support.
Reviewers: Anastasia
Subscribers: yaxunl, cfe-commits, bader
Differential Revision: https://reviews.llvm.org/D27453
llvm-svn: 289535
Summary:
This is similar to other loop pragmas like 'vectorize'. Currently it
only has state values: distribute(enable) and distribute(disable). When
one of these is specified the corresponding loop metadata is generated:
!{!"llvm.loop.distribute.enable", i1 true/false}
As a result, loop distribution will be attempted on the loop even if
Loop Distribution in not enabled globally. Analogously, with 'disable'
distribution can be turned off for an individual loop even when the pass
is otherwise enabled.
There are some slight differences compared to the existing loop pragmas.
1. There is no 'assume_safety' variant which makes its handling slightly
different from 'vectorize'/'interleave'.
2. Unlike the existing loop pragmas, it does not have a corresponding
numeric pragma like 'vectorize' -> 'vectorize_width'. So for the
consistency checks in CheckForIncompatibleAttributes we don't need to
check it against other pragmas. We just need to check for duplicates of
the same pragma.
Reviewers: rsmith, dexonsmith, aaron.ballman
Subscribers: bob.wilson, cfe-commits, hfinkel
Differential Revision: http://reviews.llvm.org/D19403
llvm-svn: 272656
It is not widely used and removed from OpenCL v2.1.
This change modifies Clang to parse the attribute for OpenCL
but ignores it afterwards.
Patch by Liu Yaxun (Sam)!
Differential Revision: http://reviews.llvm.org/D17861
llvm-svn: 265006
exactly the same as clang's existing [[clang::fallthrough]] attribute, which
has been updated to have the same semantics. The one significant difference
is that [[fallthrough]] is ill-formed if it's not used immediately before a
switch label (even when -Wimplicit-fallthrough is disabled). To support that,
we now build a CFG of any function that uses a '[[fallthrough]];' statement
to check.
In passing, fix some bugs with our support for statement attributes -- in
particular, diagnose their use on declarations, rather than asserting.
llvm-svn: 262881
Add support for opencl_unroll_hint attribute from OpenCL v2.0 s6.11.5.
Reusing most of metadata generation from CGLoopInfo helper class.
The code is based on Khronos OpenCL compiler:
https://github.com/KhronosGroup/SPIR/tree/spirv-1.0
Patch by Liu Yaxun (Sam)!
Differential Revision: http://reviews.llvm.org/D16686
llvm-svn: 261350
This change adds the new unroll metadata "llvm.loop.unroll.enable" which directs
the optimizer to unroll a loop fully if the trip count is known at compile time, and
unroll partially if the trip count is not known at compile time. This differs from
"llvm.loop.unroll.full" which explicitly does not unroll a loop if the trip count is not
known at compile time
With this change "#pragma unroll" generates "llvm.loop.unroll.enable" rather than
"llvm.loop.unroll.full" metadata. This changes the semantics of "#pragma unroll" slightly
to mean "unroll aggressively (fully or partially)" rather than "unroll fully or not at all".
The motivating example for this change was some internal code with a loop marked
with "#pragma unroll" which only sometimes had a compile-time trip count depending
on template magic. When the trip count was a compile-time constant, everything works
as expected and the loop is fully unrolled. However, when the trip count was not a
compile-time constant the "#pragma unroll" explicitly disabled unrolling of the loop(!).
Removing "#pragma unroll" caused the loop to be unrolled partially which was desirable
from a performance perspective.
llvm-svn: 244467
Specifying #pragma clang loop vectorize(assume_safety) on a loop adds the
mem.parallel_loop_access metadata to each load/store operation in the loop. This
metadata tells loop access analysis (LAA) to skip memory dependency checking.
llvm-svn: 239572
Previously loop hints such as #pragma loop vectorize_width(#) required a constant. This patch allows a constant expression to be used as well. Such as a non-type template parameter or an expression (2 * c + 1).
Reviewed by Richard Smith
llvm-svn: 219589
This patch is necessary to support constant expressions which replaces the integer value in the loop hint attribute with an expression. The integer value was also storing the pragma’s state for options like vectorize(enable/disable) and the pragma unroll and nounroll directive. The state variable is introduced to hold the state of those options/pragmas. This moves the validation of the state (keywords) from SemaStmtAttr handler to the loop hint annotation token handler.
Resubmit with changes to try to fix the build-bot issue.
Reviewed by Aaron Ballman
llvm-svn: 214432
This patch is necessary to support constant expressions which replaces the integer value in the loop hint attribute with an expression. The integer value was also storing the pragma’s state for options like vectorize(enable/disable) and the pragma unroll and nounroll directive. The state variable is introduced to hold the state of those options/pragmas. This moves the validation of the state (keywords) from SemaStmtAttr handler to the loop hint annotation token handler.
Reviewed by Aaron Ballman
llvm-svn: 214333
Dávid Bolvanský