Erik Pilkington [Thu, 10 Jan 2019 18:03:07 +0000 (18:03 +0000)]
Split -Wdelete-non-virtual-dtor into two groups
This group controls two diagnostics: deleting an abstract class with
a non-virtual dtor, which is a guaranteed crash, and deleting a
non-abstract polymorphic class with a non-virtual dtor, which is just
suspicious.
Rafael Stahl [Thu, 10 Jan 2019 17:44:04 +0000 (17:44 +0000)]
[analyzer][CrossTU][NFC] Generalize to external definitions instead of external functions
Summary: This is just changing naming and documentation to be general about external definitions that can be imported for cross translation unit analysis. There is at least a plan to add VarDecls: D46421
Richard Trieu [Thu, 10 Jan 2019 03:23:25 +0000 (03:23 +0000)]
Refactor declarations of ASTContext allocate functions into its own header.
Forward declarations of the allocate functions combine with the forward
declaration of the ASTContext class is enough information for some headers
without pulling in ASTContext.h in its entirety. Pull the existing
declarations from AttrIterator.h into a new header. Also place the default
alignment size into this header. Previously, new had its default in
AttrIterator.h while new[] had its default in ASTContext.h. Add new header
includes where it is needed. Specifically to ASTVector.h to make it a
standalone header, unlike previously which it was standalone as long as
none of its functions were called.
[Sema] Mark target of __attribute__((alias("target"))) used for C
Summary:
Prevents -Wunneeded-internal-delcaration warnings when the target has no
other references. This occurs frequently in device drivers in the Linux
kernel.
Sema would need to invoke the demangler on the target, since in C++ the
target name is mangled:
int f() { return 42; }
int g() __attribute__((alias("_Z1fv")));
Sema does not have the ability to demangle names at this time.
Alex Lorenz [Wed, 9 Jan 2019 22:31:37 +0000 (22:31 +0000)]
[ObjC] Allow the use of implemented unavailable methods from within
the @implementation context
In Objective-C, it's common for some frameworks to mark some methods like init
as unavailable in the @interface to prohibit their usage. However, these
frameworks then often implemented said method and refer to it in another method
that acts as a factory for that object. The recent change to how messages to
self are type checked in clang (r349841) introduced a regression which started
to prohibit this pattern with an X is unavailable error. This commit addresses
the aforementioned regression.
[OpenMP] Avoid remainder operations for loop index values on a collapsed loop nest.
Summary: Change the strategy for computing loop index variables after collapsing a loop nest via the collapse clause by replacing the expensive remainder operation with multiplications and additions.
[OpenMP] Add flag for preventing the extension to 64 bits for the collapse loop counter
Summary: Introduce a compiler flag for cases when the user knows that the collapsed loop counter can be safely represented using at most 32 bits. This will prevent the emission of expensive mathematical operations (such as the div operation) on the iteration variable using 64 bits where 32 bit operations are sufficient.
Shoaib Meenai [Wed, 9 Jan 2019 20:05:16 +0000 (20:05 +0000)]
[CodeGen] Clarify comment about COFF common symbol alignment
After a discussion on the commit thread, it seems the 32 byte alignment
limitation is an MSVC toolchain artifact, not an inherent COFF
restriction. Clarify the comment accordingly, since saying COFF in the
comment but using isKnownWindowsMSVCEnvironment in the conditional is
confusing. Also add a newline before the comment, which is consistent
with the local style.
Bruno Ricci [Wed, 9 Jan 2019 16:41:33 +0000 (16:41 +0000)]
[AST] Move back BasePathSize to the bit-fields of CastExpr
The number of trailing CXXBaseSpecifiers in CastExpr was moved from
CastExprBitfields to a trailing object in r338489 (D50050). At this time these
bit-fields classes were only 32 bits wide. However later r345459 widened these
bit-field classes to 64 bits.
The reason for this change was that on 64 bit archs alignment requirements
caused 4 bytes of padding after the Stmt sub-object in nearly all expression
classes. Reusing this padding yielded an >10% reduction in the size used by all
statement/expressions when parsing all of Boost (on a 64 bit arch). This
increased the size of statement/expressions for 32 bits archs, but this can be
mitigated by moving more data to the bit-fields of Stmt (and moreover most
people now care about 64 bits archs as a host).
Therefore move back the number of CXXBaseSpecifiers in CastExpr to the
bit-fields of Stmt. This in effect mostly revert r338489 while keeping the
added test.
Alexey Bataev [Wed, 9 Jan 2019 15:58:05 +0000 (15:58 +0000)]
Incorrect implicit data-sharing for nested tasks
Summary:
There is a minor issue in how the implicit data-sharings for nested tasks are computed.
For the following example:
```
int x;
#pragma omp task shared(x)
#pragma omp task
x++;
```
We compute an implicit data-sharing of shared for `x` in the second task although I think that it should be firstprivate. Below you can find the part of the OpenMP spec that covers this example:
- // In a task generating construct, if no default clause is present, a variable for which the data-sharing attribute is not determined by the rules above and that in the enclosing context is determined to be shared by all implicit tasks bound to the current team is shared.//
- //In a task generating construct, if no default clause is present, a variable for which the data-sharing attribute is not determined by the rules above is firstprivate.//
Since each implicit-task has its own copy of `x`, we shouldn't apply the first rule.
Bruno Ricci [Wed, 9 Jan 2019 15:43:19 +0000 (15:43 +0000)]
[AST] Store the results in OverloadExpr in a trailing array
Use the newly available space in the bit-fields of Stmt to pack
OverloadExpr, UnresolvedLookupExpr and UnresolvedMemberExpr.
Additionally store the results in the overload set in a trailing array.
This saves 1 pointer + 8 bytes per UnresolvedLookupExpr and
UnresolvedMemberExpr.
Ilya Biryukov [Wed, 9 Jan 2019 13:08:11 +0000 (13:08 +0000)]
[Driver] Fix libcxx detection on Darwin with clang run as ./clang
Summary:
By using '..' instead of fs::parent_path.
The intention of the code was to go from 'path/to/clang/bin' to
'path/to/clang/include'. In most cases parent_path works, however it
would fail when clang is run as './clang'.
This was noticed in Chromium's bug tracker, see
https://bugs.chromium.org/p/chromium/issues/detail?id=919761
Use DeclSpec for quals in DeclaratorChunk::FunctionTypeInfo.
Rather than duplicating data fields, use DeclSpec directly to store
the qualifiers for the functions/methods. This change doesn't handle
attributes yet and has to be extended further.
Alex Lorenz [Tue, 8 Jan 2019 23:28:37 +0000 (23:28 +0000)]
[libclang] Fix the mismatched delete operator for ExprEvalResult
The '.stringVal' field in ExprEvalResult is allocated using new[],
but was freed using a regular delete. That caused memory leaks in
the test from r350666.
[libclang] evalute compound statement cursors before trying to evaluate
the cursor like a declaration
This change fixes a bug in libclang in which it tries to evaluate a statement
cursor as a declaration cursor, because that statement still has a pointer to
the declaration parent.
Stephen Kelly [Tue, 8 Jan 2019 22:32:48 +0000 (22:32 +0000)]
Implement the TreeStructure interface through the TextNodeDumper
Summary:
This way, when the generic ASTTraverser is extracted from ASTDumper,
there can't be any problem related to ordering of class members, a
concern that was raised in https://reviews.llvm.org/D55337.
This will also preserve the property that the generic traverser does not
enforce any coupling between the NodeDumper and the TreeStructure.
Louis Dionne [Tue, 8 Jan 2019 20:26:56 +0000 (20:26 +0000)]
[Sema] Teach Clang that aligned allocation is not supported with macosx10.13
Summary:
r306722 added diagnostics when aligned allocation is used with deployment
targets that do not support it, but the first macosx supporting aligned
allocation was incorrectly set to 10.13. In reality, the dylib shipped
with macosx10.13 does not support aligned allocation, but the dylib
shipped with macosx10.14 does.
Erich Keane [Tue, 8 Jan 2019 18:44:22 +0000 (18:44 +0000)]
Limit COFF 'common' emission to <=32 alignment types.
As reported in PR33035, LLVM crashes if given a common object with an
alignment of greater than 32 bits. This is because the COFF file format
does not support these alignments, so emitting them is broken anyway.
This patch changes any global definitions greater than 32 bit alignment
to no longer be in 'common'.
Summary:
`buildASTFromCodeWithArgs()` was creating a memory buffer referencing a
stack-allocated string. This diff changes the implementation to copy the code
string into the memory buffer so that said buffer owns the memory.
Bruno Ricci [Tue, 8 Jan 2019 14:17:00 +0000 (14:17 +0000)]
[AST] Pack CXXDependentScopeMemberExpr
Use the newly available space in the bit-fields of Stmt. Additionally store
FirstQualifierFoundInScope as a trailing object since it is most of the time
null (non-null for 2 of the 35446 CXXDependentScopeMemberExpr when parsing
all of Boost).
It would be possible to move the data for the nested-name-specifier to a
trailing object too to save another 2 pointers, however doing so did actually
regress the time taken to parse all of Boost slightly.
This saves 8 bytes + 1 pointer per CXXDependentScopeMemberExpr in the vast
majority of cases.
Bruno Ricci [Tue, 8 Jan 2019 13:52:54 +0000 (13:52 +0000)]
[Sema] Diagnose array access preceding the array bounds even when the base type is incomplete.
When the type of the base expression after IgnoreParenCasts is incomplete,
it is still possible to diagnose an array access which precedes the array
bounds.
This is a follow-up on D55862 which added an early return when the type of
the base expression after IgnoreParenCasts was incomplete.
Sam McCall [Tue, 8 Jan 2019 07:29:46 +0000 (07:29 +0000)]
[ASTMatchers] Improve assert message for broken parent map.
Summary:
This assert catches places where the AST (as seen by RecursiveASTVisitor)
becomes disconnected due to incomplete traversal.
Making it print the actual parent-less node is a lot more helpful - it's
possible to work out which part of the tree wasn't traversed.
Erik Pilkington [Tue, 8 Jan 2019 00:21:05 +0000 (00:21 +0000)]
Split -Wdelete-non-virtual-dtor into -Wdelete-abstract-non-virtual-dtor
-Wdelete-non-virtual-dtor previously controlled two diagnostics: 1)
calling a non-virtual dtor from an abstract class, and 2) calling a
non-virtual dtor from a polymorphic class. 1) is a lot more severe
than 2), since 1) is a guaranteed crash, but 2) is just "code smell".
Previously, projects compiled with -Wall -Wno-delete-non-virtual-dtor,
which is somewhat reasonable, silently crashed on 1).
Stephen Kelly [Mon, 7 Jan 2019 21:57:30 +0000 (21:57 +0000)]
NFC: Replace asserts with if() in SourceLocation accessors
Summary:
Nowhere else in the AST classes assert on these kinds of accessors.
This way, we can call the accessors and check the validity of the result
instead of externally duplicating the conditions. This generality will
make it possible to introspect instances for source locations:
Alexey Bataev [Mon, 7 Jan 2019 21:30:43 +0000 (21:30 +0000)]
[OPENMP]Add call to __kmpc_push_target_tripcount() function.
Each we create the target regions with the teams distribute inner
region, we can better estimate number of the teams required to execute
the target region. Function __kmpc_push_target_tripcount() is used for
purpose, which accepts device_id and the number of the iterations,
performed by the associated loop.
Rafael Stahl [Mon, 7 Jan 2019 15:07:01 +0000 (15:07 +0000)]
[analyzer] Pass the correct loc Expr from VisitIncDecOp to evalStore
Summary: The LocationE parameter of evalStore is documented as "The location expression that is stored to". When storing from an increment / decrement operator this was not satisfied. In user code this causes an inconsistency between the SVal and Stmt parameters of checkLocation.
Bruno Ricci [Mon, 7 Jan 2019 15:04:45 +0000 (15:04 +0000)]
[AST] Store some data of CXXNewExpr as trailing objects
Store the optional array size expression, optional initialization expression
and optional placement new arguments in a trailing array. Additionally store
the range for the parenthesized type-id in a trailing object if needed since
in the vast majority of cases the type is not parenthesized (not a single new
expression in the translation unit of SemaDecl.cpp has a parenthesized type-id).
This saves 2 pointers per CXXNewExpr in all cases, and 2 pointers + 8 bytes
per CXXNewExpr in the common case where the type is not parenthesized.
Bruno Ricci [Mon, 7 Jan 2019 14:27:04 +0000 (14:27 +0000)]
[AST][NFC] Pack DependentScopeDeclRefExpr and CXXUnresolvedConstructExpr
Use the newly available space in the bit-fields of Stmt.
This saves 1 pointer per DependentScopeDeclRefExpr/CXXUnresolvedConstructExpr.
Additionally rename "TypeSourceInfo *Type;" to "TypeSourceInfo *TSI;"
as was done in D56022 (r350003) (but this is an internal detail anyway),
and clang-format both classes. NFC.
Gabor Marton [Mon, 7 Jan 2019 14:05:19 +0000 (14:05 +0000)]
[CTU] Make loadExternalAST return with non nullptr on success
Summary:
In loadExternalAST we return with either an error or with a valid
ASTUnit pointer which should not be a nullptr.
This prevents in the call site any superfluous check for being a nullptr.
Richard Smith [Mon, 7 Jan 2019 06:00:46 +0000 (06:00 +0000)]
DR674, PR38883, PR40238: Qualified friend lookup should look for a
template specialization if there is no matching non-template function.
This exposed a couple of related bugs:
- we would sometimes substitute into a friend template instead of a
suitable non-friend declaration; this would now crash because we'd
decide the specialization of the friend is a redeclaration of itself
- ADL failed to properly handle the case where an invisible local
extern declaration redeclares an invisible friend
Both are fixed herein: in particular, we now never make invisible
friends or local extern declarations visible to name lookup unless
they are the only declaration of the entity. (We already mostly did
this for local extern declarations.)
Brian Gesiak [Mon, 7 Jan 2019 03:25:59 +0000 (03:25 +0000)]
[SemaCXX] Fix ICE for unexpanded parameter pack
Summary:
The documentation for RecursiveASTVisitor::TraverseDecl states that the
Decl being traversed may be null. In fact, this is the case when a
CXXCatchStmt with no exception decl is traversed. Because the visitor
for diagnosing unexpanded parameter packs does not check for null, it
ends up crashing when it attempts to call the Decl::isParameterPack
method on a null Decl pointer.
Add a null check to prevent an ICE, and a test case that would crash
otherwise. Also, because the test requires C++ exceptions and C++14,
change the test parameters for the entire test file. (Alternatively, I
thought about adding a new test file, but went with this approach for my
own convenience.)
Co-authored-by: Andreas Molzer <andreas.molzer@gmx.de> Co-authored-by: Mara Bos <m-ou.se@m-ou.se>
Reviewers: rsmith
The autolinking extension for ELF uses a slightly different format for
encoding the autolink information compared to COFF and MachO. Account
for this in the CGM to ensure that we do not assert when emitting
assembly or an object file.
Serge Guelton [Sat, 5 Jan 2019 12:07:36 +0000 (12:07 +0000)]
[python] Make the collections import future-proof
On Python 3.7 the old code raises a warning:
DeprecationWarning: Using or importing the ABCs from 'collections' instead of from 'collections.abc' is deprecated, and in 3.8 it will stop working
class ArgumentsIterator(collections.Sequence):
Petr Hosek [Sat, 5 Jan 2019 07:57:46 +0000 (07:57 +0000)]
[CMake][Fuchsia] Enable x86 relaxation by default
This enables x86 relaxation by default. This depends on a linker new
enough to support the new reloc types but since we default to lld we
don't worry about host system linkers that might be too old to support
the new reloc types.
Joel E. Denny [Fri, 4 Jan 2019 22:12:13 +0000 (22:12 +0000)]
[OpenMP] Refactor const restriction for linear
As discussed in D56113, this patch refactors the implementation of the
const restriction for linear to reuse a function introduced by D56113.
A side effect is that, if a variable has mutable members, this
diagnostic is now skipped, and the diagnostic for the variable not
being an integer or pointer is reported instead.
Joel E. Denny [Fri, 4 Jan 2019 22:11:56 +0000 (22:11 +0000)]
[OpenMP] Refactor const restriction for reductions
As discussed in D56113, this patch refactors the implementation of the
const restriction for reductions to reuse a function introduced by
D56113. A side effect is that diagnostics sometimes now say
"variable" instead of "list item" when a list item is a variable.
Joel E. Denny [Fri, 4 Jan 2019 22:11:31 +0000 (22:11 +0000)]
[OpenMP] Replace predetermined shared for const variable
The following appears in OpenMP 3.1 sec. 2.9.1.1 as a predetermined
data-sharing attribute:
> Variables with const-qualified type having no mutable member are
> shared.
It does not appear in OpenmP 4.0, 4.5, or 5.0. This patch removes the
implementation of that attribute when the requested OpenMP version is
greater than 3.1.
One effect of that removal is that `default(none)` affects const
variables without mutable members.
Also, without this patch, if a const variable without mutable members
was explicitly lastprivate or private, it was an error because it was
predetermined shared. Now, clang instead complains that it's const
without mutable fields, which is a more intelligible diagnostic. That
should be fine for all of the above versions because they all have
something like the following, which is quoted from OpenMP 5.0
sec. 2.19.3:
> A variable that is privatized must not have a const-qualified type
> unless it is of class type with a mutable member. This restriction does
> not apply to the firstprivate clause.
reduction and linear clauses already have separate checks for const
variables. Future patches will merge the implementations.
hwasan: Implement lazy thread initialization for the interceptor ABI.
The problem is similar to D55986 but for threads: a process with the
interceptor hwasan library loaded might have some threads started by
instrumented libraries and some by uninstrumented libraries, and we
need to be able to run instrumented code on the latter.
The solution is to perform per-thread initialization lazily. If a
function needs to access shadow memory or add itself to the per-thread
ring buffer its prologue checks to see whether the value in the
sanitizer TLS slot is null, and if so it calls __hwasan_thread_enter
and reloads from the TLS slot. The runtime does the same thing if it
needs to access this data structure.
This change means that the code generator needs to know whether we
are targeting the interceptor runtime, since we don't want to pay
the cost of lazy initialization when targeting a platform with native
hwasan support. A flag -fsanitize-hwaddress-abi={interceptor,platform}
has been introduced for selecting the runtime ABI to target. The
default ABI is set to interceptor since it's assumed that it will
be more common that users will be compiling application code than
platform code.
Because we can no longer assume that the TLS slot is initialized,
the pthread_create interceptor is no longer necessary, so it has
been removed.
Ideally, lazy initialization should only cost one instruction in the
hot path, but at present the call may cause us to spill arguments
to the stack, which means more instructions in the hot path (or
theoretically in the cold path if the spills are moved with shrink
wrapping). With an appropriately chosen calling convention for
the per-thread initialization function (TODO) the hot path should
always need just one instruction and the cold path should need two
instructions with no spilling required.
Aaron Enye Shi [Fri, 4 Jan 2019 19:09:20 +0000 (19:09 +0000)]
[HIP][DRIVER][OFFLOAD] Do not unbundle unsupported file types
The offload bundler action should not unbundle the input file types that does not match the action type. This fixes an issue where .so files are unbundled when the action type is object files.
Aaron Enye Shi [Fri, 4 Jan 2019 19:05:41 +0000 (19:05 +0000)]
[HIP][DRIVER][OFFLOAD] Do not unbundle unsupported file types
The offload bundler action should not unbundle the input file types that does not match the action type. This fixes an issue where .so files are unbundled when the action type is object files.
Erik Pilkington [Fri, 4 Jan 2019 18:33:06 +0000 (18:33 +0000)]
[ObjCARC] Add an new attribute, objc_externally_retained
This attribute, called "objc_externally_retained", exposes clang's
notion of pseudo-__strong variables in ARC. Pseudo-strong variables
"borrow" their initializer, meaning that they don't retain/release
it, instead assuming that someone else is keeping their value alive.
If a function is annotated with this attribute, implicitly strong
parameters of that function aren't implicitly retained/released in
the function body, and are implicitly const. This is useful to expose
for performance reasons, most functions don't need the extra safety
of the retain/release, so programmers can opt out as needed.
This attribute can also apply to declarations of local variables,
with similar effect.
Aaron Ballman [Fri, 4 Jan 2019 17:20:00 +0000 (17:20 +0000)]
Add two new pragmas for controlling software pipelining optimizations.
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.
Aaron Ballman [Fri, 4 Jan 2019 16:58:14 +0000 (16:58 +0000)]
Refactor the way we handle diagnosing unused expression results.
Rather than sprinkle calls to DiagnoseUnusedExprResult() around in places where we want diagnostics, we now diagnose unused expression statements and full expressions in a more generic way when acting on the final expression statement. This results in more appropriate diagnostics for [[nodiscard]] where we were previously lacking them, such as when the body of a for loop is not a compound statement.
Erich Keane [Fri, 4 Jan 2019 15:24:06 +0000 (15:24 +0000)]
Prevent unreachable when checking invalid multiversion decls.
CPUSpecifc/CPUDispatch call resolution assumed that all declarations
that would be passed are valid, however this was an invalid assumption.
This patch deals with those situations by making the valid version take
priority. Note that the checked ordering is arbitrary, since both are
replaced by calls to the resolver later.
[Basic] Extend DiagnosticEngine to store and format Qualifiers.
Qualifiers can now be streamed into the DiagnosticEngine using
regular << operator. If Qualifiers are empty 'unqualified' will
be printed in the diagnostic otherwise regular qual syntax is
used.
Daniel Dunbar [Thu, 3 Jan 2019 23:24:50 +0000 (23:24 +0000)]
Adopt SwiftABIInfo for WebAssembly.
Summary:
- This adopts SwiftABIInfo as the base class for WebAssemblyABIInfo, which is in keeping with what is done for other targets for which Swift is supported.
- This is a minimal patch to unblock exploration of WASM support for Swift (https://bugs.swift.org/browse/SR-9307)
projects / clang / log