Represent pass_object_size attrs in ExtParameterInfo
The goal of this is to fix a bug in modules where we'd merge
FunctionDecls that differed in their pass_object_size attributes. Since
we can overload on the presence of pass_object_size attributes, this
behavior is incorrect.
We don't represent `N` in `pass_object_size(N)` as part of
ExtParameterInfo, since it's an error to overload solely on the value of
N. This means that we have a bug if we have two modules that declare
functions that differ only in their pass_object_size attrs, like so:
// In module A, from a.h
void foo(char *__attribute__((pass_object_size(0))));
// In module B, from b.h
void foo(char *__attribute__((pass_object_size(1))));
// In module C, in main.c
#include "a.h"
#include "b.h"
At the moment, we'll merge the foo decls, when we should instead emit a
diagnostic about an invalid overload. We seem to have similar (silent)
behavior if we overload only on the return type of `foo` instead; I'll
try to find a good place to put a FIXME (or I'll just file a bug) soon.
This patch also fixes a bug where we'd not output the proper extended
parameter info for declarations with pass_object_size attrs.
Douglas Yung [Fri, 24 Feb 2017 01:25:02 +0000 (01:25 +0000)]
Recently a change was made to this test in r294639 which fails when the
compiler is run in a mode where the default C++ standard is newer than C++03.
The reason is because one of the warnings checked is only produced when the
compiler is using C++03 or lower.
This change fixes this problem as well as adds explicit run lines to run the
test in C++03 and C++11 modes.
David L. Jones [Fri, 24 Feb 2017 00:28:01 +0000 (00:28 +0000)]
[Driver] Move architecture-specific free helper functions to their own files.
This patch moves helper functions that are CPU-specific out of Driver.cpp and to
separate implementation files. The new files are named for the architecture,
e.g. ARMArch.cpp.
The next step after this will be to move OS-specific code, which I expect will
include many of the tool implementations, to similarly separate files.
Some CPU-specific functions are not being moved just yet. In cases where the
only caller is the platform-specific tools, I plan to move them together. An
example is Hexagon, where the only caller of the architecture-specific functions
are the tools themselves. (I'm happy to revise this choice, it just seems like
less churn to me.)
This does mean that some functions which were previously static are now exposed
through the library header Driver.h.
Alex Lorenz [Fri, 24 Feb 2017 00:09:30 +0000 (00:09 +0000)]
NFC, Add a test that ensures that we don't emit helper code in copy/dispose
routines for objects that are captured with the __unsafe_unretained
ownership qualifier
This is a preparation commit that improves code-coverage in code that emits
block copy/dispose routines.
[CodeGen] Fix ExtParameterInfo bugs in C++ CodeGen code.
This patch makes use of the prefix/suffix ABI argument distinction that
was introduced in r295870, so that we now emit ExtParameterInfo at the
correct offset for member calls that have added ABI arguments. I don't
see a good way to test the generated param info, since we don't actually
seem to use it in CGFunctionInfo outside of Swift. Any
suggestions/thoughts for how to better test this are welcome. :)
This patch also fixes a small bug with inheriting constructors: if we
decide not to pass args into an base class ctor, we would still
generate ExtParameterInfo as though we did. The added test-case is for
that behavior.
[CodeGen] Don't reemit expressions for pass_object_size params.
This fixes an assertion failure in cases where we had expression
statements that declared variables nested inside of pass_object_size
args. Since we were emitting the same ExprStmt twice (once for the arg,
once for the @llvm.objectsize call), we were getting issues with
redefining locals.
This also means that we can be more lax about when we emit
@llvm.objectsize for pass_object_size args: since we're reusing the
arg's value itself, we don't have to care so much about side-effects.
Richard Trieu [Thu, 23 Feb 2017 03:25:57 +0000 (03:25 +0000)]
[ODRHash] Handle types in ODR hashing.
Fields will now have their types added to the hash, allowing for detection of
mismatched field types. This detection allows the existing ODR checking to
produce the correct message.
Richard Smith [Thu, 23 Feb 2017 02:09:03 +0000 (02:09 +0000)]
Fix tracking of whether the previous template instantiation stack matches the current one.
Rather than attempting to compare whether the previous and current top of
context stack are "equal" (which fails for a number of reasons, such as the
context stack entries containing pointers to objects on the stack, or reaching
the same "top of stack" entry through two different paths), track the depth of
context stack at which we last emitted a note and invalidate it when we pop the
context stack to less than that depth.
This causes us to emit some missing "in instantiation of" notes and to stop
emitting redundant "in instantiation of" stacks matching the previous stack in
rare cases.
Richard Smith [Thu, 23 Feb 2017 01:43:54 +0000 (01:43 +0000)]
Rename ActiveTemplateInstantiation to CodeSynthesisContext in preparation for
using it for other kinds of context (where we currently produce context notes
in a highly ad-hoc manner).
Richard Trieu [Wed, 22 Feb 2017 22:22:42 +0000 (22:22 +0000)]
[ODRHash] static_cast and Stmt hashing.
Add support for static_cast in classes. Add pointer-independent profiling for
Stmt's, sharing most of the logic with Stmt::Profile. This is the first of the
deep sub-Decl diffing for error messages.
Richard Smith [Wed, 22 Feb 2017 22:09:50 +0000 (22:09 +0000)]
PR32034: Evaluate _Atomic(T) in-place when T is a class or array type.
This is necessary in order for the evaluation of an _Atomic initializer for
those types to have an associated object, which an initializer for class or
array type needs.
Bob Haarman [Wed, 22 Feb 2017 20:29:39 +0000 (20:29 +0000)]
stop using associative comdats for SEH filter functions
Summary: We implement structured exception handling (SEH) by generating filter functions for functions that use exceptions. Currently, we use associative comdats to ensure that the filter functions are preserved if and only if the functions we generated them for are preserved. This can lead to problems when generating COFF objects - LLVM may decide to inline a function that uses SEH and remove its body, at which point we will end up with a comdat that COFF cannot represent. To avoid running into that situation, this change makes us not use associative comdats for SEH filter functions. We can still get the benefits we used the associative comdats for: we will always preserve filter functions we use, and dead stripping can eliminate the ones we don't use.
[CodeGen] Note where we add ABI-specific args in ctors. NFC.
Meta: The ultimate goal is to teach ExtParameterInfo about
pass_object_size attributes. This is necessary for that, since our
ExtParameterInfo is a bit buggy in C++. I plan to actually make use of
this Prefix/Suffix info in the near future, but I like small
single-purpose changes. Especially when those changes are hard to
actually test...
At the moment, some of our C++-specific CodeGen pretends that ABIs can
only add arguments to the beginning of a function call. This isn't quite
correct: args can be appended to the end, as well. It hasn't mattered
much until now, since we seem to only use this "number of arguments
added" data when calculating the ExtParameterInfo to use when making a
CGFunctionInfo. Said ExtParameterInfo is currently only used for
ParameterABIs (Swift) and ns_consumed (ObjC).
So, this patch allows ABIs to indicate whether args they added were at
the beginning or end of an argument list. We can use this information to
emit ExtParameterInfos more correctly, though like said, that bit is
coming soon.
Richard Smith [Wed, 22 Feb 2017 20:01:55 +0000 (20:01 +0000)]
Improve support for 'decltype(auto)' in template template parameter matching.
A 'decltype(auto)' parameter can match any other kind of non-type template
parameter, so should be usable in place of any other parameter in a template
template argument. The standard is sadly extremely unclear on how this is
supposed to work, but this seems like the obviously-correct result.
It's less clear whether an 'auto' parameter should be able to match
'decltype(auto)', since the former cannot be used if the latter turns out to be
used for a reference type, but if we disallow that then consistency suggests we
should also disallow 'auto' matching 'T' for the same reason, defeating
intended use cases of the feature.
Jonas Hahnfeld [Wed, 22 Feb 2017 06:49:10 +0000 (06:49 +0000)]
[OpenMP] Generate better diagnostics for cancel and cancellation point
checkNestingOfRegions uses CancelRegion to determine whether cancel and
cancellation point are valid in the given nesting. This leads to unuseful
diagnostics if CancelRegion is invalid. The given test case has produced:
region cannot be closely nested inside 'parallel' region
As a solution, introduce checkCancelRegion and call it first to get the
expected error:
one of 'for', 'parallel', 'sections' or 'taskgroup' is expected
This is because we were always selecting the version of
`@llvm.objectsize` that takes an i8* in CodeGen. Passing an i32* as an
i8* makes LLVM very unhappy.
(Yes, I'm surprised that this remained uncaught for so long, too. :) )
As an added bonus, we'll now also use the appropriate address space when
emitting @llvm.objectsize calls.
Richard Trieu [Wed, 22 Feb 2017 01:11:25 +0000 (01:11 +0000)]
Add more ODR checking.
Add the basics for the ODRHash class, which will only process Decl's from
a whitelist, which currently only has AccessSpecDecl. Different access
specifiers in merged classes can now be detected.
Richard Smith [Tue, 21 Feb 2017 23:49:18 +0000 (23:49 +0000)]
Fix deduction of type of pack-expanded non-type template parameter.
We need to look through the PackExpansionType in the parameter type when
deducing, and we need to consider the possibility of deducing arguments for
packs that are not lexically mentioned in the pattern (but are nonetheless
deducible) when figuring out which packs are covered by a pack deduction scope.
Jacob Gravelle [Tue, 21 Feb 2017 22:37:27 +0000 (22:37 +0000)]
Declare lgamma library builtins as never being const
Summary:
POSIX requires lgamma writes to an external global variable, signgam.
This prevents annotating lgamma with readnone, which is incorrect on
targets that write to signgam.
Taewook Oh [Tue, 21 Feb 2017 22:30:55 +0000 (22:30 +0000)]
Fix for pr31836 - pp_nonportable_path on absolute paths: broken delimiters
Summary: This is a patch for PR31836. As the bug replaces the path separators in the included file name with the characters following them, the test script makes sure that there's no "Ccase-insensitive-include-pr31836.h" in the warning message.
Dehao Chen [Tue, 21 Feb 2017 20:36:21 +0000 (20:36 +0000)]
Only enable AddDiscriminator pass when -fdebug-info-for-profiling is true
Summary: AddDiscriminator pass is only useful for sample pgo. This patch restricts AddDiscriminator to -fdebug-info-for-profiling so that it does not introduce unecessary debug size increases for non-sample-pgo builds.
Richard Smith [Tue, 21 Feb 2017 08:42:39 +0000 (08:42 +0000)]
Fix lookup through injected-class-names in implicit deduction guides in the
case where the class template has a parameter pack.
Checking of the template arguments expects an "as-written" template argument
list, which in particular does not have any parameter packs. So flatten the
packs into separate arguments before passing them in.
Richard Smith [Tue, 21 Feb 2017 06:30:38 +0000 (06:30 +0000)]
PR32010: Fix template argument depth mixup when forming implicit constructor
template deduction guides for class template argument deduction.
Ensure that we have a local instantiation scope for tracking the instantiated
parameters. Additionally, unusually, we're substituting at depth 1 and leaving
depth 0 alone; make sure that we don't reduce template parameter depth by 2 for
inner parameters in the process. (This is probably also broken for alias
templates in the case where they're expanded within a dependent context, but
this patch doesn't fix that.)
Using the constructed name for the class properties with dot syntax may
yield an inappropriate selector (i.e. if it is specified via property
attributes). Prefer the declaration for the selector, falling back to
the constructed name otherwise.
Daniel Jasper [Mon, 20 Feb 2017 14:51:16 +0000 (14:51 +0000)]
clang-format: [JS] Improve line-wrapping behavior of template strings.
Specifically, similar to other blocks, clang-format now wraps both
after "${" and before the corresponding "}", if the contained
expression spans multiple lines.
Serge Pavlov [Sat, 18 Feb 2017 06:04:15 +0000 (06:04 +0000)]
Process attributes 'ifunc' and 'alias' when checking for redefinition
These attributes effectively turn a non-defining declaration into a
definition, so the case when the declaration already has a body must
be diagnosed properly.
Vedant Kumar [Sat, 18 Feb 2017 02:02:55 +0000 (02:02 +0000)]
[profiling] Make a test more explicit. NFC.
The cxx-structors.cpp test checks that some instrumentation doesn't
appear, but it should be more explicit about which instrumentation it
actually expects to appear.
Vedant Kumar [Sat, 18 Feb 2017 01:50:14 +0000 (01:50 +0000)]
[profiling] Tighten test cases which refer to "profn" vars. NFC.
The frontend can't see "__profn" profile name variables after IRGen
because llvm throws these away now. Tighten up some test cases which
checked for the non-existence of those variables.
Richard Smith [Sat, 18 Feb 2017 00:32:02 +0000 (00:32 +0000)]
[modules] Load the ModuleOffsetMap from the module header lazily.
If we never need to map any ID within the module to its global ID, we don't
need the module offset map. If a compilation transitively depends on lots of
unused module files, this can result in a modest performance improvement.
Vedant Kumar [Fri, 17 Feb 2017 23:22:59 +0000 (23:22 +0000)]
Retry^2: [ubsan] Reduce null checking of C++ object pointers (PR27581)
This patch teaches ubsan to insert exactly one null check for the 'this'
pointer per method/lambda.
Previously, given a load of a member variable from an instance method
('this->x'), ubsan would insert a null check for 'this', and another
null check for '&this->x', before allowing the load to occur.
Similarly, given a call to a method from another method bound to the
same instance ('this->foo()'), ubsan would a redundant null check for
'this'. There is also a redundant null check in the case where the
object pointer is a reference ('Ref.foo()').
This patch teaches ubsan to remove the redundant null checks identified
above.
Testing: check-clang, check-ubsan, and a stage2 ubsan build.
I also compiled X86FastISel.cpp with -fsanitize=null using
patched/unpatched clangs based on r293572. Here are the number of null
checks emitted:
Vedant Kumar [Fri, 17 Feb 2017 23:22:55 +0000 (23:22 +0000)]
[ubsan] Pass a set of checks to skip to EmitTypeCheck() (NFC)
CodeGenFunction::EmitTypeCheck accepts a bool flag which controls
whether or not null checks are emitted. Make this a bit more flexible by
changing the bool to a SanitizerSet.
Needed for an upcoming change which deals with a scenario in which we
only want to emit null checks.
Carlo Bertolli [Fri, 17 Feb 2017 21:29:13 +0000 (21:29 +0000)]
[OpenMP] Prepare Sema for initial implementation for pragma 'distribute parallel for'
https://reviews.llvm.org/D29922
This patch adds two fields for use in the implementation of 'distribute parallel for':
The increment expression for the distribute loop. As the chunk assigned to a team is executed by multiple threads within the 'parallel for' region, the increment expression has to correspond to the value returned by the related runtime call (for_static_init).
The upper bound of the innermost loop ('for' in 'distribute parallel for') is not the globalUB expression normally used for pragma 'for' when found in isolation. It is instead the upper bound of the chunk assigned to the team ('distribute' loop). In this way, we prevent teams from executing chunks assigned to other teams.
The use of these two fields can be see in a related explanatory patch:
https://reviews.llvm.org/D29508
Jonas Hahnfeld [Fri, 17 Feb 2017 18:32:51 +0000 (18:32 +0000)]
[OpenMP] Remove barriers at cancel and cancellation point
This resolves a deadlock with the cancel directive when there is no explicit
cancellation point. In that case, the implicit barrier acts as cancellation
point. After removing the barrier after cancel, the now unmatched barrier for
the explicit cancellation point has to go as well.
This has probably worked before rL255992: With the calls for the explicit
barrier, it was sure that all threads passed a barrier before exiting.
was turned into:
ALWAYS_INLINE::std::string getName() ...
If it turns out that clang-format is failing to clean up a lot of the
existing spaces now, we can add more analyses of the identifier. It
should not currently. Cases where clang-format breaks nested name
specifiers should be fine as clang-format wraps after the "::". Thus, a
line getting longer and then shorter again should lead to the same
original code.
projects / clang / log