Richard Smith [Tue, 27 Jan 2015 03:30:40 +0000 (03:30 +0000)]
DR1902: if overload resolution recurses, and the inner overload resolution
selects a deleted function, the outer function is still a candidate even though
the initialization sequence is "otherwise ill-formed".
Pete Cooper [Mon, 26 Jan 2015 20:51:58 +0000 (20:51 +0000)]
Don't generate llvm.expect intrinsics with -O0.
The backend won't run LowerExpect on -O0. In a debug LTO build, this results in llvm.expect intrinsics being in the LTO IR which doesn't know how to optimize them.
Reid Kleckner [Mon, 26 Jan 2015 19:53:16 +0000 (19:53 +0000)]
Fix broken doc link to Clang 3.5 release notes
We appear to use 3.5.0 in the directory structure now. That's probably
unnecessary. We should probably let the micro releases update the docs
for the same minor version.
Ben Langmuir [Mon, 26 Jan 2015 19:04:10 +0000 (19:04 +0000)]
Fix assert instantiating string init of static variable
... when the variable's type is a typedef of a ConstantArrayType. Just
look through the typedef (and any other sugar). We only use the
constant array type here to get the element count.
Nico Weber [Mon, 26 Jan 2015 06:23:36 +0000 (06:23 +0000)]
Don't let virtual calls and dynamic casts call Sema::MarkVTableUsed().
clang currently calls MarkVTableUsed() for classes that get their virtual
methods called or that participate in a dynamic_cast. This is unnecessary,
since CodeGen only emits vtables when it generates constructor, destructor, and
vtt code. (*)
Note that Sema::MarkVTableUsed() doesn't cause the emission of a vtable.
Its main user-visible effect is that it instantiates virtual member functions
of template classes, to make sure that if codegen decides to write a vtable
all the entries in the vtable are defined.
While this shouldn't change the behavior of codegen (other than being faster),
it does make clang more permissive: virtual methods of templates (in particular
destructors) end up being instantiated less often. In particular, classes that
have members that are smart pointers to incomplete types will now get their
implicit virtual destructor instantiated less frequently. For example, this
used to not compile but does now compile:
template <typename T> struct OwnPtr {
~OwnPtr() { static_assert((sizeof(T) > 0), "TypeMustBeComplete"); }
};
class ScriptLoader;
struct Base { virtual ~Base(); };
struct Sub : public Base {
virtual void someFun() const {}
OwnPtr<ScriptLoader> m_loader;
};
void f(Sub *s) { s->someFun(); }
The more permissive behavior matches both gcc (where this is not often
observable, since in practice most things with virtual methods have a key
function, and Sema::DefineUsedVTables() skips vtables for classes with key
functions) and cl (which is my motivation for this change) – this fixes
PR20337. See this issue and the review thread for some discussions about
optimizations.
This is similar to r213109 in spirit. r225761 was a prerequisite for this
change.
Various tests relied on "a->f()" marking a's vtable as used (in the sema
sense), switch these to just construct a on the stack. This forces
instantiation of the implicit constructor, which will mark the vtable as used.
(*) The exception is -fapple-kext mode: In this mode, qualified calls to
virtual functions (`a->Base::f()`) still go through the vtable, and since the
vtable pointer off this doesn't point to Base's vtable, this needs to reference
Base's vtable directly. To keep this working, keep referencing the vtable for
virtual calls in apple kext mode.
Renato Golin [Sun, 25 Jan 2015 23:17:48 +0000 (23:17 +0000)]
Allows Clang to use LLVM's fixes-x18 option
This patch allows clang to have llvm reserve the x18
platform register on AArch64. FreeBSD will use this in the kernel for
per-cpu data but has no need to reserve this register in userland so
will need this flag to reserve it.
This uses llvm r226664 to allow this register to be reserved.
David Blaikie [Sun, 25 Jan 2015 07:44:05 +0000 (07:44 +0000)]
Fix the buildbots (fallout from r227028) by returning the exprloc of non-binary operator overload calls (and ->) to the start of the expression, not the location of the operator
I'll give this some further thought/discussion later, but this is enough
to unbreak the buildbots at least.
Nico Weber [Sun, 25 Jan 2015 02:04:36 +0000 (02:04 +0000)]
Rename four test files from .C to .cpp.
lit.cfg has never supported running .C files, so these tests were never
executed by check-clang. Rename them to .cpp so that they run as part of
the test suite, and minorly tweak two of them that look like they were broken
when checked in to actually pass.
David Blaikie [Sun, 25 Jan 2015 01:19:10 +0000 (01:19 +0000)]
DebugInfo: Use the preferred location rather than the start location for expression line info
This causes things like assignment to refer to the '=' rather than the
LHS when attributing the store instruction, for example.
There were essentially 3 options for this:
* The beginning of an expression (this was the behavior prior to this
commit). This meant that stepping through subexpressions would bounce
around from subexpressions back to the start of the outer expression,
etc. (eg: x + y + z would go x, y, x, z, x (the repeated 'x's would be
where the actual addition occurred)).
* The end of an expression. This seems to be what GCC does /mostly/, and
certainly this for function calls. This has the advantage that
progress is always 'forwards' (never jumping backwards - except for
independent subexpressions if they're evaluated in interesting orders,
etc). "x + y + z" would go "x y z" with the additions occurring at y
and z after the respective loads.
The problem with this is that the user would still have to think
fairly hard about precedence to realize which subexpression is being
evaluated or which operator overload is being called in, say, an asan
backtrace.
* The preferred location or 'exprloc'. In this case you get sort of what
you'd expect, though it's a bit confusing in its own way due to going
'backwards'. In this case the locations would be: "x y + z +" in
lovely postfix arithmetic order. But this does mean that if the op+
were an operator overload, say, and in a backtrace, the backtrace will
point to the exact '+' that's being called, not to the end of one of
its operands.
(actually the operator overload case doesn't work yet for other reasons,
but that's being fixed - but this at least gets scalar/complex
assignments and other plain operators right)
Larisse Voufo [Sat, 24 Jan 2015 23:09:54 +0000 (23:09 +0000)]
First steps in implementing DR1467: List-initialization of aggregate from same-type object.
Only the first two items for now, changing Sections 8.5.4 [dcl.init.list] paragraph 3 and 13.3.1.7 [over.match.list] paragraph 1,
so that defining class objects and character arrays using uniform initialization syntax is actually treated as list initialization
and before it is treated aggregate initialization.
Richard Trieu [Sat, 24 Jan 2015 02:48:32 +0000 (02:48 +0000)]
When checking the template argument list, use a copy of that list for changes
and only update the orginal list on a valid arugment list. When checking an
individual expression template argument, and conversions are required, update
the expression in the template argument. Since template arguments are
speculatively checked, the copying of the template argument list prevents
updating the template arguments when the list does not match the template.
Additionally, clean up the integer checking code in the template diffing code.
The code performs unneccessary conversions from APSInt to APInt.
Fixes PR21758.
This essentially reverts r224770 to recommits r224667 and r224668 with extra
changes to prevent the template instantiation problems seen in PR22006.
A test to catch the discovered problem is also added.
Richard Smith [Sat, 24 Jan 2015 01:07:20 +0000 (01:07 +0000)]
[modules] Sometimes we can deserialize a class member but not have yet
encountered any definition for the class; this happens when the definition is
added by an update record that is not yet loaded. In such a case, eagerly pick
the original parent of the member as the canonical definition of the class
rather than muddling through with the canonical declaration (the latter can
lead to us failing to merge properly later if the canonical definition turns
out to be some other declaration).
Daniel Jasper [Fri, 23 Jan 2015 19:37:25 +0000 (19:37 +0000)]
clang-format: Fix another crasher caused by incomplete macro code.
We did't properly mark all of an AnnotatedLine's children as finalized
and thus would reformat the same tokens in different branches of #if/#else
sequences leading to invalid replacements.
Sanjay Patel [Fri, 23 Jan 2015 16:40:50 +0000 (16:40 +0000)]
Process the -fno-signed-zeros optimization flag (PR20870)
The driver currently accepts but ignores the -fno-signed-zeros flag.
This patch passes the flag through and enables 'nsz' fast-math-flag
generation in IR.
The existing OpenCL flag for the same functionality is made into an
alias here. It may be removed in a subsequent patch.
This should resolve bug 20870 ( http://llvm.org/bugs/show_bug.cgi?id=20870 );
patches for the optimizer were checked in at:
http://llvm.org/viewvc/llvm-project?view=revision&revision=225050
http://llvm.org/viewvc/llvm-project?view=revision&revision=224583
Rafael Espindola [Fri, 23 Jan 2015 05:26:38 +0000 (05:26 +0000)]
[pr22293] Don't crash during codegen of a recursive destructor.
In ItaniumCXXABI::EmitCXXDestructors we first emit the base destructor
and then try to emit the complete one as an alias.
If in the base ends up calling the complete destructor, the GD for the
complete will be in the list of deferred decl by the time we replace
it with an alias and delete the original GV.
Richard Smith [Fri, 23 Jan 2015 00:01:13 +0000 (00:01 +0000)]
PR22299: Relocate code for handling -fmodule-map-file= so that we don't try to
produce diagnostics with source locations before the diagnostics system is
ready for them.
Richard Smith [Thu, 22 Jan 2015 23:07:47 +0000 (23:07 +0000)]
Reorganize test/Modules:
* Put all input files under Inputs/, move corresponding tests into test/Modules.
* Rename a modulemap test file to [...].modulemap, and teach lit that such files are tests.
Alexey Bataev [Thu, 22 Jan 2015 06:17:56 +0000 (06:17 +0000)]
[OPENMP] CodeGen for "omp atomic read [seq_cst]" directive.
"omp atomic read [seq_cst]" accepts expressions "v=x;". In this patch we perform
an atomic load of "x" (using builtin atomic loading instructions or a call to
"atomic_load()" for simple lvalues and "kmpc_atomic_start();load
<x>;kmpc_atomic_end();" for other lvalues), convert the result of loading to
type of "v" (using EmitScalarConversion() for simple types and
EmitComplexToScalarConversion() for conversions from complex to scalar) and then
store the result in "v".)
Differential Revision: http://reviews.llvm.org/D6431
Alexey Bataev [Thu, 22 Jan 2015 05:44:37 +0000 (05:44 +0000)]
[OPENMP] CodeGen for "omp atomic read [seq_cst]" directive.
"omp atomic read [seq_cst]" accepts expressions "v=x;". In this patch we perform
an atomic load of "x" (using builtin atomic loading instructions or a call to
"atomic_load()" for simple lvalues and "kmpc_atomic_start();load
<x>;kmpc_atomic_end();" for other lvalues), convert the result of loading to
type of "v" (using EmitScalarConversion() for simple types and
EmitComplexToScalarConversion() for conversions from complex to scalar) and then
store the result in "v".)
Differential Revision: http://reviews.llvm.org/D6431
Alexey Bataev [Thu, 22 Jan 2015 05:29:28 +0000 (05:29 +0000)]
[OPENMP] CodeGen for "omp atomic read [seq_cst]" directive.
"omp atomic read [seq_cst]" accepts expressions "v=x;". In this patch we perform
an atomic load of "x" (using builtin atomic loading instructions or a call to
"atomic_load()" for simple lvalues and "kmpc_atomic_start();load
<x>;kmpc_atomic_end();" for other lvalues), convert the result of loading to
type of "v" (using EmitScalarConversion() for simple types and
EmitComplexToScalarConversion() for conversions from complex to scalar) and then
store the result in "v".
Differential Revision: http://reviews.llvm.org/D6431
Richard Smith [Thu, 22 Jan 2015 03:50:31 +0000 (03:50 +0000)]
[modules] If we add an implicit special member to a class through an update
record, and that class declaration is not the canonical definition of the
class, be sure to add the class to the list of classes that are consulted when
we look up a special member in the canonical definition.
Richard Smith [Thu, 22 Jan 2015 02:21:23 +0000 (02:21 +0000)]
[modules] If we load two declarations with typedef names for linkage purposes
on top of a local declaration of the same entity, we still need to remember
that we loaded the first one or we may fail to merge the second one properly.
Justin Bogner [Thu, 22 Jan 2015 02:17:23 +0000 (02:17 +0000)]
InstrProf: Avoid creating profile names for symbols in system headers
We don't emit any coverage mapping for uncovered functions that come
from system headers, but we were creating a GlobalVariable with each
of their names. This is wasteful since the linker will need to dead
strip the unused symbols, and it can lead to issues when merging
coverage with others TUs that do have coverage for those functions.
Richard Smith [Thu, 22 Jan 2015 01:41:56 +0000 (01:41 +0000)]
[modules] It's possible to merge into the pattern of a class template before we
load the definition data from the declaration itself. In that case, merge
properly; don't assume the prior definition is the same as our own.
Reid Kleckner [Thu, 22 Jan 2015 01:36:17 +0000 (01:36 +0000)]
Initial support for Win64 SEH IR emission
The lowering looks a lot like normal EH lowering, with the exception
that the exceptions are caught by executing filter expression code
instead of matching typeinfo globals. The filter expressions are
outlined into functions which are used in landingpad clauses where
typeinfo would normally go.
Major aspects that still need work:
- Non-call exceptions in __try bodies won't work yet. The plan is to
outline the __try block in the frontend to keep things simple.
- Filter expressions cannot use local variables until capturing is
implemented.
- __finally blocks will not run after exceptions. Fixing this requires
work in the LLVM SEH preparation pass.
The IR lowering looks like this:
// C code:
bool safe_div(int n, int d, int *r) {
__try {
*r = normal_div(n, d);
} __except(_exception_code() == EXCEPTION_INT_DIVIDE_BY_ZERO) {
return false;
}
return true;
}
Hans Wennborg [Thu, 22 Jan 2015 00:45:20 +0000 (00:45 +0000)]
Fix compiler_builtins.m test to not rely on including system stdlib.h and malloc.h
Importing _Builtin_intrinsics.sse and avx would transitively pull in those
headers, and the test would fail when building in an environment where
they were not available on the include path.
Chris Bieneman [Wed, 21 Jan 2015 23:26:11 +0000 (23:26 +0000)]
Adopt new cl::HideUnrelatedOptions API added r226729.
Summary: cl::HideUnrelatedOptions allows tools to hide all options not part of a specific OptionCategory. This is the common use case for cl::getRegisteredOptions, which should be deprecated in the future because it exposes implementation details of command line parsing.
David Blaikie [Wed, 21 Jan 2015 23:08:17 +0000 (23:08 +0000)]
DebugInfo: Remove forced column-info workaround for inlined calls
This workaround was to provide unique call sites to ensure LLVM's inline
debug info handling would properly unique two calls to the same function
on the same line. Instead, this has now been fixed in LLVM (r226736) and
the workaround here can be removed.
Originally committed in r176895, but this isn't a straight revert due to
all the changes since then. I just searched for anything ForcedColumn*
related and removed them.
We could test this - but it didn't strike me as terribly valuable once
we're no longer adding this workaround everything just works as expected
& it's no longer a special case to test for.
David Blaikie [Wed, 21 Jan 2015 22:57:22 +0000 (22:57 +0000)]
DebugInfo: Remove distinct-call-inlining test case as this is being fixed in LLVM.
This test will start failing shortly once this bug is fixed in LLVM. At
that point this behavior is no longer required in Clang and will be
removed. In the interim, remove this test just to avoid the race between
the LLVM and Clang commits.
After the LLVM commit, I'll cleanup the workaround behavior in Clang.
Reid Kleckner [Wed, 21 Jan 2015 22:18:17 +0000 (22:18 +0000)]
Add the "thunk" attribute to MS ABI virtual member pointers
This attribute implies indicates that the function musttail calls
another function and returns whatever it returns. The return type of the
thunk is meaningless, as the thunk can dynamically call different
functions with different return types. So long as the callers bitcast
the thunk with the correct type, behavior is well defined.
This attribute was necessary to fix PR20944, where the indirect call
combiner noticed that the thunk returned void and replaced the results
of the indirect call instruction with undef.
The improved completion in call context now works with:
- Functions.
- Member functions.
- Constructors.
- New expressions.
- Function call expressions.
- Template variants of the previous.
There are still rough edges to be fixed:
- Provide support for optional parameters. (fix known)
- Provide support for member initializers. (fix known)
- Provide support for variadic template functions. (fix unknown)
- Others?
projects / clang / log