Alexander Musman [Tue, 19 Aug 2014 11:27:13 +0000 (11:27 +0000)]
[OPENMP] Extract common superclass from all the loop directives. No functional changes (having common superclass is convenient for future loop directives CodeGen implementation)
David Majnemer [Tue, 19 Aug 2014 07:29:03 +0000 (07:29 +0000)]
MS ABI: Update alias template mangling for VC "14" CTP 3
MSVC "14" CTP 3 has fixed it's mangling for alias templates when used as
template-template arguments; update clang to be compatible with this
mangling.
Alexey Samsonov [Mon, 18 Aug 2014 22:10:42 +0000 (22:10 +0000)]
Update link strategy for sanitizer runtime libraries on Linux:
1. Always put static sanitizer runtimes to the front of the linker
invocation line. This was already done for all sanitizers except UBSan:
in case user provides static libstdc++ we need to make sure that new/delete
operator definitions are picked from sanitizer runtimes instead of libstdc++.
We have to put UBSan runtime first for similar reasons: it depends on some
libstdc++ parts (e.g. __dynamic_cast function), and has to go first in
link line to ensure these functions will be picked up from libstdc++.
2. Put sanitizer libraries system dependencies (-ldl, -lpthread etc.) right
after sanitizer runtimes. This will ensure these libraries participate in
the link even if user provided -Wl,-as-needed flag. This should fix PR15823.
3. In case we link in several sanitizer runtimes (e.g. "ubsan", "ubsan_cxx"
and "san"), add system dependencies (-ldl, -lpthread, ...) only once.
Ben Langmuir [Mon, 18 Aug 2014 19:32:45 +0000 (19:32 +0000)]
Fix the rececl chain for redeclarations of predefined decls
Predefined decls like 'Protocol' in objc are not loaded from AST files,
so we cannot rely on loading the canonical decl to complete the redecl
chain for redeclarations of these decls. The broken redecl chain was
non-circular, so looping over redecls() would hang.
Ben Langmuir [Sat, 16 Aug 2014 04:54:18 +0000 (04:54 +0000)]
When loading a module with no local entities, still bump the size of the
tables that correspond to ContinuousRangeMaps, since the keys to those
maps need to be unique, or we may map to the wrong offset.
This fixes a crash + malformed AST file seen when loading some modules
that import Cocoa on Darwin, which is a module with no contents except
imports of other modules. Unfortunately I have not been able to find a
reduced test case that reproduces this problem.
Also add an assert that we aren't mapping one key to multiple values
in CRM. We ought to be able to say there are no duplicate keys at all,
but there are a bunch of 0 -> 0 mappings that are showing up, probably
coming from the source location table.
Nico Weber [Fri, 15 Aug 2014 23:21:41 +0000 (23:21 +0000)]
Make sure that vtables referenced from delay-parsed templates get referenced.
This fixes PR20671, see the bug for details. In short, ActOnTranslationUnit()
calls DefineUsedVTables() and only then PerformPendingInstantiations(). But
PerformPendingInstantiations() is what does delayed template parsing, so
vtables only references from late-parsed templates weren't marked used.
As a fix, move the SavePendingInstantiationsAndVTableUsesRAII in
PerformPendingInstantiations() up above the delayed template parsing code.
That way, vtables referenced from templates end up in the RAII object, and the
call to DefineUsedVTables() in PerformPendingInstantiations() marks them used.
Sean Silva [Fri, 15 Aug 2014 23:18:49 +0000 (23:18 +0000)]
Update for LLVM change (StringSaver)
There is more cleanup to be done here. Once
llvm::sys::Process::GetArgumentVector is switched over to StringSaver,
we can simplify this code a fair amount.
David Blaikie [Fri, 15 Aug 2014 21:11:25 +0000 (21:11 +0000)]
DebugInfo: While loop backedge should be attribute to the start of the while statement.
A little test case simplification - this could be simplified further,
though there are certainly interesting connections to the if/else
construct so I'm hesitant to remove that entirely though it does appear
somewhat unrelated.
David Blaikie [Fri, 15 Aug 2014 20:50:45 +0000 (20:50 +0000)]
DebugInfo: Fix PR19864 better - attribute the jump at the end of a range-for loop, to the start of the loop.
This avoids debuggers stepping to strange places (like the last
statement in the loop body, or the first statement in the if).
This is not the whole answer, though - similar bugs no doubt exist in
other loops (patches to follow) and attributing exception handling code
to the correct line is also tricky (based on the previous fix to
PR19864, exception handling is still erroneously attributed to the 'if'
line).
Sean Silva [Fri, 15 Aug 2014 19:23:53 +0000 (19:23 +0000)]
[cleanup] Rename this function to better reflect its purpose. NFC.
This also suggests some refactoring to simplify this code. Basically, a
ton of complexity in this argument handling code comes from the need to
save const char *'s in stable storage for pushing onto argv.
It seems like llvm::cl::StringSaver can be improved to cover all the
needs here.
Sean Silva [Fri, 15 Aug 2014 18:50:00 +0000 (18:50 +0000)]
[cleanup] Factor out some checks. NFC.
The core logic in main() is actually pretty simple, but there's lots of
stuff that has been added over time which obscures the flow of the code.
In upcoming patches, I'll be pulling more stuff out of the main
codepath.
I'm open to naming suggestions for these helper functions.
Objective-C. Do not warn if user is using property-dox syntax to name a
user provided setter name (as declared in @property attribute declaration).
rdar://18022762
Keno Fischer [Fri, 15 Aug 2014 01:39:12 +0000 (01:39 +0000)]
Make sure CallExpr::getLocStart doesn't segfault
Summary:
When the CallExpr passed to Sema::ConvertArgumentsForCall has all default parameters, and the number of actual arguments passed is zero, this function will segfault in the call to Call->getLocStart() if the Callee has an invalid getLocStart(), the reason being that since ConvertArgumentsForCall has set the correct number of arguments, but has not filled them in yet, getLocStart() will try to access the first (not yet existent) argument and thus segfaults.
This fixes that by making getLocStart return an invalid source location if the queried argument is NULL rather than segfaulting.
Reid Kleckner [Thu, 14 Aug 2014 23:34:52 +0000 (23:34 +0000)]
Limit our MSVC compat hack for nested names from dependent bases
Previously, any undeclared unqualified id starting a nested name
specifier in a dependent context would have its lookup retried during
template instantiation. Now we limit that retry hack to methods of a
class with dependent bases. Free function templates in particular are
no longer affected by this hack.
Also, diagnose this as a Microsoft extension. This has the downside that
template authors may see this warning *and* an error during
instantiation time about this identifier. Fixing that will probably
require formalizing some kind of "delayed" identifier, instead of our
ad-hoc solutions of forming dependent AST nodes when lookup fails.
Thread Safety Analysis: Move -Wthread-safety-negative out of the
-Wthread-safety umbrella flag, pending updates to documentation. The flag
works, but is likely to be confusing to existing users of -Wthread-safety.
Thread safety analysis: add -Wthread-safety-verbose flag, which adds additional notes that are helpful when compiling statistics on thread safety warnings.
Richard Smith [Thu, 14 Aug 2014 20:30:52 +0000 (20:30 +0000)]
[modules] Turn off a broken optimization: we need to pick up implicit special
members from all redefinitions of a class that have them, in case the special
member is defined in one module but only declared in another.
Artyom Skrobov [Thu, 14 Aug 2014 16:04:47 +0000 (16:04 +0000)]
Use the proper post-order traversal in LiveVariables analysis,
to recover the performance after r214064.
Also sorts out the naming for PostOrderCFGView, ReversePostOrderCFGView,
BackwardDataflowWorklist and ForwardDataflowWorklist, to match the accepted
terminology.
Also unifies BackwardDataflowWorklist and ForwardDataflowWorklist to share
the "worklist for prioritization, post-order traversal for fallback" logic,
and to avoid repetitive sorting.
Justin Bogner [Thu, 14 Aug 2014 15:44:29 +0000 (15:44 +0000)]
Revert "CodeGen: When bitfields fall on natural boundaries, split them up"
It fits better with LLVM's memory model to try to do this in the
backend. Specifically, narrowing wide loads in the backends should be
relatively straightforward and is generally valuable, whereas widening
loads tends to be very constrained.
David Majnemer [Thu, 14 Aug 2014 06:35:08 +0000 (06:35 +0000)]
Parse: Don't attempt to act on #pragma init_seg when not targeting MSVC
It doesn't really make sense to try and do stuff with #pragma init_seg
when targeting non-Microsoft platforms; notions like library vs user
initializers don't exist for other targets.
Richard Smith [Thu, 14 Aug 2014 04:11:33 +0000 (04:11 +0000)]
[modules] Factor .td-generated diagnostics stuff out into its own module; this
avoids users of AllDiagnostics.h from needing to pregenerate *all* generated
headers. Hopefully this will make my modules buildbot happier...
Richard Smith [Thu, 14 Aug 2014 03:30:27 +0000 (03:30 +0000)]
[modules] Don't assert if the same imported class template specialization
declaration has its definition instantiated in two sibling modules and they use
a partial specialization.
Justin Bogner [Thu, 14 Aug 2014 02:42:10 +0000 (02:42 +0000)]
CodeGen: When bitfields fall on natural boundaries, split them up
Currently when laying out bitfields that don't need any padding, we
represent them as a wide enough int to contain all of the bits. This
can be hard on the backend since we'll do things like represent stores
to a few bits as loading an i144, masking it with a large constant,
and storing it back.
This turns up in less pathological cases where we load and mask 64 bit
word on a 32 bit platform when we actually only need to access 32 bits.
This leads to bad code being generated in most of our 32 bit backends.
In practice, there are often natural breaks in bitfields, and it's a
fairly simple and effective heuristic to split these fields into legal
integer sized chunks when it will be equivalent (ie, it won't force us
to add any extra padding).
Richard Smith [Thu, 14 Aug 2014 02:21:01 +0000 (02:21 +0000)]
[modules] When we merge together multiple class template specialization
definitions (because some other declaration declares a special member that
isn't present in the canonical definition), we need to search *all* of them; we
can't just stop when we find the requested name in any of the definitions,
because that can fail to find things (and in particular, it can fail to find
the member of the canonical declaration and return a bogus ODR failure).
Richard Smith [Wed, 13 Aug 2014 21:15:09 +0000 (21:15 +0000)]
[modules] Fix a rejects-valid resulting from emitting an inline function
recursively within the emission of another inline function. This ultimately
led to us emitting the same inline function definition twice, which we then
rejected because we believed we had a mangled name conflict.
Objective-C. This patch is to resolve the method used in method
expression to the best method found in global method pools.
This is wip. // rdar://16808765
projects / clang / log