robust. It now uses an InstVisitor and worklist to actually walk the
uses of the Alloca transitively and detect the pattern which we can
directly promote: loads & stores of the whole alloca and instructions we
can completely ignore.

Also, with this new implementation teach both the predicate for testing
whether we can promote and the promotion engine itself to use the same
code so we no longer have strange divergence between the two code paths.

I've added some silly test cases to demonstrate that we can handle
slightly more degenerate code patterns now. See the below for why this
is even interesting.

Performance impact: roughly 1% regression in the performance of SROA or
ScalarRepl on a large C++-ish test case where most of the allocas are
basically ready for promotion. The reason is because of silly redundant
work that I've left FIXMEs for and which I'll address in the next
commit. I wanted to separate this commit as it changes the behavior.
Once the redundant work in removing the dead uses of the alloca is
fixed, this code appears to be faster than the old version. =]

So why is this useful? Because the previous requirement for promotion
required a *specific* visit pattern of the uses of the alloca to verify:
we *had* to look for no more than 1 intervening use. The end goal is to
have SROA automatically detect when an alloca is already promotable and
directly hand it to the mem2reg machinery rather than trying to
partition and rewrite it. This is a 25% or more performance improvement
for SROA, and a significant chunk of the delta between it and
ScalarRepl. To get there, we need to make mem2reg actually capable of
promoting allocas which *look* promotable to SROA without have SROA do
tons of work to massage the code into just the right form.

This is actually the tip of the iceberg. There are tremendous potential
savings we can realize here by de-duplicating work between mem2reg and
SROA.

llvm-svn: 187191

llvm-svn: 187190

llvm-svn: 187189

llvm-svn: 187188

llvm-svn: 187187

The bitcode representation attribute kinds are encoded into / decoded from
should be independent of the current set of LLVM attributes and their position
in the AttrKind enum. This patch explicitly encodes attributes to fixed bitcode
values.

With this patch applied, LLVM does not silently misread attributes written by
LLVM 3.3. We also enhance the decoding slightly such that an error message is
printed if an unknown AttrKind encoding was dected.

Bonus: Dropping bitcode attributes from AttrKind is now easy, as old AttrKinds
       do not need to be kept to support the Bitcode reader.
llvm-svn: 187186

easier to retrieve a register value.

llvm-svn: 187184

Attempt 2.  Sorry about the noise.

llvm-svn: 187183

llvm-svn: 187182

This patch provides basic support for powerpc64le as an LLVM target.
However, use of this target will not actually generate little-endian
code.  Instead, use of the target will cause the correct little-endian
built-in defines to be generated, so that code that tests for
__LITTLE_ENDIAN__, for example, will be correctly parsed for
syntax-only testing.  Code generation will otherwise be the same as
powerpc64 (big-endian), for now.

The patch leaves open the possibility of creating a little-endian
PowerPC64 back end, but there is no immediate intent to create such a
thing.

The new test case variant ensures that correct built-in defines for
little-endian code are generated.

llvm-svn: 187180

This patch provides basic support for powerpc64le as an LLVM target.
However, use of this target will not actually generate little-endian
code.  Instead, use of the target will cause the correct little-endian
built-in defines to be generated, so that code that tests for
__LITTLE_ENDIAN__, for example, will be correctly parsed for
syntax-only testing.  Code generation will otherwise be the same as
powerpc64 (big-endian), for now.

The patch leaves open the possibility of creating a little-endian
PowerPC64 back end, but there is no immediate intent to create such a
thing.

The LLVM portions of this patch simply add ppc64le coverage everywhere
that ppc64 coverage currently exists.  There is nothing of any import
worth testing until such time as little-endian code generation is
implemented.  In the corresponding Clang patch, there is a new test
case variant to ensure that correct built-in defines for little-endian
code are generated.

llvm-svn: 187179

llvm-svn: 187178

llvm-svn: 187177

Based on patch by Yunzhong Gao.

llvm-svn: 187176

Just more craziness to find crtbegin.o etc.

Patch by Stephen Kelly.

llvm-svn: 187175

This allows the ObjFW runtime to correctly implement message forwarding
for messages which return a struct.

Patch by Jonathan Schleifer.

llvm-svn: 187174

Refine the fix in r187094 to only distrust the StackID comparision when we are starting from an address with no symbols.
If we don't do that "nexti" will stop too soon when stepping past a tail call jump.

rdar://problem/14516227

llvm-svn: 187173

It turns out that Plum Hall depends on us not emitting an error on
integer literals which fit into long long, but fit into
unsigned long long.  So C99 conformance requires not conforming to C99. :)

llvm-svn: 187172

Also, fix the wording to reflect this.

llvm-svn: 187171

llvm-svn: 187170

- First, the watchpoint size was being cast to the
  wrong type.  This is primarily cosmetic, but
  annoying.

- Second, the options for the watchpoint command
  were not being initialized correctly, which led
  to the watchpoint size sometimes having absurdly
  large values.  This caused watchpoints to fail to
  be set in some cases.

<rdar://problem/12658775>

llvm-svn: 187169

They seemed to have the same implications, and this makes for one
less flag to worry about.

Differential Revision: http://llvm-reviews.chandlerc.com/D1219

llvm-svn: 187168

sufficient to only consider names visible at the point of instantiation,
because that may not include names that were visible when the template was
defined. More generally, if the instantiation backtrace goes through a module
M, then every declaration visible within M should be available to the
instantiation. Any of those declarations might be part of the interface that M
intended to export to a template that it instantiates.

The fix here has two parts:

1) If we find a non-visible declaration during name lookup during template
instantiation, check whether the declaration was visible from the defining
module of all entities on the active template instantiation stack. The defining
module is not the owning module in all cases: we look at the module in which a
template was defined, not the module in which it was first instantiated.

2) Perform pending instantiations at the end of a module, not at the end of the
translation unit. This is general goodness, since it significantly cuts down
the amount of redundant work that is performed in every TU importing a module,
and also implicitly adds the module containing the point of instantiation to
the set of modules checked for declarations in a lookup within a template
instantiation.

There's a known issue here with template instantiations performed while
building a module, if additional imports are added later on. I'll fix that
in a subsequent commit.

llvm-svn: 187167

llvm-svn: 187166

llvm-svn: 187165

llvm-svn: 187164

llvm-svn: 187163

llvm-svn: 187162

llvm-svn: 187161

Previously, we tried to avoid creating new temporary object regions if
the value to be materialized itself came from a temporary object region.
However, once we became more strict about lvalues vs. rvalues (months
ago), this optimization became dead code, because the input to this
function will always be an rvalue (i.e. a symbolic value or compound
value rather than a region, at least for structs).

This would be a nice optimization to keep, but removing it makes it
simpler to reason about temporary regions.

llvm-svn: 187160

Using a different loop induction variable than the enclosing scope.  No functional changes intended.

llvm-svn: 187159

structure not just a pointer. This implements that and thus fixes va_copy
on PPC32. Fixes #15286. Both bug and patch by Florian Zeitz!

llvm-svn: 187158

llvm-svn: 187157

Use the same filtering for assembly arguments to -cc1as as we do for
-cc1, this allows a consistent (& more useful) diagnostic experience for
users (rather than getting an error from -cc1as (which a user shouldn't
really be thinking about) about --foo, they get an error from clang
about --foo in -Wa,)

I'm sort of surprised by the separation of -cc1as & the separate
argument handling, etc, but at least this removes a little bit of the
duplication.

llvm-svn: 187156

llvm-svn: 187155

Back in r140220 we removed the autoconf code that would set LLVMCC_OPTION
since it was only used by the test-suite. This patch now removes code
that would only be used if LLVMCC_OPTION was set.

llvm-svn: 187154

Recent failures on a freebsd buildbot indicated a weakness in the
Reformatting.cpp lit test. Tweaking the test to avoid false negatives and
hopefully make the buildbot happy.

llvm-svn: 187153

llvm-svn: 187152

llvm-svn: 187151

Make sure the context field of DIType is MDNode.
Fix testing cases to make them pass the verifier.

llvm-svn: 187150