Teach lowering to correctly handle invoke statepoint and gc results tied to them. Note that we still can not lower gc.relocates for invoke statepoints.
Also it extracts getCopyFromRegs helper function in SelectionDAGBuilder as we need to be able to customize type of the register exported from basic block during lowering of the gc.result.
(Resubmitting this change after not being able to reproduce buildbot failure)

Differential Revision: http://reviews.llvm.org/D7760

llvm-svn: 231800

When tail merging it may be necessary to remove MMOs from memory operations to
ensures later passes (e.g., MI sched) conservatively compute dependencies.
Currently, we only remove the MMO from the common tail if the MMO doesn't match
with the relative instruction in the non-common tail(s).

A more robust solution would be to add multiple MMOs from the duplicate MIs to
the new MI. Currently ScheduleDAGInstrs.cpp ignores all MMOs on instructions
with multiple MMOs, so this solution is equivalent for the time being.

No test case included as this is incredibly difficult to reproduce.

Patch was a collaborative effort between Ana Pazos and myself.
Phabricator: http://reviews.llvm.org/D7769

llvm-svn: 231799

llvm-svn: 231798

llvm-svn: 231797

This fixes a machine verifier error in the salu-to-valu.ll, which
would have been exposed by a future commit.

llvm-svn: 231796

This was done by refactoring the v_cndmask_b32 tablegen definition
to use inherit from VOP2Inst.

llvm-svn: 231795

We want to replace as much custom x86 shuffling via intrinsics
as possible because pushing the code down the generic shuffle
optimization path allows for better codegen and less complexity
in LLVM.

This is the sibling patch for the Clang half of this change:
http://reviews.llvm.org/D8088

Differential Revision: http://reviews.llvm.org/D8086

llvm-svn: 231794

The corresponding LLVM commit is 231740.

llvm-svn: 231793

We want to replace as much custom x86 shuffling via intrinsics
as possible because pushing the code down the generic shuffle
optimization path allows for better codegen and less complexity
in LLVM.

This is the sibling patch for the LLVM half of this change:
http://reviews.llvm.org/D8086

Differential Revision: http://reviews.llvm.org/D8088

llvm-svn: 231792

llvm-svn: 231791

Switching back and forth between sections does nothing (other than producing
larger .s files).

llvm-svn: 231790

Fixes PR18985

Patch by Vassil Vassilev!

llvm-svn: 231789

This crash occurs due to memory corruption when trying to update dependency
direction based on Constraints.

This crash was observed during lnt regression of Polybench benchmark test case dynprog.

Review: http://reviews.llvm.org/D8059
llvm-svn: 231788

Using clang as a cross-compiler with the 'target' option could be confusing
for those inexperienced in the realm of cross compiling.

This patch would allow the use of all these four variants of the target option:
-target <triple>
--target <triple>
-target=<triple>
--target=<triple>

Patch by Gabor Ballabas.

llvm-svn: 231787

llvm-svn: 231786

Removing dead code to silence warning C4060: switch statement contains no 'case' or 'default' labels; NFC.

llvm-svn: 231785

This crash in Dependency analysis is because we assume here that in case of UsefulGEP
both source and destination have the same number of operands which may not be true.
This incorrect assumption results in crash while populating Pairs. Fix the same.

This crash was observed during lnt regression for code such as-
  struct s{
    int A[10][10];
    int C[10][10][10]; 
  } S;
  void dep_constraint_crash_test(int k,int N)  {
     for( int i=0;i<N;i++)
       for( int j=0;j<N;j++)
         S.A[0][0] = S.C[0][0][k];
  }
Review: http://reviews.llvm.org/D8162

llvm-svn: 231784

Summary:
This is part of the work to support memory constraints that behave
differently to 'm'. The subsequent patches will expand on the existing
encoding (which is a 32-bit int) and as a result in some flag words will no
longer fit into an i16. This problem only affected the MSP430 target which
appears to have 16-bit pointers.

Reviewers: hfinkel

Reviewed By: hfinkel

Subscribers: hfinkel, llvm-commits

Differential Revision: http://reviews.llvm.org/D8168

llvm-svn: 231783

Power is 64-bit but does not support leak detection,
so this test fails. Use the more robust leak-detection
predicate instead.

llvm-svn: 231782

llvm-svn: 231781

Before I discovered that NetBSD provides a permanent handle to the C
locale called LC_C_LOCALE, I also added support for this to CloudABI
under the name LC_POSIX_LOCALE. I've renamed it to LC_C_LOCALE to
improve compatibility.

llvm-svn: 231780

llvm-svn: 231779

llvm-svn: 231778

CloudABI provides the _l() functions that are part of POSIX.1-2008, but
also the extensions that are available on systems like OS X and *BSD
(scanf_l, printf_l, etc).

llvm-svn: 231777

llvm-svn: 231776

llvm-svn: 231775

llvm-svn: 231774

llvm-svn: 231773

built before running the test suite.

llvm-svn: 231770

the test when run frem a read-only checkout of the source tree.

llvm-svn: 231769

As CloudABI does not provide sysctl(), this header is not present. Make
thread.cpp build correctly (and pass all tests) by not including the header.

llvm-svn: 231768

There are a couple of places where libc++ prints log/error messages to
stdout on its own. This may of course interfere with the output
generated with applications. Log/error messages should be directed to
stderr instead.

Differential Revision:	http://reviews.llvm.org/D8135
Reviewed by:	marshall

llvm-svn: 231767

conventions. Sorry for missing this before I committed the patch!

llvm-svn: 231766

of the vector. For a vector 'v', '&v[v.size()]' isn't a valid way to
compute a pointer one-past-the-end of the vector. Instead, write the
loop in terms of iterators and save the beginning iterator. Once we have
that we can compute the beginning pointer from the beginning iterator,
and compute the distance which we should increment the beginning pointer
by subtracting the iterators.

What might be simpler would be to convert the function accepting a raw
pointer for begin and end to accept iterators or a range or some other
construct, but I wanted to keep this to a minimal bug-fix change.

This fixes a crash on any debug STL implementation which checks for
indexing out of bounds.

llvm-svn: 231765

Nuxi CloudABI (https://github.com/NuxiNL/cloudlibc) does not allow
processes to access the global filesystem namespace. This breaks
random_device, as it attempts to use /dev/{u,}random. This change adds
support for arc4random(), which is present on CloudABI.

In my opinion it would also make sense to use arc4random() on other
operating systems, such as *BSD and Mac OS X, but I'd rather leave that
to the maintainers of the respective platforms. Switching to
arc4random() does change the ABI.

This change also attempts to make some cleanups to the code. It adds a
single #define for every random interface, instead of testing against
operating systems explicitly.

As discussed, also validate the token argument to be equal to
"/dev/urandom" on all systems that only provide pseudo-random numbers.
This should cause little to no breakage, as "/dev/urandom" is also the
default argument value.

Reviewed by: jfb
Differential Revision: http://reviews.llvm.org/D8134

llvm-svn: 231764

Saves adding .str() call to any raw_ostream << SmallString usage
and a small step towards making .str() consistent in the ADTs by
removing one of the SmallString::str() use cases, discussion at

http://lists.cs.uiuc.edu/pipermail/llvm-commits/Week-of-Mon-20141013/240026.html

I'll update the Phabricator patch http://reviews.llvm.org/D6372
for review of the Twine SmallString support, it's more complex
than this one.

llvm-svn: 231763

The task region is emmitted in several steps:

Emit a call to kmp_task_t *__kmpc_omp_task_alloc(ident_t *, kmp_int32 gtid, kmp_int32 flags, size_t sizeof_kmp_task_t, size_t sizeof_shareds, kmp_routine_entry_t *task_entry).
Here task_entry is a pointer to the function:
kmp_int32 .omp_task_entry.(kmp_int32 gtid, kmp_task_t *tt) {
    TaskFunction(gtid, tt->part_id, tt->shareds);
      return 0;
}
Copy a list of shared variables to field shareds of the resulting structure kmp_task_t returned by the previous call (if any).
Copy a pointer to destructions function to field destructions of the resulting structure kmp_task_t.
Emit a call to kmp_int32 __kmpc_omp_task(ident_t *, kmp_int32 gtid, kmp_task_t *new_task), where new_task is a resulting structure from previous items.
Differential Revision: http://reviews.llvm.org/D7560

llvm-svn: 231762

llvm-svn: 231761

We failed to use a marking set to properly handle recursive types, which caused use
to recurse infinitely and eventually overflow the stack.

llvm-svn: 231760

Fix an issue in the verifier where we could try to read information out of a malformed statepoint intrinsic.

In this situation we would always have already flagged an error on the statepoint intrinsic,
but then we carry on to parse other, related GC intrinsics, and could end up crashing during that
verification when they try to access data from the malformed statepoint.

llvm-svn: 231759