get FormatManager.cpp to build on no-python platforms again.

llvm-svn: 164284

llvm-svn: 164283

Toctree was not being interlinked properly.

llvm-svn: 164282

- Rewrite/merge pseudo-atomic instruction emitters to address the
  following issue:
  * Reduce one unnecessary load in spin-loop

    previously the spin-loop looks like

        thisMBB:
        newMBB:
          ld  t1 = [bitinstr.addr]
          op  t2 = t1, [bitinstr.val]
          not t3 = t2  (if Invert)
          mov EAX = t1
          lcs dest = [bitinstr.addr], t3  [EAX is implicit]
          bz  newMBB
          fallthrough -->nextMBB

    the 'ld' at the beginning of newMBB should be lift out of the loop
    as lcs (or CMPXCHG on x86) will load the current memory value into
    EAX. This loop is refined as:

        thisMBB:
          EAX = LOAD [MI.addr]
        mainMBB:
          t1 = OP [MI.val], EAX
          LCMPXCHG [MI.addr], t1, [EAX is implicitly used & defined]
          JNE mainMBB
        sinkMBB:

  * Remove immopc as, so far, all pseudo-atomic instructions has
    all-register form only, there is no immedidate operand.

  * Remove unnecessary attributes/modifiers in pseudo-atomic instruction
    td

  * Fix issues in PR13458

- Add comprehensive tests on atomic ops on various data types.
  NOTE: Some of them are turned off due to missing functionality.

- Revise tests due to the new spin-loop generated.

llvm-svn: 164281

LGTM by Michael Spencer

llvm-svn: 164280

llvm-svn: 164279

Try not to violate conventions immediately before explaining them.

llvm-svn: 164278

If someone provides their own function called 'strdup', or 'reallocf', or
even 'malloc', and we inlined it, the inlining should have given us all the
malloc-related information we need. If we then try to attach new information
to the return value, we could end up with spurious warnings.

<rdar://problem/12317671>

llvm-svn: 164276

While we definitely want this optimization in the future, we're not
currently handling constraints on symbolic /expressions/ correctly.
These should stay live even if the SymExpr itself is no longer referenced
because could recreate an identical SymExpr later. Only once the SymExpr
can no longer be recreated -- i.e. a component symbol is dead -- can we
safely remove the constraints on it.

This liveness issue is tracked by <rdar://problem/12333297>.

This reverts r163444 / 24c7f98828e039005cff3bd847e7ab404a6a09f8.

llvm-svn: 164275

llvm-svn: 164274

non-function friend declaration. Patch by Josh Magee!

llvm-svn: 164273

llvm-svn: 164272

llvm-svn: 164268

member function templates with an rvalue ref qualifier.

llvm-svn: 164267

Add overflow check to tanh(complex) and reduce to finite answer.  Fixes http://llvm.org/bugs/show_bug.cgi?id=13874

llvm-svn: 164266

Make the TargetPrefix setting one big setting instead of being spread out
everywhere. No functional change.

llvm-svn: 164265

llvm-svn: 164264

functions.

llvm-svn: 164263

Otherwise clang can't analyze code that relies on features provided by libc++.

llvm-svn: 164262

llvm-svn: 164261

llvm-svn: 164260

A patch that allows for mach-o architectures to be specified as "<number>-<number>" where the first number is the cpu type and the second is the cpu subtype. Also added code to allow use of mach-o architectures that aren't in our tables so that symbolication and static file introspection (crashlogs) can work with them.

llvm-svn: 164258

Don't get everything when resolving the symbol context of the ObjC Class symbol, just the module + symbol.

llvm-svn: 164257

Soften the pattern-can-never-match error in TableGen into a warning. This pattern can be very useful in cases where you want to define a multiclass that covers both commutative and non-commutative operators (say, add and sub).

llvm-svn: 164256

an out-of-date compiler in certain cases.

llvm-svn: 164255

llvm-svn: 164254

llvm-svn: 164253

llvm-svn: 164252

Implement a correct copy constructor for Record.  Now that we're using the ID number as a key in maps (for determinism), it is imperative that ID numbers be globally unique, even when we copy construct a Record.
This fixes some obscure failure cases involving registers defined inside multiclasses or foreach constructs that would not receive a unique ID, and would end up being omitted from the AsmMatcher tables.

llvm-svn: 164251

A common coalescing conflict in vector code is lane insertion:

  %dst = FOO
  %src = BAR
  %dst:ssub0 = COPY %src

The live range of %src interferes with the ssub0 lane of %dst, but that
lane is never read after %src would have clobbered it. That makes it
safe to merge the live ranges and eliminate the COPY:

  %dst = FOO
  %dst:ssub0 = BAR

This patch teaches the new coalescer to resolve conflicts where dead
vector lanes would be clobbered, at least as long as the clobbered
vector lanes don't escape the basic block.

llvm-svn: 164250

This patch adds memory support functions which will later be used to implement section-specific protection handling in MCJIT.

llvm-svn: 164249

to improve compatibility with GNU as.

Based on a patch by PaX Team.

Fixed assertion failures on non-Darwin and added additional test cases.

llvm-svn: 164248

Anthony Bryant.

llvm-svn: 164247

This patch is based on the one by PaX Team.

Patch by Andy Zhang!

llvm-svn: 164246

The directive can be matched with directives other than '.rept'

Patch by Andy Zhang!

llvm-svn: 164245

is placed on a function that has no path to the exit block.

llvm-svn: 164244

- Merge the processing of LOAD_ADD with other atomic load-arith
  operations
- Separate the logic getting target constant for atomic-load-op and add
  an optimization for atomic-load-add on i16 with negative value
- Optimize a minor case for atomic-fetch-add i16 with negative operand. Test
  case is revised.

llvm-svn: 164243

LOCKS_EXCLUDED is used on a method with a name that is is not a simple
identifier.

llvm-svn: 164242

Overloaded __pad_and_output on ostreambuf_iterator and in this overload call sputn instead of dereferencing the iterator which calls sputc.  This is intended to be purely a performance optimization, especially for clients who may have overloaded the virtual function xsputn.

llvm-svn: 164241

Patch by Joey Gouly.

llvm-svn: 164239