llvm-svn: 178662

Added support to CMake and autoconf for unit tests in clang-tools-extra. A
dummy test exists for now until more meaningful tests can be written.

llvm-svn: 178661

I discussed this with Bill Schmidt on IRC, and it was decided that this is a
safe and reasonable default.

llvm-svn: 178659

llvm-svn: 178658

llvm-svn: 178657

More direct types in PowerPC AltiVec intrinsics.

This patch follows up on work done by Bill Schmidt in r178277,
and replaces most of the remaining uses of VRRC in ISEL DAG patterns.

The resulting .inc files are identical except for comments, so
no change in code generation is expected.

llvm-svn: 178656

This combines several related changes:
a) Don't break before after the variable types in for loops with a
   single variable.
b) Better indent DeclStmts defining multiple variables.

Before:
bool aaaaaaaaaaaaaaaaaaaaaaaaa =
    aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(aaaaaaaaaaaaaaaa),
     bbbbbbbbbbbbbbbbbbbbbbbbb =
         bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb(bbbbbbbbbbbbbbbb);
for (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa
         aaaaaaaaaaa = aaaaaaaaaaaaaaaa.aaaaaaaaaaaaaaa;
     aaaaaaaaaaa != aaaaaaaaaaaaaaaaaaa; ++aaaaaaaaaaa) {
}

After:
bool aaaaaaaaaaaaaaaaaaaaaaaaa =
         aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa(aaaaaaaaaaaaaaaa),
     bbbbbbbbbbbbbbbbbbbbbbbbb =
         bbbbbbbbbbbbbbbbbbbbbbbbbbbbbbb(bbbbbbbbbbbbbbbb);
for (aaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaaa aaaaaaaaaaa =
         aaaaaaaaaaaaaaaa.aaaaaaaaaaaaaaa;
     aaaaaaaaaaa != aaaaaaaaaaaaaaaaaaa; ++aaaaaaaaaaa) {
}

llvm-svn: 178641

llvm-svn: 178640

For this we need to use a libcall.  Previously LLVM didn't implement
libcall support for frem, so I've added it in the usual
straightforward manner.  A test case from the bug report is included.

llvm-svn: 178639

Summary:
It turns out that we don't need to store CommentsBeforeNextToken in the
line state, but rather flush them before we start parsing preprocessor
directives. This fixes wrong comment indentation in code blocks in macro calls
(the test is included).

Reviewers: klimek

Reviewed By: klimek

CC: cfe-commits

Differential Revision: http://llvm-reviews.chandlerc.com/D617

llvm-svn: 178638

llvm-svn: 178637

when getting the host processor information.  It emits a .byte sequence on GNUC compilers to work around lack of xgetbv support with older assemblers, and resolves a comment typo found in the previous patch.

llvm-svn: 178636

llvm-svn: 178635

llvm-svn: 178634

It's a bit of churn in the blame log, but I think there are real benefits to
the newer system so I'm making the change in one go.

llvm-svn: 178633

llvm-svn: 178632

llvm-svn: 178631

llvm-svn: 178630

llvm-svn: 178629

ptr_refs command frequently doesn't work when run in large applicaton. This was due to the default timeout of 500ms. The timeouts have now been increased and all expression evaluations have been modified.

llvm-svn: 178628

llvm-svn: 178627

No functional changes.

llvm-svn: 178626

llvm-svn: 178625

llvm-svn: 178624

llvm-svn: 178623

[modules] If a submodule has re-definitions of the same macro, only the last definition will be used as the "exported" one.

Fixes rdar://13562262

llvm-svn: 178622

The same compare instruction is used for 32-bit and 64-bit compares. It
sets two different sets of flags: icc and xcc.

This patch adds a conditional branch instruction using the xcc flags for
64-bit compares.

llvm-svn: 178621

call DNBProcessGetCPUType() to get the cputype of the process we're debugging.

llvm-svn: 178620

number in RNBRemote::HandlePacket_qProcessInfo -- add a new
GetCurrentThreadMachPort() so callers who need to make a mach
thred_get_state() call at the RNBRemote level will have a way to
get the port number.

llvm-svn: 178619

These refer to the reciprocal estimate support recently committed.

llvm-svn: 178618

When unsafe FP math operations are enabled, we can use the fre[s] and
frsqrte[s] instructions, which generate reciprocal (sqrt) estimates, together
with some Newton iteration, in order to quickly generate floating-point
division and sqrt results. All of these instructions are separately optional,
and so each has its own feature flag (except for the Altivec instructions,
which are covered under the existing Altivec flag). Doing this is not only
faster than using the IEEE-compliant fdiv/fsqrt instructions, but allows these
computations to be pipelined with other computations in order to hide their
overall latency.

I've also added a couple of missing fnmsub patterns which turned out to be
missing (but are necessary for good code generation of the Newton iterations).
Altivec needs a similar fix, but that will probably be more complicated because
fneg is expanded for Altivec's v4f32.

llvm-svn: 178617

Thanks to Argyrios for the pointer.

llvm-svn: 178616

This finally fixes the encoding. The patch also
* Removes eh-frame.ll. It was an unnecessary .ll to .o test that was checking
  the wrong value.
* Merge fde-reloc.s and eh-frame.s into a single test, since the only difference
  was the run lines.
* Don't blindly test the content of the entire .eh_frame section. It makes it
  hard to anyone actually fixing a bug and hitting a difference in a binary
  blob. Instead, use a CHECK for each field and document what is being checked.

llvm-svn: 178615

Rolling back the AVX support patch due to breaking a gcc 4.6 build bot that doesn't understand the xgetbv instruction for some reason.  Will revisit when time permits.

llvm-svn: 178614

llvm-svn: 178613

Remove an optimization where we were changing an objc_autorelease into an objc_autoreleaseReturnValue.

The semantics of ARC implies that a pointer passed into an objc_autorelease
must live until some point (potentially down the stack) where an
autorelease pool is popped. On the other hand, an
objc_autoreleaseReturnValue just signifies that the object must live
until the end of the given function at least.

Thus objc_autorelease is stronger than objc_autoreleaseReturnValue in
terms of the semantics of ARC* implying that performing the given
strength reduction without any knowledge of how this relates to
the autorelease pool pop that is further up the stack violates the
semantics of ARC.

*Even though objc_autoreleaseReturnValue if you know that no RV
optimization will occur is more computationally expensive.

llvm-svn: 178612

expressions are integer.  It can also be ValueDecl expressions

Use the type information from the TemplateParameterList instead

Patch by Olivier Goffart!

llvm-svn: 178611

When support was added for declaration arguments, the case of variadic
declaration arguments was not supported.  This patch fixes that problem by
not crashing when certain ValueDecl's are null.

Patch by Olivier Goffart!

llvm-svn: 178610

Value depenedent expressions for default arguments cannot be evaluated.
Instead, use the desugared template type to get an argument expression that
can be used.  This is needed for both integer and declaration arguements.

Also, move this common code into a separate function.

Patch by Olivier Goffart!

llvm-svn: 178609

Symbol table function names should support lookups like symbols with debug info. 

To fix this I:
- Gutted the way FindFunctions is used, there used to be way too much smarts only in the DWARF plug-in
- Made it more efficient by chopping the name up once and using simpler queries so that SymbolFile and Symtab plug-ins don't need to do as much
- Filter the results at a higher level
- Make the lldb_private::Symtab able to chop up C++ mangled names and make as much sense out of them as possible and also be able to search by basename, fullname, method name, and selector name.

llvm-svn: 178608