Re-enable branchfolding common code hoisting optimization. Fixed a liveness test bug and also taught it to update liveins.

llvm-svn: 131241

llvm-svn: 131240

Use DW_AT_APPLE_objc_class_extension attribute to identify interfaces that represent class extension.
Radar 9423077.

llvm-svn: 131239

Let Objective-C front-end identify class extension, in dwarf output, using an attribute DW_AT_APPLE_objc_class_extension.

llvm-svn: 131238

     (__m128){ p[0], p[1], p[2], p[3] }

which produces really bad code. This could be done in instcombine, but it's
probably better to do it in the front-end instead.
<rdar://problem/9424836>

llvm-svn: 131237

Construction of the RegisterReaderArg instance should happen after we have a valid base_addr, not before.

llvm-svn: 131236

llvm-svn: 131235

llvm-svn: 131234

llvm-svn: 131233

llvm-svn: 131232

llvm-svn: 131231

llvm-svn: 131230

inconsistent with autoconf, which by default set BINUTILS_INCDIR to
empty and exclude gold from target list.

Based on a patch by Haitao Li!

llvm-svn: 131229

They are actually grammatically considered definitions and parsed
accordingly.

This fixes the outstanding bugs regarding defaulting functions after
their declarations.

We now really nicely diagnose the following construct (try it!)

int foo() = delete, bar;

Still todo: Defaulted functions other than default constructors
            Test cases (including for the above construct)

llvm-svn: 131228

defaulted default constructors.

As it happens, making sure that we handle out-of-line defaulted
functions properly will involved making sure that we actually parse them
correctly, so that's coming after.

llvm-svn: 131224

We still don't parse out-of-line defaults correctly, which is needed to
get the full effect out of this patch.

llvm-svn: 131223

llvm-svn: 131222

pointers:

        virtual bool
        PrepareTrivialCall (Thread &thread, 
                            lldb::addr_t sp,
                            lldb::addr_t functionAddress,
                            lldb::addr_t returnAddress, 
                            lldb::addr_t *arg1_ptr,
                            lldb::addr_t *arg2_ptr,
                            lldb::addr_t *arg3_ptr) const = 0;

Prior to this it was:

        virtual bool
        PrepareTrivialCall (Thread &thread, 
                            lldb::addr_t sp,
                            lldb::addr_t functionAddress,
                            lldb::addr_t returnAddress, 
                            lldb::addr_t arg,
                            lldb::addr_t *this_arg,
                            lldb::addr_t *cmd_arg) const = 0;

This was because the function that called this slowly added more features to
be able to call a C++ member function that might have a "this" pointer, and 
then later added "self + cmd" support for objective C. Cleaning this code up
and the code that calls it makes it easier to implement the functions for
new targets.

The MacOSX_arm::PrepareTrivialCall() is now filled in and ready for testing.

llvm-svn: 131221

llvm-svn: 131220

llvm-svn: 131219

llvm-svn: 131218

llvm-svn: 131217

llvm-svn: 131215

markers. In some cases a register def is dead on one path, but not on
another.

This is passing Clang self-hosting.

llvm-svn: 131214

    o get_parent_frame(frame)
    o get_args_as_string(frame)

to lldbutil.py and create TestFrameUtils.py to exercise the utils.
Plus re-arrange the test/python_api/lldbutil to have three directories
for testing iteration, process stack traces, and the just added frame utils.

llvm-svn: 131213

llvm-svn: 131212

Fix PR9902: correctly substitute alias templates within the template in which they are defined: provide an empty list of arguments for each containing template context during substitution.

llvm-svn: 131211

Interleave IV simplifications. Currently involves EliminateComparison
and EliminateRemainder. Next I'll add EliminateExtend.

llvm-svn: 131210

argument deduction failures. Only implemented in C++0x, since this is
a significant change in behavior from C++98/03.

llvm-svn: 131209

llvm-svn: 131208

nested of an out-of-line declaration, only require a 'template<>'
header for each enclosing class template that hasn't been previously
specialized; previously, we were requiring 'template<>' for enclosing
class templates and members of class templates that hadn't been
previously specialized. Fixes <rdar://problem/9422013>.

llvm-svn: 131207

llvm-svn: 131206

LLVM and binutils.

With this patch, there are no functional differences between the .o
produced directly from LLVM versus the .s to .o via GNU as, for relocation tags
at least, for both PIC and non-PIC modes.

Because some non-PIC reloc tags are used (legally) on PIC, so IsPCRel flag is
necessary but not sufficient to determine whether the overall codegen mode is
PIC or not. Why is this necessary? There is an incompatibility of how relocs
are emitted in the .rodata section.  Binutils PIC likes to emit certain relocs
as section relative offsets.  Non-PIC does not do this.

So I added a hidden switch on the ELFObjectwriter "-arm-elf-force-pic" which
forces the objectwriter to pretend that all relocs are for PIC mode.


Todo: Activate ForceARMElfPIC to true if -relocation-model=pic is selected
on llc.

Todo: There are probably more issues for PIC mode on ARM/MC/ELF...

Todo: Existing tests in MC/ARM/elf-reloc*.ll need to be converted over to .s 
tests as well as expanded to cover the gamut.

llvm-svn: 131205

This means we get C++0x jump-across-intializer semantics correct.

llvm-svn: 131204

Yes, I'm aware that the diagnostics are awful.

Tests to follow.

llvm-svn: 131203

llvm-svn: 131202

In Microsoft mode, allow conversion from pointer to integral type no matter what size the integral type is. Necessary to parse MFC code.

Example:
void f(char *ptr) {
  char var = (char)ptr;
}

llvm-svn: 131201

Part of rdar://8470697

llvm-svn: 131200

Corrected some bugs in both memory and the tests.  Preparing for being able to turn on support for alias templates.

llvm-svn: 131199

llvm-svn: 131198