As suggested by Nick Lewycky, ignore debugging intrinsics when trying to decide whether a destructor is empty or not.

llvm-svn: 127985

llvm-svn: 127984

a year now.  Update the internals manual.

llvm-svn: 127983

llvm-svn: 127982

Re-apply r127953 with fixes: eliminate empty return block if it has no predecessors; update dominator tree if cfg is modified.

llvm-svn: 127981

llvm-svn: 127980

llvm-svn: 127977

llvm-svn: 127976

llvm-svn: 127975

llvm-svn: 127974

llvm-svn: 127973

gun as does. This makes it a lot easier to compare the output of both
as the addresses are now a lot closer.

llvm-svn: 127972

if '-x' flag is present it will be unused.

llvm-svn: 127971

Add an optimization to GlobalOpt that eliminates calls to __cxa_atexit, if the function passed is empty.

llvm-svn: 127970

- glibc falls back to fork+exec if a file actions object is present.
- On BSDs this saves a malloc.

llvm-svn: 127969

llvm-svn: 127967

If a class inherits from RefCountedBaseVPTR allow all its subclasses to be used with IntrusiveRefCntPtr.

llvm-svn: 127966

This can happen when multiple sibling registers are spilled after live range
splitting.

llvm-svn: 127965

llvm-svn: 127964

static archive that can be linked against. LLDB.framework/lldb.so
exports a very controlled API. Splitting the API into a static
library allows other tools (debugserver for now) to use the power
of the LLDB debugger core, yet not export it as its API is not
portable or maintainable. The Host layer and many of the other
internal only APIs can now be statically linked against.

Now LLDB.framework/lldb.so links against "liblldb-core.a" instead
of compiling the .o files only for the shared library. This fix
is only for compiling with Xcode as the Makefile based build already
does this.

The Xcode projecdt compiler has been changed to LLVM. Anyone using
Xcode 3 will need to manually change the compiler back to GCC 4.2,
or update to Xcode 4.

llvm-svn: 127963

llvm-svn: 127962

Downgrade err_mismatched_exception_spec to a ExtWarning in Microsoft mode. MSVC doesn't do any validation on exception specifications.

This remove 1 error when parsing MSVC stl lib with clang.

llvm-svn: 127961

llvm-svn: 127960

llvm-svn: 127959

Some of those POSIX <-> Windows command line conversions ended on
failure.

llvm-svn: 127958

llvm-svn: 127957

llvm-svn: 127956

to canonicalize IR", it broke a lot of things.

llvm-svn: 127954

to have single return block (at least getting there) for optimizations. This
is general goodness but it would prevent some tailcall optimizations.
One specific case is code like this:
int f1(void);
int f2(void);
int f3(void);
int f4(void);
int f5(void);
int f6(void);
int foo(int x) {
  switch(x) {
  case 1: return f1();
  case 2: return f2();
  case 3: return f3();
  case 4: return f4();
  case 5: return f5();
  case 6: return f6();
  }
}

=>
LBB0_2:                                 ## %sw.bb
  callq   _f1
  popq    %rbp
  ret
LBB0_3:                                 ## %sw.bb1
  callq   _f2
  popq    %rbp
  ret
LBB0_4:                                 ## %sw.bb3
  callq   _f3
  popq    %rbp
  ret

This patch teaches codegenprep to duplicate returns when the return value
is a phi and where the phi operands are produced by tail calls followed by
an unconditional branch:

sw.bb7:                                           ; preds = %entry
  %call8 = tail call i32 @f5() nounwind
  br label %return
sw.bb9:                                           ; preds = %entry
  %call10 = tail call i32 @f6() nounwind
  br label %return
return:
  %retval.0 = phi i32 [ %call10, %sw.bb9 ], [ %call8, %sw.bb7 ], ... [ 0, %entry ]
  ret i32 %retval.0

This allows codegen to generate better code like this:

LBB0_2:                                 ## %sw.bb
        jmp     _f1                     ## TAILCALL
LBB0_3:                                 ## %sw.bb1
        jmp     _f2                     ## TAILCALL
LBB0_4:                                 ## %sw.bb3
        jmp     _f3                     ## TAILCALL

rdar://9147433

llvm-svn: 127953

llvm-svn: 127952

not have native support for this operation (such as X86).
The legalized code uses two vector INT_TO_FP operations and is faster
than scalarizing.

llvm-svn: 127951

llvm-svn: 127950

llvm-svn: 127949

llvm-svn: 127948

description updates.

llvm-svn: 127947

update short descriptions to match those from the options, alphabetize table
of contents.

llvm-svn: 127946

llvm-svn: 127945

llvm-svn: 127944

llvm-svn: 127943

llvm-svn: 127942