llvm-svn: 56673

llvm-svn: 56672

llvm-svn: 56670

directory is shadowed by a user directory in the lookup
path, ignore the user directory not the system one.  Not
doing this can affect file lookup and the "is a system
header" bit on locations.  For example:
  clang -v -I/usr/include inc.c -E | & grep /usr/inc
now prints:

# 1 "/usr/include/i386/_types.h" 1 3 4
# 37 "/usr/include/i386/_types.h" 3 4
# 70 "/usr/include/i386/_types.h" 3 4

instead of:

# 1 "/usr/include/i386/_types.h" 1
# 37 "/usr/include/i386/_types.h"
# 70 "/usr/include/i386/_types.h"

This is part of rdar://6243860.

llvm-svn: 56669

llvm-svn: 56668

llvm-svn: 56667

llvm-svn: 56666

llvm-svn: 56665

Move VLA processing logic from LiveVariables to CFG construction.  This way all dataflow analyses "see" the VLA size expressions.

llvm-svn: 56655

This fixes <rdar://problem/6248392> clang: Error when using address of stack variable inside block.

llvm-svn: 56652

Fixes <rdar://problem/6248086>

llvm-svn: 56645

llvm-svn: 56644

llvm-svn: 56643

llvm-svn: 56642

llvm-svn: 56641

llvm-svn: 56640

llvm-svn: 56639

llvm-svn: 56638

require more work.

llvm-svn: 56637

llvm-svn: 56636

 - This really needs to be automated and configurable.

llvm-svn: 56635

llvm-svn: 56634

llvm-svn: 56633

Fixes: <rdar://problem/6248119> @finally doesn't introduce a new scope

llvm-svn: 56629

information is in an unreachable block, then it's possible that the high/low pc
values won't be set for the dwarf information. E.g., this function:

void abort(void) __attribute__((__noreturn__));
void dead_beef(void) __attribute__ ((noreturn));

int *b;

void dead_beef(void) {
  *b=0xdeadbeef;
  abort();
}

has a call to "@llvm.dbg.region.end" only in the unreachable block:

define void @dead_beef() noreturn nounwind  {
entry:
	call void @llvm.dbg.func.start(...)
	call void @llvm.dbg.stoppoint(...)
...
	call void @abort( ) noreturn nounwind 
	unreachable

return:		; No predecessors!
	call void @llvm.dbg.stoppoint(...)
	call void @llvm.dbg.region.end(...)
	ret void
}

The dwarf information emitted is something like:

0x00000084:     TAG_subprogram [5]  
                 AT_name( "dead_beef" )
                 AT_external( 0x01 )
                 AT_prototyped( 0x01 )
                 AT_decl_file( 0x01 )
                 AT_decl_line( 0x08 )

Note that this is *not* the best fix for this problem, but a band-aid for an
gaping wound. This code needs to be changed when we revamp our debugging
information.

llvm-svn: 56628

llvm-svn: 56627

load from the stub, instead of the result of the load from the stub.

llvm-svn: 56626

llvm-svn: 56625

s/ParamAttr/Attribute/g
s/PAList/AttrList/g
s/FnAttributeWithIndex/AttributeWithIndex/g
s/FnAttr/Attribute/g

This sets the stage 
- to implement function notes as function attributes and 
- to distinguish between function attributes and return value attributes.

llvm-svn: 56623

s/ParamAttr/Attribute/g
s/PAList/AttrList/g
s/FnAttributeWithIndex/AttributeWithIndex/g
s/FnAttr/Attribute/g

This sets the stage 
- to implement function notes as function attributes and 
- to distinguish between function attributes and return value attributes.

This requires corresponding changes in llvm-gcc and clang.

llvm-svn: 56622

llvm-svn: 56621

With sse3 and when the source is a load or has multiple uses, favors movddup over shuffp*, pshufd, etc. Without sse3 or when the source is from a register, make use of movlhps

llvm-svn: 56620

meaning sse_regparm (i.e. float/double values go
in XMM0 instead of ST0).  Update documentation
to reflect reality.

llvm-svn: 56619

llvm-svn: 56618

llvm-svn: 56617

llvm-svn: 56616

i1 operands are assumed to already by zero-extended.

llvm-svn: 56615

llvm-svn: 56614

llvm-svn: 56613

llvm-svn: 56612