- Steve, can you take a look at this? It seems like this code should live
   elsewhere, and there is a FIXME about having Sema validates the UTF-8 to
   UTF-16 conversion.

llvm-svn: 76915

llvm-svn: 76908

llvm-svn: 76907

llvm-svn: 76903

llvm-svn: 76898

llvm-svn: 76897

llvm-svn: 76885

We don't need to keep track of the packed alignment, just whether the struct is packed or not. Fixes PR4610.

llvm-svn: 76884

Move the LLVM field number for bit fields into the BitFieldInfo structure, since it's meaning is completely different than for non-bit fields.

llvm-svn: 76882

llvm-svn: 76878

llvm-svn: 76862

llvm-svn: 76856

Use arrays as union padding. Also, since the resulting struct will always contain a single element and either a single i8 element or an array of i8s, there's no reason to use a packed struct.

llvm-svn: 76854

llvm-svn: 76847

llvm-svn: 76846

Check in  CGRecordLayoutBuilder which is a reimplementation of the record layout code. (Yay, no more packed structs unless absolutely necessary). We currently don't use the layouts being built but that will change when the new code is mature enough :)

llvm-svn: 76845

Technically we could still do a bit more to avoid deferred generation of statics
which we know are used, but I seriously doubt this is important.

llvm-svn: 76844

This requires llvm rev. 76769 or higher.

llvm-svn: 76770

Fix a couple recent ABI regressions noticed during code review (fallout from the ObjC type system rewrite).

It's unfortunate that the mangling includes the low-level structs. Nevertheless, we need this for binary compatibility with GCC.

llvm-svn: 76755

llvm-svn: 76726

   __attribute__((address_space(1))) struct {int arr[ 3 ]; }  *p1;
   ... = p1->arr[2];  // load from address space 1

llvm-svn: 76717

llvm-svn: 76599

llvm-svn: 76534

llvm-svn: 76498

llvm-svn: 76493

llvm-svn: 76473

llvm-svn: 76365

expressions.
 - This generally catches the important case of noreturn functions.

 - With the last two changes, we are down to 152 unreachable blocks emitted on
   403.gcc, vs the 1805 we started with.

llvm-svn: 76364

llvm-svn: 76362

 - Emit variable declarations as "simple", we want to avoid forcing the creation
   of a dummy basic block, but still need to make the variable available for
   later use.

 - With that, we can now skip IRgen for other unreachable statements (which
   don't define a label).

 - Anders, I added two fixmes on calls to EmitVLASize, can you check them?

llvm-svn: 76361

llvm-svn: 76339

llvm-svn: 76327

llvm-svn: 76321

populate CGM's list directly.

llvm-svn: 76266

until Doug Gregor's Type smart pointer code lands (or more discussion occurs).
These methods just call the new Type::getAs<XXX> methods, so we still have
reduced implementation redundancy. Having explicit getAsXXXType() methods makes
it easier to set breakpoints in the debugger.

llvm-svn: 76193

Replaced Type::getAsLValueReferenceType(), Type::getAsRValueReferenceType(), Type::getAsMemberPointerType(), Type::getAsTagType(), and Type::getAsRecordType() with their Type::getAs<XXX> equivalents.

llvm-svn: 76139

in 32bit abi.

llvm-svn: 76109

llvm-svn: 76099

This method is intended to eventually replace the individual
Type::getAsXXXType<> methods.

The motivation behind this change is twofold:

1) Reduce redundant implementations of Type::getAsXXXType() methods. Most of
them are basically copy-and-paste.

2) By centralizing the implementation of the getAs<Type> logic we can more
smoothly move over to Doug Gregor's proposed canonical type smart pointer
scheme.

Along with this patch:

a) Removed 'Type::getAsPointerType()'; now clients use getAs<PointerType>.
b) Removed 'Type::getAsBlockPointerTypE()'; now clients use getAs<BlockPointerType>.

llvm-svn: 76098

llvm-svn: 76090