update comments about .objc_ symbols being generated

llvm-svn: 72708
This commit is contained in:
Nick Kledzik 2009-06-01 23:41:09 +00:00
parent 0b8ca79253
commit 94e2d973c8
1 changed files with 20 additions and 2 deletions

View File

@ -270,8 +270,26 @@ void LTOModule::addDefinedDataSymbol(GlobalValue* v, Mangler& mangler)
// add to list of defined symbols
addDefinedSymbol(v, mangler, false);
// special case i386/ppc ObjC data structures in magic sections
if ( v->hasSection() ) {
// Special case i386/ppc ObjC data structures in magic sections:
// The issue is that the old ObjC object format did some strange
// contortions to avoid real linker symbols. For instance, the
// ObjC class data structure is allocated statically in the executable
// that defines that class. That data structures contains a pointer to
// its superclass. But instead of just initializing that part of the
// struct to the address of its superclass, and letting the static and
// dynamic linkers do the rest, the runtime works by having that field
// instead point to a C-string that is the name of the superclass.
// At runtime the objc initialization updates that pointer and sets
// it to point to the actual super class. As far as the linker
// knows it is just a pointer to a string. But then someone wanted the
// linker to issue errors at build time if the superclass was not found.
// So they figured out a way in mach-o object format to use an absolute
// symbols (.objc_class_name_Foo = 0) and a floating reference
// (.reference .objc_class_name_Bar) to cause the linker into erroring when
// a class was missing.
// The following synthesizes the implicit .objc_* symbols for the linker
// from the ObjC data structures generated by the front end.
if ( v->hasSection() /* && isTargetDarwin */ ) {
// special case if this data blob is an ObjC class definition
if ( v->getSection().compare(0, 15, "__OBJC,__class,") == 0 ) {
if (GlobalVariable* gv = dyn_cast<GlobalVariable>(v)) {