[LLVMdev] More Garbage Collection Questions

[Talin]

I'm still (slowly) working on the project of creating a concurrent 
garbage collector that works with LLVM. I want to ask a little bit more 
about object tags and write barriers and so on.

Let's start with the assumption that a particular language does not use 
per-object type tags. The code generator knows the types of all objects, 
however, and can pass along that type information to llvm.gcroot. And 
since the type of each pointer field is also known, the type information 
can be determined for each traced object, as long as there is no 
polymorphism.

However, what I don't understand is how this works in the presence of 
write barriers. The llvm_gc_write intrinsic takes no type information, 
so there's no way to determine the type of the object except with a tag. 
So I'm not sure what exactly the write barrier can accomplish.

Note that in a real language, most objects would have type tags, but 
there would be a significant number of objects that did not. For 
example, a 'string' class might have a fixed-length header with a type 
tag, which in turn points to a flat array of characters with no tag. The 
character array would not be visible to the outside world, but it would 
still need to be visible to the collector.

-- Talin