Hello,

quoted text
> > But I think each memory model type should have its own way of managing
> > its page_cgroup arrays as doing for its struct page arrays.
> > It would be better rather than the sparsemem approach he said.
> > 
> My patch adds an interface. Then...
> FLATMEM support will be very easy.

Yes, that is true, it is really easy.

quoted text
> I'll ignore DISCONTIGMEM and SPARSEMEM (they will use my 'hash')

What part of this do you think is the problem to implement it in the
straight way for this model?
I think it won't be difficult to implement it since each pgdat can have
its page_cgroup array, which can care about holes in the node as well as
doing it for its struct page array.

quoted text
> SPARSEMEM_VMEMMAP support will took some amount of time. It will need
> per-arch patches.

Yes, each of ia64, powerpc and x86_64 use this memory model.

We should also care about the regular SPARSEMEM case as you mentioned.

quoted text
> Thanks,
> -Kame

--
unsubscribe notice
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/