[Cbe-oss-dev] [PATCH] Fix sparsemem on Cell

[Benjamin Herrenschmidt]

> The only other assumption is that all memory-hotplug-time pages 
> given to memmap_init_zone() are valid and able to be onlined into
> any any zone after the system is running.  The "valid" part is
> really just a question of whether or not a 'struct page' is there
> for the pfn, and *not* whether there is actual memory.  Since
> all sparsemem sections have contiguous mem_map[]s within them,
> and we only memory hotplug entire sparsemem sections, we can
> be confident that this assumption will hold.
> 
> As for the memory being in the right node, we'll assume tha
> memory hotplug is putting things in the right node.

BTW, just that people know, what we are adding isn't even memory :-) We
are calling __add_pages() to create struct page for the SPE local stores
and register space as we use them later from a nopage() handler (and no,
we can't use no_pfn just yet for various reasons, notably we need to
handle races with unmap_mapping_ranges() and thus have the truncate
logic in).

Those pages, thus, must never be onlined. Ever. It might make sense to
create a way to inform memory hotplug of that fact, but on the other
hand, I wouldn't bother as I have a plan to get rid of those
__add_pages() completely and work without struct page, maybe in a 2.6.21
timeframe.

Ben.