Re: 2.4: NFS client kmapping across I/O

From: Rik van Riel (riel@conectiva.com.br)
Date: Tue Jan 29 2002 - 13:34:55 EST


On Tue, 29 Jan 2002, Hugh Dickins wrote:
> On Mon, 28 Jan 2002, Xeno wrote:
> >
> > Now I also have time to mention the other NFS client issue we ran into
> > recently, I have not found mention of it on the mailing lists. The NFS
> > client is kmapping pages for the duration of reads from and writes to
> > the server. This creates a scaling limitation, especially under
>
> You're right that kmap users should avoid holding them for very long,
> I'd certainly not discourage anyone from pursuing that effort.

Things like this would be fixed by having a kmap variant
which maps the pages into process-private space, kind of
like kmap_atomic(), but into a (4 MB sized?) part of
address space which is only visible within this process.

This would mean that each process can have a few MB of
kmap()d pages ... should be nice for things like having
copy_{to,from}_user "cache" the hot pages while keeping
longer term mappings from filling the global pool.

regards,

Rik

-- 
DMCA, SSSCA, W3C?  Who cares?  http://thefreeworld.net/

http://www.surriel.com/ http://distro.conectiva.com/

- 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/



This archive was generated by hypermail 2b29 : Thu Jan 31 2002 - 21:01:05 EST