[Cbe-oss-dev] PATCH [0/3] spusched enhancements

[Luke Browning]

On Wed, 2007-10-24 at 10:23 +0200, Christoph Hellwig wrote:
> Hi Luke,
> 
> I like this a lot.  I remember we had some problems with pagefault
> handling with this idea, so we might have to look into that again.
> 

Thanks!  You are right.  I believe there is a simple less optimal
solution for page fault handling.  We can exploit the NOSCHED option to
force the load of the spe context so that we can process the fault under
the same conditions as today.  I will look into that as well as a better
solution that does not require the context to be loaded.  It is sort of
an arbitrary limitation due to the way we process page faults.  

BTW, can you recommend a testcase that does a lot of I/O.

Luke