On Thu, Jun 20, 2002 at 01:01:35PM +0200, Wojciech Puchar wrote:
> > > write/read/free etc.?
> > >
> > > is it CPU dependant?
> > >
> > > that's why it is possible that people doesn't see that so much as they
> > > have todays screamers like athlon.
> >
> > No, it's not a CPU problem (a 50MB queue would require less than 1000 entry,
> > it's not that large). The problem is that when you have several 10s
> > of MB in the queue to be written, it takes several seconds for the queue to
> > drain. As the queue is mostly FIFO, new requests are delayed for several
> > seconds.
> i understand. can't we have 2 queues?

This is what I'm saying since the begining :)
Yes, we need to give different priorities to different I/O, and having several
queues one way to do this (it may not be the best; if we want to keep disksort
keeping a single queue may be easier). The problem is to define the
priority policy (and eventually to gather some infos from upper layer to help
implementing this policy).

--
Manuel Bouyer, LIP6, Universite Paris VI.           Manuel.Bouyer@lip6.fr
--