Subject: Re: i/o scheduling (was Re: NEW_BUFQ_STRATEGY)
To: None <bouyer@antioche.lip6.fr>
From: YAMAMOTO Takashi <yamt@mwd.biglobe.ne.jp>
List: tech-kern
Date: 12/17/2003 20:34:58
hi,

> > > I'd like to see this too. IMHO, this can share the same queue as delayed
> > > writes. Only wake up the disk when there's too much requests in the queue, or
> > > we have something else to do.
> > > This would also prevent writes to log files to wake up the disk.
> > 
> > it isn't so easy.
> > "delayed writes" can't be delayed so much because e.g. synchronous
> > read requests can happen on the delayed-writed page and we have no way
> > to look up corresponding buf from a page.
> 
> I don't understand what you mean here. If we have a read request for a
> delayed-write page, why would the read end in the disk's queue at all ?
> UVM should find the data in memory.

please read my replies to tls@.
(maybe a term "read request" was not appropriate to use, sorry.)
i meant, e.g. getpages will find the page in the cache as you say
and then notice the page is marked PG_BUSY and wait on it.

YAMAMOTO Takashi