Skip site navigation (1)Skip section navigation (2)
Date:      Mon, 31 Aug 1998 13:18:16 -0600 (MDT)
From:      "Kenneth D. Merry" <ken@plutotech.com>
To:        luoqi@watermarkgroup.com (Luoqi Chen)
Cc:        hackers@FreeBSD.ORG, imp@village.org
Subject:   Re: 64k physio limit
Message-ID:  <199808311918.NAA07483@panzer.plutotech.com>
In-Reply-To: <199808310926.FAA13727@lor.watermarkgroup.com> from Luoqi Chen at "Aug 31, 98 05:26:35 am"

next in thread | previous in thread | raw e-mail | index | archive | help
Luoqi Chen wrote...
> > Is there any way to get a larger I/O to happen than 64k?  I have a tape
> > drive that would be happiest if I could do 256kish writes to it at a time,
> > rather than only 64k.
> > 
> > Warner
> > 
> Set the device driver's cdevsw:d_maxio to 256k?
> 
> -lq

	I'm not so sure that will work.  physio currently restricts I/O
size by using minphys().  minphys() looks at the buffer kva size
(b_kvasize), which I think it currently set to MAXPHYS, which is 128K.

	I think that at least in CAM, you should be able to do 256K writes
to a tape drive, assuming:  1.  you can get the transactions through to the
tape driver's strategy routine in that size and 2. your controller can
handle things that size.  The first one probably requires changing MAXPHYS,
or a buffer chaining scheme.

	I'm not quite sure what happens if a controller can't handle I/O
requests of a particular size.  I guess it's handled in the driver, but
I've never looked, so I don't know. :)

Ken
-- 
Kenneth Merry
ken@plutotech.com

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-hackers" in the body of the message



Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199808311918.NAA07483>