Subject: Re: CVS commit: src/share/man/man9
To: Jason Thorpe <>
From: Bill Studenmund <>
List: source-changes
Date: 10/24/2005 12:27:13
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Mon, Oct 24, 2005 at 10:40:32AM -0700, Jason Thorpe wrote:
> On Oct 24, 2005, at 10:12 AM, Bill Studenmund wrote:
> >This documentation also is the direct result of one of my SoC projects
> >stalling for about a week as the student assumed bno was in device =20
> >blocks,
> >and then nothing worked. When he switched it to DEV_BSIZE blocks, =20
> >things
> >started working.
> >
> >I strongly believe that we need enough documentation so that someone
> >looking at this man page can figure out what units to use for blkno =20
> >for
> >bread() and bwrite().
> I agree with Bill 100% here.  We might not like the current =20
> situation, but we should at least document it.
> (I also agree with Chuq that we should simply switch to using bytes =20
> for this everywhere...)

Yeah, Chuq and I have talked about this in the past. Way back when I was=20
at NASA, I worked on one get-rid-of-DEV_BSIZE method. I changed jobs=20
before finishing, and Chuq finished it differently (as we have it now).=20
It's really a matter of where we want to shove the work; it does all even=
out in the end.

My one thought about bytes vs DEV_BSIZE vs device block size is just that=
it would be nice to choose something that makes all the calculations=20
to&from whatever we use in the interface fast, since they are in the=20
fast-path. :-)

Take care,


Content-Type: application/pgp-signature
Content-Disposition: inline

Version: GnuPG v1.2.3 (NetBSD)