Subject: Re: CVS commit: src/share/man/man9
To: YAMAMOTO Takashi <>
From: Bill Studenmund <>
List: source-changes
Date: 10/27/2005 09:04:57
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
Content-Transfer-Encoding: quoted-printable

On Thu, Oct 27, 2005 at 01:36:13PM +0900, YAMAMOTO Takashi wrote:
> > Module Name:	src
> > Committed By:	wrstuden
> > Date:		Thu Oct 27 00:05:03 UTC 2005
> >=20
> > Modified Files:
> > 	src/share/man/man9: buffercache.9
> >=20
> > Log Message:
> > Adjust discussion of blkno's units to describe the units for
> > blkno for non-special files. Per comments from yamt at.
> the "For other files," part is an internal detail of each filesystems.
> i don't think it's appropriate to document it here.

I disagree. The call is here, so we should say something. As best I can=20
tell, we don't document it anywhere else, so we can't do cross-references.=

I agree each file system gets to make up its own mind, but that's why I
said, "File system blocks." My experience has been that each file system
gets to choose its own idea of what a block is (and they often all choose
differently), thus we have a kilometer of wiggle room in even this

As others may not realize "file system blocks" is such an overloaded term,=
I'll soften the text a bit more.

> there is no filesystem independent way to use bread() anyway.

I'm not saying there is. We however use bread() in a number of file=20
systems, and for the most part we use it in similar ways. I think it's=20
perfectly reasonable to document that similarity.

Take care,


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

Version: GnuPG v1.2.3 (NetBSD)