Subject: Re: kern/3460: block io unit for sd should be sector size
To: Koji Imada - je4wob/2 <>
From: Zdenek Salvet <>
List: netbsd-bugs
Date: 04/15/1997 10:50:21
> Another problem is about cd.c and cd9660 file system. In that,
> disklabel sector size and physical io unit is inconsistent. So we
> need to decide policy. My idea is:
> 	1. disklabel information is based on disklabel sector size
> 		unit. And disklabel sector size maybe integer multiple
> 		of physical sector size if position problem of
> 		disklabel is resolved.
> 	2. Block io unit for disk device should be standardized.
> 		(maybe DEV_BSIZE or disklabel sector size. It's policy 
> 		problem).
> For 2, DEV_BSIZE corresponds to my "first method" and requires file
> system modification. The disklabel sector size corresponds to my
> "second method". This is simpler but limited(enough?).
> Current cd9660 implementation assumes DEV_BSIZE case, it will require
> some modification if disklabel sector size is selected. And on the
> other hand(DEV_BSIZE case), disklabel initialization code of cd.c
> needs modification.
> If these problem is cleared, It maybe possible to make disklabel
> partitions and ffs for each partition on CD-ROM.
>  Chris> In this case, i'd say "avoid modifying the FS code where possible, and
>  Chris> require that fs sector size == disklabel sector size."

What is the relation between fs sector size (newfs -S) and disklabel sector
size? What about vnd (no disklabel) ?
Will it be possible to have CD image on disk with second method ? 

Zdenek Salvet                                     
           If God had meant for us to be in the Army,
         we would have been born with green, baggy skin.