I have more than a vague suspicion now.

I built a multi-session CD one my HP here at work (running under NT).  There
wasn't anything on the CD (except for a single 0 byte file).  I took it home
and put it in the burner.  I ran "cdrecord image.iso" and wrote my 1.5.1 ISO
image onto the disk.

I then popped a virgin blank into the drive and went back to ASC 0x89 on CDB
0x25.

I'm thinking it has to be a kernel problem.

I might (as a cheap thrill) go back to a really old version of NetBSD and
see if that helps.

> -----Original Message-----
> From: SUNAGAWA Keiki [mailto:Keiki_SUNAGAWA@yokogawa.co.jp]
> Sent: Tuesday, March 13, 2001 7:22 PM
> To: Dave Burgess
> Cc: SUNAGAWA Keiki; current-users@netbsd.org
> Subject: Re: CD Writer failure
>
>
> Dave Burgess <burgess@cynjut.neonramp.com> wrote:
>
> >> Well, I have no idea what's going on here, but rebuilding
> >> cdrecord with SONY_DEBUG defined in cdrecord/drv_sony.c
> >> would help some?
>
> Dave> Not really - SONY_DEBUG has been defined for the last
> Dave> 20 or 30 tries.
>
> Hmm, that's too bad.
>
> Dave> I'm willing to bet that it is a problem in the 'fake'
> Dave> disklabel in the kernel is either bogus or the OS is
> Dave> not handling the CD ROM correctly.  When the disk has
> Dave> a real 'available' amount, or when the CD-ROM has no
> Dave> room left (as in a pressed CD) the system processes
> Dave> the request correctly.
>
> You might want to look at the difference between 1.3 and 1.4
> of the CD-ROM handling code.  I guess we have some troubles
> with drives that change the result to the inquiry command in
> the status of the media tray.  e.g. medium type, empty, etc.
>
> Dave> I've been looking at the kernel code, but I'm not
> Dave> familiar enoght with the code yet to be able to tell
> Dave> what should be happening.
>
> /sys/dev/scsipi/scsiconf.c has quirk list of several SONY
> drives, but it doesn't seem to be related to the problem.
>
> --
> SUNAGAWA Keiki <Keiki_SUNAGAWA@yokogawa.co.jp>
>