Subject: Re: 1.3alpha MO-Drives
To: emanuel stiebler <emu@ecubics.com>
From: Jonathan Stone <jonathan@DSG.Stanford.EDU>
List: port-pmax
Date: 11/13/1997 17:23:55
On Thu, 13 Nov 1997 14:53:13 -0700,
  "emanuel stiebler" <emu@ecubics.com> writes:

>Hi all,
>
>I tried to use my Fujitsu MO-Drive on my Decststion 5000/133
>.
>The Drive is a M2513A. it can use 230, 512 and 640 MByte Magneto Optical
>disks.
>Because i have big problems to edit a right disklabel for the disks, i
>tried to use sysinst for this.
>
>THIS WORKS for the 230 MByte disks. I got many warnings and errors, but i
>could use it after sysinst
>wrote a new filesystem on it. I "told" sysinst to make a configuration like
>partition a 0 bytes, partition b 0 bytes, partition c whole disk and
>partition d whole disk. sysinst made a new filesystem on d.

Thanks for trying it out.  That it works even with warnings is good news.


Did the warnings and errors come from sysinst or from the SCSI driver?
I would not be surprised at some warnings from the SCSI subsystem; it
really should use an MO-specific sense code.

>with df i got something about 217 free MBytes, it is ok.
>
>Sounds like the automatic sensing of SCSI-2 devices is ok.

Neato...

>But ...
>
>the problems are with the 640 MByte cartriges. I got all funny .... kernel
>panics..... ??
>
>i think it is because the cartriges are hardformatted to 2048bytes/sector.

Yes, that's a good guess. The pmax drivers just won't handle 2048
bytes/sector properly. The kernel simply should not crash, even so,
but I would not expect this to work.  (I'm afraid I cannot remember
the name of the Japanese contributor passed away earlier this year,
who did so much work to support this.  But that would probably have to
wait till the pmax uses the MI scsi code.)

Could you please send a PR with as much information about the crash as
possible?  (e.g., a log of any SCSI error messages from a serial
console, if possible). I may look at avoiding the panic for the 1.3
release, so that the kernel quietly refuses to mount disks that would
cause problems.


[snip]

>thanks all,
>
>emanuel

No, thank *you* for testing NetBSD/pmax 1.3_ALPHA on magneto-optical disks.


>P.S. I have no 512 MByte Cartriges here so i could'nt test it. (They have
>512byte/sector)

I'd guess those should work, but it's just a guess.