Date: Sun, 23 Apr 1995 19:38:14 -0400 From: Thomas David Rivers <ponds!rivers@dg-rtp.dg.com> To: ref.tfs.com!julian@dg-rtp.dg.com, ponds!rivers Cc: freebsd-hackers@freefall.cdrom.com Subject: Re: Interesting SCSI cdrom problem.. Message-ID: <199504232338.TAA00879@lakes>
next in thread | raw e-mail | index | archive | help
>
> >
> This is a 'feature' of the driver
> once the door has been openned, the drive reports to the driver, that
> there has been user intervention of some kind..
> specifically a 'Unit attention' erro is reported to the driver..
>
> this tells the driver that the cdrom MAY HAVE BEEN CHANGED!.
>
> The driver therefore aborts ALL operations until ALL USERS have closed the
> device. (including the 'mount' 'user' ).
>
> When the last user has closed the device, operations on the cdrom are
> re-enabled..
>
> I'm glad to see you are getting this as it proves that the code in
> question is working, and it is vitally important to devices
> with read-write removable media that it does....
> consider..
> async write to drive...
> {change media}
> sync
> (OUCH!)
>
> the same problem exists for cdoms but it doesn't corrupt the media,
> just totally screws the internal cached copy of the filesystem nodes
> (i.e if it assumed that the same cdrom was in but wasn't.....)
>
Yes, that entirely makes sense. I was considering the situation myself;
what else could you do given that the "disk" can be jerked out from under
you?
- Thanks -
- Dave Rivers -
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?199504232338.TAA00879>