Subject: Re: More on CD changers
To: Charles M. Hannum <mycroft@MIT.EDU>
From: Laine Stump <>
List: current-users
Date: 04/06/1996 21:26:44
Charles M. Hannum writes:
> Laine Stump <> writes:
> > Yep. I tested it today and saw about 6.5 - 7 seconds for switching times
> > on the NEC 2x (MBR-7), and between 7.5 and 9 seconds switching for the
> > Pioneer 4x (DRM-624x). 
> > 
> > Since the current NetBSD probe stuff causes each disk to be loaded in
> > turn, [...]
> Is this still the case under -current?  I removed the code to probe
> CD-ROM drives at boot time for other reasons, but it should also fix
> this.

Sorry, I guess I should be careful how I use the word "current" on this
mailing list. On the machine in question I am still using 1.1.

> > On that subject - at boottime, the probe properly identifies which slots
> > have disks installed on the Pioneer, but fails to do so reliably on the
> > NEC. As far as mounting, I never get an error from the Pioneer, but
> > always from the NEC (the first try after inserting a disk)
> Does the kernel output any messages when it fails to mount?  The sense
> data would be very helpful.  

No messages are output by the kernel. Only the "mount_cd9660:
Input/Output Error" message. This is the same message that is printed if
a disk is inserted upside down, by the way.

> (To be more blunt, I can't even begin to
> look at why it does that without the sense data.)

Actually I was sending the message more as information/warning to other
people than to con somebody into fixing it. I've pretty much decided to
return the NEC and keep the Pioneer (which has no problems).

If you'd like me to do some things to it first to gather information for
you, I'd be happy to. I'll have to get -current on a machine with a SCSI
controller first, though (and I can't do that with this particular
machine, since it manages my desktop at work). Is there a recent i386
binary snapshot somewhere?

> NEC CD-ROM drives have been known to violate the SCSI-2 spec in lame
> and annoying ways, but it's possible there's something I can do about
> it.