Re: Bug in mounting routines of atapi-cdroms

Gordon Chaffee (chaffee@cs.berkeley.edu)
Sat, 12 Sep 1998 16:43:57 -0700

Chris Ricker writes:
> The problem he's describing sounds like the same one I have that we hunted
> down a couple of weeks ago (the changes in fs/isofs/inode.c between 2.1.101
> and 2.1.102).
>
> Sorry, I'd been meaning to send this to the list since I can't tell from the
> Maintainers file who does isofs (is that you, Gordon?) but then got busy
> with school and forgot about it. At any rate, the changes to
> fs/isofs/inode.c between 2.1.101 and 2.1.102 break cd-rom access for me and
> a few other people. They cause mount attempts of the cdrom to hang forever
> (on all cds, which work fine under 2.0.x / 2.1.x where x <= 101 , so it's
> not a problem with the images being bad). This is still true as of
> 2.1.122-pre2. So far, I haven't been able to narrow it down any farther than
> that set of changes. Isofs people, please ask away if you need more info.,
> have magic patches, etc.

I would certainly appreciate if you could send some data on any failing
CD images with:

dd if=/dev/cdrom bs=1024 count=2048 | gzip -c -f > cddata.gz

I have a number of images I use for testing purposes, and I'm happy to
add any images that cause problems to my tests. If you have something
that doesn't work with isofs, I would definitely appreciate getting
this data.

I just tried to mount a real CD (as opposed to loopback mounted images
that I normally test with), and I ran into trouble. What I get with
2.1.121 when I try to mount a CD is:

mount -t iso9660 /dev/hdb /cdrom
mount: /dev/cdrom has wrong major or minor number

The CD is a NEC multi-disk CD changer.
/proc/ide/hdb/model: NEC CD-ROM DRIVE:253
/proc/ide/hdb/media: cdrom

Thinking that it may have set the slot wrong somehow, I tried setting it:

% cdchange -v /dev/hdb 1
cdchange: open failed for `/dev/hdb': No such device

This coincides with the strace:

...
open("/dev/hdb", O_RDONLY|O_NONBLOCK) = -1 ENODEV (Operation not supported by device)
...

I have it the IDE CD driver compiled into the kernel, from my .config:
CONFIG_BLK_DEV_IDEDISK=y
CONFIG_BLK_DEV_IDECD=y

I haven't yet tried to determine when this started failing. It has been
awhile since I last used the cd reader, so I don't know if this is new.
It was probably back in the 2.1.80 timeframe I last tried it.

I'll see if I can come up with some better debugging information.

- Gordon

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/faq.html