Re: kern/103602: drive gets wedged on READ CD CAPACITY if no disc is in

* Josh Carroll, 2007-04-04 :

> Thinking out of the box, I decided to use sg3_utils on
both the
> semi-working 7.0-CURRENT kernel I had, and on a Linux
Knoppix 5.1 live
> CD.

OK, this is a useful experiment. It really looks to me like
some error
condition that *could* be correctly detected (because it
*is* detected
under Knoppix) is *not* correctly detected in FreeBSD,
causing various
commands to time out.

Now what I think would be really useful would be to try to
reproduce
these timeouts with ATAPI/CAM out of the loop, calling
directly the
ATAPI layer instead.

I wrote a small program that tests that, which you'll find
attached to
this message.  You can compile it using:
  gcc -o readcap readcap.c

You can then try:
  $ readcap /dev/acd0

Here's what I observe here, first with no disk in the drive
(the
important item is sk=02, NOT READY), second with a disk in
the drive.

$ ./readcap /dev/acd0
ioctl returned error 5 Input/output error
Sense data: error=70 sk=02 asc=3a ascq=00
$ ./readcap /dev/acd0
ioctl returned error 0 Unknown error: 0
Sense data: error=00 sk=00 asc=00 ascq=00
LBA of last sector in last session: 187279
Sector size: 2048

> Interestingly, the sg_vpd command fails to read the
serial # as well,
> so that is definitely a known bug with this drive I
think. Here is the
> sg_vpd output on Knoppix:

OK, so ultimately we might need to include this quirk
anyway. We'll see.

Thomas.


_______________________________________________
freebsd-scsifreebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-scsi

To unsubscribe, send any mail to
"freebsd-scsi-unsubscribefreebsd.org"

> Here's what I observe here, first with no disk in the
drive (the
> important item is sk=02, NOT READY), second with a disk
in the drive.

Here is my output (first with no disk, second with). I see
sk=02 here as well:

% ./readcap /dev/acd0
ioctl returned error 5 Input/output error
Sense data: error=70 sk=02 asc=3a ascq=00

% ./readcap /dev/acd0
ioctl returned error 0 Unknown error: 0
Sense data: error=00 sk=00 asc=00 ascq=00
LBA of last sector in last session: 185805
Sector size: 2048

So it looks like directly talking to the ATAPI layer works
ok here as
well. At least, my results look the same as yours.

> > Interestingly, the sg_vpd command fails to read
the serial # as well,
> > so that is definitely a known bug with this drive
I think. Here is the
> > sg_vpd output on Knoppix:
>
> OK, so ultimately we might need to include this quirk
anyway. We'll see.

I was able to boot into Windows XP and run sg_vpd in there
as well,
and it also fails to read the serial number. Just some
additional info
on that. 

Thanks,
Josh
_______________________________________________
freebsd-scsifreebsd.org mailing list
http://lists.freebsd.org/mailman/listinfo/freebsd-scsi

To unsubscribe, send any mail to
"freebsd-scsi-unsubscribefreebsd.org"