[4503] in linux-scsi channel archive
Re: 2.0 scsi BLIST_SINGLELUN and MBR-7 (fwd)
daemon@ATHENA.MIT.EDU (Derrick J Brashear)
Tue Aug 4 08:45:27 1998
Date: Tue, 4 Aug 1998 08:39:46 -0400 (EDT)
From: Derrick J Brashear <shadow@DEMENTIA.ORG>
To: linux-scsi@vger.rutgers.edu
In-Reply-To: <Pine.LNX.4.02.9808032352200.32209-100000@feral-gw>
On Mon, 3 Aug 1998, Matthew Jacob wrote:
>
> I actually use a Nakamichi just fine on i386 linux. I haven't
> tried it on a sparc- but it really shouldn't be different.
>
> The peculiar thing about the nakamichi is that it refuses to
> change logical block size. That's the only thing really wierd about it.
That should be fine, they're all standard ISO discs I burned, in all 7
slots.
But I try anything which does simultaneous accesses and what ends up
happening is:
Aug 4 00:22:00 jooky kernel: scsi : aborting command due to timeout : pid
4937, scsi1, channel 0, id 2, lun 0 Read (6) 00 00 19 01 00
Aug 4 00:22:30 jooky kernel: scsi : aborting command due to timeout : pid
4950, scsi1, channel 0, id 2, lun 3 Read (6) 60 00 26 01 00
Aug 4 00:22:32 jooky kernel: scsi : aborting command due to timeout : pid
4937, scsi1, channel 0, id 2, lun 0 Read (6) 00 00 19 01 00
Aug 4 00:22:32 jooky kernel: SCSI host 1 abort (pid 4937) timed out -
resetting
Aug 4 00:22:32 jooky kernel: SCSI bus is being reset for host 1 channel
0.
Aug 4 00:23:03 jooky kernel: SCSI host 1 abort (pid 4950) timed out -
resetting
Aug 4 00:23:03 jooky kernel: SCSI bus is being reset for host 1 channel
0.
Aug 4 00:23:04 jooky kernel: CD-ROM I/O error: dev 0b:02, sector 100
Aug 4 00:23:04 jooky kernel: CD-ROM I/O error: dev 0b:05, sector 152
There are only 2 places in scsi.c where single_lun protects any code.
Immediately inside both chunks of code, I inserted a line:
printk("request to singlelun %d:%d:%d: %d\n", device->channel,
device->id, device->lun, SCpnt->request.cmd);
and I never see it, so there's absolutely nothing protecting me,
apparently, which I don't understand at all. Do (the things which are
responsible for calling allocate_device and request_queueable) get set up
on a host basis, i.e. the sparc esp driver overrides them, or something?
Oh well, I'll have more time to poke at this when I get home from work
tonight; I've killed 2 other small sparc scsi problems which had irked me
in the last 48 hours, why not one more?
-D
-
To unsubscribe from this list: send the line "unsubscribe linux-scsi" in
the body of a message to majordomo@vger.rutgers.edu
