[1831] in linux-scsi channel archive
Re: PPA module weidness
daemon@ATHENA.MIT.EDU (campbell@tirian.che.curtin.edu.au)
Thu May 8 21:27:14 1997
From: campbell@tirian.che.curtin.edu.au
To: mcculley@iag.net
Date: Fri, 9 May 1997 09:23:03 +0000
Cc: linux-scsi@vger.rutgers.edu
> Date: Thu, 8 May 1997 04:36:03 -0400
> From: mcculley@iag.net
>
> I just got a parallel port ZIP drive and have been trying to compile a
> kernel to support it. I already am happily using an AHA-2940 Ultra in
> the box. I tried to compile the PPA support as a module, but leave
> the base SCSI code and the aic7xxx support compiled in. However, when
> I try to load the PPA module, it screws up the scsi /proc entries and
> who knows what else.
>
> When I load the PPA module, this happens:
>
> root://$ cd /proc
> root:/proc/$ cd scsi
> root:/proc/scsi/$ ls -la
> total 0
> dr-xr-xr-x 4 root root 0 May 7 10:10 .
> dr-xr-xr-x 18 root root 0 May 7 10:09 ..
> dr-xr-xr-x 2 root root 0 May 7 10:10 aic7xxx
> dr-xr-xr-x 2 root root 0 May 7 10:10 ppa
> -rw-r--r-- 1 root root 0 May 7 10:10 scsi
> -rw-r--r-- 1 root root 0 May 7 10:10 scsi
> root:/proc/scsi/$ cat scsi
> Attached devices:
> Host: scsi0 Channel: 00 Id: 06 Lun: 00
> Vendor: IOMEGA Model: ZIP 100 Rev: D.13
> Type: Direct-Access ANSI SCSI revision: 02
> root:/proc/scsi/$ cat ppa/0
> Version : Curtin 1-08-BETA
> Parport : parport0
> Mode : SPP
>
> Timing Parameters
> ppa_speed 6
> ppa_speed_fast 1
> root:/proc/scsi/$ cat aic7xxx/0
> Version : Curtin 1-08-BETA
> Parport : parport0
> Mode : SPP
>
> Timing Parameters
> ppa_speed 6
> ppa_speed_fast 1
> root:/proc/scsi/$ exit
>
> The aic7xxx entry gets totally baked. I am trying to compile both as
> modules right now. Is this a known problem? Is it a PPA problem with
> not handling the different cases of module registration right, or a
> more general problem affecting any SCSI module that tries to register
> when there are already SCSI devices compiled in? I'll try to figure
> it out once I get running again.
Sounds like something mid-level is broken, ppa accepts proc requests
from the main scsi driver so if you are seeing the same contents else
where then something is definitely wrong.
*ponder*
Could you do a grep for PROC_SCSI_ in the other module (aic7xxx) I
think someone has used the same proc inode number for both ppa and
aic7xxxx, this could cause the problems you are seeing (as a result
it is only limited to /proc being corrupted).
The PROC_SCSI_ entries are defined in <linux/procfs.h>.
David Campbell
