Bug 8266

Reply-To: akpm@linux-foundation.org

On Mon, 26 Mar 2007 14:50:51 -0700
bugme-daemon@bugzilla.kernel.org wrote:

> http://bugzilla.kernel.org/show_bug.cgi?id=8266
> 
>            Summary: Exabyte 1x10 autoloader breaks on upgrade
>     Kernel Version: 2.6.20.4
>             Status: NEW
>           Severity: high
>              Owner: scsi_drivers-other@kernel-bugs.osdl.org
>          Submitter: freelsjd@ornl.gov
> 
> 
> Most recent kernel where this bug did *NOT* occur: 2.6.20.3
> Distribution: debian/stable/sarge
> Hardware Environment: amd64
> Software Environment: amanda
> Problem Description:
> When upgrading from 2.6.20.3 to 2.6.20.4, the Exabyte 1x10 autoloader would be
> work anymore.  I am using the aic79xx old driver because the aic79xx new driver
> had already broken (see my previous bugs 5268 and 7616)
> 
> Steps to reproduce:
> upgrade from 2.6.20.3 to 2.6.20.4

We appear to have a regression in the stable kernel.

Probably, that regression is in 2.6.21-rc5 as well.

On Mon, Mar 26, 2007 at 03:33:46PM -0700, Andrew Morton wrote:
> On Mon, 26 Mar 2007 14:50:51 -0700
> bugme-daemon@bugzilla.kernel.org wrote:
> 
> > http://bugzilla.kernel.org/show_bug.cgi?id=8266
> > 
> >            Summary: Exabyte 1x10 autoloader breaks on upgrade
> >     Kernel Version: 2.6.20.4
> >             Status: NEW
> >           Severity: high
> >              Owner: scsi_drivers-other@kernel-bugs.osdl.org
> >          Submitter: freelsjd@ornl.gov
> > 
> > 
> > Most recent kernel where this bug did *NOT* occur: 2.6.20.3
> > Distribution: debian/stable/sarge
> > Hardware Environment: amd64
> > Software Environment: amanda
> > Problem Description:
> > When upgrading from 2.6.20.3 to 2.6.20.4, the Exabyte 1x10 autoloader would be
> > work anymore.  I am using the aic79xx old driver because the aic79xx new driver
> > had already broken (see my previous bugs 5268 and 7616)
> > 
> > Steps to reproduce:
> > upgrade from 2.6.20.3 to 2.6.20.4
> 
> We appear to have a regression in the stable kernel.
> 
> Probably, that regression is in 2.6.21-rc5 as well.

Please attach the dmesg's of both 2.6.20.3 and 2.6.20.4 in Bugzilla.

Does reverting the attached patch fix it?

cu
Adrian

Created attachment 10961 [details]
dmesg for 2.6.20.3 boot

this boot works with the 1x10 autoloader and vxa-2 tape drive  however, error
messages do appear, but I have not studied them

Created attachment 10962 [details]
several boot-up message sequences-most of them have errors

I cannot create a separate dmesg output for 2.6.20.4 at the present time,
because this machine is a production machine and needs to stay up right now. 
However, yesterday while debugging and trying to isolate the problem, I
rebooted several times with different kernels, including the failed 2.6.20.4
kernel.  I also tried the newer aic79xx driver once again to make sure it still
did not work.  I also tried several older kernels; in particular, the 2.6.18.5
kernel which is the last kernel I can get the newer aic79xx driver to work
with.  I hope this will help.  Please let me know if I can help in other ways.

I believe this bug has not been resolved in 2.7.20.7.  I no longer have this
problem and can now use this driver with the same components that caused the
error before.  I recommend this bug be listed as resolved anc closed.

Created attachment 11255 [details]
this is a /var/log/messages exerpt during the time of the scsi error that starts at the tape drive

Created attachment 12011 [details]
new aic7xxx driver dmesg output

supports bug 8266 for kernel 2.6.22.1

Created attachment 12012 [details]
old aic7xxx driver dmesg output

dmesg output from 2.6.22.1 boot and old aic7xxx driver (this one works correctly)

upgrading the kernel to 2.6.22.1 and using the aic7xxx OLD driver seems to fix the scsi problems I was having witht he exabyte autoloader.  Using the aic7xxx NEW driver still has errors.  The dmesg output from both drivers is now attached.

Thanks, I'll close this bug against the old aix7xxx driver.

Did you raise a bugzilla describing the problems in the new
driver?  if not, can you please do so?

Thanks.

Actually, I should have kept this bug open.  A few days later, it failed again with the old aic7xxx driver.  Therefore, both the new and old drivers do not work with the new kernel.  I have resorted back to the 2.6.19.8 kernel and the new aic7xxx driver with that kernel.  Sorry for not reporting earlier.

the patch that fixed bug #8366 also fixed this bug