Bug 13716 - The AIC-7892P controller does not work any more
Summary: The AIC-7892P controller does not work any more
Status: CLOSED INSUFFICIENT_DATA
Alias: None
Product: SCSI Drivers
Classification: Unclassified
Component: Other (show other bugs)
Hardware: All Linux
: P1 high
Assignee: scsi_drivers-other
URL:
Keywords:
Depends on:
Blocks: 13615
  Show dependency tree
 
Reported: 2009-07-05 19:23 UTC by Andrej Podzimek
Modified: 2009-08-10 14:50 UTC (History)
1 user (show)

See Also:
Kernel Version: 2.6.31-rc1. 2.6.31-rc1-git10
Subsystem:
Regression: Yes
Bisected commit-id:


Attachments

Description Andrej Podzimek 2009-07-05 19:23:17 UTC
The usual probe fails with an error message saying that no IRQ could be assigned to the controller, possibly due to a corrupt MP table.

The controller works fine with 2.6.30.1 and below.

This is how the adapter is normally initialized (messages from 2.6.29.6):
    aic7xxx 0000:01:03.0: PCI INT A -> GSI 28 (level, low) -> IRQ 28
    scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
    <Adaptec aic7892 Ultra160 SCSI adapter>
    aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs

With 2.6.31-rc1, the kernel freezes with an error message and tries to restart the controller periodically, which always fails.
Comment 1 Matthew Wilcox 2009-07-05 19:51:29 UTC
On Sun, Jul 05, 2009 at 07:23:18PM +0000, bugzilla-daemon@bugzilla.kernel.org wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=13716
> 
> The usual probe fails with an error message saying that no IRQ could be
> assigned to the controller, possibly due to a corrupt MP table.
> 
> The controller works fine with 2.6.30.1 and below.
> 
> This is how the adapter is normally initialized (messages from 2.6.29.6):
>     aic7xxx 0000:01:03.0: PCI INT A -> GSI 28 (level, low) -> IRQ 28
>     scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
>     <Adaptec aic7892 Ultra160 SCSI adapter>
>     aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
> 
> With 2.6.31-rc1, the kernel freezes with an error message and tries to
> restart
> the controller periodically, which always fails.

That sounds like you have an interrupt routing problem, not a problem
with the AIC driver.  Can you bisect this down to a particular commit?
Comment 2 Anonymous Emailer 2009-07-05 19:54:23 UTC
Reply-To: James.Bottomley@HansenPartnership.com

On Sun, 2009-07-05 at 19:23 +0000, bugzilla-daemon@bugzilla.kernel.org
wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=13716
> 
>            Summary: The AIC-7892P controller does not work any more
>            Product: SCSI Drivers
>            Version: 2.5
>     Kernel Version: 2.6.31-rc1. 2.6.31-rc1-git10
>           Platform: All
>         OS/Version: Linux
>               Tree: Mainline
>             Status: NEW
>           Severity: high
>           Priority: P1
>          Component: Other
>         AssignedTo: scsi_drivers-other@kernel-bugs.osdl.org
>         ReportedBy: andrej@podzimek.org
>         Regression: Yes
> 
> 
> The usual probe fails with an error message saying that no IRQ could be
> assigned to the controller, possibly due to a corrupt MP table.
> 
> The controller works fine with 2.6.30.1 and below.
> 
> This is how the adapter is normally initialized (messages from 2.6.29.6):
>     aic7xxx 0000:01:03.0: PCI INT A -> GSI 28 (level, low) -> IRQ 28
>     scsi0 : Adaptec AIC7XXX EISA/VLB/PCI SCSI HBA DRIVER, Rev 7.0
>     <Adaptec aic7892 Ultra160 SCSI adapter>
>     aic7892: Ultra160 Wide Channel A, SCSI Id=7, 32/253 SCBs
> 
> With 2.6.31-rc1, the kernel freezes with an error message and tries to
> restart
> the controller periodically, which always fails.

OK, so I have a 7896 which acts as the root controller of my SCSI test
system and it's not showing any such problems.  Does this actually work
at all, or does it never work (this could indicate an interrupt routing
problem).  

We're also going to need the kernel logs, with the initialisation
sequence and the actual errors, please.

Thanks,

James
Comment 3 Rafael J. Wysocki 2009-08-10 14:50:40 UTC
Closing due to the lack of information from the reporter.

Note You need to log in before you can comment on or make changes to this bug.