Bug 12759

Summary: scsi: aic7xxx hang since v2.6.28-rc1 ...
Product: SCSI Drivers Reporter: Rafael J. Wysocki (rjw)
Component: OtherAssignee: scsi_drivers-other
Status: RESOLVED INSUFFICIENT_DATA    
Severity: normal CC: alan, mingo, stern
Priority: P1    
Hardware: All   
OS: Linux   
Kernel Version: 2.6.28-rc1 Subsystem:
Regression: No Bisected commit-id:

Description Rafael J. Wysocki 2009-02-23 14:50:15 UTC
Subject    : scsi: aic7xxx hang since v2.6.28-rc1 ...
Submitter  : Ingo Molnar <mingo@elte.hu>
Date       : 2009-02-15 11:44
References : http://marc.info/?l=linux-kernel&m=123469852817058&w=4
Handled-By : Alan Stern <stern@rowland.harvard.edu>

This entry is being used for tracking a regression from 2.6.27.  Please don't
close it until the problem is fixed in the mainline.
Comment 1 Rafael J. Wysocki 2009-03-14 14:18:19 UTC
The problem seems to have been fixed, closing.
Comment 2 Anonymous Emailer 2009-03-14 15:52:21 UTC
Reply-To: James.Bottomley@HansenPartnership.com

On Sat, 2009-03-14 at 14:18 -0700, bugme-daemon@bugzilla.kernel.org
wrote:
> http://bugzilla.kernel.org/show_bug.cgi?id=12759
> 
> 
> rjw@sisk.pl changed:
> 
>            What    |Removed                     |Added
> ----------------------------------------------------------------------------
>              Status|NEW                         |CLOSED
>          Resolution|                            |CODE_FIX
> 
> 
> 
> 
> ------- Comment #1 from rjw@sisk.pl  2009-03-14 14:18 -------
> The problem seems to have been fixed, closing.

Actually, I don't think the problem is gone, it's just gone into hiding
again.  What seems to happen is that as we alter code in scsi_lib.c,
particularly along the completion paths, it reappears and then
disappears again.  It looks like some obscure timing related bug,
possibly in the aic7xxx driver, or possibly in the mid-layer 

I've rebuilt my main SCSI test box to be aic7xxx based to see if we can
somehow find it again and debug it.

So, I'd like this bug left open against SCSI please, but you can
probably remove it as a current regression (I think it's probably been
hiding for a very long time).

James
Comment 3 Rafael J. Wysocki 2009-03-14 17:17:51 UTC
OK, reopening but dropping from the list of regressions.
Comment 4 Alan 2012-05-30 14:14:07 UTC
Has this been seen since - should we bump the version or close it ?