AF:
NF:0
PS:10
SRH:1
SFN:
DSR:
MID:<20080807105650.12e70a15@ripper.onstor.net>
CFG:
PT:0
S:andy.sharp@onstor.com
RQ:
SSV:onstor-exch02.onstor.net
NSV:
SSH:
R:<deepak.veliath@onstor.com>,<jonathan.goldick@onstor.com>,<sandrine.boulanger@onstor.com>,<amit.bothra@onstor.com>
MAID:1
X-Sylpheed-Privacy-System:
X-Sylpheed-Sign:0
SCF:#mh/Mailbox/sent
RMID:#mh/Mailbox/perfarce	0	WEBMAILN0Kz8uNiLPYX000037e9@mail.onstor.com
X-Sylpheed-End-Special-Headers: 1
Date: Thu, 7 Aug 2008 10:57:08 -0700
From: Andrew Sharp <andy.sharp@onstor.com>
To: Deepak Veliath <deepak.veliath@onstor.com>
Cc: Jonathan Goldick <jonathan.goldick@onstor.com>, Sandrine Boulanger
 <sandrine.boulanger@onstor.com>, Amit Bothra <amit.bothra@onstor.com>
Subject: Re: PERFORCE change 30306 for review
Message-ID: <20080807105708.61eb6aa2@ripper.onstor.net>
In-Reply-To: <WEBMAILN0Kz8uNiLPYX000037e9@mail.onstor.com>
References: <WEBMAILN0Kz8uNiLPYX000037e9@mail.onstor.com>
Organization: Onstor
X-Mailer: Sylpheed-Claws 2.6.0 (GTK+ 2.8.20; x86_64-pc-linux-gnu)
Mime-Version: 1.0
Content-Type: text/plain; charset=US-ASCII
Content-Transfer-Encoding: 7bit

Do we want this for cougar release branch?

On 6 Aug 2008 23:35:02 -0700 Deepak Veliath <deepak.veliath@onstor.com>
wrote:

> Change 30306 by deepakv@pinky on 2008/08/06 23:31:39
> 
> 	   TED00025010.  Repeated "mirror pause <mirr-name> && mirror
> resume <mirr-name>" while stressing the mirror source during a mirror
> session results in various panics in sanm_ag_dmRsp() or its callees.
> 	
> 	   The problem was tracked down to the pause logic in sanm_ag
> responding to a pause event after caching the current data-state but
> before caching the pending data-event.  (Both are needed when
> resuming a mirror session). This results in a "mirror resume" using
> stale data-event information from a previous pause-resume cycle.
> Since this stale event contains an edescriptor which has been
> processed and freed, when this event is delivered again, instability
> ensues. 
> 	   The fix is to delay acknowledging the pause event until
> the pending data-event has been received.  This will result in resume
> using the correct as-yet-unprocessed data-event to restart the mirror
> I/O. 
> 	   Reviewed by JobiA.
> 
> Affected files ...
> 
> ... //depot/dev/nfx-tree/code/sm-sanm-agent/sanm-agent-sm-funcs.c#11
> edit
> 
> 
> http://liszt:1818/@md=d&cd=//depot/$c=G35@/30306?ac=10
