Thread (1 message) 1 message, 1 author, 2013-03-19

RE: [PATCH 1/1] connector: Added coredumping event to the process connector

From: Jesper Derehag <hidden>
Date: 2013-03-19 09:21:31

From: jderehag@hotmail.com
To: netdev@vger.kernel.org
Subject: RE: [PATCH 1/1] connector: Added coredumping event to the process connector
Date: Sat, 16 Mar 2013 19:08:03 +0000

----------------------------------------
quoted
Date: Sat, 16 Mar 2013 19:40:36 +0100
From: hannes@stressinduktion.org
To: jderehag@hotmail.com
CC: zbr@ioremap.net; netdev@vger.kernel.org
Subject: Re: [PATCH 1/1] connector: Added coredumping event to the process connector

On Sat, Mar 16, 2013 at 05:57:20PM +0000, Jesper Derehag wrote:
quoted
quoted
Date: Sat, 16 Mar 2013 18:03:48 +0100
From: hannes@stressinduktion.org
To: jderehag@hotmail.com
CC: zbr@ioremap.net; netdev@vger.kernel.org
Subject: Re: [PATCH 1/1] connector: Added coredumping event to the process connector

On Sat, Mar 16, 2013 at 11:50:50AM +0100, Jesper Derehag wrote:
quoted
+ ev->event_data.exit.exit_code = task->exit_code;
+ ev->event_data.exit.exit_signal = task->exit_signal;
Do these already contain meaningful values?
I have to admit that they dont.And you are correct, I should add a new event struct specific for the coredump event instead of piggybacking on the exit struct.Will re-submit a patch..
Hm, I am still unsure if such a patch is needed. Couldn't you test for
coredump by inspecting exit_code on PROC_EVENT_EXIT?
*** resubmitted message due to it got dropped by vger.kernel.org ***

 Well, what this patch adds I think is more a question of timing. 
 As an example, say you want to quickly detect process failures. In that case if we would only have the EXIT event, that would mean that we get notified after the dump is done, which could take minutes depending on how large the dump is. 
If we instead watch for both EXIT & COREDUMP events, it would mean that we would quickly catch any failing process, regardless of if its actually starting to coredump or if its exited for some other reason. 

Any other comments on this before I send a v2 patch with the exit vs the coredump event struct change?

Regards,
Jesper 		 	   		  
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help