[ovs-dev] Backport Request: dpif-netdev: Remove PMD latency on seq_mutex

[Vinod, Chegu]

Thanks Flavio for checking and Daniel for your consideration. 
Indeed the issue exists in 2.5 branch. 

We are treating this more in the bucket of a performance bug fix than a feature.

Any specific testing that you would like to see run to help reduce 
your concern related to changes to the core modules ?

Ciara, what's your opinion on these changes for a backport ?

Thanks
Vinod



-----Original Message-----
From: Daniele Di Proietto [mailto:diproiettod at vmware.com] 
Sent: Friday, July 22, 2016 3:49 PM
To: Flavio Leitner <fbl at redhat.com>; Vinod, Chegu <chegu_vinod at hpe.com>
Cc: krister at redhat.com; ovs-dev <dev at openvswitch.org>; Loftus, Ciara <ciara.loftus at intel.com>
Subject: Re: Backport Request: dpif-netdev: Remove PMD latency on seq_mutex

I'm not sure I'm 100% comfortable back porting this to branch-2.5

I see the change more as a feature rather than a bugfix.

Also it touches some core modules (seq and rcu) in a non trivial way.


What do you guys think?

Thanks,

Daniele

On 22/07/2016 15:03, "Flavio Leitner" <fbl at redhat.com> wrote:

>(adding ovs-dev mailing list and more people interesting on the 
>backport to CC)
>
>On Mon, Jul 18, 2016 at 05:31:52AM +0000, Vinod, Chegu wrote:
>> Hi Flavio, Karl,
>> 
>> Is there a version of the following fix available that is compatible with OVS 2.5?
>> 
>> https://github.com/openvswitch/ovs/commit/9dede5cff553d7c4e074f04c525
>> c1417eb209363
>> 
>> If yes can it backported to the 2.5 branch ?
>
>branch-2.5 is affected by the same issue.  I tested the patch from 
>branch master (cherry-pick) and it solved the issue.
>
>Daniele,
>
>What do you think? If you agree, do you need me to post the backported 
>patch or is it enough for you to cherry-pick?
>
>Thanks,
>--
>fbl