[ovs-dev] [PATCH 1/3] dpif-netdev: only poll enabled vhost queues

[David Marchand]

On Fri, Apr 12, 2019 at 12:00 PM Ilya Maximets <i.maximets at samsung.com>
wrote:

> On 11.04.2019 17:13, David Marchand wrote:
> >
> >
> > On Thu, Apr 11, 2019 at 4:02 PM David Marchand <
> david.marchand at redhat.com <mailto:david.marchand at redhat.com>> wrote:
> >
> >     We currently poll all available queues based on the max queue count
> >     exchanged with the vhost peer and rely on the vhost library in DPDK
> to
> >     check the vring status beneath.
> >     This can lead to some overhead when we have a lot of unused queues.
> >
> >     To enhance the situation, we can skip the disabled queues.
> >     On rxq notifications, we make use of the netdev's change_seq number
> so
> >     that the pmd thread main loop can cache the queue state periodically.
> >
> >     $ ovs-appctl dpif-netdev/pmd-rxq-show
> >     pmd thread numa_id 0 core_id 1:
> >       isolated : true
> >       port: dpdk0             queue-id:  0  pmd usage:  0 %  polling:
> enabled
> >     pmd thread numa_id 0 core_id 2:
> >       isolated : true
> >       port: vhost1            queue-id:  0  pmd usage:  0 %  polling:
> disabled
> >       port: vhost3            queue-id:  0  pmd usage:  0 %  polling:
> disabled
> >     pmd thread numa_id 0 core_id 15:
> >       isolated : true
> >       port: dpdk1             queue-id:  0  pmd usage:  0 %  polling:
> enabled
> >     pmd thread numa_id 0 core_id 16:
> >       isolated : true
> >       port: vhost0            queue-id:  0  pmd usage:  0 %  polling:
> disabled
> >       port: vhost2            queue-id:  0  pmd usage:  0 %  polling:
> disabled
> >
> >
> > This change broke the tests, since it matches the exact command output.
> > Is the output format something we must maintain ?
>
> I don't think so.
> You just need to fix tests to match with the new output format.
> However, in general, the output is already overloaded. Maybe it's
> worth to print polling state only if it's disabled?
> This should also save you from changing most of the tests.
>

Well, either updating the test or doing this is fine for me.
I will go with your suggestion.


-- 
David Marchand