Re: [PATCH 1/2] e1000: fix netpoll with NAPI

[Mitch Williams <mitch.a.williams@intel.com>]

On Sun, 2006-06-11 at 17:13 -0700, Neil Horman wrote:
> Any further thoughts on this guys?  I still think my last solution
> solves all of
> the netpoll problems, and isn't going to have any noticable impact on
> performance.
> 
I haven't had time to evaluate performance on your patch (sorry!), but
after thinking about it, I agree that it should not have any noticeable
impact.  OTOH, performance tuning is a funny thing, and things you think
won't cause problems often do.

Anyway, I'm still not quite ready to ACK this because it's just not
future-proof.  Eventually, we will need to support multiple RX queues,
and this solution will not work in that situation.

A simpler short-term solution is just to schedule our NAPI polling on
the "real" netdev instead of our polling netdev.  This is a trivial
change and works correctly with a single queue.  But, like your patch,
it isn't future-proof.

So, I'm still thinking and pondering on this one.

If we get a patch in to fix the recursive loop in netpoll, my original
patch will work, right?  Or is there still another issue?

-Mitch