<html><body>
<p><tt>oland Dreier <rdreier@cisco.com> wrote on 02/27/2007 02:41:44 PM:<br>
<br>
> > So the IBV_CQ_REPORT_MISSED_EVENTS has been part of OFED-1.2 already? I can<br>
> > generate the patch for all ULPs to use this for review. Do you need me to<br>
> > do that?<br>
> <br>
> No, it's not in OFED 1.2 or the upstream kernel. And no one has<br>
> implemented it for userspace (and I'm somewhat reluctant to break the<br>
> ABI at this point without some performance numbers to motivate making<br>
> this API change).<br>
> <br>
> Have the NAPI performance problems with ehca been resolved? We could<br>
> probably merge IPoIB NAPI for 2.6.22 then, which would pull in the<br>
> kernel changes at least.<br>
> <br>
> - R.<br>
We have addressed the NAPI performance issues with ehca driver. I believe the patches have been upper stream. However the test results show that it's better to delay poll again to next NAPI interval, something like this:</tt><br>
<br>
<tt>poll-cq</tt><br>
<tt>notify-cq, if missed_event && netif_rx_reschedule()</tt><br>
<tt>return 1</tt><br>
<br>
<tt>vs.</tt><br>
<tt>poll-cq,</tt><br>
<tt>notify-cq, if missed_event && netif_rx_reschedule()</tt><br>
<tt>poll again</tt><br>
<tt>return 0</tt><br>
<br>
It seems ehca delivering packet much faster than other HCAs. So poll again would stay in the loop for many many times. So the above changes doesn't impact other HCAs, I would recommand it. I saw same implementations on other ethernet drivers.<br>
<br>
Thanks<br>
Shirley Ma</body></html>