<html><body>
<p><tt>> > This is a digression from spidernet, but what if a device is able to<br>
> > generate separate MSIs for TX and RX? Some people from IBM have<br>
> > suggested that it is beneficial for throughput to handle TX work and<br>
> > RX work for IP-over-InfiniBand in parallel on separate CPUs, and<br>
> > handling everything through the ->poll() method would defeat this.<br>
> <br>
> The TX work is so incredibly cheap, relatively speaking, compared<br>
> to the full input packet processing path that the RX side runs<br>
> that I see no real benefit.<br>
> <br>
> In fact, you might even get better locality due to the way the<br>
> locking can be performed if TX reclaim runs inside of ->poll()<br>
</tt><br>
<tt>We are still comparing these two approaches IPoIB performance under different configurations. Hopefully the performance work can be done soon. So we can have some data to share.</tt><br>
<br>
Thanks<br>
Shirley Ma<br>
IBM Linux Technology Center<br>
15300 SW Koll Parkway<br>
Beaverton, OR 97006-6063<br>
Phone(Fax): (503) 578-7638<br>
<br>
<br>
<br>
</body></html>