<br><font size=2><tt>"Michael S. Tsirkin" <mst@mellanox.co.il>
wrote on 05/09/2006 01:20:41 PM:<br>
<br>
> Quoting r. Shirley Ma <xma@us.ibm.com>:<br>
> > My understanding is NAPI handle interrutps CQ callbacks on the
same CPU.<br>
> <br>
> My understanding is NAPI disables interrupts under high RX load. No?<br>
> <br>
> -- <br>
> MST<br>
</tt></font>
<br><font size=2><tt>Yes, NAPI disables the interrupts based on the weight.
In IPoIB case, it doesn't </tt></font>
<br><font size=2><tt>send out the next completion notification under heavy
loading. </tt></font>
<br><font size=2><tt>The similiar CQ polling is still in NAPI on same CPU,
but it's not a callback</tt></font>
<br><font size=2><tt>anymore. </tt></font>
<br>
<br><font size=2><tt>What I find that the send completion and recv completion
are not </tt></font>
<br><font size=2><tt>that fast, which means RX load is not that heavy in
IPoIB. That might be</tt></font>
<br><font size=2><tt>the reason compared to multiple threads implementation
NAPI is not good.</tt></font>
<br>
<br><font size=2 face="sans-serif">Thanks</font>
<br><font size=2 face="sans-serif">Shirley Ma<br>
IBM Linux Technology Center<br>
15300 SW Koll Parkway<br>
Beaverton, OR 97006-6063<br>
Phone(Fax): (503) 578-7638</font>