On Fri, 2007-11-30 at 15:28 -0800, Shirley Ma wrote: > I just touch tested ofed-1.3 beta IPoIB. And found there was a kernel > parameter hw_csum being added in IPoIB. I have several questions here: > 1. Why not using ethtool to set up these HW_CSUM flags? There is no adequate interface in Ethtool for doing it so we use a module parameter. This is because we see this as a static configuration per host.
> 2. I haven't looked at the detailed code yet, is that possible with this > flag, TCP/IP will not do CSUM for HCA which has no TCP/IP offload support? Yes, the HCA need not have checksum offload support. the idea is the IB ICRC provides the insurance that the packets are not corrupt. > If so, then these packets should be limited to be IB network. Routing to > ethernet network, the packets would be dropped. If not, I tested IPoIB-UD, > why I saw 30% improvement with hw_csum set in none connectX mthca SDR > environment? > 3. I saw switching between IPoIB-cm and IPoIB-ud corrupted interface IP > address (unicast address, subnet mask, broadcast address. Anybody saw the > same problem? I did not notice this but it happened to me a few times that when I wanted to change the mtu and forgot to specify "mtu" in if config, then ifconfig interpreted the mtu as the IP address and ended up in a wrong IP address. > > Thanks > Shirley > _______________________________________________ > general mailing list > [email protected] > http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general > > To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general _______________________________________________ general mailing list [email protected] http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
