[ofa-general] [PATCH RFCv2 1/2] RDMA: Add protocol statistics attributes to the RDMA sysfs.

[Roland Dreier]

 > The "Retransmission count" is per netdev and we support multiple netdev
 > per physical port.  We have designed our silicon to support stats at
 > netdev level and it doesn't make sense for us to aggregate stats from
 > all netdevs and report at a device level.  Perhaps we can compromise,
 > Chelsio can give device level stats at /sys/class/infiniband/$device and
 > for NetEffect, we can provide netdev level stats below it?

Hmm, kind of a mess if we can't achieve some uniformity...

how do you handle stats if I change routing so an offloaded connection
goes from physical port 1 to physical port 2?  Or presumably you have
support for port aggregation in hardware so an offloaded connection can
use both physical ports at the same time?

Actually it seems that you attach a struct ib_device to each netdev (eg
a two-port NIC gets two RDMA devices with iw_nes) so Steve's scheme
should work fine for you too?  (I just looked on a system with a
two-port NIC and I do see nes0 and nes1 devices)

 - R.