[ofa-general] Re: [PATCH v2] rdma_cm: Add debugfs entries to monitor rdma_cm connections

Moni Shoua monis at Voltaire.COM
Mon Apr 27 06:10:46 PDT 2009


Or Gerlitz wrote:
> Moni Shoua wrote:
>> Create a virtual file under debugfs for each cma device and use it to print
>> information about each rdma_id that is attached to this device.
> 
> If you create virtual file for each device, where are you going to print
> listener IDs which aren't bind to any specific device?
A listener that listens an all will appear on all devices
> 
>> Here is an example of 'cat /sys/kernel/debug/rdma_cm/mthca0_rdma_id'
>> TYPE DEVICE  PORT NET_DEV SRC_ADDR                                            DST_ADDR                                            SPACE STATE          QP_NUM
>>      mthca0  0            0.0.0.0:7174                                                                                            TCP   LISTEN         0    
>> IB   mthca0  1    ib0     192.30.3.249:46079                                  192.30.3.248:7174                                   TCP   CONNECT        132102
>> IB   mthca0  1    ib0     192.30.3.249:7174                                   192.30.3.248:42561                                  TCP   CONNECT        132103
> 
> First, if by definition this file relates to device mthca0, why printing mthca0 in the output? 
You are right. It is not necessary. I'll remove it.
> second, if a listener is binded to mthca0 then its src address can't be 0.0.0.0 - correct?
> 
> I'd like to see how we come up with a solution with somehow less space-ing between columns, 
> maybe take a look at the netperf code, e.g that generates output as below. 
I reserved enough space for IPV6 addresses. The output below is good for IPV4 addresses  only. 
> 
> Applying something like the -p option of netperf would be cool, for that 
> end you can have the rmda_cm keep track of rdma_create_id caller's current->pid 
PID looks like a nice addition to the information

> 
>> # netstat -natup
>> Active Internet connections (servers and established)
>> Proto Recv-Q Send-Q Local Address               Foreign Address             State       PID/Program name
>> tcp        0      0 127.0.0.1:2208              0.0.0.0:*                   LISTEN      5979/hpiod
>> tcp        0      0 0.0.0.0:802                 0.0.0.0:*                   LISTEN      5708/rpc.statd
>> tcp        0      0 0.0.0.0:111                 0.0.0.0:*                   LISTEN      5667/portmap
>> tcp        0      0 0.0.0.0:58385               0.0.0.0:*                   LISTEN      -
>> tcp        0      0 127.0.0.1:631               0.0.0.0:*                   LISTEN      6020/cupsd
>> tcp        0      0 127.0.0.1:25                0.0.0.0:*                   LISTEN      6045/sendmail: acce
>> tcp        0      0 127.0.0.1:2207              0.0.0.0:*                   LISTEN      5984/python
>> tcp        0      0 172.30.49.1:47609           172.30.49.2:1000            ESTABLISHED 18567/qperf
>> tcp        0      0 172.30.49.1:58935           172.30.49.2:22              ESTABLISHED 18123/ssh
>> tcp        0      0 :::1003                     :::*                        LISTEN      18562/qperf
>> tcp        0      0 :::22                       :::*                        LISTEN      6004/sshd
>> tcp        0      0 ::ffff:172.30.49.1:22       ::ffff:172.25.5.138:60437   ESTABLISHED 17734/0
>> tcp        0      0 ::ffff:172.30.49.1:1003     ::ffff:172.30.49.3:45282    ESTABLISHED 18563/qperf
>> tcp        0      0 ::ffff:10.10.5.157:22       ::ffff:10.10.0.90:58802     ESTABLISHED 18226/2
>> tcp        0      0 ::ffff:172.30.49.1:22       ::ffff:172.25.5.138:56035   ESTABLISHED 11932/1
> 
> _______________________________________________
> general mailing list
> general at lists.openfabrics.org
> http://lists.openfabrics.org/cgi-bin/mailman/listinfo/general
> 
> To unsubscribe, please visit http://openib.org/mailman/listinfo/openib-general
> 




More information about the general mailing list