New: 2.6.30 -> 2.6.31: very high number of RX dropped in bnx2

Bugme-new Andrew Morton 3 participants 5 messages Nov 25, 2009 – Dec 2, 2009 1 message not in the archive

Subsystems (2)

Networking Network drivers

lore

Share & tools

lore

A message that is not in the archive (sent off-list or not yet received): [email protected]/

Andrew Morton

author

2009-11-25 20:47 UTC #1 lore raw

(switched to email. Please respond via emailed reply-to-all, not via the bugzilla web interface).

Quoted 32 lines · On Wed, 25 Nov 2009 00:16:00 GMT [email protected] wrote:

http://bugzilla.kernel.org/show_bug.cgi?id=14685

           Summary: 2.6.30 -> 2.6.31: very high number of RX dropped in
                    bnx2
           Product: Drivers
           Version: 2.5
    Kernel Version: 2.6.31.6
          Platform: All
        OS/Version: Linux
              Tree: Mainline
            Status: NEW
          Severity: normal
          Priority: P1
         Component: Network
        AssignedTo: [email protected]
        ReportedBy: [email protected]
        Regression: Yes

After upgradig my kernel from 2.6.30 to 2.6.31 I'm observing very high number of dropped packets on BCM5708(Dell PowerEdge 1950)/BCM5716(Dell PowerEdge R410) NICs.

# uptime ; ifconfig eth0|grep "RX p"; ethtool -S eth0|grep disc
 01:07:31 up 6 min,  1 user,  load average: 0.00, 0.07, 0.04
          RX packets:2234 errors:0 dropped:527 overruns:0 frame:0
     rx_discards: 0
     rx_fw_discards: 527

Everything seems to work just fine so I suspect that the driver/firmware is counting something wrong.

Michael Chan

2009-11-30 11:33 UTC #2 lore raw

Quoted 6 lines · Andrew Morton

Quoted 6 lines

# uptime ; ifconfig eth0|grep "RX p"; ethtool -S eth0|grep disc
 01:07:31 up 6 min,  1 user,  load average: 0.00, 0.07, 0.04
          RX packets:2234 errors:0 dropped:527 overruns:0 frame:0
     rx_discards: 0
     rx_fw_discards: 527

Please report rx_bcast_packets and rx_mcast_packets together with rx_fw_discards. Thanks.

Michael Chan

2009-11-30 17:56 UTC #4 lore raw

Quoted 5 lines · Krzysztof Oledzki

# ethtool -S eth0|egrep "(rx_bcast_packets|rx_mcast_packets|
rx_fw_discards)"
      rx_mcast_packets: 2806
      rx_bcast_packets: 790016
      rx_fw_discards: 790016

This tells me that all the dropped packets are occurring in the iSCSI ring.

We have recently enabled iSCSI in the kernel and you probably don't have the userspace components to handle the iSCSI ring. All broadcast packets are also received in the iSCSI ring and they are all dropped in this case without the userspace components. Networking is not affected. If you rmmod the bnx2i and cnic drivers, the drop counter should stop counting.

We have newer firmware that will not include dropped iSCSI packets in the networking counters. It should appear in the 2.6.33 kernel time-frame.

Thanks.

Krzysztof Olędzki

2009-12-02 13:38 UTC #5 lore raw

Quoted 17 lines · Michael Chan

Quoted 5 lines · Krzysztof Oledzki
# ethtool -S eth0|egrep "(rx_bcast_packets|rx_mcast_packets|
rx_fw_discards)"
      rx_mcast_packets: 2806
      rx_bcast_packets: 790016
      rx_fw_discards: 790016
This tells me that all the dropped packets are occurring in the iSCSI ring.
We have recently enabled iSCSI in the kernel and you probably don't have the userspace components to handle the iSCSI ring. All broadcast packets are also received in the iSCSI ring and they are all dropped in this case without the userspace components. Networking is not affected. If you rmmod the bnx2i and cnic drivers, the drop counter should stop counting.

Quoted 3 lines

We have newer firmware that will not include dropped iSCSI packets in the networking counters. It should appear in the 2.6.33 kernel time-frame.

Yep, it was the cnic driver. Now I feel silly. Thank you for you help and sorry for the noise. I going to close it with RESOLVED-DOCUMENTED.

Best regards,

Krzysztof Olędzki