Hello,
I just tried with Q6 and the network behaviour seems odd :
- my ssh connection freezes a bit,
- communication between the two PCs sharing the GigE connection is difficult
to start :
13:17:37.965962 arp who-has djerba-gig tell canopus-gig
13:17:37.966220 IP djerba-gig > canopus-gig: udp
13:17:37.966230 IP djerba-gig > canopus-gig: udp
13:17:37.966313 IP djerba-gig > canopus-gig: udp
13:17:38.965906 arp who-has djerba-gig tell canopus-gig
13:17:38.966040 IP djerba-gig > canopus-gig: udp
This is strange since canopus is the one with Q6
- ping does not start for a moment since all packets are RX-ERROR, then they
start being received.
Here is the output of ethtool -S eth1
NIC statistics:
rx_packets: 581
tx_packets: 552
rx_bytes: 545746
tx_bytes: 66508
rx_errors: 112950
tx_errors: 0
rx_dropped: 90
tx_dropped: 0
multicast: 0
collisions: 0
rx_length_errors: 0
rx_over_errors: 0
rx_crc_errors: 0
rx_frame_errors: 0
rx_fifo_errors: 112860
rx_missed_errors: 112860
tx_aborted_errors: 0
tx_carrier_errors: 0
tx_fifo_errors: 0
tx_heartbeat_errors: 0
tx_window_errors: 0
tx_abort_late_coll: 0
tx_deferred_ok: 0
tx_single_coll_ok: 0
tx_multi_coll_ok: 0
rx_long_length_errors: 0
rx_short_length_errors: 0
rx_align_errors: 0
tx_tcp_seg_good: 0
tx_tcp_seg_failed: 0
rx_flow_control_xon: 0
rx_flow_control_xoff: 0
tx_flow_control_xon: 45
tx_flow_control_xoff: 112905
rx_long_byte_count: 545746
rx_csum_offload_good: 1
rx_csum_offload_errors: 0
I rebooted with Q5 (+ netdev_max_backlog=32) and everything was fine. Is there
anything I can do to make my report more useful for you ?
P.O. Gaillard
* P.O. Gaillard <[email protected]> wrote:
> Hello,
>
> I just tried with Q6 and the network behaviour seems odd :
> - my ssh connection freezes a bit,
> - communication between the two PCs sharing the GigE connection is
> difficult to start :
please try the -Q9 patch i just released, and if you still see any
problems does the following:
echo 300 > /proc/sys/net/core/netdev_backlog_granularity
help?
Ingo