wendy wrote:
Hello,
I was doing a stress test of my GLD2 driver, I started 20 windows each pinging
another PC. However, after a while, ping looks slower and finally some ping
stops but does not have any packet loss. Other windows' are still pinging and
even faster, as if the remaining sessions gain more CPU resource.
"
64 bytes from 192.168.0.10: icmp_seq=2917. time=9.180 ms
64 bytes from 192.168.0.10: icmp_seq=2918. time=6.028 ms
64 bytes from 192.168.0.10: icmp_seq=2919. time=15.373 ms
64 bytes from 192.168.0.10: icmp_seq=2920. time=9.335 ms
64 bytes from 192.168.0.10: icmp_seq=2921. time=1014.238 ms
[b]/*stops at here, I pressed ctrl-C to quit*/[/b]
^C
----192.168.0.10 PING Statistics----
2922 packets transmitted, 2922 packets received, 0% packet loss
round-trip (ms) min/avg/max/stddev = 0.476/2.960/1014.238/26.547 "
I don't why it stops, if the OS keeps sending and driver does not respond, then
there should be some packet lost, however, there is no packet loss. So, why
Ping stops ?
I suspect that the driver is "stuck" in the transmit section... perhaps
the transmit thread cannot acquire a shared lock? I'd have a look at
the threadlist with mdb.
-- Garrett
Wendy
This message posted from opensolaris.org
_______________________________________________
networking-discuss mailing list
[email protected]
_______________________________________________
networking-discuss mailing list
[email protected]
