Pktgen has intermittent stalls playing about 770mbits/sec traffic through an
intel e1000 nic. We are using count = clone_skb=1000000, looping it with a
script. Each run takes 10 seconds assuming it doesnt hang. After putting
in some code diagnostic code, pktgen is hanging after all packets are sent
and it is looping waiting for the skb->users to get to 1. So you have to
hit control c to get it out of that loop. We got tired of that so we made
it go through the wait loop a max of 1000 times with a per loop mdelay(10)
and then automatically break out.
We ran it overnight and got 49 instances where it had to give up after
waiting 10 seconds for users to get to 1. Each time, skb->users was stuck
at 2. This is a dual p3 600.
I suppose we can work around this by simply making the test infinite and
break it when needed but the behavior was a little concerning...at least I
dont know if these skbs are ever being finally released...
Help STOP SPAM with the new MSN 8 and get 2 months FREE*
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to firstname.lastname@example.org
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/
This archive was generated by hypermail 2b29 : Tue Jul 15 2003 - 22:00:31 EST