Alexander Motin wrote: >I have uploaded that patch back. Not sure is it correct at this moment, >there was some changes, but that time it worked fine.
Thanks a lot. I've installed the patch with some changes. I'm using the
7-STABLE (ng_base.c,v 1.135.2.10) at a production NAS. Corresponding patch
is in the attachment.
I can't say anything of performance yet. Not enough statistics data is
collected at the moment. But there is a strangeness with ng_queue threads:
under high network load top display too low WCPU (while TIME has a normal
value) for these threads. The maximum WCPU I saw under lower network load
was 35-40% per ng_queue thread. Are there any ideas? Am I using correctly
the kthread_create() in the patch?
last pid: 21000; load averages: 1.52, 1.40, 1.02 up 0+00:25:39
14:24:28
79 processes: 5 running, 63 sleeping, 11 waiting
CPU: 0.3% user, 0.0% nice, 41.4% system, 0.6% interrupt, 57.7% idle
Mem: 40M Active, 11M Inact, 166M Wired, 112K Cache, 77M Buf, 1731M Free
Swap: 2048M Total, 2048M Free
PID USERNAME THR PRI NICE SIZE RES STATE C TIME WCPU COMMAND
12 root 1 171 ki31 0K 16K RUN 2 18:36 58.50% idle: cpu2
11 root 1 171 ki31 0K 16K RUN 3 18:02 63.87% idle: cpu3
13 root 1 171 ki31 0K 16K CPU1 1 17:56 41.26% idle: cpu1
14 root 1 171 ki31 0K 16K CPU0 0 17:20 59.96% idle: cpu0
2 root 1 96 - 0K 16K sleep 0 7:16 0.20% ng_queue0
3 root 1 96 - 0K 16K sleep 1 7:13 0.20% ng_queue1
25 root 1 -68 - 0K 16K - 1 5:44 36.67% em0 taskq
26 root 1 -68 - 0K 16K - 2 5:26 35.60% em1 taskq
16 root 1 -32 - 0K 16K WAIT 0 1:30 3.86% swi4:
clock
2538 root 1 48 0 36720K 15228K select 3 0:25 0.59% mpd5
--
Vladimir Kobal
netgraph.threads.patch
Description: Binary data
_______________________________________________ freebsd-net@freebsd.org mailing list http://lists.freebsd.org/mailman/listinfo/freebsd-net To unsubscribe, send any mail to "[EMAIL PROTECTED]"
