I'm running 7.1 on an IXP425 embedded SBC (Avila). The SBC is used to host a pt2pt Wifi bridge. I have been experiencing periodic packet errors (10-6) where I believe the qmgr0 and the ath0 intr are colliding (occasional mbuf exhaustion). I recently turned on PREEMPTION and all errors cease but after several minutes the system stops passing packets and when I enter anything on the serial port, it resumes, thus I can't see what had happened. I am a bit disparate here and will reward a solution provider handsomely. I am also wondering how to ensure the interrupts are prioritized appropriately and if remote debugging would be of use..
Jim Carlson 707 223 4000
Code:
top -S
PID USERNAME THR PRI NICE SIZE RES STATE TIME WCPU COMMAND
10 root 1 171 ki31 0K 1952K RUN 21:14 87.21% idle
20 root 1 -68 - 0K 1952K WAIT 1:22 5.13% intr3: ixpqmgr0
18 root 1 -68 - 0K 1952K WAIT 0:51 2.93% intr27: ath0
7 root 1 -68 - 0K 1952K - 1:01 1.03% ath0 taskq
733 root 1 8 0 4076K 1112K nanslp 0:07 0.00% sentry
24 root 1 -64 - 0K 1952K WAIT 0:04 0.00% intr29: ata_avi
11 root 1 -32 - 0K 1952K WAIT 0:04 0.00% swi4: clock
723 root 1 8 0 3692K 984K nanslp 0:01 0.00% dials
728 root 1 96 0 4080K 1700K mqrecv 0:01 0.00% kahuna
4 root 1 -8 - 0K 1952K - 0:01 0.00% g_down
3 root 1 -8 - 0K 1952K - 0:00 0.00% g_up
2 root 1 -8 - 0K 1952K - 0:00 0.00% g_event
14 root 1 -16 - 0K 1952K - 0:00 0.00% yarrow
946 root 1 96 0 5048K 2480K select 0:00 0.00% mini_httpd
45 root 1 -8 - 0K 1952K mdwait 0:00 0.00% md0
29 root 1 -16 - 0K 1952K - 0:00 0.00% schedcpu
1063 root 1 20 0 3768K 2144K pause 0:00 0.00% csh