[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: memory (mbuf) leak in fxp driver.
> * the problem seems to exist in 2.7, 2.8, and 2.9 (beta). I haven't checked
> -current.
I don't have problems with 2.7-stable, this (bridge) machine filters a 4Mbit
link.
pengo:/home/camield $ netstat -m
541 mbufs in use:
527 mbufs allocated to data
13 mbufs allocated to packet headers
1 mbuf allocated to socket names and addresses
256/450 mapped pages in use
967 Kbytes allocated to network (59% in use)
0 requests for memory denied
0 requests for memory delayed
0 calls to protocol drain routines
pengo:/home/camield $ vmstat -m
Memory statistics by bucket size
Size In Use Free Requests HighWater Couldfree
16 1646 146 22140416 1280 0
32 1117 291 14008431 640 8
64 1986 510 147712885 320 523
128 2690 7454 1545351716 160 39083583
256 3412 92 9568403 80 49
512 84 76 7648996 40 202999
1024 162 22 22963168 20 303586
2048 9 7 27599 10 0
4096 29 10 85801 5 80182
8192 2 0 2 5 0
16384 9 0 9 5 0
32768 6 0 322 5 0
65536 1 0 1 5 0
131072 1 0 1 5 0
netstat -in output, heavily edited. Counters overflowed.
Name Mtu Ipkts Ierrs Opkts Oerrs Coll
fxp0 1500 -1946629836 242864 1506465332 0 21774535
fxp1 1500 1507653787 27 1628895185 0 5780973
bridge0 1500 -436448576 0 -1159605186 0 0
pengo:/home/camield $ uptime
8:29AM up 146 days, 22:24, 1 user, load averages: 1.10, 1.10, 1.08
Kernel is not GENERIC, but has these changes:
+option NMBCLUSTERS=8192
+option NKMEMCLUSTERS=8192
+option MAX_KMAP=40
+option BUFCACHEPERCENT=20
+
+option IPSTATE_MAX=16000
+option IPSTATE_SIZE=12289
-maxusers 32 # estimated number of users
+maxusers 64 # estimated number of users
--
Cam