Date: Tue, 28 Dec 2004 07:50:32 GMT From: "Rob Swindell" <rob@synchro.net> To: freebsd-bugs@FreeBSD.org Subject: Re: misc/75510: panic: kmem_malloc(4096): kmem_map too small Message-ID: <200412280750.iBS7oWVB013500@freefall.freebsd.org>
next in thread | raw e-mail | index | archive | help
The following reply was made to PR misc/75510; it has been noted by GNATS.
From: "Rob Swindell" <rob@synchro.net>
To: <freebsd-gnats-submit@FreeBSD.org>,
"Rob Swindell" <rob@synchro.net>
Cc:
Subject: Re: misc/75510: panic: kmem_malloc(4096): kmem_map too small
Date: Mon, 27 Dec 2004 23:57:05 -0800
Here's another capture of netstat -m (shortly before the panic), top (at the
time of the panic), /var/crash/info and vmcore kgdb backtrace.
[devil:~]# netstat -m
162 mbufs in use
33/4800 mbuf clusters in use (current/max)
0/3/1456 sfbufs in use (current/peak/max)
106 KBytes allocated to network
0 requests for sfbufs denied
0 requests for sfbufs delayed
0 requests for I/O initiated by sendfile
105 calls to protocol drain routines
last pid: 10117; load averages: 0.01, 0.04, 0.16
up 2+08:09:41 22:50:26
77 processes: 2 running, 50 sleeping, 25 waiting
CPU states: 1.6% user, 0.0% nice, 4.7% system, 1.6% interrupt, 92.2% idle
Mem: 26M Active, 3536K Inact, 46M Wired, 840K Cache, 22M Buf, 40M Free
Swap: 231M Total, 8K Used, 231M Free
PID USERNAME PRI NICE SIZE RES STATE TIME WCPU CPU COMMAND
7757 root 20 0 11908K 9988K kserel 1:58 0.00% 0.00% sbbs
8161 root 96 0 6092K 2288K select 0:00 0.00% 0.00% sshd
7985 root 96 0 6092K 2276K select 0:02 0.00% 0.00% sshd
654 root 96 0 6092K 2152K select 0:01 0.00% 0.00% sshd
377 root 96 0 3360K 1612K select 0:00 0.00% 0.00% sshd
7992 root 8 0 2208K 1600K wait 0:00 0.00% 0.00% bash
8176 root 5 0 2208K 1600K ttyin 0:00 0.00% 0.00% bash
657 root 8 0 2212K 1584K wait 0:00 0.00% 0.00% bash
7997 root 96 0 2292K 1336K RUN 0:10 0.00% 0.00% top
215 root 96 0 1784K 1060K select 0:06 0.00% 0.00% dhclient
390 root 8 0 1356K 808K nanslp 0:02 0.00% 0.00% cron
459 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
461 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
460 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
462 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
465 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
464 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
466 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
463 root 5 0 1280K 692K ttyin 0:00 0.00% 0.00% getty
265 root 96 0 1312K 676K select 0:02 0.00% 0.00% syslogd
426 root 107 0 1232K 600K select 0:00 0.00% 0.00% moused
341 root 96 0 1236K 592K select 0:01 0.00% 0.00% usbd
164 root 20 0 1184K 508K pause 0:00 0.00% 0.00% adjkerntz
1 root 8 0 740K 236K wait 0:00 0.00% 0.00% init
245 root 111 0 516K 224K select 0:00 0.00% 0.00% devd
11 root 98 0 0K 12K RUN 54.3H 92.97% 92.97% idle
405 root 8 0 0K 12K 90idle 0:57 0.15% 0.15% smbiod0
27 root -44 -163 0K 12K WAIT 0:33 0.05% 0.05% swi1: net
28 root -28 -147 0K 12K WAIT 77:42 0.00% 0.00% swi5: clock
sio
43 root 20 0 0K 12K syncer 1:01 0.00% 0.00% syncer
30 root 76 0 0K 12K - 0:26 0.00% 0.00% yarrow
4 root -8 0 0K 12K - 0:21 0.00% 0.00% g_down
2 root -8 0 0K 12K - 0:21 0.00% 0.00% g_event
21 root -68 -187 0K 12K WAIT 0:20 0.00% 0.00% irq10: de0
uhci0
3 root -8 0 0K 12K - 0:17 0.00% 0.00% g_up
45 root -8 0 0K 12K - 0:16 0.00% 0.00% schedcpu
40 root 171 52 0K 12K pgzero 0:13 0.00% 0.00% pagezero
32 root -24 -143 0K 12K WAIT 0:03 0.00% 0.00% swi6:+
42 root -4 0 0K 12K vlruwt 0:03 0.00% 0.00% vnlru
41 root -16 0 0K 12K psleep 0:02 0.00% 0.00% bufdaemon
25 root -64 -183 0K 12K WAIT 0:02 0.00% 0.00% irq14: ata0
Good dump found on device /dev/ad0s1b
Architecture: i386
Architecture version: 1
Dump length: 134205440B (127 MB)
Blocksize: 512
Dumptime: Mon Dec 27 22:50:28 2004
Hostname: devil.synchro.net
Versionstring: FreeBSD 5.3-RELEASE #0: Sat Dec 25 02:42:54 PST 2004
root@devil.synchro.net:/usr/src/sys/i386/compile/MYKERNEL
Panicstring: kmem_malloc(4096): kmem_map too small: 40898560 total allocated
Bounds: 1
(kgdb) bt
#0 doadump () at pcpu.h:159
#1 0xc05fdc49 in boot (howto=260) at ../../../kern/kern_shutdown.c:397
#2 0xc05fdf05 in panic (
fmt=0xc07c38d3 "kmem_malloc(%ld): kmem_map too small: %ld total allocated")
at ../../../kern/kern_shutdown.c:553
#3 0xc06fc8d9 in kmem_malloc (map=0xc103a0c0, size=4096, flags=258)
at ../../../vm/vm_kern.c:300
#4 0xc070b7c2 in page_alloc (zone=0xc1044d80, bytes=4096, pflag=0x0, wait=258)
at ../../../vm/uma_core.c:935
#5 0xc070b301 in slab_zalloc (zone=0xc1044d80, wait=258)
at ../../../vm/uma_core.c:805
#6 0xc070ca54 in uma_zone_slab (zone=0xc1044d80, flags=2)
at ../../../vm/uma_core.c:1962
#7 0xc070cc74 in uma_zalloc_bucket (zone=0xc1044d80, flags=2)
at ../../../vm/uma_core.c:2071
#8 0xc070c900 in uma_zalloc_arg (zone=0xc1044d80, udata=0x0, flags=2)
at ../../../vm/uma_core.c:1889
#9 0xc06455ba in cache_enter (dvp=0xc1dc2a50, vp=0xc1bb4528, cnp=0xcc855a48)
at uma.h:274
#10 0xc1593310 in ?? ()
#11 0xc1dc2a50 in ?? ()
#12 0xc1bb4528 in ?? ()
#13 0xcc855a48 in ?? ()
As Stephen stated, the application (Synchronet, aka "sbbs") doesn't use KVM, so
I don't know how it could be "causing [the] kernel to run out of KVM" as Kris
claimed. The application requires between 10 and 20MB of memory, there is 128MB
physical RAM and 256MB of swap available. If the OS can't run the application
in this configuration without a kernel panic, I hardly see "add more RAM" as a
viable solution.
Thanks,
-Rob
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200412280750.iBS7oWVB013500>