memory leak in 2.4 kernel - ways to debug ??
Hi,
Since I've upgraded my 164LX 533Mhz-Alpha to 384 MB RAM (from 128MB),
kernel 2.4 sometimes "eats up" about 160MB which is obviously not the expected
behaviour.
At the moment I'm using kernel 2.4.5-pre1, but the problem also occurs in
earlier versions of 2.4 (I updated to the pre-version, hoping that the
problem would have been fixed).
Testing with 2.2 is rather difficult, since many of my filesystems are
reiserfs now and the 2.2.19-reiserfs-patch can't mount these partitions.
Has anybody else experienced this problem (sorry, if this is a FAQ) ??
What possibilities are there to debug this behaviour (i'm far away from
a kernel hacker) ??
Here are some diagnostic outputs, only init (+gettys), the logging-daemons, 2x
bash, gpm and the software to write this mail are running (mutt+vi).
free:
total used free shared buffers cached
Mem: 382304 237424 144880 0 1776 50200
-/+ buffers/cache: 185448 196856
Swap: 0 0 0
ps axv:
PID TTY STAT TIME MAJFL TRS DRS RSS %MEM COMMAND
1 ? S 0:03 0 1069 1442 880 0.2 init
2 ? SW 0:00 0 0 0 0 0.0 [keventd]
3 ? SW 0:01 0 0 0 0 0.0 [kswapd]
4 ? SW 0:00 0 0 0 0 0.0 [kreclaimd]
5 ? SW 0:00 0 0 0 0 0.0 [bdflush]
6 ? SW 0:00 0 0 0 0 0.0 [kupdated]
66 ? SW 0:00 0 0 0 0 0.0 [kreiserfsd]
157 ? S 0:00 0 1065 5062 1184 0.3 /sbin/syslogd
159 ? S 0:00 0 1053 1514 888 0.2 /sbin/klogd
312 tty3 S 0:00 0 1044 1435 872 0.2 /sbin/getty 115200 tty3
313 tty4 S 0:00 0 1044 1435 872 0.2 /sbin/getty 115200 tty4
314 tty5 S 0:00 0 1044 1435 872 0.2 /sbin/getty 115200 tty5
315 tty6 S 0:00 0 1044 1435 872 0.2 /sbin/getty 115200 tty6
322 tty2 S 0:00 0 1746 6509 2216 0.5 -bash
1183 tty1 S 0:00 0 1746 6493 2192 0.5 -bash
1213 tty1 S 0:00 0 1782 6809 2608 0.6 mutt
1218 tty1 S 0:00 0 1505 5590 2176 0.5 /usr/bin/editor /tmp/mut 1225 ? S 0:00 0 1119 1440 840 0.2 /usr/sbin/gpm -m /dev/ps 1226 tty2 R 0:00 0 1112 7775 2120 0.5 ps axv
The 0.-percentages show, that the huge (for me) memory usage of 185M isn't
produced by a running program. Calculated with bc, 167M are left for the
kernel - even too much if you loaded hundreds of modules.
head of top:
7:32pm up 1:28, 2 users, load average: 0.00, 0.00, 0.00
19 processes: 18 sleeping, 1 running, 0 zombie, 0 stopped
CPU states: 0.2% user, 0.2% system, 0.0% nice, 99.4% idle
Mem: 382304K av, 238448K used, 143856K free, 0K shrd, 1776K buff
Swap: 0K av, 0K used, 0K free 50184K cached
/proc/meminfo:
total: used: free: shared: buffers: cached:
Mem: 391479296 243105792 148373504 0 1818624 51388416
Swap: 0 0 0
MemTotal: 382304 kB
MemFree: 144896 kB
MemShared: 0 kB
Buffers: 1776 kB
Cached: 50184 kB
Active: 4272 kB
Inact_dirty: 39752 kB
Inact_clean: 7936 kB
Inact_target: 0 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 382304 kB
LowFree: 144896 kB
SwapTotal: 0 kB
SwapFree: 0 kB
Thanks in advance for any tips,
Marius Hillenbrand
--
living a RISCy life
PGP PublicKey: http://sites.inka.de/sirius/marius.key.asc
Reply to: