Hi all,
I have a Toshiba laptop (DynaBook SS/S5: PIIIM 800MHz, 256MB, 20GB 1.8" IDE) that has been running
all kernels since 2.6.0 (vanilla with occasional small patches) as a Gentoo machine.
Now, recently (a few months) I have been using it more and even as a desktop (leaving it up most of
the time). What I use 95% of the time is xorg/windowmaker/{xterm,firefox,thunderbird,gaim}...
The problem is that after some time memory is consumed (as `free -m` shows), response (kbd, mouse)
slows down to a stall and hard disk spins ad infinum (noise/ide LED). Sometimes I am able to switch
to a xterm/console and see that there is 2-3MB free memory, 1-2MB cached and 0-2MB buffers. Killing
all apps (X/WM included), `telinit 1` gives me something like 100-150MB used memory (not counting
buffers/cache) and virtually no apps are running. Restarting the machine, gives about 10-15MB used.
BTW, adding swap just slows down the process, but result is the same. Usually I work without swap.
Who uses that memory?
Why cannot be reclaimed by kernel?
WTF is the disk reading(writing?) all the time? (no swap)
So basically, how do I debug what process gets memory and doesn't release it?
At present (2.6.13.3) I need to reboot about once a day if I use the machine. And sometimes the only
way is coldboot or Alt+SysRq+{S;U,S,B}...
I am running almost the same config on a few other boxes but the problem does not seem to exist
(they have more memory though). The special thing on this laptop is the IDE and VGA subsystem:
00:04.0 IDE interface: ALi Corporation M5229 IDE (rev c3)
01:00.0 VGA compatible controller: Trident Microsystems CyberBlade XPAi1 (rev 82)
The 1.8" HDD is: Device Model: TOSHIBA MK2003GAH
All advice is appreciated!
Kalin.
--
|[ ~~~~~~~~~~~~~~~~~~~~~~ ]|
+-> http://ThinRope.net/ <-+
|[ ______________________ ]|
> Who uses that memory?
the programs
slabtop
xrestop
cat /proc/meminfo
are probably good starting points to find the leak; they give a detailed
overview of various memory pools
Arjan van de Ven wrote:
>>Who uses that memory?
>
>
> the programs
>
> slabtop
> xrestop
> cat /proc/meminfo
>
> are probably good starting points to find the leak; they give a detailed
> overview of various memory pools
Thank you for the info, I didn't know about xrestop :-)
So here is a present state of the machine:
ss ~ # date
Thu Dec 15 12:19:16 JST 2005
ss ~ # free -m
total used free shared buffers cached
Mem: 233 228 5 0 64 8
-/+ buffers/cache: 155 78
Swap: 0 0 0
ss ~ # cat /proc/meminfo
MemTotal: 239460 kB
MemFree: 5588 kB
Buffers: 66212 kB
Cached: 8920 kB
SwapCached: 0 kB
Active: 53708 kB
Inactive: 28328 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 239460 kB
LowFree: 5588 kB
SwapTotal: 0 kB
SwapFree: 0 kB
Dirty: 0 kB
Writeback: 0 kB
Mapped: 9752 kB
Slab: 148528 kB
CommitLimit: 119728 kB
Committed_AS: 16316 kB
PageTables: 260 kB
VmallocTotal: 794324 kB
VmallocUsed: 2276 kB
VmallocChunk: 791880 kB
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 4096 kB
ss ~ # ps -A v k-size
PID TTY STAT TIME MAJFL TRS DRS RSS %MEM COMMAND
5153 tty2 Ss+ 0:01 10 515 4384 2044 0.8 -bash
4140 ? Ss 0:00 298 247 4724 860 0.3 /usr/sbin/cupsd
6572 pts/0 Ss 0:01 2 515 4124 2832 1.1 -bash
6720 pts/0 R+ 0:00 0 55 2676 760 0.3 ps -A v k-size 50
3448 ? Ss 0:00 17 133 1894 1044 0.4 /usr/sbin/cannaserver
6567 ? Ss 0:00 6 292 6087 2112 0.8 sshd: root@pts/0
4044 ? Ss 0:00 4 319 1796 736 0.3 /sbin/dhclient -cf /etc/dhcp/dhclient.co
4680 ? Ss 0:00 60 21 2146 500 0.2 login -- kalin
4485 ? Ss 0:00 3 292 3103 404 0.1 /usr/sbin/sshd
3381 ? Ss 0:00 28 100 1747 564 0.2 /usr/sbin/syslog-ng
4363 ? Ss 0:00 0 33 1694 120 0.0 /sbin/rpc.statd
4681 tty3 Ss+ 0:00 0 11 1452 64 0.0 /sbin/agetty 38400 tty3 linux
4704 tty4 Ss+ 0:00 0 11 1452 64 0.0 /sbin/agetty 38400 tty4 linux
4709 tty5 Ss+ 0:00 0 11 1452 60 0.0 /sbin/agetty 38400 tty5 linux
6015 tty1 Ss+ 0:00 1 11 1452 520 0.2 /sbin/agetty 38400 tty1 linux
4288 ? Ss 0:00 0 10 1681 116 0.0 /sbin/portmap
4714 tty6 Ss+ 0:00 0 11 1448 60 0.0 /sbin/agetty 38400 tty6 linux
841 ? S<s 0:00 1 22 1445 212 0.0 udevd
1 ? S 0:00 345 22 1441 160 0.0 init [3]
4215 ? Ss 0:07 739 16 2083 212 0.0 /usr/sbin/fnfxd
4550 ? Ss 0:00 186 9 1454 284 0.1 /usr/sbin/uptimed
2 ? SN 0:00 0 0 0 0 0.0 [ksoftirqd/0]
3 ? S< 0:00 0 0 0 0 0.0 [events/0]
4 ? S< 0:00 0 0 0 0 0.0 [khelper]
5 ? S< 0:00 0 0 0 0 0.0 [kthread]
7 ? S< 0:47 0 0 0 0 0.0 [kacpid]
95 ? S< 0:00 0 0 0 0 0.0 [kblockd/0]
134 ? S 0:00 0 0 0 0 0.0 [pdflush]
137 ? S< 0:00 0 0 0 0 0.0 [aio/0]
136 ? S 0:01 0 0 0 0 0.0 [kswapd0]
724 ? S< 0:00 0 0 0 0 0.0 [kseriod]
753 ? S< 0:00 0 0 0 0 0.0 [reiserfs/0]
1245 ? S< 0:00 0 0 0 0 0.0 [khubd]
1410 ? S 0:00 0 0 0 0 0.0 [pccardd]
1477 ? S 0:00 0 0 0 0 0.0 [pccardd]
4399 ? S< 0:01 0 0 0 0 0.0 [rpciod/0]
4400 ? S 0:00 0 0 0 0 0.0 [lockd]
5330 ? S 0:00 0 0 0 0 0.0 [pdflush]
According to slabtop:
Active / Total Objects (% used) : 1806688 / 1826914 (98.9%)
Active / Total Slabs (% used) : 36958 / 36958 (100.0%)
Active / Total Caches (% used) : 64 / 106 (60.4%)
Active / Total Size (% used) : 138777.10K / 143195.01K (96.9%)
Minimum / Average / Maximum Object : 0.02K / 0.08K / 128.00K
OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
1768936 1768884 99% 0.07K 34018 52 136072K size-64
19642 16616 84% 0.06K 322 61 1288K buffer_head
10152 4828 47% 0.14K 376 27 1504K dentry_cache
ss ~ # cat /proc/slabinfo |grep "size-64 "
size-64 1767932 1768000 76 52 1 : tunables 32 16 0 : slabdata 33999
34000 0 : globalstat 21513120 1767964 34005 0 0 1 84 0 : cpustat 614583639
1370079 612951759 1234047
ss ~ # lsmod
Module Size Used by
nfs 114540 1
lockd 66920 2 nfs
nfs_acl 3680 1 nfs
sunrpc 152004 4 nfs,lockd,nfs_acl
usbhid 27876 0
yenta_socket 26028 2
rsrc_nonstatic 14112 1 yenta_socket
ehci_hcd 33224 0
ohci_hcd 21988 0
usbcore 125660 4 usbhid,ehci_hcd,ohci_hcd
e100 37984 0
mii 5568 1 e100
OK, now stop whatever I can, unload most modules and again:
ss ~ # free -m
total used free shared buffers cached
Mem: 233 227 6 0 63 10
-/+ buffers/cache: 153 80
Swap: 0 0 0
ss ~ # ps -A v k-size
PID TTY STAT TIME MAJFL TRS DRS RSS %MEM COMMAND
5153 tty2 Ss+ 0:01 10 515 4384 2044 0.8 -bash
6572 pts/0 Ss 0:01 2 515 4124 2840 1.1 -bash
7352 pts/0 R+ 0:00 0 55 2676 756 0.3 ps -A v k-size
6567 ? Ss 0:00 6 292 6087 2112 0.8 sshd: root@pts/0
4044 ? Ss 0:00 4 319 1796 736 0.3 /sbin/dhclient -cf /etc/dhcp/dhclient.co
4680 ? Ss 0:00 60 21 2146 500 0.2 login -- kalin
4485 ? Ss 0:00 3 292 3103 404 0.1 /usr/sbin/sshd
3381 ? Ss 0:00 28 100 1747 564 0.2 /usr/sbin/syslog-ng
4681 tty3 Ss+ 0:00 0 11 1452 64 0.0 /sbin/agetty 38400 tty3 linux
4704 tty4 Ss+ 0:00 0 11 1452 64 0.0 /sbin/agetty 38400 tty4 linux
4709 tty5 Ss+ 0:00 0 11 1452 60 0.0 /sbin/agetty 38400 tty5 linux
6015 tty1 Ss+ 0:00 1 11 1452 520 0.2 /sbin/agetty 38400 tty1 linux
4714 tty6 Ss+ 0:00 0 11 1448 60 0.0 /sbin/agetty 38400 tty6 linux
841 ? S<s 0:00 1 22 1445 256 0.1 udevd
1 ? S 0:00 345 22 1441 160 0.0 init [3]
4550 ? Ss 0:00 186 9 1454 284 0.1 /usr/sbin/uptimed
2 ? SN 0:00 0 0 0 0 0.0 [ksoftirqd/0]
3 ? S< 0:00 0 0 0 0 0.0 [events/0]
4 ? S< 0:00 0 0 0 0 0.0 [khelper]
5 ? S< 0:00 0 0 0 0 0.0 [kthread]
7 ? S< 0:47 0 0 0 0 0.0 [kacpid]
95 ? S< 0:00 0 0 0 0 0.0 [kblockd/0]
134 ? S 0:00 0 0 0 0 0.0 [pdflush]
137 ? S< 0:00 0 0 0 0 0.0 [aio/0]
136 ? S 0:01 0 0 0 0 0.0 [kswapd0]
724 ? S< 0:00 0 0 0 0 0.0 [kseriod]
753 ? S< 0:00 0 0 0 0 0.0 [reiserfs/0]
1245 ? S< 0:00 0 0 0 0 0.0 [khubd]
1410 ? S 0:00 0 0 0 0 0.0 [pccardd]
1477 ? S 0:00 0 0 0 0 0.0 [pccardd]
5330 ? S 0:00 0 0 0 0 0.0 [pdflush]
ss ~ # cat /proc/slabinfo |grep "size-64 "
size-64 1770831 1770860 76 52 1 : tunables 32 16 0 : slabdata 34055
34055 0 : globalstat 21551153 1770860 34060 0 0 1 84 0 : cpustat 615607009
1372499 613972465 1236227
ss ~ # lsmod
Module Size Used by
yenta_socket 26028 2
rsrc_nonstatic 14112 1 yenta_socket
usbcore 125660 1
e100 37984 0
mii 5568 1 e100
Now, I cannot `kill -9 1410` or 1477 (the pccard) or unload yeanta_socket and rsrc_nonstatic (that
never worked).
ss ~ # reboot
Now let's how things changed... From 153MB we are back down to 14MB in use... that is a leak of more
than half my pshysical RAM...
ss ~ # free -m
total used free shared buffers cached
Mem: 233 29 204 0 3 11
-/+ buffers/cache: 14 219
Swap: 0 0 0
ss ~ # cat /proc/meminfo
MemTotal: 239460 kB
MemFree: 209544 kB
Buffers: 3324 kB
Cached: 12276 kB
SwapCached: 0 kB
Active: 14448 kB
Inactive: 5720 kB
HighTotal: 0 kB
HighFree: 0 kB
LowTotal: 239460 kB
LowFree: 209544 kB
SwapTotal: 0 kB
SwapFree: 0 kB
Dirty: 0 kB
Writeback: 0 kB
Mapped: 7996 kB
Slab: 6564 kB
CommitLimit: 119728 kB
Committed_AS: 13456 kB
PageTables: 252 kB
VmallocTotal: 794324 kB
VmallocUsed: 2276 kB
VmallocChunk: 791880 kB
HugePages_Total: 0
HugePages_Free: 0
Hugepagesize: 4096 kB
ss ~ # lsmod
Module Size Used by
nfs 114540 1
lockd 66920 2 nfs
nfs_acl 3680 1 nfs
sunrpc 152004 4 nfs,lockd,nfs_acl
usbhid 27876 0
yenta_socket 26028 2
rsrc_nonstatic 14112 1 yenta_socket
ehci_hcd 33224 0
ohci_hcd 21988 0
usbcore 125660 4 usbhid,ehci_hcd,ohci_hcd
e100 37984 0
mii 5568 1 e100
ss ~ # cat /proc/slabinfo |grep "size-64 "
size-64 1264 1300 76 52 1 : tunables 32 16 0 : slabdata 25 25
0 : globalstat 6357 1280 31 1 0 1 84 0 : cpustat 151935 427 150804 317
ss ~ # ps -A v k-size
PID TTY STAT TIME MAJFL TRS DRS RSS %MEM COMMAND
4065 ? Ss 0:00 10 247 4724 1872 0.7 /usr/sbin/cupsd
4651 pts/0 Ss 0:00 8 515 3992 2804 1.1 -bash
4678 pts/0 R+ 0:00 0 55 2676 760 0.3 ps -A v k-size
3969 ? Ss 0:00 0 319 1796 996 0.4 /sbin/dhclient -cf /etc/dhcp/dhclient.co
4646 ? Ss 0:00 2 292 5927 2060 0.8 sshd: root@pts/0
4410 ? Ss 0:00 0 292 3103 1508 0.6 /usr/sbin/sshd
3373 ? Ss 0:00 0 133 1634 764 0.3 /usr/sbin/cannaserver
3306 ? Ss 0:00 0 100 1747 880 0.3 /usr/sbin/syslog-ng
4288 ? Ss 0:00 0 33 1698 824 0.3 /sbin/rpc.statd
4213 ? Ss 0:00 0 10 1685 648 0.2 /sbin/portmap
4619 tty3 Ss+ 0:00 0 11 1452 528 0.2 /sbin/agetty 38400 tty3 linux
4621 tty5 Ss+ 0:00 0 11 1452 528 0.2 /sbin/agetty 38400 tty5 linux
801 ? S<s 0:00 0 22 1449 480 0.2 udevd
4614 tty1 Ss+ 0:00 0 11 1448 524 0.2 /sbin/agetty 38400 tty1 linux
4618 tty2 Ss+ 0:00 0 11 1448 520 0.2 /sbin/agetty 38400 tty2 linux
4620 tty4 Ss+ 0:00 0 11 1448 524 0.2 /sbin/agetty 38400 tty4 linux
4622 tty6 Ss+ 0:00 0 11 1448 524 0.2 /sbin/agetty 38400 tty6 linux
1 ? S 0:00 16 22 1441 504 0.2 init [3]
4140 ? Ss 0:00 0 16 2083 560 0.2 /usr/sbin/fnfxd
4475 ? Ss 0:00 0 9 1450 508 0.2 /usr/sbin/uptimed
2 ? SN 0:00 0 0 0 0 0.0 [ksoftirqd/0]
3 ? S< 0:00 0 0 0 0 0.0 [events/0]
4 ? S< 0:00 0 0 0 0 0.0 [khelper]
5 ? S< 0:00 0 0 0 0 0.0 [kthread]
7 ? S< 0:00 0 0 0 0 0.0 [kacpid]
95 ? S< 0:00 0 0 0 0 0.0 [kblockd/0]
134 ? S 0:00 0 0 0 0 0.0 [pdflush]
135 ? S 0:00 0 0 0 0 0.0 [pdflush]
137 ? S< 0:00 0 0 0 0 0.0 [aio/0]
136 ? S 0:00 0 0 0 0 0.0 [kswapd0]
724 ? S< 0:00 0 0 0 0 0.0 [kseriod]
753 ? S< 0:00 0 0 0 0 0.0 [reiserfs/0]
1243 ? S< 0:00 0 0 0 0 0.0 [khubd]
1411 ? S 0:00 0 0 0 0 0.0 [pccardd]
1490 ? S 0:00 0 0 0 0 0.0 [pccardd]
4324 ? S< 0:00 0 0 0 0 0.0 [rpciod/0]
4325 ? S 0:00 0 0 0 0 0.0 [lockd]
ss ~ # uname -a
Linux ss 2.6.13.3-ss #1 Tue Oct 18 14:08:05 Local time zone must be set--see zic manu i686 Intel(R)
Pentium(R) III Mobile CPU 800MHz GenuineIntel GNU/Linux
** Hmm, have to check my timezone :-) **
So, given all this info, does anybody see anything strange? What else can I do?
Or, I should put kdb in action...
What else can I compare before and after a reboot?
Will /proc/config.gz provide any info (just trying to save bandwith, can post it anytime)?
This is a vanilla 2.6.13.3 + mppe-mppc-1.3.patch (as module), but the patch/module was never
used/loaded for this session.
NB: In this situation, X was shut down, so xrestop is irrelevant.
Thank you,
Kalin.
--
|[ ~~~~~~~~~~~~~~~~~~~~~~ ]|
+-> http://ThinRope.net/ <-+
|[ ______________________ ]|
On 12/14/05, Kalin KOZHUHAROV <[email protected]> wrote:
<snip>
> According to slabtop:
> Active / Total Objects (% used) : 1806688 / 1826914 (98.9%)
> Active / Total Slabs (% used) : 36958 / 36958 (100.0%)
> Active / Total Caches (% used) : 64 / 106 (60.4%)
> Active / Total Size (% used) : 138777.10K / 143195.01K (96.9%)
> Minimum / Average / Maximum Object : 0.02K / 0.08K / 128.00K
>
> OBJS ACTIVE USE OBJ SIZE SLABS OBJ/SLAB CACHE SIZE NAME
> 1768936 1768884 99% 0.07K 34018 52 136072K size-64
> 19642 16616 84% 0.06K 322 61 1288K buffer_head
> 10152 4828 47% 0.14K 376 27 1504K dentry_cache
>
> ss ~ # cat /proc/slabinfo |grep "size-64 "
> size-64 1767932 1768000 76 52 1 : tunables 32 16 0 : slabdata 33999
> 34000 0 : globalstat 21513120 1767964 34005 0 0 1 84 0 : cpustat 614583639
> 1370079 612951759 1234047
>
> ss ~ # lsmod
> Module Size Used by
> nfs 114540 1
> lockd 66920 2 nfs
> nfs_acl 3680 1 nfs
> sunrpc 152004 4 nfs,lockd,nfs_acl
> usbhid 27876 0
> yenta_socket 26028 2
> rsrc_nonstatic 14112 1 yenta_socket
> ehci_hcd 33224 0
> ohci_hcd 21988 0
> usbcore 125660 4 usbhid,ehci_hcd,ohci_hcd
> e100 37984 0
> mii 5568 1 e100
<snip>
size-64 is pretty tiny, and I don't think e100 makes any allocations
that tiny, so you should be able to eliminate e100. If you try an
older kernel and it works you might be able to get an idea of where
changes have been made to start looking for kmallocs of size <= 64
bytes
your unloading module case is a bit of a misnomer because the item
that leaked may not know it leaked and therefore could unload without
freeing the things it leaked.
I'm not sure how to help you from here, have you tried a kernel.org
kernel like 2.6.14.3? You're likely to get more help from here if you
do.
Jesse