Package: linux-image-2.6.25-2-686
Version: 2.6.25-7

When Linux 2.6.25 has been running for a while (say, several hours)
xorg starts to respond more slowly and sporadically. Everythin seems
okay for a while after boot, or at least the effects are less
noticable. This is on a Dell laptop with a Core Duo CPU (not Core 2).

Just typing in an xterm right now I frequently get several characters
ahead of the display before it catches up. Logging into a vt instead
seems okay, so I guess it's just X that is affected.

System load average seems higher than it should be also.

kmshanah@kulgan:~$ uptime
06:40:35 up 12:53, 8 users, load average: 0.88, 0.93, 0.93

And this is what top is showing for me, no obvious CPU hogs or io wait
to cause that load:

top - 06:41:30 up 12:54, 8 users, load average: 0.70, 0.87, 0.91
Tasks: 136 total, 1 running, 135 sleeping, 0 stopped, 0 zombie
Cpu(s): 1.6%us, 1.6%sy, 0.0%ni, 96.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
Mem: 1546824k total, 749328k used, 797496k free, 83024k buffers
Swap: 2097144k total, 0k used, 2097144k free, 285324k cached
PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
3511 root 20 0 410m 63m 9064 S 1 4.2 4:36.25 Xorg
1124 root 15 -5 0 0 0 S 0 0.0 0:28.88 kcryptd
3864 kmshanah 20 0 47424 16m 8828 S 0 1.1 0:05.34 gnome-settings-
3918 kmshanah 20 0 23040 10m 8056 S 0 0.7 0:26.90 nm-applet
1 root 20 0 2100 720 620 S 0 0.0 0:02.02 init

This behaviour doesn't happen on 2.6.24. For my day-to-day work I've
gone back to using linux-image-2.6.24-1-686 (2.6.24-6) with no problems.

Apologies for the fairly vague bug report, but I'm at a bit of a loss
as to whan information to provide to assist.

Regards,
Kevin.



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org

On Wed, 23 Jul 2008, Kevin Shanahan wrote:

> Package: linux-image-2.6.25-2-686
> Version: 2.6.25-7
>
> When Linux 2.6.25 has been running for a while (say, several hours)
> xorg starts to respond more slowly and sporadically. Everythin seems
> okay for a while after boot, or at least the effects are less
> noticable. This is on a Dell laptop with a Core Duo CPU (not Core 2).
>
> Just typing in an xterm right now I frequently get several characters
> ahead of the display before it catches up. Logging into a vt instead
> seems okay, so I guess it's just X that is affected.
>
> System load average seems higher than it should be also.
>
> kmshanah@kulgan:~$ uptime
> 06:40:35 up 12:53, 8 users, load average: 0.88, 0.93, 0.93
>
> And this is what top is showing for me, no obvious CPU hogs or io wait
> to cause that load:
>
> top - 06:41:30 up 12:54, 8 users, load average: 0.70, 0.87, 0.91
> Tasks: 136 total, 1 running, 135 sleeping, 0 stopped, 0 zombie
> Cpu(s): 1.6%us, 1.6%sy, 0.0%ni, 96.7%id, 0.0%wa, 0.0%hi, 0.0%si, 0.0%st
> Mem: 1546824k total, 749328k used, 797496k free, 83024k buffers
> Swap: 2097144k total, 0k used, 2097144k free, 285324k cached
> PID USER PR NI VIRT RES SHR S %CPU %MEM TIME+ COMMAND
> 3511 root 20 0 410m 63m 9064 S 1 4.2 4:36.25 Xorg
> 1124 root 15 -5 0 0 0 S 0 0.0 0:28.88 kcryptd
> 3864 kmshanah 20 0 47424 16m 8828 S 0 1.1 0:05.34 gnome-settings-
> 3918 kmshanah 20 0 23040 10m 8056 S 0 0.7 0:26.90 nm-applet
> 1 root 20 0 2100 720 620 S 0 0.0 0:02.02 init
>
> This behaviour doesn't happen on 2.6.24. For my day-to-day work I've
> gone back to using linux-image-2.6.24-1-686 (2.6.24-6) with no problems.
>
> Apologies for the fairly vague bug report, but I'm at a bit of a loss
> as to whan information to provide to assist.

the relevant sched options are set for 2.6.26, please checkout trunk
daily snapshots:
http://wiki.debian.org/DebianKernel

thanks

--
maks



--
To UNSUBSCRIBE, email to debian-kernel-REQUEST@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmaster@lists.debian.org