> I keep running free on the 2nd xterm and got messages like that (disk
> spinning like crazy):
>
> $ free
> free: can't map '/lib/libc.so.5'
> At some time, the system got not responsive: no mouse, cant switch VCs,
> NumLock and CapsLock don't even lit the LEDs.
> System: the same as above with linux-2.1.124+your patch
>
> $ leak
> Bus error [ something different :-) ]
> $ leak
>
> The disk was not spinning like crazy, the system was more responsive.
>
> Suddenly, one xterm died, the other died, KDE interface died, X died and
> the kdm (xdm) appeared. I was able to log in again and go to TEST 3
This is the side effect of Andrea's patch I mentioned
before. It can even happen far _before_ you actually
run out of swap, but only with Andrea's patch applied.
A normal system will somewhat deadlock (looping in
kswapd). I plan to implement a test in kswapd testing
if it used over 30% of CPU over the last 10 seconds
(weighted gliding average) and then test for OOM.
When we are indeed OOM, and only then, we should proceed
with killing the worst offender.
Rik.
+-------------------------------------------------------------------+
| Linux memory management tour guide. H.H.vanRiel@phys.uu.nl |
| Scouting Vries cubscout leader. http://www.phys.uu.nl/~riel/ |
+-------------------------------------------------------------------+
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/