On Tue, 29 Sep 1998, Hubert Tonneau wrote:
> I have noticed that with 2.0.35 (or 2.0.36pre12) kernel, heavy swapping
> in a process may generate a general protection fault in another one !
I'm running 2.0.36pre12 right now. I'm also running swapper.c:
#include <stdio.h>
#include <stdlib.h>
main()
{
char *mem;
char *ptr;
mem=malloc(73400320); /* 70M */
while(1)
{
ptr=mem+(( rand() / 2048 ) * 70 ); /* pick a byte */
/* printf("%d\n",(rand()/2048)*70);*/
/* printf("%d\n",(int)ptr);*/
*ptr=(rand()/RAND_MAX) * 255; /* dirty it */
}
}
(ugly, I know :)
I have a 64M machine with 128M swap. I'm currently using 62592K of
physical and 88156K swap. Swapper currently has a size of 70812K and 49M
RSS. The load is 1.25, 1.03, 0.53.
Um... I also have KDE, Navigator and some other stuff running.
My machine feels a little lagged, but it's otherwise working normally.
> I can repoduce the problem at any time:
>
> 1) i open two nxterm session.
Yup.
> 2) in the first one, i run a program that will consume very much
> memory (130 Mb whereas the physical memory is 64 Mb), then access
> randomly all the memor.
If you want to post your program (or a pointer to it if it's a 'real'
application, I'll try that too.
> 3) in the second session, i just do some 'ls'
> It will generate segmentation fault as long as the first session
> is heavily swapping.
>
>
> Please find below the report of the second nxterm session.
>
> bash# ls -l /database/
> total 60726
> -rw-r--r-- 1 root root 5094086 Sep 25 19:28 heliosam.linux
> -rwxr-xr-x 1 root root 56843448 Sep 23 12:03 heliosam.os2
> bash# ls -l /database/
> Segmentation fault (core dumped)
> bash# ls -l /database/
> total 56731
> -rw-r--r-- 1 root root 1019904 Sep 29 16:31 heliosam.linux
> -rwxr-xr-x 1 root root 56843448 Sep 23 12:03 heliosam.os2
> bash#
Not for me.
> While typing the first 'ls' the machine was not very much loaded,
> while typping the second, a process was swapping intensively,
> while typing the third, the heavy swapping process had been killed.
>
>
> Extra information:
>
> - i am running on an SMP box
I am on a UP PII-266.
> - the problem does not occure with 2.1.12x kernel
I haven't tried any 2.1.x kernels.
Best Regards,
Alex Butcher. (swapper now has an RSS of 53M :)
-
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/